New Carbon Based Materials For Electrochemical and Fuel Cells Energy Storage Systems: Batteries, Supercapacitors

New Carbon Based Materials for Electrochemical
Energy Storage Systems: Batteries, Supercapacitors

and Fuel Cells
NATO Science Series
ASeries presenting the results of scientific meetings supported under the NATO Science
Programme.
The Series is published by IOS Press, Amsterdam, and Springer in conjunction with the NATO
Public Diplomacy Division
Sub-Series
I. Life and Behavioural Sciences IOS Press

II. Mathematics, Physics and Chemistry Springer
III. Computer and Systems Science IOS Press
IV. Earth and Environmental Sciences Springer
The NATO Science Series continues the series of books published formerly as the NATO ASI Series.
The NATO Science Programme offers support for collaboration in civil science between scientists of
countries of the Euro-Atlantic Partnership Council. The types of scientific meeting generally supported
are “Advanced Study Institutes” and “Advanced Research Workshops”, and the NATO Science Series
collects together the results of these meetings. The meetings are co-organized bij scientists from
NATO countries and scientists from NATO’s Partner countries – countries of the CIS and Central and
Eastern Europe.
Advanced Study Institutes are high-level tutorial courses offering in-depth study of latest advances
in a field.
Advanced Research Workshops are expert meetings aimed at critical assessment of a field, and
identification of directions for future action.
As a consequence of the restructuring of the NATO Science Programme in 1999, the NATO Science
Series was re-organised to the four sub-series noted above. Please consult the following web sites for
information on previous volumes published in the Series.
http://www.nato.int/science
http://www.springer.com
http://www.iospress.nl
Series II: Mathematics, Physics and Chemistry – Vol. 229

New Carbon Based Materials
for Electrochemical Energy
Storage Systems: Batteries,
Supercapacitors and Fuel Cells
edited by
Igor V. Barsukov
Superior Graphite Co., Chicago, IL, U.S.A.
Christopher S. Johnson
Argonne National Laboratory, Argonne, IL, U.S.A.
Joseph E. Doninger
Dontech Global, Inc., Lake Forest, IL, U.S.A.
and
Vyacheslav Z. Barsukov
Kiev National University of Technologies and Design,
Kiev, Ukraine
Published in cooperation with NATO Public Diplomacy Division

NATO - Carbon Advanced Research Workshop and Conference
New Carbon Based Materials for Electrochemical Energy Storage Systems
Argonne, Illinois, U.S.A.
19-24 October 2003
A C.I.P. Catalogue record for this book is available from the Library of Congress.
ISBN-10 1-4020-4811-4 (PB)

ISBN-13 978-1-4020-4811-1 (PB)
ISBN-10 1-4020-4810-6 (HB)
ISBN-13 978-1-4020-4810-4 (HB)
ISBN-10 1-4020-4812-2 (e-book)
ISBN-13 978-1-4020-4812-8 (e-book)
Published by Springer,
P.O. Box 17, 3300 AA Dordrecht, The Netherlands.
www.springer.com
Printed on acid-free paper
All Rights Reserved

© 2006 Springer
No part of this work may be reproduced, stored in a retrieval system, or transmitted in
any form or by any means, electronic, mechanical, photocopying, microfilming,
recording or otherwise, without written permission from the Publisher, with the exception
of any material supplied specifically for the purpose of being entered and executed on a
computer system, for exclusive use by the purchaser of the work.
Printed in the Netherlands.

TABLE OF CONTENTS
Preface ......................................................................................................... xi
Letter from Dr. Hermann Grunder – ANL.................................................. xiii
Photographs & Comments .......................................................................... xiv
Organization............................................................................................... xvii
List of Participants .................................................................................... xviii
Chapter 1
New Carbon Materials for Supercapacitors ............................................. 1
Subject Overview............................................................................... 3
Novel Carbonaceous Materials for Application in the

Electrochemical Supercapacitors ................................................ 5
E. Frackowiak, J. Machnikowski and F. Béguin
Effect of Carbonaceous Materials on Performance of

Carbon-Carbon and Carbon-Ni Oxide Types of
Electrochemical Capacitors with Alkaline
Electrolyte ................................................................................. 21
A. I. Belyakov
Hybrid Supercapacitors Based on a-MnO2/Carbon

Nanotubes Composites.............................................................. 33
V. Khomenko, E. Raymundo-Piñero and F. Béguin
Development of Supercapacitors Based on

Conducting Polymers ................................................................ 41
V. Khomenko, E. Frackowiak, V. Barsukov, and F. Béguin
Supercapacitors: Old Problems and New Trends...................... 51

Y. Malein, N. Strizhakova, V. Izotov, A. Mironova,
S. Kozachkov, V. Danilin and S. Podmogilny
Modeling Porosity Development During KOH

Activation of Coal and Pitch-Derived Carbons for
Electrochemical Capacitors....................................................... 63
K. Kierzek, G. Gryglewicz and J. Machnikowski
v
vi
General Properties of Ionic Liquids as Electrolytes

for Carbon-Based Double Layer Capacitors.................................... 73
A. Lewandowski and M. Galinski
Chapter 2:
Carbon Materials for Gas Diffusion Electrodes, Metal Air
Cells and Batteries.................................................................................. 85
Subject Overview............................................................................. 87
New Concept for the Metal-Air Batteries Using Composites:

Conducting Polymers/Expanded Graphite as Catalysts................... 89
V. Z. Barsukov, V. G. Khomenko, A. S. Katashinskii
and T. I. Motronyuk
Mechanically Rechargeable Magnesium-Air Cells with

NaCl-Electrolyte ............................................................................ 105
A. Kaisheva and I. Iliev
Application of Carbon-Based Materials in Metal-Air

Batteries: Research, Development, Commercialization ................ 117
A. Kaisheva and I. Iliev
Metal – Air Batteries with Carbonaceous Air Electrodes and

Nonmetallic Catalysts .................................................................... 137
N. Korovin
Chapter 3:
Carbon Anodes for Lithium-Ion Batteries............................................ 151
Subject Overview........................................................................... 153
Carbonaceous Materials for Batteries............................................ 157

T. Takamura and R. J. Brodd
Anode-Electrolyte Reactions in Li Batteries:

The Differences Between Graphitic and Metallic Anodes ............ 171
H. J. Santner, K. C. Möller, W. Kohs, C. Veit, E. Lanzer,
A. Trifonova, M. R. Wagner, P. Raimann, C. Korepp, J. O.
Besenhard and M. Winter
vii
Performance of Novel Types of Carbonaceous Materials in

the Anodes of CLAiO’s Lithium-Ion Battery Systems.................. 189
M. Walkowiak, K. Knofczynski, D. Waszak, M. Kopczyk,
M. Rusinek and J. Machnikowski
Why Graphite Electrodes Fail in PC Solutions: An

Insight from Morphological Studies .............................................. 197
D. Aurbach, M. Koltypin, H. Teller and Y. S. Cohen
New Developments in the Advanced Graphite for

Lithium-Ion Batteries..................................................................... 213
F.-X. Henry, I. V. Barsukov, J. E. Doninger, S. Anderson,
P. R. Booth, P. L. Zaleski, R. J. Girkant, D. J. Derwin,
M. A. Gallego, T. Huerta and G. Uribe
Mechanisms of Reversible and Irreversible Insertion in

Nanostructured Carbons Used for Li-Ion Batteries ....................... 231
F. Béguin, F. Chevallier, M. Letellier, C. Vix, C. Clinard,
J. N. Rouzaud and E. Frackowiak
Some Thermodynamics and Kinetics Aspects of the Graphite-

Lithium Negative Electrode for Lithium-Ion Batteries.................. 245
R. Yazami, A. Martinent and Y. Reynier
Characterization of Anodes Based on Various Carbonaceous

Materials for Application in Lithium-Ion Cells ............................. 259
A. N. Kozhevnikov, Y. A. Podalinski, O. R. Yakovleva,
V. S. Kotlyar, M. E. Petropavlovski, V. G. Smirnov and
V. V. Dzhurzha
A Carbon Composite for the Negative Electrode of Li-Ion

Batteries ......................................................................................... 269
A. V. Churikov, N. A. Gridina and N. V. Churikova
Electrochemical Intercalation of PF 6 and BF 4

into Single-Walled Carbon Nanotubes .......................................... 277
R. Yazami, I. V. Goncharova and V. N. Plakhotnik
viii
Surface Treated Natural Graphite as Anode Material

for High-Power Li-Ion Battery Applications................................. 283
J. Liu, D. Vissers, K. Amine, I. B. Barsukov and J. E. Doninger
Chapter 4 :
Emerging Metal/Carbon Composite Anodes for Next
Generation Lithium-Ion Batteries ........................................................ 293
On The Theoretical Prerequisites for Application of Novel

Materials in Promising Energy Systems ........................................ 297
V. Z. Barsukov and J. E. Doninger
Capabilities of Thin Tin Films as Negative Electrode Active

Materials for Lithium-Ion Batteries ............................................... 309
Y. O. Illin, V. Z. Barsukov and V. S. Tverdokhleb
Composite Anode Materials for High Energy Density

J. S. Gnanaraj, M. K. Gulbinska, J. F. DiCarlo, I. V. Barsukov,
N. Holt, V. Z. Barsukov and J. E. Doninger
Electrochemical Activity of Carbons Modified by d-Metal

Complexes with Ethanolamines..................................................... 333
L. G. Reiter, V. A. Potaskalov, A. A. Andriiko,
V. S. Kublanovsky, M. A. Chmilenko, Y. K. Pirskiy,
V. I. Lisin and S. M. Chmilenko
Metal-Graphite Composites as Materials for Electrodes of

L. Matzui, M. Semen’ko, M. Babich and L. Kapitanchuk
Electrochemical Performance of Ni/Cu-Metallized & Carbon-

Coated Graphites for Lithium Batteries......................................... 357
C. S. Johnson, K. Lauzze, N. Kanakaris, A. Kahaian,
M. M. Thackeray, K. Amine, G. Sandí-Tapia, S. A. Hackney
and R. O. Rigney
ix
Chapter 5:
New Nano- Through Macro-Carbons for Energy Systems:
Synthesis, Modeling, Characterization................................................. 377
Stabilization of Graphite Nitrate via Co-intercalation of Organic

Compounds .................................................................................... 381
M. V. Savoskin, A. P. Yaroshenko, R. D. Mysyk
and G. E. Whyman
Electrochemical Stability of Natural, Thermally Exfoliated and

Modified Forms of Graphite towards Electrochemical
Oxidation ....................................................................................... 389
I. O. Kovalenko, M. L. Barsukov, V. Z. Barsukov
and Y. V. Myshko
Low Temperature Synthesis of Graphite from Iron

Carbide .......................................................................................... 399
S. Dimovski, A. Nikitin, H. Ye and Y. Gogotsi
High Resolution Transmission Electron Microscopy Image

Analysis of Disordered Carbons Used for Electrochemical
Storage of Energy .......................................................................... 411
J.-N. Rouzaud, C. Clinard, F. Chevallier, A. Thery and F. Béguin
Electrolytes of Carbamide-Chloride Melts at Inert

Electrodes ...................................................................................... 425
S. A. Kochetova and N. Kh. Tumanova
Graphite Intercalation as a Way to Carbon-Carbon

Composites and Carbon Nanoscrolls ............................................. 433
M. V. Savoskin, A. P. Yaroshenko, V. N. Mochalin,
N. I. Lazareva and T. E. Konstantinova
Chapter 6:
Carbons in the Cathodes of Lithium-Ion Batteries; Alternative
Forms of MnO2, Cathode/Carbon Modeling ........................................ 441
x
Diagnostic Evaluation of Power Fade Phenomena and

Calendar Life Reduction in High-Power Lithium-Ion
Batteries ......................................................................................... 445
R. Kostecki and F. McLarnon
Modeling of Electrochemical Processes in the Electrodes

Based on Solid Active Reagents and Conductive Carbon
Additives........................................................................................ 453
V. V. Matveyev
On the Optimal Design of Amorphous Mangaense Oxide

For Applications in Power Sources................................................ 473
S. A. Kirillov, T. V. Lesnichaya, N. M. Visloguzova,
S. A. Khainakov, O. I. Pendelyuk, D. I. Dzanashvili,
T. A. Marsagishvili, V. Z. Barsukov, V. G. Khomenko,
A. V. Tkachenko, and S. I. Chernukhin
Investigation of Cathodic Materials Based on Different

Types of MnO2/Carbon.................................................................. 481
I. S. Makyeyeva, N. D. Ivanova, and G. V. Sokolsky
Investigation of Thin-Film Electrode Materials as

Cathodic Actives for Power Sources ............................................. 487
N. E. Vlasenko, N. D. Ivanova, E. I. Boldyrev, and
O. A. Stadnik
Synthesis of Mixed Oxides using Polybasic Carboxylic

Hydroxy-and Amino-Acid Routes: Problems and
Prospects ........................................................................................ 495
S. A. Kirillov, I. V. Romanova, and I. A. Farbun
Improved Electrochemical Properties of Surface-Coated

Li(Ni,Co,Mn)O2 Cathode Material for Li Secondary
Batteries ......................................................................................... 505
S. H. Kang and K. Amine
Index ....................................................................................................... 513

PREFACE
Carbonaceous materials play a fundamental role in electrochemical

energy storage systems. Carbon in the structural form of graphite is widely
used as the active material in lithium-ion batteries; it is abundant, and
environmentally friendly. Carbon is also used to conduct and distribute
charge effectively throughout composite electrodes of supercapacitors,
batteries and fuel cells. The electronic conductive pathways are critical to
delivering and extracting current out of the device. However, many
challenges and the understanding of the role of carbon and its stability and
efficiency in charge storage applications still exists. This NATO-ARW
volume contains a diverse collection of papers addressing the role of carbon
in some key electrochemical systems, both conventional and emerging.
These papers discuss the latest issues associated with development,
synthesis, characterization and use of new advanced carbonaceous materials
for electrochemical energy storage. Such systems include: metal-air primary
and rechargeable batteries, fuel cells, supercapacitors, cathodes and anodes
of lithium-ion and lithium polymer rechargeable batteries, as well as
nanocarbon materials of the future.
In the present volume, these papers originate from most of the
lectures given at the NATO-Carbon Advanced Research Workshop and
Conference (NATO-CARWC). Forty-one papers and six chapters are
presented in this book. The order of papers in each chapter in this volume
followed the order of lectures during the conference. This workshop and
conference featured papers from leading researchers in the field of
supercapacitors and batteries as well as papers from scientists and engineers
from the former Soviet Union (Russia and Ukraine) and other Eastern
European countries, such as Bulgaria and Poland. The workshop and
conference was held at Argonne National Laboratory in Argonne, Illinois
from October 20 to 23, 2003. A total number of registrants of approximately
90 were in attendance. Of these, approximately 21 Ukrainian and Russian
delegates were present. Roughly 16 other presenters were from NATO and
Mediterranean countries, and an additional 10 from the United States of
America. The remainder participants came from national laboratories,
government agencies, academia, and industry around the United States
whom are interested in energy storage research and technologies.
We thank all the presenters who contributed to this book and
appreciate their time spent in preparing their presentations for the conference
and manuscripts for this volume. We acknowledge the National University
of Technologies and Design (Kiev, Ukraine), Center for Research on
Divided Materials (CRDM-Orleans, France), and Poznan University of
xi
xii
Technology (Poland) for their participation and assistance in identifying

speakers from the Ukraine and Russia. We also thank all other registrants for
coming and making the workshop and conference a success.
Finally we acknowledge the NATO Science Committee, and are
grateful to the Civilian Research and Development Foundation (CRDF),
which provided funding in the form of individual travel grants for the invited
scientists and engineers from the Ukraine and Russia. In addition we would
like to thank the NATO-Science for Peace Program, Superior Graphite Co.,
Argonne National Laboratory, the National Renewable Energy Laboratory,
Lawrence Berkeley National Laboratory, Illinois Institute of Technology,
Gas Technology Institute and Broddarp of Nevada, Inc. for financial
sponsorship assistance.
We would like to specifically acknowledge logistical support from
the following individuals: Laurie Carbaugh and Joan Brunsvold (Argonne
National Laboratory), Maritza Gallego, Gabriela Uribe and Peter R. Carney
(Superior Graphite Co.), Dr. Chris De Wispelaere (Science for Peace
program, NATO), Dr. Fausto Pedrazzini (ARW Program, NATO), Dr.
Michael Onischak and Michael Romanco (Gas Technology Institute),
Professor Jai Prakash and Dr. Said Al-Hallaj (Illinois Institute of
Technology), Ross Campbell (Next Steps to Market Program, U.S. Civilian
Research and Development Foundation), Chris Hilton (Travel Grants
Program, U.S. Civilian Research and Development Foundation).
Last but not least, the editors wish to express gratitude to those loved
members of our families and friends, who inspired and supported us during
preparation of this volume for press.
Igor V. Barsukov
Christopher S. Johnson
Joseph E. Doninger
Vyacheslav Z. Barsukov
ARGONNE NATIONAL LABORATORY
9700 SOUTH CASS AVENUE, ARGONNE, ILLINOIS 60439-4832
OFFICE OF THE DIRECTOR (630) 252-2481 / FAX (630) 252-7923
October 20, 2003
Dear Participants,
On behalf of Argonne National Laboratory, I would like to welcome you to Argonne

National Laboratory and the NATO Carbon Advanced Research Workshop and
Conference. This event, which brings together innovative scientists and engineers
from many organizations and nations, is part of the NATO Science for Peace
program initiative to encourage fruitful collaborations between former Warsaw pact
countries and the United States and other Western nations. International
collaborations like these have become especially important in today’s climate, and I
am pleased that so many leading experts from government, industry, and academia
are in attendance.
I invite you to take advantage of the tours of the laboratory. As the nation’s first
national laboratory, Argonne has a unique mix of history and leading-edge
technology throughout its 1,500-acre site. Events will come to a close with a private
tour and dinner at the Museum of Science and Industry, one of Chicago’s great
treasures.
I am confident that this meeting will be a stimulating and productive one. As

scientists and engineers dedicated to peaceful collaboration, I trust your vision and
efforts will carry electrochemical technology to new heights.
Sincerely,
Hermann A. Grunder
Director of Argonne National Laboratory
OPERATED by THE UNIVERSITY of CHICAGO for THE UNITED

STATES DEPARTMENT of ENERGY
xiii
PHOTOGRAPHS AND COMMENTS
Group Photograph, Argonne National Laboratory, Argonne, IL USA (Courtesy C. Johnson)
Technical discussions during the poster ses-

sion. Left to right: Professor R. Yazami
(CALTEC / CNRS, USA-France), Dr. I.
Goncharova (Dniepropetrovsk National
Univ. of Railway Transport, Ukraine) and
Professor M. Winter (Graz Technical Univ.,
Austria). Courtesy M. Gallego. o
Eastern European participants of the NATO-

ARW during the downtown Chicago city
tour. Courtesy Dr. V. Matveev. p
A satellite meeting during the NATO-ARW. nPlanning of what later resulted in a start of a
high profile U.S. Department of Energy Initiative for Proliferation Prevention project
between the Argonne National Laboratory, Superior Graphite Co. (Chicago, IL, USA) and
three ex-weapons institutes, and other organizations in Ukraine (Headquarters SGC,
downtown Chicago).
xiv
xv
PHOTOGRAPHS AND COMMENTS

m At the renowned Chicago’s
Museum of Science and Industry,
posing in the Space Exhibit are
from left: Prof. J. Machnikowski
(Wroclaw Univ. of Technology,
Poland), M. Gallego and J. Chavez
(Superior Graphite, USA), Y. Il’in
(Kiev National University of
Technology & Design, Ukraine),
Prof. F. Beguin (CNRS, France),
Dr. E. Frackowiak (Inst. of
Chemistry & Technical Electro-
chemistry, Poland) and Dr. Neal
Courtesy Prof. Doninger. White (Sion Power Research, Great
Britain).
m Conducting a technical tour of
Superior Graphite’s Peter R.
Carney Technology Center is one
of the inventors of expanded
graphite for batteries, Peter Zaleski,
Lab Director (center); Listening
left to right are: Prof. Y. Maletin
(National Technical Univ. of
Ukraine “KPI”), Dr. N. White
(Sion Power Research, Great
Britain), Dr. A. Hull (Argonne
National Lab, USA), Mr. M.
Walkowiak (CLAiO, Poland), and
Mr. V. Khomenko (Kiev National
University of Technology &
Design, Ukraine).
A honorary NATO-CARWC award for Dr. N. Vlasenko (left) of the Institute of

significant contribution to the science of General & Inorganic Chemistry, Ukraine and
carbon for power sources is being given by Dr. V. Kotlyar of Research Institute
Dr. I. Barsukov (workshop co-organizer) to “Istochnik”, Russia observe a high tempe-
Prof. D. Aurbach (Bar-Ilan Univ., Israel). rature fuel cell, as shown by Dr. R. Remick
Sitting are other awardees: S. Dimovski during a tour to Gas Technology Institute (a
(Drexel Univ., USA (left), and Prof. R. leading fuel cell technology organization in
Yazami (Caltech/CNRS – USA/France). greater Chicago area). Courtesy V. Matveev.
ORGANIZATION
(i)USA Director:
Dr. Joseph E. Doninger– President, Dontech Global, Inc.
Dontech Global, Inc.
427 East Deerpath Rd., Lake Forest, IL 60045, USA
(224) 436-4835 (phone); +1 (847) 234-4835 (fax);
[email protected]
(ii) USA Co-director:

Dr. Christopher S. Johnson – Staff Chemist
Electrochemical Technology Program
Chemical Engineering Division
Argonne National Laboratory
9700 S. Cass Ave., Argonne, IL 60565, USA
(630)-252 4787 (phone); (630)-252 4176 (fax);
[email protected]
(iii) Ukraine Co-director:

Prof. Dr. Vyacheslav Z. Barsukov – Head of Chemistry Department
National University of Technologies and Design, Department of
Chemistry, 2, Nemirovich-Danchenko street, Kiev, 252011, Ukraine
+ 38 (044) 291-2102 (phone); + 38 (044) 290-1603 (fax);
[email protected]
Local Organizing Committee
(i) Dr. Igor Barsukov Superior Graphite Co., Peter R. Carney

Technology Center, 201 West 36th Street, Chicago, IL,
60632, USA, Phone: (773) 890-4117, Fax: (773) 890-4121,
E-mail: [email protected]
(ii) Dr. Christopher S. Johnson, Chair (Argonne National
Laboratory, ANL)
(iii) Dr. Joe Doninger (Superior Graphite Co.)
(iv) Mr. Gary Henriksen (ANL), phone (630) 252-4591
(v) Dr. James Miller (ANL), phone (630) 252-4537
xvii
LIST OF PARTICIPANTS
Oleksandr Andriiko Chemical Engineering Faculty, Kiev National

Technical University, Polytechnik Institute (NTUU
KPI), 7 Peremogy Prospekt, 03056 Kiev, Ukraine,
[email protected]
Terrill Atwater Army Power, U.S. Army CECOM, AMSEL-RD-C2-
AP-B, Fort Monmouth, NJ 07703, USA,
[email protected]
Doron Aurbach Department of Chemistry, Bar-Ilan University
Remat ban 53900, Israel, [email protected]
James Barnes U.S. Department of Energy, EE-2G, 1000
Independence Avenue, SW, Washington, DC 20585,
USA, [email protected]
Igor Barsukov Department of Research & Development, Superior
Graphite Company, 4201 West 36th Street, Chicago,
IL 60629, USA, [email protected]
Vyacheslav Barsukov Electrochemical Power Engineering & Chemistry
Kiev National University of Technologies an Design
Nemirovich-Danchenko str. 2 02011 Kiev, Ukraine
[email protected]
James Beckwith Power Sources Center, 3206 Kinross Circle, Herndon,
VA 20171, USA, [email protected]
Francois Béguin CRMD, CNRS – University of Orléans
1b, rue de la Férollerie, 45071 Orléans, France
[email protected]
Alexey Belyakov Management Department, ELIT Electrochemical
Power Sources Co., 40, Rpospect Leninscogo
Komsomola, 305026 Kursk, Russia
[email protected]
Ira Bloom Chemical Engineering Division, Argonne National
Laboratory, Bldg. 205, 9700 South Cass Avenue,
Argonne, IL 60439, USA
[email protected]
Peter Booth Technology Department, Superior Graphite Company
10 South Riverside Plaza, Chicago. IL 60606 USA
[email protected]
Ralph Brodd Broddarp of Nevada, Inc., 2161 Fountain Springs
Drive, Henderson, NV 89074
[email protected]
Ross Campbell Next Steps to Market, U.S. Civilian Research &
Development Foundation (CRDF), 1530 Wilson
Boulevard, Arlington, VA 22209, USA
[email protected]
xviii
xix
Peter Carney Executive Department, Superior Graphite Company

10 South Riverside Plaza, Chicago, IL 60606, USA
Alexi Churikov Chemistry Department, Saratov State University
83, Astrakhanskaya, 410012 Saratov, Russia
[email protected]
Svetlana Dimovski Department of Materials Science and Engineering
Drexel University, Lewbow 344, 3141 Chestnut
Street, Philadelphia, PA 19104, USA
[email protected]
Joseph Doninger Dontech Global, Inc., 427 East Deerpath Rd., Lake
Forest, IL 60045, USA
[email protected]
Charles Drews EMS, Inc., 165 King Street, Elk Grove Village, IL
60007, USA, [email protected]
Harvey Drucker Energy & Environmental Science & Technology
Argonne National Laboratory, Bldg. 202
9700 South Cass Avenue, Argonne, IL 60439, USA
[email protected]
Elzbieta Frackowiak Institute of Chemistry & Technical Electrochemistry
Poznan University of Technology, Piotrowo 3, 60-965
Poznan, Poland, [email protected]
Arkadiy Furman EMS, Inc., 165 King Street, Elk Grove Village, IL
Maritza Gallego Department of Research & Development, Superior
Graphite Company, 4201 West 36th Street, Chicago,
Rex Gerald Chemical Engineering Division, Argonne National
Argonne, IL 60439, USA, [email protected]
Steve Gerts R&D Department, Amtek Research International
250 North Hansard Ave., Lebanon, OR 97355, USA
[email protected]
Yury Gogotsi Department of Materials Science & Engineering
Drexel University, Lebow 344, 3141 Chestnut Street
Philadelphia, PA 19104, USA, [email protected]
Irina Goncharova Department of Chemistry & Engineering Ecology
Dnieproppetrovsk National University of Railway
Transport, 2 Academika Lazaryana St., 49010
Dniepropetrovsk, Ukraine
[email protected]
Diane Graziano Chemical Engineering Division, Argonne National
Gary Henriksen Chemical Engineering Division, Argonne National
xx
Francois-Xavier Henry Department of Research & Development,

th
Superior
Graphite Company, 4201 West 36 Street, Chicago,
Amy Hull Energy Technology Division, Argonne National
Ilia Illiev Electrochemistry of Biocatalytic & Metal-Air
Systems, Central Laboratory of Electrochemical
Power Sources, Acad. G. Bonchev Str., bl. 10, 1113
Sofia, Bulgaria, [email protected]
Yevgen Ilyin Electrochemical Energetics & Chemistry, Kiev
National University of Technologies & Design
Nemirovich-Danchenko str., 2, 02001 Kiev, Ukraine
[email protected]
Andrew Jansen Chemical Engineering Division, Argonne National
Christopher Johnson Chemical Engineering Division, Argonne National
Donald Joyce Office of the Laboratory Director, Argonne National
Arthur Kahaian Chemical Engineering Division, Argonne National
Anastassia Kaisheva Electrochemistry of Biocatalytic & Metal-Air
Systems, Central Laboratory of Electrochemical
Power Sources, Acad. G. Bonchev str., bl. 170, 1113
Sofia, Bulgaria, [email protected]
Sun-Ho Kang Chemical Engineering Division, Argonne National
Volodimir Khomenko Electrochemical Power Engineering & Chemistry
Kiev national University of Technologies & Design,
Nemirovich-Danchenko Str., 2, 02011 Kiev, Ukraine
[email protected]
Jeom-Soo Kim Chemical Engineering Division, Argonne National
Sviatoslav Kirillov Institute for Sorption & Problems of Endoecology,
National Academy of Scienes, 13, Naumov St., 03164
Kiev, Ukraine, [email protected]
Arkadiy Klementov JME, Inc., 17210 Parkland Drive, Shaker Heights,
OH 44120, USA, [email protected]
xxi
Svitlana Kochetova N11-High Temperature Electrochemical Synthesis,

National Academy of Sciences, Institute of General &
Inorganic Chemistry, 32/34 Palladin Avenue, 03680
Kiev, Ukraine, [email protected]
Robert Kostecki Environmental Energy Technologies, Lawrence
Berkeley National Laboratory, 1 Cyclotron Road,
Berkeley, CA 94720, USA, [email protected]
Vladimir Kotlyar Reliability Department, Accumulator Research &
Design Institute “Istochnik”, 10, Dahl Street, 197376
Saint-Petersburg, Russia, [email protected]
Igor Kovalenko Ministry for Emergency Management of Chemistry,
Ukrainian Fire Safety Research Institute, 18,
Rybarskya Street, 01011 Kiev, Ukraine
[email protected]
Alexander N20, Acumulator Research & Design Institute
Kozhevniko “Istochnik”, 10, Dahl Street, 197376 Saint-Petersburg,
Russia, [email protected]
David Laughton Sales Department, Superior Graphite Company
10 South Riverside Plaza, Chicago, IL 60606, USA
[email protected]
Andrzej Lewandowski Faculty of Chemical Technology, Poznan University
of Technology, ul. Piotrowo 3, 60-965 Poznan,
Poland, [email protected]
David Lewis Chemical Engineering Division, Argonne National
Jun Liu Chemical Engineering Division, Argonne National
Laboratory, Bldg.205, 9700 South Cass Avenue,
Jacek Machnikowski Institute of Chemical & Technology of Petroleum and
Coal, Wroclaw University of Technology, Gdanska
7/9, 50 344 Wroclaw, Poland,
[email protected]
Irina Najteteva Electrochemical Power Engineering & Chemistry,
Kiev National University of Technologies & Design,
Nemirovich-Danchenco str., 2, 02011 Kiev, Ukraine,
[email protected]
Yurii Maletin Department of Physical Chemistry, National
Technical University of Ukraine “KPI”, 37 Prospect
Peremohy, 03056 Kiev, Ukraine, [email protected]
kpi.kiev.ua
Vadym Matveyev Scientific Research Laboratory of Storage-Battery
System, Ukrainian State University of Chemical
Engineering, 8, Gagarin Avenue, 49005
Dniepropetrovsk, Ukraine, [email protected]
xxii
Luidmila Matzui Department of Biophysics, Kyiv National

Shevchenko University Volodymyrska Str., 64, Kyiv,
Ukraine, matzui@mail/univ/kiev/ua
Dayal Meshri Advance Research Chemicals, Inc., 1110 West
Keystone Avenue, Catoosa, OK 74015, USA,
[email protected]
James Miller Chemical Engineering Division, Argonne National
Vadim Mochalin Chemistry of Molecular Complexes, National
Academy of Sciences, L.M. Litvinenko Institute of
Physical Organic & Coal Chemistry, R. Luxemburg
Str., 70, 83114 Donetsk, Ukraine,
[email protected]
Michael Onischak Department of Hydrogen & Distributed Energy, Gas
Technology Institute, 1700 South Mount Prospect
Road, Des Plaines, IL 60018, USA,
[email protected]
Oksana Pendelyuk National Academy of Sciences, Institute for Sorption
and Problems of Endeocology, 13, Naumov St.,
03164 Kiev, Ukraine, [email protected],ua
Iouri Prokofiev Nuclear Engineering Division, Argonne National
Encarnacion CRMD, CNRS – University of Orléans, 1B rue de la
Raymundo-Pinero Férollerie, 45071 Orleans, France, raymundo@cnrs-
orleans.fr
Michael Romanco Department of Hydrogen & Distributed Energy, Gas
Technology Institute, 1700 South Mount Prospect
Road, Des Plaine, IL 60018, USA,
[email protected]
Jean-Noel Rouzaud CRMD, CNRS – University of Orléans, 1B rue de la
Férollerie, 45071 Orleans, France, rouzand@cnrs-
orleans.fr
Giselle Sandi-Tapia Chemistry Division, Argonne National Laboratory,
Bldg. 200, 9700 South Cass Avenue, Argonne, IL
Michael Savoskin Chemistry of Molecular Complexes, National
Academy of Sciences, L.M. Livinenko Institute of
Physical Organic & Coal Chemistry, R. Luxemburg
Str., 70, 83114 Donetsk, Ukraine, m-
[email protected]
Sergiy Yu Sayenko Institute of Solid State Physics, National Science
Center – KIPT, Materials Science and Technologies,
1, Akadamicheskaya, 61108 Kharkov, Ukraine,
[email protected]
xxiii
Sergiy Sazhin Department of Lithium & Rechargeable

Technologies, Rayovac Corporation, 630 Forward
Drive, Madison, WI 53711-2497, USA,
[email protected]
Kathryn Striebel Environmental Energy Technologies Division,
Lawrence Berkeley National Laboratory, MS:
70R0108B, 1 Cyclotron Road, Berkeley, CA 94720,
Jeffrey Swoyer Research Department, Valence Technology, Inc., 301
Conestoga Way, Henderson, NV 89015, USA,
[email protected]
Michael Thackeray Chemical Engineering Division, Argonne National
Thanh Tran Sentech, Inc., 4733 Bethesda Avenue, Bethesda, MD
Gabriela Uribe Research & Development, Superior Graphite
Company, 4201 West 36th Street, Chicago, IL 60632,
Jack Vaughey Chemical Engineering Division, Argonne National
Donald Vissers Chemical Engineering Division, Argonne National
Natasha Vlasenko Photoelectrochemistry & Chemical Power Sources,
National Academy of Sciences, Institute of General &
Inorganic Chemistry, 32/34 Palladin Prospect, 03680
Kiev – 142, Ukraine, [email protected]
Mariusz Walkowiak Lithium Batteries Department, Central Laboratory of
Batteries & Cells, Ul. Forteczna 12, 61-362 Poznan,
Poland, [email protected]
Qingzheng Wang Chemical Engineering Division, Argonne National
Robert Waterhouse Department of Research & Development, Amtek
Research International, 250 North Hansard Avenue,
Lebanon,OR, 97355, USA, rwaterhouse@amtek-
research.com
Neal White Sion Power Research, Inc., Sion Power Corporation,
Tanfield Lea Industrial Park, Stanley, Durham DH9
9QF, United Kingdom,
[email protected]
Martin Winter Institute of Chemical Technology of Inorganic
Materials, Graz University of Technology,
Stremayrgasse 16, A-8010 Graz, Austria,
[email protected]
xxiv
Mathis Wissler Product Management, Superior Graphite Europe,

Kungsmatt 8, Ch 5643 Sins, Switzerland,
[email protected]
Rachid Yazami Department of Materials Science, CRNS/Caltech,
M/C 183-78, Pasadena, CA 91125, USA,
[email protected]
Peter Zaleski Laboratory Department, Superior Graphite Company,
4201 West 36th Street, Chicago, IL 60632, USA,
[email protected]
CHAPTER 1:
NEW CARBON MATERIALS FOR

SUPERCAPACITORS
Chapter 1: Subject Overview
A classic definition of electrochemical ultracapacitors or super-
capacitors summarizes them as devices, which store electrical energy via
charge in the electrical double layer, mainly by electrostatic forces, without
phase transformation in the electrode materials. Most commercially available
capacitors consist of two high surface area carbon electrodes with graphitic
or soot-like material as electrical conductivity enhancement additives.
Chapter 1 of this volume contains seven papers with overview presentations,
and development reports, as related to new carbon materials for this
emerging segment of the energy market.
In the first paper by E. Frackowiak et al., a French-Polish team of
authors describes properties of new materials produced by KOH activation
of microporous carbon precursors, as well as carbons obtained through the
so-called template synthesis. The authors also devise a carbon-carbon
composite, which has been synthesized via a process of carbonization of
carbon nanotubes with a polyacrylonitrile conducting polymer. It is claimed
that a synergistic effect has been observed with the above composite, while
its individual ingredients show instead negligible capacitance values. An
attempt made by the authors to explain the reported synergistic capacitance
enhancement effect by, exclusively, nitrogen-based surface groups on
carbon, is an interesting, and, perhaps, debatable concept. It may likely
stimulate further discussions and investigative research on the subject.
The second paper, by A. Belyakov, comes from the Russian
supercapacitor industry. This applied work highlights modern industrial
approaches used for boosting the power output characteristics of aqueous
commercial supercapacitors. The author discusses two supercapacitor
chemistries: carbon-carbon and NiOx-carbon. A key emphasis is given to the
role of conductive additives, which constitute up to 40wt% of electrode
composition. Several classes of conductive diluents were investigated, e.g.
Ni powder, colloidal carbon black, expanded and boron-doped flake
graphite. The author concludes that for the carbon-carbon type of
supercapacitors, the most preferable conductive additive should be carbon
material, but the exact type of carbon is of a secondary importance.
However, in the case of asymmetric NiOx-carbon capacitor, boron-doped
graphite showed the biggest promise, due to it being prone to oxidation.
In the third paper by French and Ukrainian scientists (Khomenko et
al.), the authors focus on high performance a-MnO2/carbon nanotube
composites as pseudo-capacitor materials. Somewhat surprisingly, this paper
teaches to use carbon nanotubes for the role of conductive additives, thus
suggesting an alternative to the carbon blacks and graphite materials – low
cost, widely accepted conductive diluents, which are typically used in
todays’ supercapacitors. The electrochemical devices used in the report are
full symmetric and optimized asymmetric systems, and are discussed here
3
4
for the first time. The editors tend to agree with author’s conclusions that
carbon nanotubes may become quite efficient conductive diluents at low
loadings in the electrode matrixes, due to their unique structure, especially
after the costs of carbon nanotubes are brought in line with alternative
conductive additives.
The fourth contribution (Khomenko et al.) is a product of an
extensive international study as presented by an alliance of academia from
France, Poland and Ukraine. In this interesting work, the authors furnish
carbon nanotubes and/or expanded graphite as a backbone, highly
conductive electrode network, onto which the following conducting
polymers have been deposited: polyaniline, polypyrrole, and poly(3,4-
ethylenedioxythiophene). In so doing, these authors propose an interesting
approach to solving a major drawback of modern pseudocapacitor
technologies, e.g. stability of their performance characteristics over time.
Ukraine’s Y. Maletin et al. presented a comprehensive overview
describing state of the art as well as future development trends in
supercapacitors, as the fifth paper in this chapter. The authors establish key
performance bars for supecapacitors; upon meeting those, supercapacitors
may start to compete with batteries. Also, this paper highlights so-called
hybrid applications where supercapacitors complement operation of batteries
and/or fuel cells. Optimization of supercapacitor performance through
varying electrode thickness is contemplated in length.
In the sixth paper of this chapter, Kierzek et al., mainly focus on
modeling of pore formation vs surface area growth phenomena upon
activation of coal and pitch-derived carbon precursors. These authors briefly
touch on other precursor carbons as well. The properties of newly
synthesized materials are being looked at from the point of view of their
application as active materials in the supercapacitor electrodes. Editors
thought this work by the Institute of Chemistry and Technology of
Petroleum and Coal in Poland, could be of genuine interest to the practical
developers of carbon materials for the supercapacitor industry.
In the last paper, A. Lewandowski et al. of Poland, examines the role
of ionic liquids as new electrolytes for carbon-based supercapacitors.
Although not directly addressing the role of new carbon materials (the area
of major focus of this book), this interesting theoretical work seeks to
optimize electrolyte media, which is in contact with carbon electrodes.
We hope you will enjoy reading this chapter.
NOVEL CARBONACEOUS MATERIALS FOR
APPLICATION IN THE ELECTROCHEMICAL
SUPERCAPACITORS
1
E. Frackowiak*, 2J. Machnikowski, 3F. Béguin
1
Poznan University of Technology, 60-965 Poznan, ul. Piotrowo 3, Poland
2
Wroclaw University of Technology, 50-344 Wroclaw, ul. Gdanska 7/9, Poland
3
CRMD, CNRS-University, 45071 Orléans, 1B rue de la Férollerie, France
Abstract
Three types of novel carbon materials, such as KOH-activated

microporous carbons, highly ordered carbons obtained by a template
technique, and a special carbon/carbon composite based on carbon nanotubes
with carbonized polyacrylonitrile have been used as electrodes for
supercapacitors in aqueous and organic electrolytes. The various carbon
materials used allow their capacitance properties to be correlated with the
total surface area, micro and meso pore volume, elemental composition.
Especially, the role of mesopores in charge propagation was strongly
underlined. The capacitance values widely range from 100 to 300 F/g
depending on the carbon microtexture but also on the heteroatom (nitrogen)
content responsible for pseudocapacitance effects. Comparison of these three
classes of carbon materials gives a basic knowledge for optimization of
supercapacitor electrode materials depending on electrolytic solutions used.
1. INTRODUCTION
Electrochemical capacitors are power storage devices, whose
performance is based on the charge accumulation from an electrolytic
solution through electrostatic attraction by polarized electrodes. The
capacitance of this system is directly proportional to the electrode surface,
therefore carbons are very efficient for this application because of various
possibilities of their modification and creation of a controlled pore size
distribution [1-3]. The electrostatic attraction of ions takes place mainly in
micropores, however, the presence of mesopores is necessary for efficient
*
Corresponding author. E-mail: [email protected]
5
I.V. Barsukov et al. (eds.), New Carbon Based Materials for Electrochemical Energy Storage Systems, 5–20.
© 2006 Springer. Printed in the Netherlands.
6
charge propagation. It is noteworthy that the electrochemically active surface

is not the same as the physically described surface area from nitrogen
adsorption data, and is determined by the ability of ions to be trapped
(function of the size of ions, wetability, conductivity of carbon, etc.). Hence,
in general, well-balanced micro/meso porosity is an important criterion for
the selection of the carbon material.
In supercapacitors, apart from the electrostatic attraction of ions in
the electrode/electrolyte interface, which is strongly affected by the
electrochemically available surface area, pseudocapacitance effects
connected with faradaic reactions take place. Pseudocapacitance may be
realized through carbon modification by conducting polymers [4-7],
transition metal oxides [8-10] and special doping via the presence of
heteroatoms, e.g. oxygen and/or nitrogen [11, 12].
In this article, we will pay a special attention to three novel types of
carbonaceous materials:
1) Carbons activated with KOH, featuring a highly developed

surface area;
2) highly ordered carbons prepared from silica templates;
3) carbon/carbon composite based on multi-walled carbon
nanotubes (MWNTs) and polyacrylonitrile (PAN).
1.1. Chemical Activation by KOH

The first type of materials, i.e. KOH activated carbons constitute an
interesting class of capacitor electrodes due to their highly developed surface
area of the order of 3000 m2/g. Especially, inexpensive natural precursors are
well adapted for this process. The activation process is strongly affected by
the C:KOH ratio, temperature and time. The optimal ratio seems to be 4:1
and the temperature for activation ca. 800°C. The total activation process is
quite complicated and proceeds via different pathways and by-products. The
following reactions can be considered:
2KOH ĺ K2O + H2O (1)
C + H2O ĺ CO + H2 (2)
CO + H2O ĺ CO2 + H2 (3)
K2O + H2 ĺ 2K + H2O (4)
K2O + C ĺ 2K + CO (5)
7
Under severe conditions (above 700qC), a potassium vapor is formed. It

plays a special role in the activation of carbonaceous materials, easily
penetrating in the graphitic domains that form cage-like micropores. The
efficient development of micropores, which often gives a few-fold increase
of the total specific surface area, is very useful for the application of these
materials in supercapacitors [13-14].
Capacitance of this type of material reaches extremely high values
over 300 F/g. Our target is to correlate it with the total specific surface area,
pore size distribution, particle size and elemental composition of carbon.
1.2. Carbons Produced by the Template Technique

The second class of materials, which we will consider herein are
carbons with a highly ordered porosity prepared by a template technique [15-
18]. The pores are characterized by a well-defined size determined by the
wall thickness of the silica substrate used as substrate for carbon infiltration.
They can be also interconnected, that is very useful for the charge diffusion
in the electrodes. Figure 1 presents the general principle of the carbon
preparation by a template technique, where the silica matrix can be, for
example, MCM-48 or SBA-15.
Infiltration Dissol. SiO2 2
carbon
Porosity Silica matrix SiO2/C Carbon
Figure 1. Schematic representation of carbon elaboration by a template technique.
(a) (b)
Figure 2. Two examples of silica matrices: (a) MCM-48 and (b) SBA-15.
8
Schematic representations of organization of the porous network in

MCM-48 and SBA-15 are given in Fig. 2 [16]. The mesoporous silica
material referred to as MCM-48 displays a cubic structure, which consists of
two independent and intricately interwoven networks of mesoporous
channels. The structure of the material named SBA-15 is composed of a
hexagonal arrangement of 1-D silica rods forming cylindrical pores
interconnected through smaller pores, which are randomly located
perpendicular to the 1-D channels [15]. Both silica materials exhibit a highly
ordered porous structure, a high specific surface area and a narrow pore size
distribution centered at 2.6 nm and 5.2 nm for MCM-48 and SBA-15,
respectively.
The pores of the silica template can be filled by carbon from a gas or
a liquid phase. One may consider an insertion of pyrolytic carbon from the
thermal decomposition of propylene or by an aqueous solution of sucrose,
which after elimination of water requires a carbonization step at 900°C. The
carbon infiltration is followed by the dissolution of silica by HF. The main
attribute of template carbons is their well sized pores defined by the wall
thickness of the silica matrix. Application of such highly ordered materials
allows an exact screening of pores adapted for efficient charging of the
electrical double layer. The electrochemical performance of capacitor
electrodes prepared from the various template carbons have been determined
and are tentatively correlated with their structural and microtextural
characteristics.
1.3. Carbon/Carbon Composites

Another interesting type of novel carbons applicable for
supercapacitors, consists of a carbon/carbon composite using nanotubes as a
perfect backbone for carbonized polyacrylonitrile. Multiwalled carbon
nanotubes (MWNTs), due to their entanglement form an interconnected
network of open mesopores, which makes them optimal for assuring good
mechanical properties of the electrodes while allowing an easy diffusion of
ions.
Due to their moderate specific surface area, carbon nanotubes alone
demonstrate small capacitance values. However, the presence of heteroatoms
can be a source of pseudocapacitance effects. It has been already proven that
oxygenated functional groups can significantly enhance the capacitance
values through redox reactions [11]. Lately, it was discovered that nitrogen,
which is present in carbon affects also the capacitance properties [12].
In this work, MWNTs will be mixed with polyacrylonitrile (PAN)
and the MWNTs/PAN composite will be carbonized, giving rise to a new
C/C composite with interesting capacitance properties. It is striking that the
composite components alone give negligible capacitance values below 20
9
F/g, whereas the composite in the optimal proportions can supply quite
interesting capacitance behavior explained only by a positive effect of
nitrogen.
2. EXPERIMENTAL
2.1. Preparation and Characteristics of the Carbon Materials

2.1.1. KOH Activated Carbons
The following natural precursors have been selected for KOH

activation: coal (C), coal semi-coke (CS), pitch semi-coke (PS) and pitch
mesophase (PM). An industrial activated carbon (AC) was also used.
Activation was performed at 800qC in KOH with 4:1 (C:KOH) weight ratio,
for 5 hours, followed by a careful washing of the samples with 10% HCl and
distilled water. The activation process supplied highly microporous carbons
with BET specific surface areas from 1900 to 3150 m2/g. The BET surface
area together with the micro and the total pore volume of the KOH-activated
carbons are presented in Table 1. The mean micropore width calculated from
the Dubinin equation is designed as LD.
Table 1. Porosity parameters of the KOH activated carbons

(A-C means activated carbon from coal C etc.)
Sample SBET VDR Vtotal LD
m2 g-1 cm3 g-1 cm3 g-1 nm
A-C 3150 0.951 1.612 1.39
A-CS 3190 0.936 1.446 1.36
A-PM 2660 0.839 1.209 1.37
A-PS 2750 0.859 1.227 1.34
A-AC 1900 0.609 1.051 1.29
2.1.2. Template Carbons
Two types of organized mesoporous silica, MCM-48 and SBA-15,

which display different crystallographic structure, size and shape of the
pores, were selected as templates for carbons.
Pyrolytic carbon was inserted into the pores of these silica matrices
by chemical vapor infiltration (CVI). The silica template was contacted with
a flow of propylene Pr, (2.5 vol%) diluted in argon at 750qC during 15
hours. A quite uniform pore filling can be obtained by CVI. At the end, the
carbon represents about 50 wt% of the C/SiO2 material. Since the deposition
10
time was the same for the two silica precursors, the fraction of the pore
volume filled with carbon is function of the template. The C/SiO2 material
was then treated with hydrofluoric acid (HF) to remove the silica template
and the carbons obtained are designed CPr48 and CPr15.
The second type of template carbon was formed by liquid
impregnation of the pores of MCM-48 and SBA-15 silica by a sucrose (S)
solution, followed by a carbonization treatment at 900°C under vacuum.
After a second impregnation step, the amount of carbon introduced in the
silica matrix represents 33wt% which is not far of the theoretical amount
expected, taking into account the pore volume of the silica matrix and the
concentration of the sucrose solution, i.e. 36wt% of the total mass of the
C/SiO2 composite. The silica template is eliminated by dissolution in HF to
recover the carbon material denominated here as CS48 and CS15.
Silica matrices MCM-48 and SBA-15, as well as carbons obtained
by replica method are characterized by nitrogen adsorption at 77K and X-ray
diffraction (Table 2).
Table 2. Characteristics of the silica templates and the corresponding carbon materials;
a: unit cell parameter; SBET: specific surface area;Vp: total pore volume (at P/Po=0.95);
Pore size determined according to the BJH method - Maximum value of the BJH pore size
distribution peak calculated from the adsorption branch of the N2 isotherm.
Materials a (nm) SBET (m²/g) Vp (cm3/g) Pore size

(nm)
MCM-48 8.6 1330 1.07 2.6
CPr48 8.4 850 0.56 -
CS48 7.6 2000 1.10 2.7
SBA-15 11.3 807 0.97 5.2
CPr15 10.5 713 0.60 3.4
CS15 9.4 1470 1.11 3.1
2.1.3. CNTs/PAN Composites
The high purity carbon nanotubes (CNTs) used in this study were
obtained by decomposition of acetylene over a powdered CoxMg1-xO solid
solution catalyst [19]. Different proportions of CNTs from 15 to 70% and
polyacrylonitrile (PAN, Aldrich) have been mixed in an excess of acetone to
obtain a slurry. After evaporation of acetone, precursor electrodes were
formed by pressing the CNTs/PAN mixture at 1-2 tons/cm2. The C/C
composites were formed by carbonisation of the pellets at 700-900°C for 30-
420 min under nitrogen flow [20]. The optimal capacitance properties of the
composite were obtained for a mixture CNTs/PAN (30/70 wt%) treated at
700qC. Such C/C composite remains still quite rich in nitrogen (9 at% of N)
demonstrating that PAN is an efficient nitrogen carrier. On the other hand,
11
the noticeable amount of oxygen (4.1 at%) in the composite could be due to
oxygen addition on the dangling bonds when the C/C composite is exposed
to air after its formation.
During the pyrolysis of the CNTs/PAN mixture, the pellets keep
their original shape while becoming rigid, without any noticeable cracks or
defects. The carbonised C/C composites are quite compact, with a strong
bonding between CNTs and the carbon matrix. Hence, CNTs play a very
important role during the C/C composite formation, by creating a stable
network which prevents the dimensional changes and shrinkage. However,
when the CNTs content is less than 15%, the weight loss increases
dramatically and shrinkage appears. Such self-standing composite pellets
were used as electrodes for capacitance investigations.
2.2. Electrochemical Measurements

The capacitor electrodes were pellets formed by pressing a mixture
of the active carbon material (85wt%) with a PVDF binder (10wt%) and
acetylene black (5wt%). In the case of the C/C composite, PVDF and
acetylene black were not used because the composite forms self-standing
electrodes. The capacitance measurements were performed in 6 mol l-1 KOH,
1 mol l-1 H2SO4 and 1 mol l-1 TEABF4 in acetonitrile. The values of
capacitance were estimated by voltammetry (scan rate of potential from 2 to
10 mV/s) and galvanostatic charge-discharge cycling (VMP-Biologic,
France and ARBIN BT2000, USA). Impedance spectroscopy (AUTOLAB-
ECOCHEMIE BV) allows the capacitance (F/g) dependence versus
frequency (Hz), series resistance, time constant and charge propagation to be
evaluated. For some samples a careful analysis of leakage current, self-
discharge and capacitors cycleability was performed.
3. RESULTS AND DISCUSSION

3.1. Capacitance Properties of the KOH Activated Carbons
From three electrochemical techniques described in the previous

section, the most reliable capacitance results are obtained from galvanostatic
discharge, however each method supplies complementary information. A
typical galvanostatic charge-discharge characteristic for carbon A-PM is
shown in Figure 3. The curve presents a correct triangular shape without a
significant ohmic drop. The values of specific capacitance per mass of
carbon material and per surface area estimated by all three electrochemical
methods using 1 mol l-1 sulfuric acid solution are given in Table 3. All the
carbons present a very satisfactory capability of charge accumulation in the
12
electrical double layer with capacitance values ca. 300 F/g, especially for the
activated carbons from coal (A-C) and mesophase pitch (A-PM). The highest
capacitance value for A-C well correlates with a maximum total pore volume
of 1.6 cm3/g and its BET surface area of 3150 m2/g. However, a careful
comparison of the carbons characteristics (Table 1) with the capacitance
values (Table 3) shows that there is not a proportional relation between the
surface area or the pore volume and the electrochemical behavior. On the
other hand, from the low values of capacitance per surface area (7 to 11
µF/cm2) one can assume that not all the micropores take part in the charge
accumulation. It is clear that the micropores not adapted to the size of the
solvated ions will not take part in the double layer charging. The charging of
the double layer is very complex and depends also on other parameters such
as the pore size distribution, the affinity to the electrolytic solution, the
hydrophobic-hydrophilic character, the particle conductivity and their size.
0.9
0.8
0.7
0.6
0.5
E/V
0.4
0.3
0.2
0.1
0
-0.1
0 1000 2000 3000 4000 5000 6000 7000
t/s
Figure 3. Galvanostatic charge/discharge characteristics of a capacitor built from KOH

activated carbon A-PM (mass of electrodes 12.2 mg/12.8 mg) I = 2 mA. Electrolytic solution:
1mol l-1 H2SO4.
Table 3. Capacitance values (F g-1) of the KOH activated carbons (A-C; A-CS; A-PM; A-PS;
A-AC) estimated by galvanostatic discharge, cyclic voltammetry and impedance spectroscopy.
Specific capacitance µF cm-2 calculated per surface area of carbon.
Galvanostatic Cyclic Impedance Specific
Sample discharge voltammetry spectroscopy capacitance
-1 -1 -1
C/Fg C/Fg C/Fg µF cm-2
A-C 312 317 282 9.9
A-CS 223 235 206 7.0
A-PM 294 299 273 11.0
A-PS 261 241 227 9.5
A-AC 198 198 193 10.4
13
Even if the KOH activated carbons supply high capacitance values,

the practical application of such materials is determined by the
supercapacitor cycleability, quick charge propagation at different loads, a
low self-discharge. Highly microporous carbons supply always some
diffusion limitation. This effect can be observed at quick scan rates during
voltammetry experiments and during impedance spectroscopy measure-
ments. Figure 4 shows the impedance characteristic for the carbon A-PM
with almost perpendicular dependence of imaginary part to real one that is a
proof for capacitive response; however, a small diffusion slope is slightly
marked.
160
140
120
100
Z'' / ohm
80
60
40
20
0
-20
0 40 80 120 160
Z' / ohm
Figure 4. Impedance spectroscopy characteristic for the carbon sample A-PM.
Mass of electrodes: 12.2/12.8 mg. C = 273 F g-1 (at 1mHz).
It is important to stress that the capacitive behaviour of the

microporous carbons could be further improved by enhancing the mesopore
volume. The presence of mesopores plays a crucial role for the ion
transportation to the active surface. Hence, a development of mesopores in
these materials, and a strict control of the micropore-mesopore volume ratio
is necessary.
3.2. Template Carbons as Capacitor Electrodes

The total surface area of the template carbons prepared by sucrose
impregnation is significantly higher than the surface area of the
corresponding silica template (Table 2), that confirms the formation of
micropores during the carbonization. Just an opposite tendency is observed
14
when propylene is used as carbon precursor. In the latter case, the pore
volume of the template carbon is consistent with an uniform pore filling, and
the values of surface area are easily justified by different pores and walls
diameters in the pristine template. The higher pore volume and surface area
of sucrose derived template carbons is attributed to the release of water and
gases during the thermal treatment.
The capacitance values determined by the three techniques are
reported in Table 4 for the different types of carbons (CS48, CS15, CPr48,
CPr15) depending on the electrolytic medium (acidic, basic and organic).
For each kind of capacitor, it can be seen that a satisfactory correlation is
found between the three techniques. The various electrode materials give rise
to different values of capacitance due to their different physico-chemical
characteristics as seen in Table 2. The voltammograms of CS15 and CS48
for the acidic and organic media have a box-like shape which is
characteristic for an ideal capacitor with quick charge propagation. As seen
in Table 4, the two materials exhibit high values of capacitance both in
aqueous and organic medium; the capacitance values of CS48 reach 200F/g
and 110 F/g in acidic and organic medium, respectively. These values are
promising compared to other reported in the literature for various activated
carbons with comparable area. Hence, the additional porosity created during
the carbonization step is profitable for the performance of these template
carbons. Comparing the performance of the electrodes made from CS48 and
CS15, the specific capacitance per surface unit in acidic medium is roughly
the same for the two materials, ca.10 PF/cm2 and 12 PF/cm2, respectively.
This would lead to the conclusion that the main parameter controlling the
electrochemical performance in these materials is the total surface area
(Table 2). However, taking into account that the surface area determined by
nitrogen adsorption and the electrochemically active surface area are
different, other factors might explain the different capacitance values
observed for CS48 and CS15, e.g. the different pore arrangement which
affects the diffusion of the solvated ions, the presence of a more marked
secondary microporosity (1.5 nm < D <1.8 nm) for CS48 and consequently,
smaller pore size compared to CS15.
The electrochemical performance of CS15 looks to be similar in
basic and acidic medium, while smaller values are observed for CS48 in
basic medium (Table 4). Such a difference might be related to the different
surface functionality of the two materials, the latter controlling the
wettability of the materials. In the case of CS48, surface oxygenated groups
which dissociate differently depending on the electrolyte pH, could be
responsible of different adsorption of the ions on the carbon surface in both
media. On the other hand, CS15 behaves rather as a material, which has a
poor surface functionality. A careful analysis of the surface properties of
these carbons is necessary to support these hypotheses.
15
Table 4. Capacitance values in F/g (per gram of carbon material) for the carbons CS48,
CS15, CPr48, CPr15 in acidic, basic and organic media†.
Material: CS48 C (F/g) C (F/g) C (F/g)

1M H2SO4 6M KOH TEABF4 in CH3CN
G (2mA) 206 184 115
G (5mA) 188 170 116
P (2mV/s) 193 181 118
P (10mV/s) 170 159 114
I (1mHz) 210 177 96
Material: CS15 C (F/g) C (F/g) C (F/g)

G (2mA) 169 174 93
G (5mA) 158 161 90
P (2mV/s) 162 170 98
P (10mV/s) 144 156 94
I (1mHz) 171 172 78
Material: CPr48 C (F/g) C (F/g) C (F/g)

G (2mA) 101 74 65
G (5mA) 99 87 62
P (2mV/s) 113 82 64
P (10mV/s) 99 75 62
I (1mHz) 113 81 57
Material: CPr15 C (F/g) C (F/g) C (F/g)

G (2mA) 62 58 48
G (5mA) 61 55 46
P (2mV/s) 62 56 50
P (10mV/s) 58 51 49
I (1mHz) 78 59 42
Figure 5 presents the capacitance-frequency dependence from

impedance spectroscopy measurements for CS48 and CS15 in acidic and
organic medium. In the low frequency region (from 1mHz to 100mHz)
nearly a complete penetration of the ions into the pores is allowed and the
quite stable values indicate the domination of the capacitive behavior at the
electrolyte/carbon interface. All the curves show a typical drop of
†
In the table, G, P, I stand for measurements by galvanostatic, potentiodynamic and
impedance spectroscopy, respectively.
16
capacitance with frequency, which becomes quite significant at frequencies

higher than 1Hz. In the case of the sample CS15, the decrease of capacitance
is shifted to slightly higher frequency compared to CS48, both in aqueous
and organic medium. The ions access to the active surface, although the pore
diameter is larger than the size of the solvated ions in both cases.
It means that, for a given frequency, the ion diffusion to the active
surface is more efficient in the case of CS15. Comparing CS15 and CS48,
this better frequency behavior of CS15 could be due to the presence of
slightly larger mesopores (Table 2), or mesopores of appropriate shape,
which favor a better efficiency. However, one must be careful with the
interpretation of adsorption data, because the pores of the template carbons,
which are the walls replica, are no longer cylindrical, and the determination
of the pore size by the BJH method (which assumes cylindrical pores) might
be slightly misleading.
250
200
(a)
Capacitance (F/g)
150
(b)
100
(c)
50
(d)
0
1.0E-04 1.0E-03 1.0E-02 1.0E-01 1.0E+00 1.0E+01 1.0E+02 1.0E+03 1.0E+04 1.0E+05
Frequency (Hz)
Figure 5. Capacitance vs frequency for carbons from CS48 (a,c) and CS15 (b,d) in 1M H2SO4
(a,b) and 1M TEABF4 in acetonitrile (c,d).
The capacitance values of the carbons from propylene, i.e. CPr in the
three electrolytic media are lower than for the materials from sucrose CS
(Table 4), and they decrease with the total surface area of the carbon
materials, i.e. with the filling rate of carbon in the silica porosity [18]. These
results are not surprising since the carbon filling is quite uniform during the
CVI process, and consequently, the fraction of micropores formed is much
17
less important than for CS. The same as for the CS materials, the carbon
made from the MCM-48 template gives higher capacitance values than the
carbons resulting from the SBA-15 template. Since the wall thickness of
MCM-48 (1 to 1.5 nm) is smaller than that of SBA-15 (4 nm) templates, the
template carbon CPr48 possess a more important amount of secondary
micropores than CPr15, that is profitable for charging the double layer. The
presence of a well-balanced micro/mesoporosity in CPr48 could explain the
relatively high capacitance per surface unit, e.g. 12 PF/cm2, against only 8
PF/cm2 for CPr15.
In conclusion, although the porous texture of these materials is of
limited interest for getting high capacitance values, it allows to clearly
demonstrate the beneficial effect of mesopores on the capacitor performance.
3.3. C/C Composite for Supercapacitor Electrodes

The composites based on multiwalled carbon nanotubes (CNTs) and
N-enriched carbons are novel materials, which may be interesting for
supercapacitor application. The example of the composite material prepared
with the nitrogenated carbon from polyacrylonitrile (PAN) is analyzed in this
paper. The strongly entangled network of the nanotubes reinforces the
composite, preventing its shrinkage during the carbonization process. The
observations by Scanning Electrone Microscopy technique (Figure 6) show
that the C/C composites are formed of a nanotube backbone and individual
or interconnected aggregates of pyrolysed PAN.
The capacitance of cells built from the C/C composites were
compared with the values given by the single components of the composite.
The pristine CNTs gave 19 F/g and the value for carbonized PAN was
negligible. The formation of pores, due to PAN decomposition and
conversion into microporous carbon matrix, as well as the nitrogen presence
could be the reasons of the higher capacitance values. Table 5 summarizes
the values of capacitance measured on supercapacitors based on the C/C
composite electrodes prepared in different conditions, without oxidative
stabilization of PAN. By comparison with the single components, the most
striking fact for the C/C composites, was a noticeable increase of capacitance
ranging from 50 to 100 F/g. The highest capacitance of 100 F/g was obtained
for the composite with 30wt% of CNTs and 70 wt% of PAN, carbonized at
700°C in inert atmosphere for 180 min. Comparable values of atomic
percentages of nitrogen (7 and 9 at%) were found from XPS and elemental
analysis, respectively on this C/C composite. It demonstrates that nitrogen is
well distributed on the surface as well as in the bulk of electrode. On the
other hand the BET surface area of this composite is only 210 m2/g.
These extraordinary properties of the C/C composite prepared with
the CNTs/PAN (30wt%/70wt%) mixture are the consequence of several
18
effects provided by the combination of the two components. The PAN

matrix is auto-activated by the gas evolution during the thermal
decomposition, giving rise to the creation of pores with the help of the CNTs
backbones which prevent the matrix from shrinkage. However, taking into
account the relatively high value of capacitance compared with the BET
surface area, the most important contribution to capacitance in this material
seems to be the pseudocapacitance effects provided by the nitrogen
functionality.
Figure 6. SEM images of the C/C composite showing individual (a) or interconnected
(b) aggregates of PAN.
Table 5. Relation between the C/C composites composition, HTT, time

of carbonisation and the capacitance values (F/g).
NTs (%) PAN (%) HTT (0C) Time (min.) C (F/g)
70 30 900 30 48
70 30 700 30 53
30 70 900 30 55
30 70 900 90 48
30 70 700 30 60
30 70 700 90 76
30 70 700 180 100
30 70 700 420 60
15 85 700 180 58
50 50 700 180 57
4. CONCLUSION
The KOH-activated carbons give excellent capacitance values
reaching ca. 300 F g-1 despite a rather moderate specific capacitance per
surface area, being in the range of 7-11 PF cm-2. Apparently, this can be
19
considered as due to a restricted accessibility to the micropores active

surface in the case of the KOH activated materials with extremely developed
porosity. It seems that the carbon A-PM from mesophase pitch, because of a
good precursor organization, gives pores of more adapted size as a
consequence of intercalation phenomena during KOH activation. Hence, A-
PM is a definitively better candidate for capacitor electrode than the other
investigated carbons. A detailed analysis proved that wide micropores and
narrow mesopores play a crucial role for ion transportation during charging
of the electrical double layer using carbons with extremely developed
surface area. Additionally, the low cost of the carbons based on natural
precursors makes them extremely attractive for capacitor application.
However, for a long-term capacitor performance, the carbons with a low
leakage current and limited self-discharge should be selected to fulfil the
practical demand.
Interesting carbon materials were prepared by a templating
procedure from MCM-48 or SBA-15 highly ordered mesoporous silica. The
perfectly controlled texture of the template carbons allows a better
understanding of the electrochemical performance in supercapacitors.
Capacitance values as high as 200F/g and 110F/g in aqueous and organic
media are reached with the carbons CS from sucrose presenting secondary
micropores. Taking into account these high values, it is obvious that the
presence of interconnected mesopores and secondary micropores makes the
active surface more available for charging the electrical double layer than in
a strictly microporous material. An improved capacitance-frequency
dependence has been found for the essentially mesoporous carbons. To our
knowledge, it is the first time that the beneficial effect of mesopores on the
diffusion of ions to the active surface is so clearly demonstrated. The
template carbons, with a well-tailored pore size and pore distribution are of
prime interest for designing high performance supercapacitors.
The N-doped carbons with a nanotube backbone combine a
moderate presence of micropores with the extraordinary effect of nitrogen
that gives pseudocapacitance phenomena. The capacitance of the PAN/CNts
composite (ca. 100 F/g) definitively exceeds the capacitance of the single
components (5-20 F/g). The nitrogen functionalities, with electron donor
properties, incorporated into the graphene rings have a great importance in
the exceptional capacitance behavior.
ACKNOWLEDGMENTS
This research was supported by the NATO Science for Peace project
number 973849. Thanks are due to C. Vix from ICSI, Mulhouse, for
providing some of the materials investigated in this paper.
20
REFERENCES
1. Conway B.E. Electrochemical Supercapacitors. Kluwer Academic/Plenum Publishers,

New York, 1999.
2. Frackowiak E., Béguin F. Carbon materials for the electrochemical storage of energy in
capacitors. Carbon 2001; 39: 937-50.
3. Mayer S.T., Pekala R.W., Kaschmitter J.L. The aeorocapacitor: an electrochemical
double-layer energy-storage device. J Electrochem Soc 1993; 140:446-51.
4. Laforgue A., Simon P., Sarrazin Ch., Fauvarque J-F. Polythiophene-based
supercapacitors. J Power Sourc 1999; 80:142-8.
5. Frackowiak E., Jurewicz K., Delpeux S., Béguin F. Nanotubular materials for
supercapacitors. J Power Sourc 2001; 97-98:822-5.
6. Jurewicz K., Delpeux S., Bertagna V., Béguin F., Frackowiak E. Supercapacitors from
nanotubes/polypyrrole composites. Chem Phys Lett 2001; 347: 36-40.
7. Arbizzani C., Mastragostino M., Soavi F. New trends in electrochemical supercapacitors.
J Power Sourc 2001; 100:164-170.
8. Miller J.M., Dunn B., Tran T.D., Pekala R.W. Morphology and electrochemistry of
ruthenium/carbon/aerogel nanostructures. Langmuir 1999; 15:799-806.
9. Toupin M., Brousse T., Bélanger D. Influence of microtexture on the charge storage
properties of chemically synthesized manganese dioxide, Chem Mater 2002; 14:3946-52.
10. Wu N.L. Nanocrystalline oxide supercapacitors. Mater Chem Phys 2002; 75:6-11.
11. Frackowiak E., Méténier K., Bertagna V., Béguin F. Supercapacitor electrodes from
multiwalled carbon nanotubes. Appl Phys Lett 2000; 77:2421-3.
12. Jurewicz K, Babel K., Ziolkowski A., Wachowska H. Ammoxidation of active carbons
for improvement of supercapacitor characteristics. Electrochim Acta 2003; 48:1491-8.
13. Kierzek K., Frackowiak E., Lota G., Gryglewicz G., Machnikowski J. Electrochemical
capacitors based on highly porous carbons prepared by KOH activation Electrochim
Acta 2004; 49:515-23.
14. Frackowiak E., Delpeux S., Jurewicz K., Szostak K., Cazorla-Amoros D., Béguin F.
Enhanced capacitance of carbon nanotubes through chemical activation. Chem Phys Lett
2002; 361: 35-41.
15. Galarneau A., Cambon H., Di Renzo F., Fajula F. True microporosity and surface area of
mesoporous SBA-15 silicas as a function of synthesis temperature. Langmuir 2001;
17:8328.
16. Ryoo R., Joo S.H., Kruk M., Jaroniec M. Ordered mesoporous carbons Adv Mater 2001;
13:677-81.
17. Fuertes A.B. Template synthesis of mesoporous carbons with a controlled particle size, J.
Mater. Chem. 2003; 13:3085-8.
18. Jurewicz K., Vix C., Frackowiak E., Saadallach S., Reda M., Parmentier J., Patarin J.,
Béguin F. Capacitance properties of ordered porous carbon materials prepared by a
templating procedure. J Phys Chem Solids (2004) in press.
19. Delpeux S., Szostak K., Frackowiak E., Bonnamy S., Beguin F. High yield carbon
nanotubes from the catalytic decomposition of acetylene on in-situ formed Co
nanoparticles. J Nanosci Nanotech 2002; 2: 481-4.
20. Szostak K., Frackowiak E., Delpeux S., Béguin F. CD-ROM G-1 Carbon’03 Oviedo
Spain (2003).
EFFECT OF CARBONACEOUS MATERIALS
ON PERFORMANCE OF CARBON-CARBON
AND CARBON-Ni OXIDE TYPES OF
ELECTROCHEMICAL CAPACITORS
WITH ALKALINE ELECTROLYTE
Alexey I. Belyakov*
ELIT JSC (Electrochemical Power Sources)
40, Prospekt Leninskogo Komsomola, Kursk, 305026, Russian Federation
1. INTRODUCTION
Carbonaceous materials play a key role in achieving the necessary
performance parameters of electrochemical capacitors (EC). In fact, various
forms of carbon constitute more than 95% of electrode composition [1].
Double layer capacity and energy storage capacity of the capacitor is directly
proportional to the accessible electrode surface, which is defined as surface
that is wetted with electrolyte and participating in the electrochemical
process.
Different types of activated carbon are among the most suitable
materials for this purpose. For this reason specialists, involved in
development of active materials for EC try to increase carbon’s specific
surface as much as possible and to optimize the internal structure of the
carbon porous structure.
Activated carbons possess sufficient volumetric conductivity for
electrolyte/collector current interchange. However, contact resistance
between carbon particles in the electrode limits charge/discharge currents of
the porous volumetric system and therefore EC’s power capability.
It should be also mentioned that an increase of activated carbon’s
specific surface (over activation) always leads to increase of its specific
resistance. Different methods of obtaining activated carbon with an
optimized volumetric structure were developed [2, 3], but they have not yet
found industrial application.
At this point of time, key improvements of volumetric conductivity
are being achieved through introduction of highly conductive, finely
*
21
22
dispersed additives, such as carbon black, graphite, metal powders, carbon

fibers, other powdered materials into activated carbon’s active mass.
Depending on the power demand of a particular EC, percentage of the
required conductive additive in the electrode varies in the range of 5 –
40wt%.
Finely dispersed carbon substances introduced into the active mass
of pseudocapacitive electrodes of “hybrid” (asymmetric) capacitors play the
most important role in operation of an efficient volumetric collector.
A classic case is an EC of a faradic type in which an electrode is
comprised of: Ni(OH)2, MnOOH, etc. active materials. Since in these
chemistries the conductivity depends on electrode state-of-charge charge
level, they require presence of additional stable conductive skeletons in their
structure. Noteworthy mentioning that besides traditional forms of carbon or
other conductors that may form such a skeleton, the latest progressive
investigations demonstrate the possibility of application of different
nanostructured forms of carbon, such as single-wall and multi-wall carbon
nanotubes [4, 5]. Yet, for the industrial application, highly conductive carbon
powders, fibers and metal powders dominate at present.
Summarizing the above, it may be stated that activated carbons and
pseudocapacitive materials in EC electrode structure are responsible for the
energy storage parameters (specific energy), while non-active highly
conductive carbon additives are responsible for the electrode internal
resistance (EC specific power).
The present work describes the results of investigations of carbon
materials used by us for creation of conductive skeletons in the EC electrode
body. Investigations were carried out in full-size Ni Oxide and carbon-
carbon supercapacitor systems with aqueous solution of KOH.
2. EXPERIMENTAL
2.1. Conductivity of Feed Materials and Their Dry Mixtures

Before measuring conductivity, carbon materials and their mixtures
were dried to residual moisture content of less than 5wt%. A sample of the
investigated material was put into a measuring cell, representing a dielectric
cylinder closed at both ends by the measuring electrodes.
Electrode pressure on the sample material approximately
corresponded to the pressure used in a practical EC in its operational mode
and was equal to 8 kgf·cm-2 (this pressure was applied for purposes of
minimization of particle contact resistance).
Conductivity measurements were carried out by an alternating
current method using 1 kHz, and measurements were taken at a current
density in the range of 2,2-270 mA·cm-2. Six samples of each formula were
23
analyzed in order to ensure repeatability of data points. In order to reduce the

influence of the current density on resistance, as well as for comparability of
results with different materials, a diagram for extrapolation of “Zero” current
density (J=0) was developed.
2.2. Testing in the Carbon-Carbon type of Electrochemical

Capacitors
In our tests, we used pasted mixtures of carbon-carbon electrode
components with KOH solution having a density of 1,26 g·cm-3. Positive and
negative electrodes were pasted onto the conductive polymer film, separated
by ionoconductive separator, made out of special paper, pressed between
external collectors of nickel-plated copper with pressure of 8 kgf·cm-2.
Each cell was sealed via welding all around its edges. The welding
seal has been formed on the part of the cell, where conductive polymer
makes a boundary with the layer of nonconductive polymer. Electrode
thickness did not exceed 150Pm; separator thickness was approximately
50Pm; active electrode surface dimensions were: 128 x 148 mm. Electric
capacity was measured within the voltage range of 0,8v-0,4v by a constant
current method. ESR (electrode specific resistance) was measured at a
frequency of 1 kHz at measuring current value of 1A. Before the test, all the
cells had been kept during 24 hours under working voltage and additionally
had been subjected to 10 charge-discharge cycles. Test temperature was kept
at 22-24°C.
While choosing an optimum conductive additive, dozens of various
carbon and metal powders were researched. In this report, the characteristics
of three main representatives are given. They are: Nickel (produced in
Russian Federation by means of carbonyl method), colloidal graphite,
refined graphite and refined Boron-doped graphite. Superior Graphite Co.,
Chicago, IL, USA supplied all grades, of high purity graphite mentioned in
this paper.
2.3. Construction of NiOx-Carbon Ultra Capacitors

2.3.1. NiOx Electrode
For electrode manufacturing, powder of Ni(OH)2 was used,

produced by H.C.Starck, USA. The product has mean particle size of
d50~6Pm. It was mixed with conductive additive, such as, for instance,
refined graphite powder.
24
The quantity of conductive component varied from 15wt% to

30wt%. Solution of thermoplastic polymer in organic solvent was prepared
separately. Then, a mixture of dry components Ni(OH)2 + graphite was
added to that solution and mixed one more time. Content of polymer binder
in composition of a dry electrode was 5-10wt%.
Electrode film was formed from a mixture of components in organic
solvent by means of casting, which after drying (removing of solvent) was
rolled by dry method to the desirable thickness (150-200Pm). As a result,
stable forms of self-supported electrodes were obtained. Then these
electrodes were subjected to formation during two charge-discharge cycles
in the solution of electrolyte, consisting of 40wt% KOH and 10g·l¯1 LiOH.
2.3.2. Carbon Electrode
Activated carbon cloth of 1-1.1mm thickness was used as the initial

material for most of experiments. Ribbon, made of carbon cloth, was soaked
in the solution with 40wt% KOH and 10 g·1¯¹ LiOH for 72 hours, and
afterwards, was cut into electrodes of the appropriate dimensions.
Electrolyte surplus was removed by pressing.
2.3.3. Electrochemical Cells
An electrode block, which included a positive NiOx electrode, a

separator, made of special paper (60Pm), and a negative electrode, were
placed between two sheets of conductive plastic of 50-60Pm thick.
The upper and the lower plastic sheets were welded along the edges
through the insolating layer with an exception of the area which was used
preliminary as a ventilation outlet.
The assembled cell was put between two metal collectors and
pressed to the level of operating pressure for purposes of reduction of the
contact resistance of assembly parts between power covers of the clamp unit.
3.1. Conductivity of Electrodes

Following the procedure described in section 2.1, initially,
conductivity parameters of dozens of materials were analyzed. Below, we
present data for conductivity of typical representatives of families of
conductive materials, which we used in our study. Data is given in Table 1.
25
The table shows, that resistance of dry activated carbon is more than
20 times higher than that of the least resistive graphite. In this table, Boron-
doped natural graphite, 2939APH–RG follows next after activated carbon.
More interesting data was found in the mixtures: we have seen that
carbon additives tend to reduce specific resistance of the mixture, while
active material (activated carbon) increases resistance (54 10-3 ohm·m vs. 49
10-3 ohm·m). This phenomenon suggests that carbon-carbon component
mixtures have better “packing” characteristics.
Table 1. Specific Resistance of Dry Powdered Materials.

SPECIFIC
TYPE OF CONDUCTIVE MATERIALS RESISTANCE,
j = 0, ohm·m
Activated Carbon (grade SKT) 49 · 10-3
Carbon Black (grade P- 267) 2,1 · 10-3
Expanded Graphite (grade ABG1005, Superior Graphite) 1,1 · 10-3
Purified Flake Graphite (grade 2939APH, Superior Graphite) 2,6 · 10-3
Boron-doped purified flake graphite (grade 2939APH-RG,
0,9 · 10-3
Superior Graphite)
Ni Powder (Carbonylic Method) 0,75 · 10-3
ABG1005 (8wt%) + Activated Carbon (92wt%) 35 · 10-3
Ni Powder (10wt%) + Activated Carbon (90wt%) 54 · 10-3
Carbon Black (15wt%) + Activated Carbon (85wt%) 11 · 10-3
Notes for the Table: Material particle size was as follows:
- activated carbon: < 60Pm (10-20)
- carbon black: < 5Pm
- nickel powder: < 5 Pm
- graphite powders “Superior Graphite”: <20Pm.
It is noteworthy mentioning that above experiment allows drawing

only a preliminary conclusion, as far as selection of an optimum conductive
additive is concerned.
It should be noted, that ESR of electrode, having minimum
resistance in dry state may be significantly increased during impregnating it
with electrolyte solution (said resistance increase is typical in n-type solid
phase composites). It is caused by the expansion effect due to formation of
electrolyte films during wetting of the surface of electrode particles. Thus,
“dry” conductivity data may serve only as an initial estimation during
optimization of volumetric structure of an EC electrode.
3.2 Effect of Carbonaceous Materials on Performance of

Carbon-Carbon EC in presence of an Alkaline Electrolyte
In active material of both electrodes of the EC with purely “double
electric layer”, volume changes do not take place during charge-discharge
processes. That’s why it is not expedient to add binder into active material.
26
However, that is observed only within the range of the so-called

“safe” voltage window of operation – for alkaline aqueous electrolytes it
covers the range of 1,12-1,24v, and depends on material purity and current
density. Exceeding this voltage causes gassing, electrode swelling and loss
of electric contact between particles in the electrode volume.
Summarized “as measured” and “calculated” data for
electrochemical cells and active electrodes with practical dimensions of the
electrodes of 128x148mm are shown in Table 2 (please see cell assembly
detail in section 2.2).
Table 2. Parameters of capacitors as a function of conductive additive

used in electrode composition.
Conductive Additive and its Concentration in the Electrode
Boron-
Natural
Expanded Doped Carbon
PARAMETERS Flake Ni
Graphite, Flake Black,
Graphite, Powder,
ABG1005, Graphite, P-267,
2939APH, 10wt%
8wt% 2939APH- 5wt%
8wt%
RG, 8wt%
Capacity, F 90 88 92,4 98 104
ESR, 1 kHz, Ohm 0,92 1,03 0,84 0,75 1,07
Energy Density, kj·kg-1 2,12 2,04 2,29 2,5 2,30

RC-time constant,
83 91 78 70 111
ms (milliseconds)
Power Density, W·g-1 12,8 11,3 14,8 15,2 10,7
Analyzing the data one may notice, that cells with carbon as
conductive powders have advantage over cells using metal as conductive
powders, possessing 1,5 times higher specific power at comparable specific
energy level. Carbon black P-267 and graphite 2939APH-RG have the best
performance indexes. Its higher specific surface area in comparison with that
of graphite explains the higher carbon black indexes. Higher surface area
provides higher specific sorption of electrolyte and improvement of
electrode conductivity according to the conductivity of the second kind.
As far as cycle life of the ECs is concerned, with the above
conductive additives, and the active mass, consisting of activated carbon, for
all groups of EC we built for this work, it exceeded 10,000,000 cycles (at
depth of discharge of 30%); that is quite sufficient for the main spheres of
application.
27
3.3 Effect of Carbonaceous Materials on Performance of

Carbon-Ni Oxide Type of EC with Alkaline Electrolyte
The character of carbonaceous material’s influence on performance
of asymmetric EC systems (carbon – NiOx) is more complicated than that of
carbon-carbon system. It is determined by the higher operating potential of
NiOx electrode, due to which oxidation and volume changes in electrode
active mass upon charge-discharge processes are taking place as the reaction
of transformation of Ni(OH)2 into NiOOH is occurring.
1 - Current Collector;
2 - Volume Collector - Chain Structures;
3 - Particles of Active Material;
Rki - Contact Resistance:

Volume Collector / Current Collector;
Rci - Contact Resistance between particles of
Volume Collectors;
Rvi - Bulk Resistance of materials of particles;
Rmi - Contact Resistance:
Volume Collector / Active material;
Rb - Resistance, determined by binder,
increased the all Ri;
AM - Active material.
1 2 3
Rk Rv1 Rv2 Rvi

Rc1 Rc2 Rci
Rk 2
Rvi Rm
Rm
Rki Rci Rvi Rmi Rb

AM
Figure 1. Simplified model of agglomerate electrode.

28
First of all let us consider the morphological structure of an agglo-

merate electrode [6] by way of example of the model shown in Figure 1.
This schematic represents a multiphase system with no fixed connection
between its components. As a rule, the active mass of an electrode is a
mixture of Nickel hydroxide (oxyhydroxide) with conductive carbon or a
metal, which are well dispersed mechanically in the matrix.
The NiOx particles, when in operation, are closely packed with
conductive additive by means of pressing of the electrode block or addition
of a binder into the active mass slurry. The quantity of conductive additive
should be chosen pursuing the necessity to create a spatially ramified grid
providing current transmission from the active material particles to an
external collector. In dependence on reciprocal dispersion of conductive
additive and active material, the quantity of conductor in the mass may vary
from 10wt% to 70wt% of the total electrode mass.
The conductive particles in the volume of electrode make a so-called
“chain” structures. As it is seen from the model, the electrical circuit
assumes interconnection of the active mass particles in the solid phase of the
electrode, while the external collector consists of a long chain of resistors.
In accordance with this model, the current path is made through the
contact resistance: external collector – volume collectors (Rki), then goes
through many other resistances, such as: bulk resistance of the conductive
particles (Rvi) and the contact resistance between the conductive particles in
the volume of the electrode (Rci); at the end, there is contact resistance
between the conductive particles and active mass particles (Rmi).
Considering this chain of electrical resistors, we should take into
account the influence of the binder (dielectric) effecting on all the above
contact resistances in the circuit and increasing their value (Rb). Evidently,
the main influence on the ESR of electrode is this circuit because of a big
number of contact resistances distributed in the bulk of the electrode.
We consider that the practical electrode’s efficiency at operation
under high current density conditions and during service life is determined
by the state of the particles of conductive additive’s surface. With reference
to this, we can point out two main factors effecting fundamentally reliable
operation of the NiOx electrode.
(a) Electrochemical oxidation of the particles’ surface of the
conductive additive. Potential of charged NiOx in KOH electrolyte is 0,45v
to 0,5v vs SHE (standard hydrogen electrode). This value is sufficient for
fast oxidation of carbon materials and appreciable for metal powder. Even
insignificant oxidation of particle surfaces causes dramatic ESR increase on
electrode, because we have here a long series circuit of contact resistances
between the particles.
(b) Electrode swelling. The agglomerate electrode, when it even has
binder inside, is not a solid structure. Volumetric changes in morphology (Į,
ȕ-Ni(OH)2; ȕ, Ȗ-NiOOH) arise thickening of electrode. It is evident even in
29
tightly squeezed electrode stacks. Thus, a mechanical movement of the

volumetric conductor particles from the active mass particles, and from each
other takes place. Simultaneously, the contact area between the particles
diminishes with time, the growth of the contact resistance increases the total
ESR of the electrode.
Combined influence of both of the above factors significantly limits
the performance and operation capability of the NiOx electrodes of
agglomerate type in the asymmetric C/NiOx capacitors. Good initial
performance may change for accelerated failure either in the floating mode
or in heavy duty cycling mode of operation.
3.4. Electrochemical Testing

In the experiments discussed hereafter, carbon black was not taken
into account as the loss of the performance parameters of EC due to ESR
growth appeared soon after several charge–discharge cycles with this type of
conductive additive.
Initially, the amount of conductive additives was matched up to the
active mass of Ni Oxide electrode, and then its amount was fine-tuned going
by the criteria of having a minimum ESR and reaching highest oxidation
resistance. Thus, for the colloidal graphite the optimum quantity of
conductor became 12wt%, for Ni powder – 60wt%, for the refined graphite –
25wt%. The design of experimental cells are thoroughly described in our
previous papers [7], and also detail is given in section 2.3.
Assembled cells were cycled in the most “severe” mode; known as
“continuous day cycling”. This is an uninterruptible (without pauses) cycling
during 8 hours in the voltage window 1,5 – 1,1v.
The upper level of charging voltage, 1,5v is a critical value for the
electrochemical system C–NiOx in aqueous electrolyte. Here, formation of
some quantity of atomic oxygen on the NiOx electrode takes place, which
oxidizes electrode components more rapidly. Figure 2 shows a typical test
cyclogram. Upon such cycling, an electrochemical device goes through
approximately 1400 cycles in a day.
Figure 3 shows changes of the ESR and capacity during cycling.
High stability of conductive carbon, represented by Boron-doped refined
graphite (2939APH–RG) in this critical test conditions was observed.
For comparison, less superior test results of ECs with conductive
additives of colloidal graphite and powdered nickel are shown in Figure 4.
ESR of EC with refined graphite has increased by 56% after 36,000 cycles,
and with nickel powder by 81% already after 1,400 cycles. Resistance of EC
colloidal graphite increased by as high as 4,5 times compared to initial
values already by the 500th cycle.
30
Capacity loss for EC with graphite 2939APH–RG was 13% during

the 36,000th cycle, while with nickel powder, it reached 43% by
approximately the 1,000th cycle; and with colloidal graphite the capacity loss
came to be 32% by the 500th cycle.
1,4 15
1,5 Current
1,3 10
1,2 5
Current, A
1,1 0
Voltage, V
Voltage
1,0 -5
0,9 -10
0,8 -15
0,7 Cycle -20
0,6 -25
0 5 10 15 20 25 30
time, sec.
Figure 2. Cycling diagram in the “day-cycling” mode.
Carried out testing confirmed the supposition about the degradation

mechanism of electrode in NiOx-carbon ultra capacitors (Figure 1).
Obviously, the smaller are the dimensions of particles of conductive additive
(colloidal graphite) and the more active its surface is (nickel powder), the
faster oxidation processes are taking place on the surface of these particles.
Figure 3. Change Capacity and ESR vs cycle number;

Capacitor C/NiOx, conductive additive: 25wt% 2939APH-RG;
Cycling regime: 1,400 cycles/day plus 650 hours of floating.
31
600 7
Capacity,
500 12% Colloidal 6
Graphite
5
400 ESR,
ESR, mOhm
Capacity, F
ESR, 60% Ni Powder 4

12% Colloidal
300
Graphite
3
DEVELOPEMENT
200 OF HIGH PERFORMANCE
2
100 1
Capacity,
0 60% Ni Powder 0
0 250 500 750 1000 1250 1500
Cycles
Figure. 4. Change of Capacity and ESR during cycling.
These parameters appear to be causing avalanche-type growth of

contact resistances and accelerated failure of EC due to high ESR and
capacity loss during operation. Instead, doping of flake graphite with Boron
(2939APH-RG sample) appears to significantly improve oxidation resistance
of conductive additive, thus resulting in superior cycle life of NiOx-carbon
supercapacitors.
4. CONCLUSIONS
1. For electrochemical capacitors of the system carbon-carbon, in
spite of the fact that the electrode body consists of conductive activated
carbon, it is always necessary to use highly conducive additives, preferably
those selected from the group of carbon materials. Exact type of additives
plays a secondary role for this type of an EC, as it operates in the so-called
“safe” voltage interval (1,12-1,24v), where limited to no oxidation of carbon
takes place in this range of voltage.
2. For the electrochemical capacitors of Carbon-Ni Oxide system
with aqueous KOH solution it is expedient to use carbonaceous graphite
materials with expanded structure and modified surface. The best results
were achieved with carbon surface doped with Boron, which makes this
carbon superior than other conductive additives used in this study, due to its
32
increased resistance to oxidation. In this case, a number of contact

resistances in the model presented by Figure 1 is decreased. Boron doping
appears to lead to higher stability of conductive material under conditions of
electrochemical oxidation (1,1-1,5v). Increased oxidation resistance enables
long service life of an EC even in critical operation conditions.
3. Presently and for the near future, the greatest demand during
development and production of electrochemical capacitors will lie in the
sphere of optimization of the highly conductive micronized carbon
structures, with modified surfaces against oxidation.
REFERENCES
1. B.E. Conway, “Electrochemical Supercapacitors. Scientific Fundamentals and
Technological Applications”, 1999, Kluwer Academic /Plenum Publishers, NY.
2. T. Extrom, S. Gordeev and V. Kuznetsov, in Proceedings of the 6th International
Seminar on DLC and Similar Energy Storage Devices, Dec. 9 – 11, 1996, Deerfield
Beach, FL, USA.
3. S. Dietz, V. Nquyen, in Proceedings of the 11th International Seminar on DLC and
Similar Energy Storage Devices, Dec. 3 – 5, 2001, Deerfield Beach, FL, USA.
4. Lipka, D. Reisner, J. Dai, R. Ge, R. Cepulis, in Proceedings of the 10th International
Seminar on DLC and Similar Energy Storage Devices, Dec. 4 – 6, 2000, Deerfield
Beach, FL, USA.
5. D. Firsich, S. Lipka, in Proceedings of the 10th International Seminar on DLC and
Similar Energy Storage Devices, Dec. 4 – 6, 2000, Deerfield Beach, FL, USA.
6. A. Beliakov, in Proceedings of the 11th International Seminar on DLC and Similar
Energy Storage Devices, Dec. 3 – 5, 2001, Deerfield Beach, FL, USA.
7. A. Beliakov, in Proceedings of the 12th International Seminar on DLC and Similar
Energy Storage Devices, Dec. 9 – 11, 2002, Deerfield Beach, FL, USA.
HYBRID SUPERCAPACITORS BASED ON
α-MnO2/CARBON NANOTUBES COMPOSITES
Volodymyr Khomenko, Encarnacion Raymundo-Piñero

and François Béguin*
Centre de Recherche sur la Matière Divisée. CNRS-University, 1B Rue de la Férollerie,
45071 Orléans Cedex 02, France
1. INTRODUCTION
In recent years, attention has been focused on use of various metal

oxides as supercapacitor electrodes for high-power applications1-4. The
energy storage mechanism with these materials is based mainly on fast
faradaic redox reactions, which occur at the interface between the oxide and
the electrolyte4-5, giving rise to the so-called “pseudo-capacitance”. The great
advantage in this case is that the specific capacitance which can be obtained
is higher than for electrochemical capacitors where the charge is only stored
in the double layer. The characteristics required for the use of a metal oxide
as capacitor electrode are a pseudo-capacitive behavior, a large specific
surface area, a high conductivity and electrochemical stability. Among the
various metal oxides studied, amorphous and hydrated manganese dioxide
(a-MnO2·nH2O) is the most promising owing to the low cost of the raw
material. However, due to the low electrical conductivity of a-MnO2·nH2O, a
conducting additive is required to realize a composite electrode for
supercapacitors. In this study, taking into account that carbon nanotubes
(CNTs) are strongly entangled, forming an interconnected conducting
network, we propose to use them as alternative agent to carbon black. Since
the electrochemical behavior of manganese oxide has been previously
studied only with three-electrode cell systems and in the positive range of
potential6-12, we also propose to test real supercapacitors, i.e. two electrode
cells, based on the a-MnO2/CNTs composites. Therefore, for the first time,
the problems of using manganese oxide in symmetric capacitors will be
extensively discussed. Finally, the asymmetric construction, previously
proposed by Hong et al.13, will be optimized in order to allow both
*
Corresponding author: E-mail: [email protected]
33
34
electrodes to operate in an optimal potential range, that broadens the voltage

window of the device up to 2V in aqueous medium.
2. EXPERIMENTAL
Amorphous and hydrated manganese oxide (a-MnO2·nH2O) was
prepared by chemical co-precipitation from KMnO4 and Mn(OAc)2.4H2O in
water medium6-7. For obtaining an homogeneous active material in view of
the preparation of composite electrodes, predetermined amounts of carbon
nanotubes (0, 15, 20, 25, 30 wt%) were added to the solution before D-
MnO2·nH2O precipitation. The active material was also mixed with carbon
black for comparison purposes. The high purity carbon nanotubes (CNTs)
used in the present study were prepared by decomposition of acetylene over
a powdered CoxMg1-xO solid solution catalyst14. The amorphous character of
the oxide was confirmed by XRD measurements. Applying the BET
equation to its N2 adsorption isotherms at 77 K, specific surface areas of
around 250 m2/g were obtained.
The capacitor electrodes were films formed by a mixture of the
active material (a-MnO2·nH2O/carbon composite, 90 wt%) with a binder
solution (Teflon, 10%). The mixture was thoroughly homogenized in an
agate mortar and dried at room temperature before being rolled into a thin
film of uniform thickness. Electrode pellets were cut with a surface of 1 cm2.
Two electrode cell capacitors were built with a glassy fibrous separator and
gold current collectors, using a Teflon Swagelok® type system. Different
aqueous electrolytic solutions have been used: 1 mol.L-1 of Na2SO4 or 2
mol.L-1 KNO3 with a pH of 6.4 or with a modified pH of 10 by the addition
of NaOH. The values of capacitance were estimated by voltammetry (scan
rate of potential from 2 to 10 mV/s) and galvanostatic charge/discharge
cycling (VMP-Biologic, France). Additional three electrode cell experiments
were done by using a Hg/HgO reference electrode.

Table 1 contains the values of specific capacitance obtained when
different amounts of carbon nanotubes are added to a-MnO2·nH2O for the
preparation of the composite electrodes.
It is clearly demonstrated that adding carbon nanotubes improves the
behavior of a-MnO2·nH2O as capacitor electrode. The specific capacitance
values referred to the mass of a-MnO2·nH2O increase with the amount of
CNTs. However, when the specific capacitance is referred to the total mass
of the composite electrode material, to be realistic, it can be noticed that a
carbon nanotubes loading higher than 10-15wt% does not improve the
35
electrodes performance. Therefore, 10-15wt% of CNTs as conductive

additive are enough to increase the capacitance of the a-MnO2·nH2O based
electrodes from 0.1 to ca. 140 F/g. Due to their entangled network forming
open mesopores, the CNTs favor the ions accessibility from the electrolyte to
an important part of the active a-MnO2·nH2O.
Table 1. Specific capacitance in 1 mol.L-1 Na2SO4 medium for a-MnO2·nH2O loaded with
different weight percentages of carbon nanotubes.
Carbon Nanotubes Specific Capacitance Specific Capacitance
loading (wt%) (F/gcomposite) (F/goxide)
0 0.1 0.1
5 19 20
10 137 154
15 137 161
20 139 174
25 141 188
30 138 197
100 20 20
250
200
Carbon Black
Carbon Nanotubes
150
100
50
C(F/g)
0
-0.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
-50
-100
-150
-200
E(V)
Figure 1. Cyclic voltammograms at 2mV/s in 2 mol.L-1 KNO3 medium for two-electrode

capacitors based on a-MnO2·nH2O loaded with 15 wt% of carbon black or carbon nanotubes.
Figure 1 presents a comparative voltamperometric analysis of a-

MnO2·nH2O electrodes loaded with 15wt% of CNTs or 15wt% of carbon
black, which is conventionally used as conductivity additive for electrodes.
Using carbon nanotubes instead of carbon black produces a remarkable
improvement of the capacitive behavior of the a-MnO2·nH2O based
electrode. The shape of the voltammetry curve demonstrates that the
resistance of the electrode with CNTs is much lower than in the case of using
the conventional carbon black, finding a performance closer to an ideal
36
capacitor. An additional highly valuable effect of CNTs is the possibility to

extract more energy from the a-MnO2·nH2O electrodes, obtaining specific
capacitance values of 140 F/g in comparison with only 70 F/g when carbon
black is used.
Another important aspect for a material to be used as electrode for
supercapacitors is its electrochemical stability. In Figure 2, presenting the
specific discharge capacitance versus the cycle number for the optimized a-
MnO2·nH2O/CNTs composite in 2 mol.L-1 KNO3 (pH=6.5), it can be
observed that the specific capacitance loss after 200 cycles is about 20%.
180
Specific capacitance (F/g)
160
140
120
pH 6.5
100 pH 10
80
0 50 100 150 200 250
Cycles
Figure 2. Specific discharge capacitance for the a-MnO2·nH2O composite with 15 wt% of
.
CNTs in 2 mol L-1 KNO3 medium at two different pH.
For explaining this decay, it has to be considered that the pseudocapacitance

of hydrous oxides like a-MnO2·nH2O is attributed to redox transitions with
exchange of protons and/or cations with the electrolyte following the
equation12:
MnOa(OH)b + nH+ + ne- l MnOa-n(OH)b+n,
where MnOa(OH)b and MnOa-n(OH)b+n indicate interfacial a-MnO2·nH2O in

higher and lower oxidation states, respectively. For the stability of the
system these reactions must be reversible. After investigating the working
potential range for the positive and negative electrodes, it is possible to
observe that the negative electrode operates in the range of the irreversible
Mn(IV) to Mn(II) reaction, which results in the dissolution of the electrode,
being the reason of the poor cyclability. Taking into account that the
potential for the equilibrium between Mn(IV) and Mn(II) depends on the
electrolyte pH, the latter was increased to a value of 10 in order to avoid
37
such an irreversible reaction. Figure 2 shows that the stability is remarkably

increased, finding a loss of only 4 % of the initial specific capacitance after
200 cycles. However, even solving this problem, an important drawback of
this kind of symmetric capacitor is the value of the working voltage which
cannot be higher than 0.6 V.
The limitations of manganese dioxide in a two electrode capacitor
were overcome by using activated carbon at the negative electrode. Such an
asymmetric system was previously proposed13, without sufficient
explanation for the performance observed. In the present study, a deep study
of the mechanism of charge storage for both electrodes allowed the system to
be optimized.
Activated carbons (AC) are usually used for symmetric
electrochemical double layer capacitors owing to their high specific surface
area5. However, pseudocapacitance properties of AC in aqueous medium
were shown when a negative polarization is applied15. Figure 3 presents the
voltammetric analysis of AC in 2 mol.L-1 KNO3 showing that high
overpotential values for di-hydrogen H2 evolution could be reached, up to ca.
0.5 V (the thermodynamic value in this medium is around –0.37 V vs NHE).
Hydrogen produced by water decomposition at potentials lower than the
thermodynamical values is immediately adsorbed in the carbon porosity15,
being the reason of the pseudocapacitive behavior.
450
300 Activated Carbon Manganese Oxide
150
0
C,F/g
-1.2 -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4
-150
-300
-450
-600
E vs NHE/ V
Figure 3. Cyclic voltammograms in three-electrode cells for activated carbon and for a-
MnO2·nH2O loaded with 15wt% of carbon nanotubes in 2 mol.L-1 KNO3 medium using Pt as
auxiliary electrode.
Figure 3 also shows that manganese oxide has a very interesting

behavior under positive polarization giving a high electrochemical
38
reversibility up to 1.2 V. Therefore, by combining these two materials in an

asymmetric system with activated carbon and amorphous manganese oxide
as negative and positive electrodes, respectively, it is theoretically possible
to obtain a working voltage window of 2.2 V in aqueous medium. Figure 4
presenting the voltammetry curve for this kind of asymmetric capacitor,
confirms a quite rectangular shape characteristic of an ideal capacitive
behavior up to 2.2 V.
2
I/mA
-2
-4
-6
-8
-0.1 0.1 0.3 0.5 0.7 0.9 1.1 1.3 1.5 1.7 1.9 2.1 2.3
Cell voltage/ V
Figure 4. Cyclic voltammogram of an asymmetric capacitor with activated carbon and

D-MnO2·nH2O as positive and negative electrodes, respectively, in 2 mol.L-1 KNO3 medium.
In practicality, after optimizing the system, it was possible to obtain

a working voltage window of 2 V in aqueous medium, while keeping an
excellent cyclability. In these conditions, the energy density of the hybrid
capacitor reaches 12.6 Wh.kg-1 that is an order of magnitude higher than in
symmetric systems with activated carbon or manganese oxide in aqueous
medium, and relatively close to the values of electric double layer capacitors
working in organic electrolytes. Due to the various advantages of the
aqueous electrolyte medium, it is worth to deeply investigate the systems
based on this concept.
4. CONCLUSION
For the first time, a-MnO2·nH2O based composites have been studied
in real two electrode capacitors. The a-MnO2·nH2O/CNTs electrodes
demonstrate an ideal capacitive behavior and high values of capacitance.
Compared to the conventional carbon black, multi-walled CNTs are a very
promising conductivity additive for capacitor or battery electrodes.
39
The performance of manganese oxide in a real two-electrode

capacitor is limited by the irreversible reaction Mn(IV) to Mn(II) taking
place at the negative electrode, which potential depends on the electrolyte
pH.
A hybrid capacitor was optimized by combining two materials with
a pseudocapacitive behavior in aqueous medium, i.e. the a-
MnO2·nH2O/CNTs composite and an activated carbon as positive and
negative electrodes, respectively. The optimized system gives a practical cell
voltage of 2 V in aqueous medium, with energy densities close to the values
obtained with electric double layer capacitors working in organic
electrolytes.
ACKNOWLEDGMENTS
This research was supported by a Marie Curie fellowship of the
European Community program “Improving Human Research Potential and
the Socio-Economic Knowledge Base” under contract number HPMF-CT-
2001-01453 and by a NATO Science for Peace grant number 973849.
REFERENCES
1. Zheng JP, Jow TR. A new charge storage mechanism for electrochemical capacitors. J.
Electrochem. Soc. 1995; 142(1):L6.
2. Long JW, Swider KE, Merzbacher CI, Rolison DR. Voltammetric characterization of
ruthenium oxide-based aerogels and other RuO2 solids: The nature of capacitance in
nanostructured materials. Langmuir 1999; 15(3):780-5.
3. Wu NL. Nanocrystalline oxide supercapacitors. Mat. Chem. Phys. 2002; 75:6-11.
4. Conway BE, Birss V, Wojtowicz. The role and utilization of pseudocapacitance for
energy storage by supercapacitors. J. Power Sources 1997; 66:1-14.
5. Frackowiak E, Béguin F. Carbon materials for the electrochemical storage of energy in
capacitors. Carbon 2001; 39:937-50.
6. Lee HY, Goodenough JB. Supercapacitor Behaviour with KCl Electrolyte. Journal of
Solid State Chem. 1999; 144:220-3.
7. Lee HY, Kim SW, Lee HY. Expansion of active site area and improvement of kinetic
reversibility in electrochemical pseudocapacitor electrode. Electrochemical and solid
state letters 2001; 4(3):A19-A22.
8. Pang SC, Anderson MA, Chapman TW. Novel Electrode Materials for Thin-Film
Ultracapacitors: Comparison of Electrochemical properties of Sol-Gel Derived and
Electrodeposited Manganese Dioxide. J. Electrochem. Soc. 2000; 147(2):444-50.
9. Jiang J, Kucernak A. Electrochemical supercapacitor material based on manganese
oxide: preparation and characterization. Electrochim. Acta 2002; 47:2381-6.
10. Hu CC, Tsou TW. Capacitive and textural characteristics of hydrous manganese oxide
prepared by anodic deposition. Electrochim. Acta 2002; 47:3523-3532.
11. Jeong YU, Manthiram A. Nanocrystalline manganese oxides for electrochemical
capacitors with neutral electrolytes. J. Electrochem. Soc. 2002; 149(11):A1419-A1422.
12. Hu CC, Tsou TW. Ideal capacitive behaviour of hydrous manganese oxide prepared by
anodic deposition. Electrochem. Comm. 2002; 4:105-9.
40
13. Hong MS, Lee SH, Kim SW. Use of KCl aqueous electrolyte for 2 V manganese
oxide/activated carbon hybrid capacitor. Electrochemical and solid state letters 2002;
5(10):A227-A230.
14. Delpeux S, Szostak K, Frackowiak E, Bonnamy S, Beguin F. High yield carbon
nanotubes from the catalytic decomposition of acetylene on in-situ formed Co
nanoparticles. J.Nanosci.Nanotec. 2002; 2: 481-84.
15. Jurewicz K, Frackowiak E, Béguin F. Towards the mechanism of electrochemical
hydrogen storage in nanostructured carbon materials. Appl. Phys. A 2003; in press.
DEVELOPMENT OF SUPERCAPACITORS
BASED ON CONDUCTING POLYMERS
V. Khomenko1, 3, E. Frackowiak2, V. Barsukov1*, F. Béguin3

1
Kiev National University of Technology and Design
Nemorovicha-Danchenko 2, Kiev 02001, Ukraine
2
Poznan University of Technology, ul. Piotrowo 3, 60-965 Poznan, Poland
3
CRMD, CNRS-University, 1B rue de la Férollerie, 45071 Orléans Cedex 02, France
Abstract
Electronically conducting polymers (ECPs) such as polyaniline

(PANI), polypyrrole (PPy) and poly(3,4-ethylenedioxythiophene) (PEDOT)
have been applied in supercapacitors, due to their excellent electrochemical
properties and lower cost than other ECPs. We demonstrated that multi-
walled carbon nanotubes (CNTs) prepared by catalytic decomposition of
acetylene in a solid solution are very effective conductivity additives in
composite materials based on ECPs. In this paper, we show that a successful
application of ECPs in supercapacitor technologies could be possible only in
an asymmetric configuration, i.e. with electrodes of different nature.
1. INTRODUCTION
The most common electrical energy storage devices are capacitors
and batteries. Rechargeable batteries generally provide a high energy and a
rather low specific power. Usually, capacitors can supply high specific
power, but the amount of energy stored is very low. Electrochemical
capacitors (often called supercapacitors or ultracapacitors) [1,2] can store the
electric charge and the corresponding energy at high density with an
excellent reversibility, as a regular capacitor, and hence can be operated at
substantially greater specific power than most batteries. Supercapacitors can
be classified basically into two different categories depending on the charge
storage mechanism. Thus, there are the electric double layer capacitors
(EDLCs) generally based on activated carbon materials and the so-called
pseudocapacitors based on electronically conducting polymers (ECPs) [3-6]
or metal oxides [7,8]. ECPs are promising materials for the realization of
*
41
42
high performance supercapacitors, as they are characterized by high specific

capacitances. Indeed, the charge storage with ECPs involves all the polymer
mass and not only the surface, as in the case of double-layer capacitors based
on activated carbons. Furthermore, the charge–discharge processes are
generally fast with ECPs. However, compared to the activated carbons,
which are commonly used as electrode materials in commercial
supercapacitors, the ECPs still present some important drawbacks (e.g. poor
stability), which preclude an industrial utilization. In order to make a
progress towards the application of ECPs in supercapacitor technologies, it is
necessary to improve the materials quality and their electrochemical
properties.
In this paper, we report on the preparation of ECP composites based
on carbon materials. In parallel with the development of the preparation
processes and the electrochemical characterization of composites, we have
performed an analysis of the supercapacitor cell design based on ECPs.
2. EXPERIMENTAL
Electronically conducting polymers (ECPs) such as polyaniline
(PAN), polypyrrole (PPy), and poly(3,4-ethylenedioxythiophene) (PEDOT)
and their composites have been chemically polymerized using an oxidative
polycondensation of the corresponding monomers. Carbon nanotubes
(CNTs) and commercial thermally exfoliated graphite (TEG) from Superior
Graphite Co., Chicago, USA, have been used for making the composites.
The carbon nanotubes were synthesized by catalytic decomposition of
acetylene on a solid solution precursor [9]. ECPs were efficiently deposited
by chemical polymerization of the monomer on the carbon support well
dispersed in the reaction mixture under sonification, leading to
homogenous composite materials. The proportion of the components was
estimated by weighing the final product in the dried state. In some cases,
acetylene black has been used for making the composite in order to
determine if there is a beneficial effect of CNTs or TEG compared to this
conducting additive. Scanning Electron Microscopy (SEM, Hitachi 4200)
and Transmission Electron Microscopy (TEM, Philips CM20) have been
used for a detailed observation of the composites.
The electrodes were 1 cm2 pellets obtained by pressing the
composite material. The pellets were pressed onto a gold current collector.
The electrolytes were 1 M H2SO4 or 2M KNO3 aqueous solutions, due to
their high ionic conductivity. Two and three-electrode cells were used for the
measurements. The two-electrode cells were built using a Swagelok® type
construction, with a glassy fibrous separator and gold current collectors. In
the case of the three-electrode cell, a Pt wire and Hg/Hg2SO4 were used as
counter and reference electrodes, respectively. The electrochemical
43
measurements were performed using a VMP multichannel potentiostat-

galvanostat (Biologic, France).
Scanning and transmission electron microscopies show that CNTs

were coated with a homogeneous layer of ECPs (up to 50 nm). In the case of
the composites based on TEG, ECP fibers with diameters up to 1 Pm are
observed on the TEG particles.
The electrochemical properties of all types of ECPs have been
investigated using cyclic voltammetry, galvanostatic charge-discharge
cycling and electrochemical impedance spectroscopy. It was established that
an amount of carbon material ca. 20 wt% provides a good electronic
conduction in the electrode when the polymer is in its isolating state.
Acetylene black and especially TEG supply quite good electrochemical
performance of the composites, however, CNTs as electrode component
gave definitively a more homogenous dispersion of ECPs, that provided a
better charge propagation. Also the CNT-based nanocomposites materials
presented good mechanical properties, and swelling or shrinkage of the
electroactive polymer during cycling did not lead to any electrode
degradation.
Our previous investigations [10, 11] of the current-producing
mechanism for ECP electrodes have shown that ECPs present a pseudo-
faradaic behavior, at least within the potential range where they are at their
low oxidation degree, while at higher oxidation degree, the introduced
charged centers have a tendency to be delocalized, giving a metal-like
character to the ECPs. In the latter case, the energy is stored in the
conducting polymer by the accumulation of charges in the electrical double
layer, mainly by electrostatic forces, without phase transformation in the
electrode material. According to our data, it is possible to estimate an
extremely high capacitance value (about 1000 F/g) for the potential range
where the Faradaic process takes place, and a significantly lower value for
the other potential range. Since the electrodes of a real supercapacitor work
in a different potential range, the doped state of each electrode is different.
Hence, the capacitance values are different for both electrodes. Taking into
account the formula 1/C = 1/C1 + 1/C2 for capacitors in series, one can see
that the total capacitance of the device is imposed by the smallest value.
Figure 1 shows as an example the galvanostatic discharge of the
PPy/CNTs composite in a three-electrode cell. It is easily seen that the slope
of the curve strongly depends on the potential range. Based on the above
equation for capacitors in series and taking into account the values of
specific capacitance of the PPy composite materials according to Fig. 1, i.e.
250 F/g (C1) calculated for the potential range from 0.2 V to –0.3 V and 903
44
F/g (C2) for the negative range from -0.3 V to –0.6 V, the overall capacitance
combining two electrodes working in these potential ranges will be the
following:
C1 u C 2 250 u 903
C 196 F / g
C1 C 2 250 903
This value of capacitance is well consistent with the experimental

value of 192 F/g measured in a two-electrode cell. Hence, one must be
extremely careful that the values of capacitance reported for ECPs in
literature are generally obtained from three-electrode cell measurements. The
above example shows clearly that in the case of a real two-electrode
capacitor the capacitance values are always smaller than in a three-electrode
cell configuration.
Figure 1. Galvanostatic discharge curve (three-electrode cell), at 2 mA,

for a PPy/CNTs pellet electrode; m= 9.6 mg.
The cycling stability upon charge/discharge of a symmetric capacitor

based on PPy composite electrodes was tested over several hundred cycles
under constant current, and the variation of specific capacitance is illustrated
in Fig. 2. These tests demonstrate that using the same material for both
electrodes provides a poor cycling stability of the supercapacitor if the
voltage range exceeds some limit. The cycleability of the PPy based
capacitor is good when the cell voltage is 0.4 V during cycling. Upon
cycling at U = 0.6 V, the capacitance is 20% less than the initial one after
45
500 cycles, and the loss reaches almost 50% when U=0.8V. A similar
behaviour was observed for symmetric capacitors based on PANI or
PEDOT.
Traditionally, many authors attribute the poor cycling life of ECPs
based capacitors to the ECPs degradation process, mainly by oxidation.
However, the analysis of the polyaniline electrochemical behaviour [11]
proved that the redox processes taking place at the limits and beyond a given
potential range can lead to the quick formation of thin passive layers of poor
conductivity. Such new phases (e.g. leucoemeraldine salt, leucoemeraldine
base etc.) rapidly inhibit and discontinue the electrochemical process playing
the role of a switch limiting the whole supercapacitor performance.
Therefore, a high cycleability (more than one million cycles according to
some literature claims [12, 13] without significant change of the ECPs can
be only realized using a three electrode cell in a strict potential range. By
contrast, in an ECP-based two-electrode capacitor configuration, one cannot
strictly control the potential of the electrodes, and the potential shift of one
electrode to the region of ECP electrochemical unstability can provoke a
capacity fade.
The galvanostatic tests, using a three electrode cell, on the negative
electrode after the galvanostatic cycling of the PPy based supercapacitor at
U=0.8 V (see Fig. 2), confirmed a noticeable capacity fade. Figure 3 (curve
1) shows that the capacitance of the negative electrode is about 160 F/g in
the potential range from -0.1 V to -0.7 V vs. Hg/Hg2SO4 where this electrode
works, while it was 610 F/g (curve 2 of Fig. 3) before cycling of the
supercapacitor. It means that the potential of the negative electrode reached
such a value that it achieved a full ECP undoping to the isolating state, being
at the origin of the observed capacity fade due to the isolation of a noticeable
part of the ECP from the charge/discharge process.
In order to solve this drawback, one could propose to recover the
electrode capacitance by applying a more positive polarisation to the
negative electrode that would allow the isolating phase to be converted into
the conducting one. Figure 3 (curve 3) shows the galvanostatic charge of the
negative electrode containing the isolating phase up to ca. 0.2 V. According
to this curve the doping of the ECP mainly takes place (at least 80% of the
total charge) at potentials above -0.1 V. Afterwards, the electrode was
discharged, giving the curve 4 (Figure 4) with a capacity comparable to the
initial state (curve 2), which confirms the re-doping of the electrode.
In a symmetric capacitor, where the cell voltage is imposed during
galvanostatic cycling, it is impossible to realise such a potential shift for the
negative electrode. Thus, the isolating phase accumulates in the negative
electrode upon cycling of the supercapacitor. This process is strongly
accelerated by an increase of the operating voltage, because the potential of
the negative electrode is shifted to more negative values, what is
experimentally supported by the data presented in Figure 2.
46
47
It is interesting to note in Figure 2 that the value of the initial

specific capacitance is more important at high voltage, but it rapidly
decreases with cycling.
This suggests that the pseudocapacitance effects, which are at the
origin of these high values gradually disappear with cycling (already after 50
cycles). Impedance spectroscopy was used in order to find an interpretation
to this phenomenon, and the spectra of a supercapacitor based on PPy/CNTs
composite electrodes before galvanostatic cycling and after 500 cycles at a
voltage of 0.8 V are reported in Figure 4. The supercapacitor resistance at 10
kHz increased from an initial value of 0.35 Ohm to 1.43 Ohm after 500
cycles, that proves that the electrodes resistance also increased. Hence, one
can conclude that irreversible redox transitions to the isolating state are at the
origin of the high values of capacitance observed under high voltage during
the first galvanostatic cycles.
Figure 4. Impedance spectra of a supercapacitor based on the PPy/CNTs composite before

(a) and after performing 500 galvanostatic charge/discharge cycles (b).
Mass of the working electrodes: 8 mg.
The appearance of a semicircle at high frequency after cycling

(Figure 3b) suggests a parallel existence of a charge transfer resistance,
which is related with the already mentioned pseudofaradaic processes, i.e.
with the formation of the isolating phase.
In our opinion, the above experiments proved that, for an optimal
usage of the conducting polymer, the potential range of the negative
electrode must be strictly controlled in order to prevent the ECP to reach its
48
insulating state. In a symmetric supercapacitor this might occur only when

the cell voltage is very low, that consequently limits the energy stored.
An asymmetric configuration, with two electrodes of different nature
[4,5,14], is an excellent way to overcome the problems mentioned before.
The main target, designing these systems, is to optimize the potential range
of each electrode, taking into account the need to obtain a high cell voltage
and a good cycling life of the supercapacitor as a whole. Thus, we have
investigated the combination of different ECP electrodes or of an ECPs
electrode and another electrode from an activated carbon or a metal oxide.
For example, an asymmetric configuration was built where the ECPs/CNTs
composite was the positive electrode and the commercial activated carbon
(PX-21) was selected for the negative electrode. We have also tested
amorphous manganese oxide, a-MnO2, as positive electrode material in
asymmetric supercapacitors where the ECPs/CNTs composites served as
negative electrode. The advantages of a-MnO2 are a pseudo-faradaic
mechanism of charge storage allowing high capacitance values to be
expected, a low cost and an environment friendly nature [7].
Figure 5. Cyclic voltammograms at 10 mV/s in three-electrode cells for PPy and for D-MnO2
based electrodes in 2 mol.L-1 KNO3 medium using Pt as auxiliary electrode.
Typical cyclic voltammograms of a-MnO2 and PPy in 2 mol.L-1

KNO3 are shown in Fig. 5. The electrochemical potential window of
manganese dioxide is spread to more positive values than for activated
carbons and ECPs, keeping high electrochemical reversibility up to 1.2 V.
The electrochemical activity of PPy is in the range which should not include
its redox undoping reaction at potentials below -0.2 V vs. NHE. Therefore,
by combining these two materials in an asymmetric system, it is possible to
obtain a working voltage window of 1.4V (U=EMnO2 – Eppy= 1.2 - (-0.2) =
1.4), which will ensure a safe performance of both electrodes during long
cycling. Thus, the above mentioned situation of deactivation of the negative
49
ECP based electrode of a symmetric capacitor would not occur when a-

MnO2 will be used as a positive electrode in an asymmetric configuration.
Table 1. Electrochemical characteristics of some asymmetric supercapacitors. U is the

maximum available cell voltage. Cs is the specific capacitance of a pellet electrode calculated
from Cs = Ct*4/M, where Ct is the capacitance of the asymmetric supercapacitor, M is the
total mass of both electrodes.
Electrode materials Supercapacitor characteristics
Positive Negative Cs (F/g) U (V)
PANI PPy 280 0.6
PANI PEDOT 160 0.8
PANI Carbon (PX21) 330 1.0
D-MnO2 PANI 160 1.4
PPy Carbon (PX21) 220 1.0
MnO2 PPy 140 1.6
PEDOT Carbon (PX21) 110 1.2
MnO2 PEDOT 120 1.8
The electrochemical performance of all types of asymmetric

supercapacitors has been evaluated by galvanostatic charge/discharge tests,
and the results are presented in Table 1. Combining an activated carbon for
the negative electrode and the ECP/CNTs composite for the positive
electrode significantly extends the practical voltage performance up to 1.2 V
in aqueous electrolyte and provides a long cycling life of the supercapacitor.
The most interesting results, which confirm our prediction presented above,
were obtained with the newly designed hybrid electrochemical capacitor
using the ECP/CNTs composite as negative electrode and manganese
dioxide as positive electrode. Almost all of the charge residing on the ECP
electrode can be used and the cell voltage reaches up to 1.8 V in aqueous
electrolyte. Using such an asymmetric construction, the values of specific
energy and specific power in aqueous electrolyte increase by more than 4
times compared with symmetric ECPs based capacitor. It should be
additionally mentioned that the asymmetric supercapacitor was cycled over
1000 cycles without any noticeable decrease of the capacitance value.
4. CONCLUSION
Our data demonstrate that the symmetric configuration of ECP-
based capacitor is not realistic, mainly because one cannot control the
potential of the negative electrode during cycling. When the voltage range
exceeds ca. 0.4 V, the negative electrode progressively shifts to its isolating
state that leads quickly to a capacity fading of the system. Consequently, for
a safe cycling of a symmetric ECP-based capacitor, the voltage should not
exceed 0.4V, that makes the system of poor interest in term of energy stored.
50
In order to circumvent these drawbacks, we propose an asymmetric

configuration with two different materials working in their optimal potential
range for both electrodes. The combination of two different materials allows
the operating voltage range to be extended up to 1.8 V in aqueous medium,
keeping a good cycleability of the material without a noticeable aggravation
of the capacitance values. Due to their environment friendly nature, these
new high energy systems are quite promising in replacement of organic
electrolyte capacitors.
ACKNOWLEDGMENTS
This work was supported by the NATO Science for Peace “Carbon”
programme No. 973849.
REFERENCES
1. Conway B.E. in Electrochemical Supercapacitors. Kluwer Academic/Plenum Publishers,
New York, 1999.
3. Laforgue A., Simon P., Sarrazin Ch., Fauvarque J-F. Polythiophene-based
supercapacitors. J Power Sources 1999; 80:142-8.
4. Mastragostino M., Arbizanni C., Soavi F. Polymer-based supercapacitors. J Power
Sources 2001; 97-98:812-5.
5. Mastragostino M., Arbizzani C., Soavi F. Conducting polymers as electrode materials in
supercapacitors. Solid State Ionics 2002; 148:493-8.
6. Jurewicz K., Delpeux S., Bertagna V., Béguin F., Frackowiak E. Supercapacitors from
nanotubes/polypyrrole composites. Chem Phys Lett 2001; 347: 36-40.
7. Toupin M., Brousse T., Bélanger D. Influence of microtexture on the charge storage
properties of chemically synthesized manganese dioxide, Chem Mater 2002; 14:3946-52.
8. Wu N.L. Nanocrystalline oxide supercapacitors. Mater Chem Phys 2002; 75:6-11.
9. Delpeux S., Szostak K., Frackowiak E., Bonnamy S., Béguin F. High yield of pure
multiwalled carbon nanotubes from the catalytic decomposition of acetylene on in-situ
formed cobalt nanoparticles. J. Nanosc. Nanotech. 2002; 2: 481-4.
10. Barsukov V., Chivikov S. The “Capacitor” concept of the current-producing process
mechanism in polyaniline-type conducting polymers. Electrochim. Acta 1996; 41: 1773–
79.
11. Barsukov V.Z., Khomenko V.G., Chivikov S.V., Barsukov I.V., Motronyuk T.I. On the
faradaic and non-faradaic mechanisms of electrochemical processes in conducting
polymers and some other reversible systems with solid-phase reagents. Electrochim.
Acta 2001; 46: 4083–94
12. Kobayashi T., Yoneyama H., Tamura H. Polyaniline film-coated electrodes as
electrochromic display devices J.Electroanal.Chem., 1984; 161: 429-23
13. Kobayashi T., Yoneyama H., Tamura H. Electrochemical reactions concerned with
electrochromism of polyaniline film-coated electrodes. J.Electroanal.Chem 1984; 177:
281-91
14. Arbizzani C., Mastragostino M., Soavi F. New trends in electrochemical supercapacitors.
J Power Sourc 2001; 100:164-170.
SUPERCAPACITORS:
OLD PROBLEMS AND NEW TRENDS
Yurii Maletin*, Natalie Strizhakova, Vladimir Izotov, Antonina Mironova,

Sergey Kozachkov, Valerie Danilin, Sergey Podmogilny
National Technical University of Ukraine “KPI”, Department of Physical Chemistry,
37 Pobedy Prospekt, Kiev 03056, Ukraine
1. INTRODUCTION
Supercapacitors (SC) attract much attention as a good alternative or
supplement to batteries. Over the past decade a considerable breakthrough
took place in our understanding of how the supercapacitors function [1-3].
Though they are inherently behind the batteries by the specific energy stored,
they are way ahead of them by their peak output power, efficiency and
cycling life. As a result they can be used as pulse power sources, for load
leveling the batteries, in hybrid power supply units (e.g., being combined
with a fuel cell), etc. This paper is an attempt to analyze what are the
prospects for SC to occupy their niche in the market of energy storage
devices, what performance can be anticipated in near future, and how this
performance might be achieved with the special emphasis on the SC power
output and its dependence on the electrode properties.
2. PRINCIPLE OF OPERATION AND KEY

PERFORMANCE CHARACTERISICS
Active electrochemical part of a typical SC device represents two
conductive porous (normally carbonaceous) electrodes with a dielectric
porous film (separator) interposed between them; the system being
impregnated with a liquid electrolyte. The double electric layer forms at the
electrode/electrolyte interface on charge, and this layer discharges during
operation, making it possible for the electric energy to be stored and then to
be delivered to the load in due time. The electrolyte can be either aqueous or
organic aprotic one, and this differs SC of Generation 1 from Generation 2, if
John Miller’s [4] SC classification is used. As an alternative, hybrid devices
*
Corresponding Author. E-mail: [email protected]
51
52
can also be designed wherein one of carbonaceous “double layer” electrodes

is replaced by an “electrochemical” electrode taken from the battery
technology, e.g. NiO(OH) or PbO2 in aqueous electrolyte (Generation 3 [4])
or Li4Ti5O12 in aprotic electrolyte (Generation 4 [4]).
The specific energy stored and the specific power that can be
delivered to the load are no doubt the crucial characteristics of a SC device
along with others, such as its cycling life, self discharge current and
efficiency.
Specific energy is proportional to the capacitance and voltage
squared, and it should be noted that both of these parameters are close to
their limit in modern SC devices comprising two carbonaceous electrodes
and organic electrolyte. Double layer capacitance value is ca. 12(r2)
PF/sq.cm for porous carbon materials [5], and their accessible specific
surface area has practically reached the reasonable limit of ca. 1200 sq.m/g
(readers should note that ultramicropores with the diameter less than 0.7÷0.8
nm are beyond the reach of organic electrolytes comprising relatively large
species). This gives the specific capacitance of about 150 F/g for a carbon
electrode in organic electrolytes, and this value has almost been reached in
the best SC prototypes. As regards to the rated voltage, it can hardly be
higher than 3 V for a SC unit cell with known electrolytes, and the value of
2.3y2.5 V looks even more feasible if the long cycling life is required [6]. As
a result, over a few past years the specific energy of the best SC devices with
organic electrolytes has not exceeded 5 W.h/kg [7], and the value of 10
W.h/kg can be considered as the practical limit. For hybrid devices with
dissimilar electrodes (vide supra) a bit larger specific energy can be obtained
[4], though at the expense of lower power output and shorter cycling life.
High specific power value accompanied by long cycling life seems
to be the most attractive advantage of SC devices since this is the
performance which enables them to be effectively used for load leveling the
batteries in electric or hybrid vehicle applications. However, to be
competitive in the battery market, SC must demonstrate the specific power of
about or higher than 10 kW/kg for a charge/discharge at ca. 95% efficiency.
This power, p, can roughly be estimated using the equation (1) [7,8] as
follows:
9(1 K )U 2
p , (1)
16 RM
where K is the efficiency, U is the voltage in V, R is the equivalent series

resistance in Ohm, and M is the mass in kilos.
This means the equivalent series resistance of SC should notably be
reduced in order to increase their specific power. Hence, as can be shown,
the electrode resistivity should be reduced significantly – down to 0.5
53
Ohm.cm2 (of visible surface area) or even less. Some visible ways to lower
the electrode resistivity are: (a) fabricating thin electrodes, (b) increasing the
conductivity of electrolyte and electrode material, (c) matching the electrode
porosity with the size of species in the electrolyte in order to facilitate the ion
transport along the pores, and (d) assembling bipolar devices.
3. OPTIMIZATION OF THE ELECTRODE THICKNESS

In order to see how the electrode thickness might be optimized in
order to provide the lowest electrode resistivity, we have developed a
theoretical model to describe the charge/discharge processes in porous
carbon electrodes. As a first approximation, let us consider an electrode
having two sets of cylindrical pores, namely, nanopores (NP) of less than 3
nm in diameter and transport channels (TC) of more than 20 nm in diameter,
with each nanopore having an exit to only one TC‡.
The distribution of potential in TC is practically the same as that
near the flat surface if the electrolyte concentration is about 1 mol/l [2]. So
the discharge of TC may be considered as that of a double electric layer
formed at the flat electrode surface/electrolyte solution interface, and hence,
an equivalent circuit for the TC discharge may be presented as an RC circuit,
where C is the double layer capacitance and R is the electrolyte resistance.
On the other hand, the gradient of potential is negligible within the
NP [2]. That is why, here we will consider the process of NP charging or
discharge as a diffusion process caused by the concentration gradient, which
results from the concentration change within the diffused part of a double
layer near the region where the NP face the surface of TC. As a result, an
equivalent circuit for this electrode section may be represented by two RC
circuits connected in parallel (see the corresponding portion of Fig. 2 below).
Now let us consider a model for a SC device that comprises two
electrodes (anode and cathode), each of them being electrically connected to
a current collector fabricated of Al foil. Let two of such collectors have a
certain thickness of GAl. As an electrode material, an activated carbon powder
is considered below. Anode and cathode are interposed with a separator of
thickness Gs. The electrodes and separator are impregnated with electrolyte.
In this paper we mostly focus on the optimization of SC performance by
varying the electrode thickness, while some other effects will briefly be
considered in the next section.
As mentioned above, the electrode material comprises two types of
pores – NP and TC, NP being mainly responsible for the electrode
capacitance value. On the other hand, TC, which contribute far less to the
‡
Such a choice of pore shape and distribution leads to a relatively simple
mathematical formulation without loss of generality.
54
capacitance, are responsible for the ion transport through the electrode body
up to the boundary between the electrode and separator. Let the electrode
thickness be of Ga or Gc for anode or cathode correspondingly, and let the TC
and NP be equally distributed within the electrode body.
Capacitance of a single pore is proportional to its surface area.
Therefore, for a cylindrical NP it is proportional to 2Srl, where r is the NP
radius, and l is its length. Inner resistance of NP is proportional to l as well.
The NP parameters depend on how the activated carbon powder was
obtained but they are normally not affected by treating the powder when
fabricating the electrode.
On the contrary, the shape and dimensions of TC depend on both the
method for obtaining the porous carbon powder and the method for making
electrode. Within the framework of our model, TC are considered to be the
straight channels passing through an electrode such that their length is equal
to the electrode thickness. As another model approximation, all the NP are
assumed to face the TC forming a sort of tier structure, i.e. the NP exits are
equally distributed along the TC circle and length – see Fig. 1. Let the
distance between tiers be of 4r. This approximation simplifies mathematical
derivations significantly though does not distort a real physical picture.
Number of NP, n, facing TC in a tier may be determined from the volumetric
concentration of NP, [, and from the number of TC, P, coming through the
unit electrode section. Thus the n value can be determined as n = 4r[/P.
Figure 1. Model presentation of a few nanopore tiers facing a transport channel.
Now an equivalent circuit, which takes into account both the ion
transport along the TC and the charge transfer through the carbon electrode
material to the current collector, may be represented as in Fig. 2, wherein N =
Ga(c)/4r, Cm and Rm are the total NP capacitance and resistance in a unit
electrode volume (defined here as a product of a unit electrode area and the
tier thickness), Re is the electrical resistance of an electrode in the same unit
55
volume, and Ri is the electrolytic resistance in TC located in the same unit

volume. It should be noted that Rm >> Re, Ri in real systems.
Figure 2. An equivalent circuit of a porous carbon electrode in SC.
Electrical parameters of the system under consideration have been

estimated using an equivalent circuit presented by Fig. 2. Thus, the anode
capacitance can be written as follows:
Ca = Cm(a) Ga/4r (2)
However, it is not so easy to derive a similar functional relationship

between the electrode resistance and its thickness that is why a computer
simulation has been carried out to study the dependence of electrode
resistance on its thickness. In the calculations, Rm was chosen as a unit
resistance, and the distance between the NP tiers (4r) was chosen as a unit
thickness. Figure 3 illustrates the results of calculations at various ratios
between Re, Ri and Rm, namely, Ri = Re = 0.002Rm (curve 1); Ri = 0.004Rm, Re
= 0.0001Rm (curve 2). In these and other cases checked, the plots of
resistance vs. thickness display a sharp drop followed by a slower rise after
reaching a minimum. Let us consider these regions in more detail.
For two limiting cases – dip or rise of curve or, in other words, thin
or thick electrode – the problem can be solved asymptotically. In case of a
thin electrode, i.e. if G 4 rR m , the following expression for electrode
Ri R e
resistance can be derived:
4r . (3)
R Rm
G
4 rR m
In another limiting case, when G !! , the dependence of R on the
Ri R e
electrode thickness G can be presented as follows:
R U ef G , (4)
56
Figure 3. Plot of ESR of an electrode vs. its thickness measured in number of tiers.
where Uef is the effective resistivity including the electrical resistivity of

electrode material and the resistance of ion transport along the TC. These
two conjugated components of Uef cannot practically be separated. On the
other hand, this is the ion transport, which normally limits the electrode
resistance since not one porous electrode impregnated with electrolyte and
charged, can then be discharged until counter-ions leave it. This means that
conductivity of the electrolyte and its diffusion along the pores must play a
crucial role for reducing the electrode resistivity (vide infra).
From the practical point of view, this is the discharge of a SC device
under constant power conditions that is normally of the most interest. That is
why the present work is aimed at determining the optimum electrode
thickness that enables one to obtain the maximum energy, E, output (referred
to as unit of volume or mass) if the discharge with a fixed power takes place.
For the sake of simplicity we will speak about the energy density (E) and
power density (p), but all the expressions derived below can easily be
transformed to obtain the specific energy or power, if the volume is
substituted by mass.
As a first step we have examined the discharge of SC having the
volume of V, capacitance of C and inner resistance of Rin. For such a
capacitor charged up to the voltage U, and then discharged to a nonlinear
load to keep the constant power P until the voltage decreases to U/2 the
expression (5) for the energy density (( = E/V) and power density (p = P/V)
can be derived [3].
In the equation (5) the inner resistance Rin includes not only the
electrode resistance, but also the separator resistance Rs, defined as Rs =
57
UsGs/S, where Us is the resistivity of separator to ion transport and S is the

electrode (or separator) area. The quantities in Eq. (5) have the dimensions as
follows: ( in W.h/l, p in W/l, Rin in Ohm, C in F, and U in V.
U U 2 4 Rin pV 3 U U2
C (4 Rin pV ln( ) U U 2 4 Rin pV U 2 4 Rin pV )
U U2 4 2 4
4 Rin pV
ƪ
2 4 (5)
14400V
So far, this model differs from a real system because it does not take
into account a decrease in capacitance with an increase in current (charge or
discharge). As a result, it can be shown that calculations according to Eq. (5)
start deviating from experimental Ragone plots in the range where the
current values are high enough. Therefore, the dependence of capacitance on
current was introduced into the model.
From experimental measurements it was found that there was a fairly
good linear relation between the SC capacitance and current, the slope being
dependent on the electrode material and electrolyte. If the dependence is
known, it can be used to make the Ragone plot calculations more accurate.
Therefore, the following method was employed in our evaluations.
If the power value is constant on discharge, the current varies
evidently from the starting value is to the final one if. If one neglects the
potential drop due to inner resistance of a SC, the is and if values may be
determined as follows: is = p/U and if = 2p/U. The starting and final values
of capacitance (C(is) and C(if), correspondingly) may be evaluated from the
experimental C-i plot, and then the average value CC = (C(is) + C(if))/2 may
be put into Eq.(5). Some of the results of Ragone plot calculations using this
equation and the experimentally determined C-i dependencies are presented
by Figure 4.
As can be seen from Eq. (5), the maximum power density delivered
to the load is:
U 2c
p max (6)
16W
where c is the capacitance density, and W is the SC time constant. For further
calculations the following relationship may be used:
V = S(Gs + GAl + Ga + Gc) (7)
For the sake of simplicity let Ga = Gc = G below. Subsequently,

expressions (2), (3) and (7) can be put into Eq. (5) and the maximum of ( as
a function of G can be found, all the other quantities being constant. As a
58
result it can be shown that for “thin electrodes” ( G 4 rR m ) and under the
Ri R e
constant power discharge the maximum energy delivery can be achieved for
the following electrode thickness:
U G s G Al G s G Al
G (8)
4 pG s U s 2
4 rR m
In another limiting case ( G !! ) the maximum energy can be
Ri R e
delivered if the electrode thickness is as follows:
2 1
U 3
(G Al G s ) 3 U sG s 2 U ef (G Al G s )
G 0.3 1
(9)
( pU ef ) 3 6 U ef
4. VERIFICATION OF THE MODEL

If one substitutes now Eq. (6) into Eqs. (8) and (9) and then chooses
the values§ of GAl = 20 Pm, Gs = 30 Pm, Us = 150 :.cm, Uef = 60 :.cm, c = 15
F/cc, W = 0.2-0.3 s for thin electrodes and W = 1-2 s for thick ones, the
optimum electrode thickness G might be evaluated within the range of 50-150
Pm. That is why a series of experiments was carried out with the electrode
thickness being varied within this range. As an electrolyte, 1.3 M Et4NBF4 in
acetonitrile was used since, as was shown by various researchers and also in
our experiments, acetonitrile-based electrolytes were able to provide the
highest conductivity at room temperature as compared with other solvents -
of ca. 60 mS/cm. Composite electrodes of various thickness were fabricated
using the carbon powders produced by Skeleton Technologies as a result of
TiC chlorination at high temperature, Al foil of 10 Pm thick was used as a
current collector, and the Celgard 2400 material of 30 Pm thick was
employed to separate the electrodes. The Ragone plot for a SC prototype
having the electrodes of 120 Pm thick is shown in Fig. 4. As it can be seen,
the performance of the prototype is fairly good, and the suggested model is
able to describe the SC characteristics with a reasonable accuracy.
§
These values were estimated by us from preliminary experiments and
from the known data.
59
10
E, Wh/kg 1
1
100 1000 10000
P, W/kg
Figure 4. Ragone plots calculated for 2.3 V (1) and 3.0 V (2);
experimental dots were obtained in Dr. A. Burke's lab at 2.3 V.
5. ELECTROLYTE TRANSPORT THROUGH THE

CARBON ELECTRODES
As it has been mentioned above, this is the ion transport through the
carbon electrodes that limits the rate of SC charge or discharge, and hence
limits the specific power of a SC device.
0,8
L2,cm2
0,7
4
0,6
3
0,5
0,4
0,3
2 1
0,2
0,1
0
0 500 1000 1500 2000 2500 3000 3500 4000
t,s
Figure 5. Length (squared) of the electrode impregnation with acetonitrile vs. time
for various carbon materials produced by Skeleton Technologies: 1 – C/SiC; 2 –
C/TiC-1; 3 – C/TiC-2; 4 – C/Mo2C.
In accordance with the known equation: CL2 = 2Dt, the mean ion
displacement squared,CL2, at time t is proportional to diffusion coefficient,
D. Since we could not find any reliable data on diffusion coefficients for
60
organic electrolytes in porous carbon materials, we have measured the

electrolyte and/or solvent migration through the electrode body. With certain
reservations the transport of liquid and transport of ions through the porous
electrode body can be correlated similar to transport processes in porous
membranes [9]. Therefore, we have developed a simple visual method to
evaluate the rate of electrolyte transport through the electrode.
According to the method developed, a flat circular electrode is
placed between two sheet glasses under the controlled pressure, one of the
glasses having a tiny hole (a feeder) with a capillary near the center of the
electrode. A solvent or an electrolyte can be fed to the electrode through the
feeder, which is sealed to avoid the solvent evaporation. Then the migration
of solvent or electrolyte (impregnation of the electrode) can be observed with
microscope, and Fig. 5 illustrates the length of migration (squared) vs. time
for different electrode materials.
2,5
Vsolv ,cc/g
1,5
1
y = 1,37x + 0,24
2
R = 0,99
0,5
0
0 0,5 1 1,5
VDFT,cc/g
Figure 6. Plot of the specific electrode pore volume vs. the specific pore volume measured for
the corresponding carbon powders.
As can be seen from Fig. 5, a fairly good linear correlation takes

place for various carbon materials, the slope 2D being equidimensional with
the diffusion coefficient and showing the rate of electrolyte migration
through the electrode (more precisely, the volumetric rate of flow multiplied
by the length L). In accordance with the Hagen-Poiseuille equation, for
porous materials of the type presented in Fig. 1 the volumetric rate of flow
must be proportional to the pr2 product, where p is the porosity of the
material, and r is its mean pore radius [10]. The gravimetric porosity (the
total pore volume per 1 gram of the material) was measured by two methods.
The first method was employed for carbon powders and was based on the
analysis of N2 adsorption/desorption isotherms within the framework of the
Density Functional Theory (DFT) using the ASAP2000M unit (produced by
Micromeritics). The second method included the simple determination of
61
solvent volume absorbed by the corresponding electrode using the same two
glasses and a feeder as described above. The porosity values determined by
the two methods for various carbon powders and electrodes made of them
are plotted in Fig. 6, wherein the very left dot represents a non-activated
anthracite. A reasonably good correlation indicates that electrodes possess
about 0.24 cc/g of extra porosity due to macropores formed by carbon grains
when the electrode is manufactured.
If we assume a quasi-cylinder pore structure of the electrode material
as in Fig. 1, an average effective pore radius r can be evaluated from the
relationship r = 2V/A, where V is the total pore volume, and A is the total
pore surface that can also be determined using the DFT method (see also
[5]). Then the migration coefficients determined as shown in Fig. 5 can be
plotted vs. the pr2 product – see, e.g., Fig. 7 for five electrodes, which were
made of various porous carbons produced by Skeleton Technologies.
140
120
100
-5
D, cm /sx10
80
y = 0,139x - 17,735
2
60 2
R = 0,991
40
20
0
0 200 400 600 800 1000 1200
2
pr
Figure 7. Plot of the migration coefficient (determined from the slope as in Fig. 5) vs. the pr2
product of a carbon electrode.
Figure 7 illustrates the dynamics of fluid migration through porous

carbon electrodes to obey the Hagen-Poiseuille equation that is normally
used to describe the transport through membranes having the pores of
cylinder-like shape. Therefore, this method can probably be used for express
analysis of the electrolyte dynamics in different porous carbon materials.
6. CONCLUSIONS
1. In order to be competitive in the market of energy storage devices
SC must demonstrate the specific power output of about 10 kW/kg for a
charge/discharge at 95% efficiency.
62
2. In contrast to battery technology, ultra thin electrodes cannot provide

an increase in SC power density. Our theoretical model shows an optimum
electrode thickness to exist in the range of 50-150 Pm, the exact value
depending on the porosity and nature of the electrode material.
3. This is the electrolyte mobility in pores that crucially influences the
rate of electrode discharge and, hence, the SC power output.
4. A simple visual method to evaluate the rate of solvent/electrolyte
migration through the porous electrode body has been developed. The
method enables an express selection of carbon materials according to the
mobility of electrolyte in their pore structure.
ACKNOWLEDGEMENTS
This work was supported by the Science-and-Technology Center in
Ukraine (project STCU 1830). Financial support was also provided within
the framework of INTAS 00-761 research project. Porous carbon materials
derived from various carbides in accordance with methods, which were
developed by Skeleton Technologies, were received from Central Research
Institute for Materials (St. Petersburg, Russia).
REFERENCES
1. Conway, Brian, Electrochemical supercapacitors: scientific fundamentals and
technological applications. New York: Kluwer/ Plenum, 1999.
2. Maletin YA, Strizhakova NG, Izotov VY, Mironova AA, Kozachkov SG, Danilin VV,
Ekström T. Supercapacitor stacks of high energy & power density. Proceedings of the
7th International Seminar on Double Layer Capacitors and Similar Energy Storage
Devices, Deerfield Beach, Florida, USA, Dec. 8-10, 1997.
3. Maletin YA, Strizhakova NG, Izotov VY, Mironova AA, Kozachkov SG, Danilin VV,
Podmogilny SN. How to tune the EDLC components to improve its performance.
Proceedings of the 10th International Seminar on Double Layer Capacitors and Similar
Energy Storage Devices, Deerfield Beach, Florida, USA, Dec. 4-6, 2000.
4. Miller JR. The Evolution of Electrochemical Capacitor Technology: Future Performance
Projections. 2003 International Conference on Advanced Capacitors, Kyoto, Japan, May
29-31, 2003.
5. Calas-Blanchard C., Comtat M., Marty J.-L., Mauran S. Textural characterisation of
graphite matrices using electrolchemical methods. Carbon 2003; 41: 123-130.
6. Xu K., Ding S.P., Jow T.R. Towards Reliable Values of Electrochemical Stability Limits
for Electrolytes. J. Electrochem. Soc. 1999; 146: 4172-4178.
7. Burke AF, Miller M. Characteristics and Applications of Advanced Ultracapacitors.
Proceedings of the 12th International Seminar on Double Layer Capacitors and Similar
Energy Storage Devices, Deerfield Beach, Florida, USA, Dec. 9-11, 2002.
8. Burke AF. Ultracapacitors: why, how and where is the technology. J. Power Sources
2000; 91: 37-50.
9. Mulder, Marcel, Basic Principles of Membrane Technology. Kluwer Acad. Publ., 1995
(Russian Transl., Moscow, Mir Publ., 1999).
MODELING POROSITY DEVELOPMENT
DURING KOH ACTIVATION OF COAL
AND PITCH-DERIVED CARBONS
FOR ELECTROCHEMICAL CAPACITORS
Krzysztof Kierzek, Grazyna Gryglewicz, Jacek Machnikowski*
Institute of Chemistry and Technology of Petroleum and Coal

Wrocáaw University of Technology, GdaĔska 7/9, 50 344 Wrocáaw, Poland
Abstract
High porosity carbons ranging from typically microporous solids of

narrow pore size distribution to materials with over 30% of mesopore
contribution were produced by the treatment of various polymeric-type
(coal) and carbonaceous (mesophase, semi-cokes, commercial active carbon)
precursors with an excess of KOH. The effects related to parent material
nature, KOH/precursor ratio and reaction temperature and time on the
porosity characteristics and surface chemistry is described. The results are
discussed in terms of suitability of produced carbons as an electrode material
in electric double-layer capacitors.
Keywords
KOH-activated carbons, porosity characteristics, chemical structure,

electrochemical capacitors.
1. INTRODUCTION
Porous carbons are among the most attractive electrode materials for
electric double layer capacitors (EDLC), where the charge accumulation
occurs mainly by electrostatic attraction forces at the electrode/electrolyte
interface [1-3]. Advantages of this class of materials include high surface
*
63
64
area, versatility in terms of porosity development and surface chemistry,

availability and a reasonable price and a lack of negative environmental
impact. In most general terms the target carbon material for EDLC should be
characterized by an extensive surface area accessible by a given electrolyte
and high specific capacitance per unit surface. The accessibility depends on
both the pore size distribution and the affinity of a given electrolyte to the
surface. Heteroatoms present on the surface can further improve the
wettability by the electrolyte. The intrinsic capacitance of carbon surface is
reported to vary from 15 to 50 PF/cm2 [3]. Some authors [4] link the value
with the structural aspect of carbon matter as directly related to the electric
conductivity. Heteroatoms like oxygen or nitrogen are believed to enhance
the charge accumulated at the unit surface due to pseudocapacitance effect.
While keeping in mind all these implications, the primary
requirement in an attempt to store a huge charge based on the electrostatic
forces seems to be high surface area of an activated carbon used. Among
different ways of porosity development in carbons, the treatment with an
excess of potassium hydroxide is most efficient in terms of microporous
texture generation. Porous materials with BET surface areas in excess of
3000 m2/g could be prepared using various polymeric and carbonaceous type
precursors [5,6].
The recent reports [7,8] suggest that KOH activated carbons
produced from very common precursors like bituminous coal and coal-tar
pitch can be a suitable electrode material for EDLC. The excellent
capacitance values approaching in some cases 300 F/g were measured for
extremely high surface area (SBET=2700-3200 m2/g) materials produced
using severe activation conditions [8]. No simple relationship between the
BET surface area or pore volume and the electric capacitance was found.
The carbons show so high capacitance despite the limited accessibility of the
porous systems as indicated by a relatively low specific capacitance
(calculated per BET surface area), being in the range of 7-11 PF/cm2. For
comparison, the maximum capacitance values of about 160 F/g and 16
PF/cm2, respectively, were determined for a series of steam activated
carbons, which were studied using exactly the same technique [9].
Moreover, a fairly good cycleability has been determined for pitch and
activated carbon derived high surface area materials in contrast to those
originated from coal [8]. Clearly, the electrode performance of KOH
activated carbons depends on pore size distribution and intrinsic properties
of the surface as related to the nature of precursor used. It is therefore
anticipated that the porosity development and surface properties can be
further tailored to the EDLC application by selection of parent material and
adjusting process variables.
Hence, the objective of the present study is an advanced
understanding of the effect of parent material nature and process variables on
65
the porosity development and surface chemistry of KOH activated carbons

with respect to their behavior as an electrode material in EDLC.
2. EXPERIMENTAL
2.1. Materials
Materials used in the activation with KOH include high volatile
bituminous coal C, coal semi-coke CS, pitch mesophase PM, pitch semi-
coke PS and commercial activated carbon AC. The semi-cokes CS and PS
were produced by the heat treatment of corresponding parent materials at
520qC with a heating rate of 5qC/min and 2 hours soaking time. The
preparation of mesophase PM comprised the soaking of coal-tar pitch at
450qC for 7 h with a continuous stirring. All the treatments were performed
under argon in a vertical Pyrex retort of 45 mm diameter.
2.2. Activation
Raw materials were ground to a particle size less than 630 Pm.
Physical mixtures of solid KOH and a given precursor at 2:1, 3:1, 4:1 and
5:1 weight ratios were activated in the temperature range of 600-900qC for
1-5 h under argon flow of 15 dm3/h. The treatments were carried out in a
horizontal furnace using a nickel boat placed in a 36 mm diameter quartz
tube. The heating rate was 10qC/min. The resultant materials were washed
with 10% solution of HCl and distilled water and next were dried at 110qC
for 6 h.
2.3. Analyses
Ash and volatile matter contents were determined according to the
respective standard procedures. Elemental composition (CHNS) was
analyzed using VarioEl elemental analyzer. Oxygen content was calculated
by difference. The optical texture was evaluated using polarized light optical
microscopy of polished surface of samples. N2 adsorption/desorption
isotherms at 77 K of the activated carbons were obtained using a gas
adsorption analyser NOVA 2200 (Quantachrome). Samples were degassed
overnight at 300qC in vacuum prior to the analysis. The relative pressure up
to p/po d 0.15 was used for the calculation of the mean micropore width (LD)
by applying the Dubinin equation. The mesopore size distribution was
evaluated using Kelvin equation.
66

3.1. Parent Materials Characteristics
Characteristics of carbonaceous materials used in the activation with
KOH are given in Table 1.
Table 1. Properties of the precursors used.
Elemental analysis
Sample Ash Volatile matter Optical texture
(C/H)at (N + S + O)
wt % wt%
wt%
C 3.1 31.9 1.4 8.9 Isotropic
CS 3.8 15.7 2.2 8.8 Isotropic
PM 0.5 17. 4 2.1 2.9 Anisotropic
PS 0.3 6.6 2.4 2.8 Anisotropic
AC 7.1 1,7 21.2 5.6 Isotropic
The parent materials differ from each other in many aspects; the
differences are being related both to their origin (coal or pitch) and heat
treatment temperature. Clearly, coal should be classified as a polymeric type
precursor while the others, such as carbonaceous precursors of relatively
low, except for AC, carbonization degree. Specific of pitch-derived materials
is distinctly lower mineral matter and heteroatoms content. Anisotropic
appearance with predominating flow type texture proves the superior extent
of structural ordering in pitch-derived materials.
3.2. Effect of Activation Variables on Porosity Development

The reagents ratio, reaction temperature, reaction time and inert gas
flow have been reported as the variables of KOH activation process, which
influence effectively the porosity development on the treatment of a given
carbonaceous material with potassium hydroxide [6].
Figures 1-3 demonstrate the effect of KOH/precursor ratio, reaction
temperature and reaction time, respectively, on porous structure parameters
of carbon produced by KOH activation. While the presented relationships
concern mostly carbonaceous mesophase, basically they are typical of all
coal and pitch-derived materials of the study.
Varying KOH ratio in the mixture is a very effective way of
controlling porosity development in resultant activated carbons. The trend in
the pore volume and BET surface area increase seems to be similar for
various precursors (Fig. 1a). It is interesting to note, however, a sharp
widening of pores, resulting in clearly mesoporous texture, when a large
excess of KOH is used in reaction with coal semi-coke (Fig. 1b). Increase in
the reaction temperature within 600-900qC results in a strong development
67
of porosity up to 700qC and rather moderate above this temperature

(Fig. 2a).
3500 2,5
a
3000
2,0
Total pore volume (cm /g)

BET surface area (m /g)
2500
3
2
1,5
2000
1500 1,0
1000
PM - , 0,5
500 CS - ,
0 0,0
2:1 3:1 4:1 5:1
KOH/precursor ratio
0,50
b
0,40
Mesopore ratio
0,30
0,20
0,10
PM -
CS -
0,00
2:1 3:1 4:1 5:1
KOH/precursor ratio
Figure 1. Effect of reagents ratio on pore volume and BET surface area (a) and mesopore
ratio (b) in activated carbons produced on KOH activation of PM and CS at 800qC for 1 h.
The treatment of semi-coke or mesophase (3:1 mixture) for 2 h at

600qC gives a material of pore volume VT about 0.7 cm3/g and surface area
SBET about 1700 m2/g which is typically microporous with rather narrow
micropores (average size LD below 1.2 nm). Activated carbons produced
within the temperature range of 700-800qC have fairly similar porosity
characteristics, VT about 1 cm3/g and SBET near 2500 m2/g. It is interesting to
note that within the wide temperature range of 600-800qC the burn-off is at a
reasonably low level of 20-23 wt%.
68
3000 2,0
a
2500

1,5
3
2
2000
1500 1,0
1000
0,5
500
0 0,0
600 650 700 750 800 850 900
O
Temperature ( C)
0,20 2,0
b
Mean micropore width (nm)
1,8
0,15
Mesopore ratio
1,6
0,10
1,4
0,05
1,2
0,00 1,0
600 650 700 750 800 850 900
O
Temperature ( C)
Figure 2. Variation of pore volume and BET surface area (a) and mesopore ratio and mean
micropore width (b) with reaction temperature
on the treatment of KOH/PM 3:1 mixture for 2h.
69
3000 2.0
a
2500

1.5
3
2
2000
1500 1.0
1000
0.5
500
0 0.0
0 1 2 3
Time (h)
0,15 2,0
b Mean micropore width (nm)
1,8
0,10
Mesopore ratio
1,6
1,4
0,05
1,2
0,00 1,0
0 1 2 3
Time (h)
Figure 3. Variation of pore volume and BET surface area (a) and mesopore ratio and mean
micropore width (b) with reaction time on the treatment of KOH/PM 3:1 mixture at 800qC.
70
A sharp increase in burn-off above 800qC is associated with a

limited enhancement of porosity. Pore widening (reflected by a slight but
continuous increase in the mesopore ratio and micropore width), seems to be
most characteristic effect of increasing reaction temperature (Fig. 2b).
The soak time seems to have a less pronounced effect on porosity
development during treatment at 800qC. A considerable porosity (VT = 1.04
cm3/g and SBET = 2570 m2/g) is created within the early period of reaction
and a limited change of the parameters occurs with prolonged heating to 3 h
(Fig. 3a). Again, the widening of pores, as evidenced by an increase of both
the mesopore ratio and the micropore width, is the most noticeable result of
extended reaction time (Fig. 3b).
3.3. Effect of Parent Material on Porosity Development

In order to evaluate the potential to porosity development, which is
intrinsically associated with a given precursor nature, all parent materials
were treated using drastic conditions: reagent ratio 4:1, temperature 800qC,
time 5 hours. Porosity parameters of resultant carbons are given in Table 2.
Nitrogen adsorption/desorption isotherms of all the activated
carbons are of Type I, i.e. characteristic of basically microporous solids.
There is a lack of adsorption/desorption hysteresis. More careful analysis
permits to notice significant differences in the porous texture parameters
depending on precursor origin.
Table 2. Porosity parameters of KOH activated carbons produced from different precursors
under the same reaction conditions (800oC, 5 hours, KOH/precursor ratio 4:1)
Pore volume distribution, % Average
SBET Vtotal
Precursor 2 3 V<2nm V2-3nm V3-5nm V5-50 nm pore size
m /g cm /g
nm
C 3150 1.612 73.8 15.0 6.1 5.1 2.04
CS 3190 1.445 81.9 13.9 2.6 1.6 1.82
PM 2660 1.209 83.6 10.9 3.3 2.2 1.82
PS 2750 1.227 85.4 9.5 2.7 2.4 1.79
AC 1900 1.051 68.6 12.3 8.3 10.8 2.21
At the exhaustive activation conditions, the polymeric-type

precursor (coal C) gives an activated carbon of bigger pore volume and
wider pore size distribution with a larger burn-off. The behavior can be
attributed to a large proportion of non-carbon elements, which are readily
removed from the system and a loose packing of structural units at the stage
when the primary pyrolysis reactions are completed. Activation of low
temperature carbons (mesophase PM, semi-cokes CS and PS) allows
preparing carbons of very high SBET (2700-3200 m2/g) and rather similar
pore size distribution. The commercial active carbon (AC) gives less porous
material (SBET = 1900 m2/g) but with a considerable contribution of
71
mesopores (above 30%), including those over 5nm. Apparently, the strange
porosity characteristics appear as an effect of both a lower extent of reaction
with alkali due to enhanced carbonization degree (Table 1) and overlapping
of the original porous system with that generated by KOH activation.
3.4. Chemical Structure of KOH-Activated Carbons

Elemental analysis of activated carbons made from raw materials of
this study under various reaction conditions is presented in Table 3.
Compared to the conventional steam activated carbons,
characteristic of all the KOH activated carbons is the enhanced oxygen
content. The content varies quite widely from one activated carbon to
another. The data for a series of carbons activated for 5h at 800qC show
rather surprisingly, the highest oxygen content in pitch derived materials. A
lack of correlation in elemental composition between parent material and
activation product suggests that the origin of the oxygen groups is the high
temperature reaction with KOH and the post-activation treatment. A reduced
oxygen content in carbons produced under severe treatment conditions is an
effect of partial decomposition of the groups. Therefore, an enhanced oxygen
content in the pitch derived activated carbons can prove that more thermally
stable functionalities are created during reaction of KOH with material of
higher structural ordering.
Table 3. Composition of activated carbons made by KOH activation of different precursors.

Precursor Activation variables C H N+S Odiff
wt% wt% wt% wt%
C 800oC, 5 h, 4:1 96.1 0.23 0.2 3.5
CS 800oC, 5 h, 4:1 92.8 0.20 0.3 6.7
PM 800oC, 5 h, 4:1 91.2 0.21 0.2 8.4
PS 800oC, 5 h, 4:1 91.3 0.18 0.2 8.4
AC 800oC, 5 h, 4:1 94.8 0.21 0.4 4.6
CS 800oC, 1 h, 3:1 86.1 0.36 0.7 12.8
PM 800oC, 1 h, 3:1 90.5 0.16 0.3 9.0
CS 600oC, 2 h, 3:1 79.3 1.02 1.0 18.7
PM 600oC, 2 h, 3:1 82.8 0.82 0.8 15.6
3.5. Potential of KOH-Activated Carbons as EDLC Electrode

Activation with KOH was recognized originally as an efficient way
of producing microporous carbons with relatively narrow pore size
distribution and extremely high surface area. The results of present study
demonstrate a considerable flexibility of the process in terms of porosity
development and, to some extent, surface properties.
With respect to the producing porous carbons for electrochemical
capacitors the activation using potassium hydroxide represents several
72
interesting aspects. Primarily, activated carbons of very high surface area

and different mesopore ratio can be produced at a reasonable burn-off from a
variety of precursors. In contrast to the agents used in conventional physical
and chemical activation, KOH reacts efficiently with the whole range of
polymeric and carbonaceous materials, including those of high extent of
structural ordering as distinguished by optical anisotropy. This peculiarity
can appear meaningful for improving the electric conductivity of electrode
material. Another attractive aspect of KOH activated carbons is the presence
of thermally stable oxygen functionalities, which can positively contribute to
the electric capacitance by both enhancing wettability of surface by
electrolyte and inducing pseudocapacitance effect. Oxygen content is
controlled by both the parent material nature and reaction variables.
4. CONCLUSION
The activation with KOH of selected parent materials under
appropriate process conditions (temperature, time, reagent ratio) can provide
highly porous carbons of controlled pore size distribution and surface
chemistry, also suitable for use as electrode materials in supercapacitors.
ACKNOWLEDGMENTS
Work was supported by the Polish Ministry of Scientific Research &
Information Technology (Project T09B 13325), NATO SfP project 973849.
REFERENCES
1. Sarangapani S., Tilak B.V., Chen C.P., Materials for electrochemical capacitors –
theoretical and experimental constraints, J. Electrochem. Soc. 1996; 143:3791-3799.
2. Qu D, Shi H. J. Power Sourc. 1998; 74:99-107.
3. Frackowiak E., Béguin F., Carbon materials for the electrochemical storage of energy in
capacitors, Carbon 2001; 139, 937-950.
4. Qu D., Studies of the activated carbons used in double-layer supercapacitors, J. Power
Sourc. 2002; 109:403-411.
5. Ahmadpour A., Do D.D., The preparation of active carbons from coal by chemical and
physical activation, Carbon 1996; 34:471-479.
6. Lozano-Castello D., Lillo-Rodenas M.A., Cazorla-Amoros D., Linares-Solano A.,
Preparation of activated carbons from Spanish anthracite. Carbon 2001; 39:741-749.
7. Weng T.C., Teng H., Characterization of high porosity carbon electrodes derived from
mesophase pitch for electric double-layer capacitors, J. Electrochem. Soc. 2001;
148:368-373.
8. Kierzek K., Frackowiak E., Lota G., Gryglewicz G., Machnikowski J., Electrochemical
capacitors based on highly porous carbons prepared by KOH activation, Electrochim.
Acta 2004; 49:515-523.
9. Gryglewicz G., Lorenc-Grabowska E., Lota G., Frackowiak E., Machnikowski J., Coal-
based mesoporous activated carbons for double-layer capacitors, Carbon’03,
International Conference on Carbon, CD-ROM Proc., paper 292, Oviedo 2003.
GENERAL PROPERTIES OF IONIC LIQUIDS AS
ELECTROLYTES FOR CARBON-BASED
DOUBLE LAYER CAPACITORS
A. Lewandowski* and M. Galinski
Faculty of Chemical Technology, PoznaĔ University of Technology,

PL-60 965 PoznaĔ, Poland
1. INTRODUCTION
Activated carbons have been used for a long time in many industrial
applications. One of the recent applications is using activated carbon as an
active electrode material in double-layer capacitors (DLC). Electrochemical
capacitors, based on the capacity of the double layer formed at the
carbon/electrolyte interface, have received considerable attention as they can
be used as high power-density energy-storage devices. Batteries have high
energy density, but they suffer from low power density and low cyclability
(usually <1000). On the other hand, electrochemical capacitors offer an order
of magnitude higher power density and at least two orders of magnitude
higher number of charge-discharge cycles (~100 000). However, the energy
density is an order of magnitude lower in comparison to that characteristic
for batteries. There are a number of electrochemical capacitor potential
applications, but it seems that they are being investigated mainly for hybrid
vehicles as a short-term power accumulator and supply. Therefore, the DLC
research aims to develop an energy-storage device having energy density
close to that typical of batteries and the power characteristic typical of
capacitors. This goal can be obtained by optimisation of carbon materials as
well as electrolyte. Carbon materials, for the use in double-layer capacitors,
have been developed in various systems: powders, fabrics, tissues or
aerogels [1-3]. It is possible to obtain activated carbons of high specific
surface area at the level of 2500 m2/g. Similarly, many aqueous and organic
liquid solutions of electrolytes as well as a limited number of polymer
electrolytes have been developed for the use in carbon-based electrochemical
capacitors [4].
*
73
74
The ideal electrolyte should:
1. Produce high capacity of the carbon/electrolyte double electric layer;

2. Enter the pores of the high-surface area activated carbons;
3. Show possibly broad electrochemical stability window at an interface of
carbon electrode;
3. Show high electrical conductivity, at the level of t10 mS/cm;
4. Be liquid in a broad temperature range;
5. Emit no volatile vapours.
One of the most important factors is the electrolyte resistance for

electrochemical degradation (oxidation or reduction) at possibly high
voltages applied to the capacitor. Let us assume, that the highest practical
specific capacity of activated carbon in organic electrolytes is of the order of
160 F/g and that the device contains 25 wt.% of the carbon material.
Therefore, the specific capacity is 40F/g. However, the capacity C of the
device consisting of two identical electrodes is twice lower: (1/C = 1/C1 +
1/C2; C = Ci/2), and the final specific capacity of the capacitor equals ca. 20
kF/kg. The energy E which can be stored in the device is proportional to the
voltage U in the second power: E = ½CU2. Figure 1 shows the energy
accumulated by such a capacitor at different voltages applied to it (level of
the specific energy of batteries is indicated along energy axis).
500
Li-Ion
400
E [kJ/kg]
300
Ni-MH
200
Ni-Cd
PbO2-Pb
100
0
0 2 4 6 8
U [V]
Figure 1. Energy accumulated in a capacitor having specific capacity of 20 kF/kg as a

function of voltage (E = ½CU2).
75
It can be seen that an energy of ca. 150 kJ/kg, comparable to that

accumulated in PbO2-Pb or Ni-Cd batteries, can be obtained at voltages of
4V. Somewhat lower energy (100 kJ/kg) is accumulated at a voltage of 3V.
Consequently, the searched system carbon/electrolyte should be
characterised by: (i) specific capacity. t 160 F per gram of activated carbon
and (ii) electrochemical stability window at the level of ca. t3V.
The general aim of this work is to show the properties of activated
carbon-ionic liquid interface (AC-IL) as well as performance of capacitors
based on activated carbons as electrode materials and ionic liquids as
electrolytes.
2. IONIC LIQUIDS
Conventional electrolytes applied in electrochemical devices are

based on molecular liquids as solvents and salts as sources of ions. There are
a large number of molecular systems, both pure and mixed, characterized by
various chemical and physical properties, which are the liquids at room
temperatures. This is the reason why they dominate both in laboratory and
industrial scale. In such a case, solid salt is reacted with a molecular solvent
and if the energy liberated during the reaction exceeds the lattice energy of
the salt, the solid is ‘liquified’ chemically below its melting point, and forms
the solution. Water may serve as an example of the cheapest and most
widely used molecular solvent.
However, the lattice energy of the salt may be counterbalanced by
providing heat to the solid; in such a case the solid is melted down at its
melting point and forms a liquid phase, which does not contain any
molecular solvent, but only ions and their charged and neutral combinations.
Such a system is called ionic liquid or fused salt. Classical ionic liquids exist
at relatively high, uncomfortable temperatures, determined by the melting
point of salts. These temperatures can be lowered by an addition of other
salts forming a eutectic. Even in such a case considerable heat is necessary to
maintain the system in the liquid state. Aluminum electrolysis from Al2O3-
Na3AlF6 eutectic fused at ca. 1000°C may serve as rather unique example of
using fused salts in a large-scale industrial practice. Highly reactive metals,
such as Li, Na or Mg are also produced from high-temperature fused salts.
Ionic liquids have been used in thermally activated batteries, applied in
military applications, when an extremely long storage time of the device,
providing high power for a short time, is necessary. The battery is activated
with the help of an explosive; the heat liberated during the explosion melts
down the salt. Applying ionic liquids in commercial energy storage devices,
such as batteries or electrochemical capacitors, has been not practically
realized. On the other hand, salts characterized by a low melting point, and
therefore being liquid at room temperature, or even below, can play a role of
76
electrolytes in electrochemical systems. Room temperature ionic liquids

(RTIL) are not new, but only within the past few years an increasing interest
in their synthesis and application has been observed [5,6]. Their basic
physicochemical properties are the same as high temperature molten salts,
but the practical aspects of their application are different enough to merit a
distinction. One of the first discovered RTIL (1914) seems to be EtNH3NO3,
but the first extensively studied was 1-ethyl-3-methylimidazolium
chloroaluminate (EMImAlCl4) [5,6]. Ionic liquids based on the
chloroaluminate ion seem to be a promising medium for reactions catalyzed
by Lewis acids, such as for example Friedel-Crafts reaction. However,
chloroaluminate ionic liquids are hydrolyzed by traces of moisture, which
limits their possible practical application. Lately, ion-conducting systems
based on nonchloroaluminate room temperature ionic liquids have been
developed (Figure 2) for application as electrolytes [7-15].
+ +
N N N N +
R1 R2 R1 R3 N
R2 R1 R2
R1 R1
+ +
R4 P R2 N+ R4 N R2
R3 R1 R2
BF4 PF6 AlCl 3 CF3SO3

SO2 CF3 SO2 CF2 CF3
- -
N N
SO2 CF3 SO2 CF2 CF3
Figure 2. RTIL’s typical cations and anions.
RTILs consist of large, unsymmetrical ions, such as 1,3-dialkyl-

imidazolium, 1-alkylpyridinium, 1-alkylpyrazolium, tetralkylammonium or
tetralkyl-phosphonium cations and tetrachloroaluminate, tetrafluoroborate,
hexafluorophosphate, trifluoromethanesulfonate or bis((trifluoromethyl)
sulfonyl)imide anions.
Table 1 shows examples of melting and freezing points of a number
of organic salts, which may be regarded to be ionic liquids. As can be seen,
the compounds are characterized by tendency for supercooling. The
77
difference between melting and freezing points may be as high as 60°C. As a

result, the same compound can be described in the literature as solid or liquid
at room temperature. On the other hand, the term ‘ionic liquid’ is quite often
extended to salts molten below 100°C. In such a case heat must be provided
to keep the salt in the liquid state, and IL cannot be handled as an ordinary
solvent.
Table 1. Freezing and melting points [oC] of some representative ionic liquids [9].
Ionic liquid Tf Tm 'T
EMImBF4 -63 11 74
EMImPF6 5 62 57
EMImN(Tf)2 -50 -15 35
DMPImPF6 18 78 60
DMPIm N(Tf)2 -130 -81 49
Commonly is used a short term ‘ionic liquid’ instead of ‘room

temperature ionic liquid’ or ‘room temperature molten salt’, which makes no
distinction between salts liquid at room temperature and those liquid below
100°C.
3. CONDUCTIVITY
The room temperature specific conductivity of a number of ionic
liquids is reported in Table 2, together with their viscosity. It can be seen
from the table, that the conductivity varies in the range of 0.1–13 mS/cm.
The most conductive IL in the series shown in the table is EMImBF4.
Table 2. Concentration c, conductivity V, viscosity K and Walden product /K of some neat

ionic liquids at 25 oC [9].
c V K /K
mol/dm3 mS/cm cP cm2cP/mol
EMImBF4 6.5 13 43 86
EMImPF6 5.7 5.2* 15 66*
EMImN(Tf)2 3.9 8.4 28 60
DMPImPF6 4.8 0.5 34 63
DMPIm N(Tf)2 3.5 3.0 60 51
HTEA N(Tf)2 [7] - 0.7 17 -
*
Data reported for the temperature of 80oC
The temperature dependence of the conductivity can be described by

the classical Arrhenius equation V = Voexp(-E#/RT), where E# is the
activation energy for the conduction process. According to the Arrhenius
equation the lnV versus 1/T plot should be linear. However, in numerous
ionic liquids a non-linearity of the Arrhenius plot has been reported; in such
a case the temperature dependence of the conductivity can be expressed by
the Vogel-Tammann-Fuller (VTF) relationship: V = Voexp{-B/(T-To)},
78
where To represents the ideal glass transition temperature. The parameter B

has no sense of activation energy. However, even when the lnV = f(1/T) plot
is not ideally linear, it is quite often treated according to the Arrhenius
equation. Activation energies, determined from the Arrhenius plot, are
usually in the range of 10-40 kJ/mol.
Table 3. Conductivity of 2mol/dm3 solutions of EMImBF4 and EMImPF6

in molecular liquids at 25 oC [9].
EMImBF4 EMImPF6
- 13 5.2
PC 16, 18* 15
EC 23 22
DMF 25 33
Acetone 34 43
AN 47 60
Water 85* -
*
Data reported as quoted from ref. [16]
Table 3 shows conductivity of 2mol/dm3 solutions of EMImBF4 and

EMImPF6 in a number of molecular solvents. A high increase of
conductivity, in comparison to neat ionic liquids, can be observed after
dilution with electrically neutral molecular liquids. However, solutions of
ionic liquids in molecular liquids are simply conventional solutions of
organic salts in nonaqueous solvents, and no distinction can be seen between
them and commonly employed solutions of (C2H5)4NBF4.
4. POLYMER ELECTROLYTES WITH IONIC LIQUIDS

A polymer electrolyte may be generally defined as a system formed
by dissolution of an electrolyte in a polymer matrix. The resulting composite
shows ionic conduction and mechanical properties characteristic for
polymeric materials. Polymer electrolytes may be prepared as thin foils
showing rubber-like elasticity. However, the process may be reversed by
dissolution of a polymer in ionic liquid. In such a case the polymer acts both
as a diluent of the liquid salt, as well as the matrix providing mechanical
strength to the system. The first polymer electrolytes incorporating ionic
liquid were based on 1-butylpyridinium halides (chloride or bromide) and
aluminum chloride [17,18]. More recently, ion conducting solid systems
based on nonchloroaluminate room temperature ionic liquids have been
synthesized [3]. Poly(vinylidene fluoride-hexafluoropropylene) (PVdF-HFP)
[19,20] and NAFION [21] membranes were impregnated with room
temperature, non-chloroaluminate ionic liquids. Ionic conductivity of a
composite PVdF-HFP-ionic liquid was in the range of 3.2-0.06 mS/cm.
79
Polymeric conducting systems were also prepared by in situ polymerization

of vinyl monomers in ionic liquids [22], with a conductivity of 1 mS/cm. A
conductive polymer electrolytes were also prepared by polymerization in
liquid EMIm(HF)nF leading to a composite poly(2-hydroxyethyl
methacrylate)-EMIm(HF)nF. Recently, polymer electrolytes were prepared
in the form of thin foils, by incorporating ionic liquids in a polymer matrix
[13-15]. Conductivities of polymer-IL or polymer-IL-solvent systems are
collected in Table 4.
Table 4. Specific conductivity (25°C) of some polymer electrolytes based on ionic liquids.
Composition [wt.%] V
Polymer IL TMS [mS/cm]
25.4% PEO 74.6% BMImPF6 0.0 % 0.3
12.3% PAN 33.2% BMImPF6 54.5% 6.3
67.2% PEO 32.8% EMImBF4 0.0% 0.6
39.8% PVA 60.2% EMImBF4 0.0% 1.6
26.6% PAN 73.4% EMImBF4 0.0% 6.6
12.9% PAN 48.3% EMImBF4 38.8% 15.0
It can be seen from the table that the system PAN-EMImBF4-TMS,

containing a plasticizer (TMS), shows the best conductivity of 15 mS/cm.
5. SPECIFIC CAPACITY OF CARBON/IL INTERFACE

It is difficult to compare the specific double-layer capacitance
characteristic for capacitors based on highly porous activated carbons with
that determined for non-porous electrodes (solid glassy-carbon and liquid
mercury) as appropriate literature data on ionic liquids are limited to just a
few studies [24,25]. Apart from a report on ionic liquid based on the
chloroaluminate ion [25], one paper [24] about specific capacitance at the
electrode/IL interface for glassy carbon (GC), mercury and a commercial
carbon and four ionic liquids based on 1-ethyl-3-methyl imidazolium cation
can be found in the literature. The integral capacitance of various ionic
liquids at the dropping mercury electrode (DME) ranged between 10.6 and
12.4 PF/cm2. For the EMImN(Tf)2 electrolyte the capacitance was 11.7
PF/cm2 at DME and 12.0 at glassy carbon, but only 3.7 PF/cm2 at a high-
surface SpectraCarb 2220 yarn (ca. 2000 m2/g). This may indicate that the
accessibility of the electrolyte to the carbon electrode surface limits the
realizable specific capacitance of the device.
6. ELECTROCHEMICAL STABILITY WINDOW

It is essential from the point of view of high power-density to ensure
the electrochemical stability of the system at possibly high voltages. Broad
electrochemical stability windows are typical if ionic liquids, however, the
80
practical voltage, which can be applied to capacitor carbon electrodes seems

to be considerably lower. This is because the stability window is usually
determined at glassy carbon electrodes while the capacitor electrodes are
formed from different kinds of activated carbons, possessing functional
groups at their surface. Moreover, a polymer used as an additive of the
electrode, or being a component of a polymer electrolyte, may also undergo
red-ox electrode reactions, which can additionally limit the stability window.
Table 5. Electrochemical stability window of ionic liquids (IL) at the glassy carbon
(potentials [V] expressed versus Ag/Ag+ 0.01M in DMSO reference) [26 ].
Stability window
Ionic liquid cathodic anodic
EMImBF4 -2.1 2.1
BMImBF4 -2.1 2.1
BMImPF6 -2.1 2.0
EMImN(Tf)2 -2.1 2.0
BMPyN(Tf)2 -1.0 2.0
Table 5 shows cathodic and anodic limits of electrochemical

stability windows of a number of ionic liquids. The cathodic limit of the
stability window of the ILs based on the EMIm+ and BMIm+ cations,
investigated at the glassy carbon electrode, is –2.1 V against the Ag/Ag+
(0.01M in DMSO) reference. The BMPy+ cation is reduced at the glassy
carbon at considerably more positive potential, at ca. –1.0 V.
The anodic oxidation of the tetrafluoroborate anion occurs at
potentials higher than 2.1 V and the remaining hexafluorophosphate and
imide anions are oxidised at potentials higher than 2.0 V. Hence, the stability
window of the EMImBF4 and BMImBF4 is 4.2 V. Ionic liquids BMImPF6
and EMImN(Tf)2 shows a similar stability window of ca. 4.1 V. However,
the window of the BMPyN(Tf)2, is considerably lower: ca. 3.0 V. This is
consistent with data (ca. 4.1-4.2 V) found for a series of ionic liquids based
on EMIm+ and DMPIm+ (1,2-dimethyl-3-propylimidazolium) cations [12].
Usually, an addition of a diluent (solvent), increases the
conductivity, but also influences the stability window. It has been found, that
among the most popular organic solvents, 2 M solutions in acetonitrile
showed the highest conductivity level (EMImN(Tf)2: 47 mS/cm and
EMImPF6: 60 mS/cm) and a broader stability window (ca. 4 V) [12].
Carbonate solvents decrease the cathodic limit by ca. 0.2-0.4 V, depending
on the carbonate structure. It is worth noting that different authors give
different stability windows for the same ionic liquid [10]; this is partially due
to different electrode materials.
81
7. CARBON/IL CAPACITORS
Ionic liquids show a broad electrochemical stability window of ca.
4V, conductivity at a level of ca. 10 mS/cm as well as an acceptable specific
capacity of ca. 10 PF/cm2, which makes them promising candidates for
electrolytes. However, there are only a few papers about capacitors based on
carbon materials and ionic liquids. In 1997 some results on applying IL
solutions in supercapacitors were reported [12]. Ionic liquids, based on
EMIm+ cation (EMImBF4 and EMImPF6) were dissolved in cyclic
carbonates. Capacitors showed specific capacity, expressed versus mass of
active carbon, of ca. 130 F/g and the electrochemical stability within the
voltage range of ca. 3.5 V. One of the main problems in supercapacitors
construction is a reduction of the equivalent series resistance (ESR) to low
values. This may be done in two ways: by applying electrolytes of high
conductivity or by developing a technology of ultra-thin electrodes. The best
would be suppressing the ESR to very low values by combining high
electrolyte conductivity with ultra-thin electrodes. In the paper [12], the
conductivity was enhanced to values of ca. 26 mS/cm (2mol/dm3 solution of
EMImPF6 in a mixture EC + DMC). However, solutions of ILs in molecular
liquids are classical salt solutions, and they do not show any advantages in
comparison to solutions of tetraalkylammonium salts in acetonitrile.
Table 6. Specific capacities of an activated carbon (2000 m2/g)

interface with 2mol/dm3 solution of EMImPF6 in cyclic carbonates [12]
AC 2000 m2/g
Solvent F/g PF/cm2
EC 123 6.2
PC 127 6.4
DMC 126 6.3
EMC 123 6.2
PC+EMC 129 6.5
Capacitors made of high-surface carbon with PVdF-co-HFP as a

binder and filled with neat ionic liquids, without any molecular solvent,
show high specific capacity, up to 180 F/g [13].
Table 7. Specific capacities of activated carbon interface with ionic liquids [13]
AC 870 m2/g AC 2600 m2/g
2
Ionic liquid F/g PF/cm F/g PF/cm2
EMImBF4 55 6.3 178 6.8
BMImBF4 50 5.7 155 5.7
BMImPF6 45 5.2 160 6.1
EMImN(Tf)2 51 5.9 180 6.9
BMPyN(Tf)2 48 5.8 167 6.4
Similar or even better capacity was obtained in systems with

polymer electrolytes based on ionic liquids. A series of electrochemical
82
capacitors, based on activated carbons, applying polymer-IL solid

electrolytes were prepared and tested [14]. Both polymer electrolytes as well
as capacitor electrodes, in the form of thin foils, were prepared by a casting
technique. All-plastic capacitors, having a coin-like shape, were constructed
by sandwiching a polymer electrolyte between two electrodes. Ionic liquids:
EMImBF4, BMPyN(tf)2 and BMImPF6 served both as sources of ions as well
as polymer plasticizers. In some systems sulpholane (TMS) was added to the
system as an additional plasticizer. In other words, polymer was used as the
ionic liquid diluent. Specific capacity was in the range of 4.2-7.7 PF/cm2,
calculated per total surface of the carbon materials, and up to 200 F/g
expressed versus the mass of carbon material.
8. CONCLUSIONS
The main goal of the research on electrochemical capacitors is the
development of an energy-storage device having energy density close as far
as possible to that typical of batteries as well as the power characteristic and
cyclability typical of capacitors. This can be obtained by optimization of
carbon materials as well as electrolyte. Theoretical analysis shows that an
electrochemical capacitor of specific capacity of 20 kF/kG and practical
operation voltage of 3V would store energy comparable to that characteristic
for the lead-acid battery (ca. 100 kJ/kg). The specific capacity of the device
of ca. 20 kF/kG is equivalent to specific capacity expressed versus mass of
activated carbon at the level of ca. 160 F/gC. Ionic liquids seem to be
promising candidates for electrolytes in high energy capacitors, as they (i)
produce high specific capacity at the AC/IL interface of the order t160 F/gC
and (ii) they show broad electrochemical stability window determined at GC
of the order of 4V. Till now, the literature of the subject is limited to a few
papers on ionic liquid based capacitors.
REFERENCES
1. E. Frackowiak and F. Béguin, Carbon, 39 (2001) 937
2. G. Salitra, A. Soffer, L. Eliad, Y. Cohen, D. Aurbach, J. Electrochem. Soc., 147 (2000)
2486
3. J. Gamby, P.L. Taberna, P. Simon, J.F. Favarque, M. Chesneau, J. Power Sources, 101
(2001) 109
4. B.E. Conway, Electrochemical Supercapacitors, Kluwer Academic/Plenum Publisher,
New York, (1999)
5. J.S. Wilkes, Green Chemistry, 4 (2002) 73
6. T. Welton, Chem. Rev., 99, (1999) 2071
7. J. Sun, M. Forsyth, D.R. MacFarlane, J. Phys. Chem. B. 102 (1998) 8858
8. D.R. MacFarlane, P. Meakin, J. Sun, N. Amini, M. Forsyth, J. Phys. Chem. B. 103
(1999) 4164
83
9. A.B. McEwen, H.L. Ngo, K. LeCompte, J.L. Goldman, J. Electrochem. Soc., 146 (1999)
1687
10. R. Hagiwara, Y. Ito, J. Fluorine Chem., 105 (2000) 221
11. T. Nishida, Y. Tashiro, M. Yamamoto, J. Fluorine Chem., 120 (2003) 135
12. A.B.McEwen, S.F.McDevitt, V.R.Koch, J. Electrochem. Soc., 144 (1997) L84
13. A. Lewandowski, M. GaliĔski, J. Physics and Chemistry of Solids, in press
14. A. Lewandowski, A. Swiderska, Solid State Ionics, 161 (2003) 243
15. A. Lewandowski, A. Swiderska, Solid State Ionics, in press
16. A. Lewandowski, S. Krajewski, in preparation for publication
17. M. Watanabe, S. Yamada, K. Saniu, N. Ogata, J. Chem. Soc. Chem. Commun. 929
(1993)
18. M. Watanabe, S. Yamada, N. Ogata, Electrochim. Acta, 40 (1995) 2285
19. J. Fuller, A.C.Breda, R.T. Carlin, J.Electrochem Soc., 144 (1997) L67
20. J. Fuller, A.C.Breda, R.T. Carlin, J. Electroanal. Chem., 459 (198) 29
21. M. Doyle, S.K. Choi, G.Proulx, J.Electrochem Soc., 147 (2000) 34
22. A. Noda, M. Watanabe, Electrochim Acta. 45 (2000) 1265
23. T. Tsuda, T. Nohira, Y. Nakamori, K. Matsumoto, R. Hogiwara, Y. Ito, Solid State
Ionics, 149 (2002) 295
24. C. Nanyunduiah, S.F. Mcdevitt, V.R., Koch, J Electrochem Soc. 144 (1997) 3392
25. R.J. Gale, R.A. Osteryoung, Electrochim Acta, 25 (1980) 1527
26. A. Lewandowski, M.GaliĔski upublished data
LIST OF ABBREVIATIONS
AC activated carbon
AN acetonitrile
DMPIm+ (1,2-dimethyl-3-propylimidazolium) cation
DMC dimethyl carbonate
DMF NN-dimethylformamide
DME dropping mercury electrode
HTEA+ (n-C6H13)(C2H5)3+ cation
EC ethylene carbonate
EMC Ethyl-methyl carbonate
EMIm+ 1-ethyl-3-methyimidaziolium cation
ESR Equivalent series resistance
GC glassy carbon
IL ionic liquid
N(tf)2- -
N(CF3SO3)2 anion
PAN poly(acrylonitrile)
PC propylene carbonate
PEO poly(ethyleneoxide)
PVA poly(vinylalcohol)
PVdF-co-HFP poly(vinylidenefluoride-co-hexafluoropropylene)
RTIL room temperature ionic liquid
TMS sulpholane (tetramethylene sulphone)
CHAPTER 2:
CARBON MATERIALS FOR

GAS DIFFUSION ELECTRODES,
METAL AIR CELLS AND BATTERIES
The following collection of papers focuses on the role of carbon
materials in gas diffusion electrodes, currently used for the most part in
Zinc-air batteries and cells. Such batteries use oxygen from ambient
atmosphere to produce electrochemical energy. The air oxygen diffuses into
the cell and is used as a cathode reactant. The cathode contains few forms of
carbon: in the active material, as conductive additive, and as part of the
current collector coating. A classic example of a household Zinc-air system
is a so-called “hearing aid” cell. Besides that, these systems of power sources
occupy several other niches of the energy market (discussed in length in this
chapter). Editors believe that this chapter should be of interest to all: metal-
air battery and most fuel cell system developers, due to obvious similarities
between the construction of and requirements to the gas diffusion electrodes
used by these chemistries. The chapter contains four articles.
The first paper by Ukraine’s V. Barsukov et al. introduces a novel
concept, which allowed constructing primary and rechargeable metal-air
batteries of a new kind. The new chemistry uses a gas-diffusion electrode
made out of a polyaniline (conducting polymer) / expanded graphite
composite. The authors propose to put such an electrode against stabilized
zinc or magnesium counter electrodes in the acidic solution. While featuring
an apparent drawback of being a relatively low current density battery
system, authors report achieving energy densities of 140 Wh/kg and 200
Wh/kg for the cells based on zinc and magnesium anodes, respectively. The
data has been reported in fully functional AAA size cells.
The second and the third papers of this chapter were contributed by
Bulgaria’s National Academy of Sciences distinguished developers, Dr. A.
Kaisheva and Professor I. Iliev. The two papers address magnesium-air and
zinc-air chemistries, respectively. Concepts of mechanical recharging, new
non-KOH based electrolytes, and the role of carbon materials in cell design
have been thoroughly studied. Authors have reduced to practice their many
developments. Systems created work(ed) in many specialty applications.
The last paper in this chapter is by Professor N. Korovin of Moscow
Power Engineering Institute, Russian Federation. The author could not
attend the NATO-CARWC in Chicago due to a last minute cancellation. His
work, contributed to this volume is a comprehensive overview of various
metal-air battery technologies, with a heavy focus on the role of carbon
materials in these electrochemical systems.
While during the NATO-CARWC, the session on metal-air batteries
featured other important contributions, it is, perhaps, by coincidence that
only presenters from the former Eastern Block submitted their papers for this
volume. In this regard, the editors wish to note, that papers below may be of
special interest to the western reader, as they offer a rather unique insight
into the science and application of the metal-air battery systems from the
87
88
point of view of an Eastern European battery developer. Editors wish to

respectfully refer the reader to [1]* for a “western” look into the subject.
Professor Ilia Iliev (In Memoriam)

Editors deeply regret that one of the authors, Professor Iliev passed
away at the time of preparation of this volume. We have remembered his
highly professional keynote talk, given during the NATO-CARWC, also his
enthusiasm and high profile sense of humor. We hope that his laboratory in
Bulgaria continues to operate and contribute to the world’s body of
electrochemical knowledge.
Below, we have published a memoriam to Professor Iliev by one of
his former students, and now Assistant Professor at the Department of
Chemical & Nuclear Engineering, University of New Mexico in
Albuquerque, NM, USA, Dr. Plamen Atanassov.
Dr. Ilia Iliev was from the generation of Bulgarian scientists who
entered the professional life in the late 50s. Broth in by the excitement of the
physics in the mid-twentieth century and the growth of the post-WWII
Eastern Europe, Ilia Iliev changed fields and become one of the key
developers of the metal-air systems at the Central Laboratory of
Electrochemical Power Sources, Bulgarian Academy of Sciences.
During the 70s he was engaged in the engineering of one of the
largest electric vehicle propulsion programs being most contributive in the
area of gas-diffusion electrodes of hydrophobic type. Associated
hydrophobized carbon materials, electrode fabrication and device
manufacturing were brought by his energy to a most successful fruition and
market deployment. He contributed to development and commercialization
of zinc-air power sources that lightened buys in the Baltic Sea, radioed
messages from Mount Everest and circumnavigated the Earth.
Ilia Iliev embraced the new field of biosensors in the mid 80s and
brought the most prominent Bulgarian contribution to it through his proven
materials design approach. During the turmoil of the 90s he managed to
sustain and grow these two technical venues continuing his association with
the University of Sofia, where he taught electrochemistry to the new
generation of Bulgarian chemical physicists. His contributions to the carbon-
based materials and systems are well documented in many research
publications and treatises.
His sudden death in 2004 became a shock. Professor Iliev will be
remembered by all who knew him as a man of radiant energy, inventiveness
and drive for innovation. Ilia Iliev is survived by his wife, Dr. Anastasia
Kaisheva, collaborator and partner in all of the above research endeavors,
and daughter Mila.
*
[1] Handbook of Batteries / D. Linden, T.B. Reddy.-3rd ed. Chapt. 38, McGraw-Hill (2002).
NEW CONCEPT FOR THE METAL-AIR
BATTERIES USING COMPOSITES:
CONDUCTING POLYMERS / EXPANDED
GRAPHITE AS CATALYSTS
V.Z. Barsukova*, V.G. Khomenkoa, A.S. Katashinskiia and T.I. Motronyukb

a
Electrochemical Power Engineering & Chemistry Department, Kiev National University of
Technologies and Design, 2 Nemorovich-Danchenko str., Kiev 02001, Ukraine
b
National Technical University of Ukraine “KPI”, 37 Pobedy prosp., Kiev 03056, Ukraine
Abstract
The mechanisms and reasons of catalytic activity of polyaniline

(PANI)-type conducting polymers toward oxygen reduction in acidic and
saline solutions are investigated by electrochemical and quantum-chemical
methods. The PANI/thermally expanded graphite compositions were
developed for realization of fully functional air gas-diffusion electrodes.
Principally new concept for creation of rechargeable metal-air batteries with
such type of catalysts is proposed. The mockups of primary and rechargeable
metal-air batteries with new type of polymer composite catalysts were
developed and tested.
Keywords
Conducting polymers, polyaniline, catalytic activity,

PANI/expanded graphite composites, metal-air batteries, primary &
rechargeable cells.
1. INTRODUCTION
Oxygen reduction electrodes, where molecular oxygen is
electrocatalytically reduced, are vital to metal/air batteries and fuel cells.
Recently, noble metals such as platinum have played a dominating role as
*
89
90
electrocatalysts for the oxygen reduction reaction. However a disadvantage

of Pt is its high cost. There is a serious concern for wide commercial
application of electrocatalysts based on Pt. Search for low-cost
electrocatalysts for the oxygen reduction reaction has been an important goal
of many investigations. Here we report a polyaniline (PANI) as a low-cost
electrocatalyst for oxygen reduction in acidic and neutral solutions. To
profoundly study and to understand the mechanism of the catalytic activity
of PANI, we have carried out quantum-chemical calculations of the
electronic structure of PANI and its adsorption complexes with molecular
oxygen. The above catalytic activity phenomena has found a practical
application for development of primary and rechargeable metal-airbatteries
with low costs PANI/Expanded Graphite composite catalysts.
2. EXPERIMENTAL
The investigations in this work were mainly aimed at studying a
polymer prepared chemically. The synthesis of PANI was carried out in a
1M solution of hydrochloric acid. The aniline concentration was not higher
than 0.5M. Potassium dichromate was used as chemical oxidant. The
polymer was extracted from the reaction medium by filtering the solution.
The precipitate was then washed several times with distilled water. Final
washing was performed in ethyl alcohol and acetone. PANI was dried at a
temperature of no more than 100°C. To build a laboratory-scale model of
polymer electrode, a polyaniline paste based on polytetrafluoroethylene
(PTFE) emulsion was prepared. The dry PTFE residue content was 5% of
PANI weight. The experimental PANI electrodes were made by applying the
polymer paste to a current collector of graphite with 0.5cm2 area. The
electrochemical measurements were made in a standard three-electrode cell
using a silver chloride reference electrode and a platinum counter electrode.
The measurements were made on a PI-50-1.1 potentiostat. All potentials in
the paper are given with respect to a normal hydrogen electrode (NHE).
With respect to calculations of the electronic structure of molecular
clusters and their complexes with oxygen, the PM 3 program of the
computer-driven system called “MOPAC” was used.
3.1. Assessment of PANI Catalytic Activity (the Experiment)

The catalytic activity of PANI for the oxygen reduction reaction was
first characterized by using a thin film electrode as the working device in
oxygen-and argon-saturated 1 M HCl solution.
91
Figure 1. Cyclic voltamogramms (2 mV/s) of a PANI electrode in argon (1) and oxygen (2)
saturated 1 M HCl solution.
Figure 2. Galvanostatic discharge of a PANI electrode with current density at 20 mA/g (1)
and 200 mA/g (2) in oxygen-saturated 1M HCl.
Figure 1 shows cyclic voltammograms in these two solutions.

A reduction current is observed with two reduction peaks in the oxygen-
saturated electrolyte. In the argon-saturated electrolyte, a single reduction
peak at lower current density is observed. From the difference between
the reduction currents in the two solutions, the reduction current in the
92
oxygen-saturated solution must be due to catalyzed oxygen reduction

reaction. On the other hand, the reduction of oxygen on PANI occurs at
potentials that are close to the process of change of electrically conductive
emeraldine (EM) into nonconducting leucoemeraldine (LEM). The
formation of LEM leads to a sharp increase in electrode resistance and to
loss of the electrochemical activity of the electrode. In the presence of
oxygen, much electricity is spent on its electroreduction, and the process of
nonconducting-phase formation slows down. Owing to this, the electrode
loses its electroactivity at more negative potentials and is characterized by
the appearance of an additional cathodic peak at E ~ 0 V.
Figure 2 shows that in the case of galvanostatic discharge of a PANI
electrode with relatively low currents, a region of stable potential value (E ~
0.1 V) appears (curve 1 in Figure 2). This region actually corresponds to the
electroreduction of oxygen at the PANI electrode.
At the same time, the cathodic polarization of the electrode is
accompanied by the formation of a nonconducting phase, leading to an
increase in electrode resistance. Thus, at increased current densities, the
region of stable potential value cannot be obtained as a result of electrode
resistance increase and ohmic drop, as is shown by curve 2 in Figure 2.
3.2. Assessment of PANI Catalytic Activity (the Theory)
The above experimental data (Figs. 1 and 2) allow us to estimate the

possibility for taking place the oxygen reduction process on PANI catalyst. If
the special procedure is not taken to withdraw oxygen out of solution, PANI
electrode in 1 M HCl usually demonstrates a potential of about 0.6 V (SHE).
In compliance with the Fig. 1 the position of the reduction peak might be at
0. 28 V. From the data of the standard potentials for the following reaction:
O2 + 2H+ + 2e- = H2O2, E0= 0.682 (1)

O2 + H+ + 2e- = HO2-, E0= 0.338 (2)
it is possible to assume that oxygen reduction should occurs according to the

above equations and includes an initial two-electron reduction of O2 to H2O2
and HO2-. We have tested this approach by the qualitative reaction on
hydrogen peroxide in the solutions with PANI electrode under cathodic
polarization. The 0.5 M KI solution was used as an indicator. This
qualitative reaction has proved the formation of H2O2 or HO2- in the solution.
In order to explain the catalytic activity of PANI, we have modeled
the electronic structure of doped molecular PANI clusters and its adsorption
complexes with oxygen and hydrogen. The geometric and electronic
93
structures of PANI were modeled by a molecular cluster (Figure 3), which

contained two phenyl rings in the polymer chain.
H21 H23 H11 H10

C 20 C 22 C3 C9
H57 H59 H60
C 46 N 19 C 18 C 13 N 12 C1 C7 N8 C 24
Cl61 Cl62
C 16 C 14 C2 C5
H17 H15 H4 H6
Figure 3. Fragment of a molecular cluster of PANI.
The model clusters were chosen in accordance with the structure of

PANI proposed in Refs. [1, 3]. The present model of PANI also takes into
account that under the influence of dopants (in this particular case, protons
and anions (chlorine ions) which form bonds to PANI nitrogen), the spins of
the highest occupied molecular orbital (HOMO) become unpaired, and PANI
changes to triplet state. It should be noted that only in this state there is a
considerable increase in PANI conductivity.
In constructing the initial model of PANI, standard values of C-H,
C-C, and C-N valence bond lengths were originally adopted. The valence
angles were taken to be 120 degrees. The results of calculations of PANI and
AC’s “PANI-oxygen” are listed in Tables 1 and 2.
The calculation of the electronic structure of PANI and adsorption
complexes (AC) “PANI- oxygen” has been carried out with complete
optimization of the geometric bond parameters.
Calculations show that between the dopant anion (Cl-) and nitrogen
atom there may form a weak ionic-covalent chemical bond (the bond order is
no more then 0.3) with considerable contribution of the ionic component. As
a result of doping, the coordination number (CN) and the positive-charge
value of nitrogen atoms increase (Table 1). The large positive charges on the
nitrogen atoms N8 and N19 at CN=4 suggest that under the influence of
dopants, the nitrogen atom changes into cationic form. This conclusion
agrees with the results of Ref. [2].
As it follows from Table 1, a ʌ-charge of over 1e is concentrated on the
carbon atoms. Therefore, each carbon atom can be in principle an adsorption
site. The optimization of the position of oxygen atoms relative to the surface
was performed at “frozen” geometric parameters of the PANI cluster.
Calculations of the adsorption of molecular oxygen on PANI showed that the
carbon atoms of the phenyl ring, which are shown by Figure 4, are the most
stable surface adsorption sites (AS).
94
Table 1.Charges on atoms and ʌ-electron density distribution on PANI and PANI-oxygen
AC’s (for the adsorption sites C13 and C22).
PANI-O2 PANI
ʋ Atom Charge ʌ-electron Charge ʌ-electron
density density
1 C -0.0412 1.0576 -0.0424 1.0545
2 C -0.1547 1.0348 -0.1550 1.0356
3 C -0.1441 1.0343 -0.1502 1.0326
4 H 0.1217 - 0.1267 -
5 C -0.0202 1.0357 -0.0214 1.0316
6 H 0.1402 - 0.1421 -
7 C -0.2203 1.0495 -0.2092 1.0506
8 N 0.4609 0.3320 0.4408 0.1031
9 C -0.0630 1.0405 -0.0674 1.0330
10 H 0.1062 - 0.1055 -
11 H 0.1165 - 0.1136 -
12 N 0.1101 0.3484 0.1405 0.3767
13 C 0.1059 0.2882 -0.0572 1.0528
14 C -0.1642 1.0294 -0.1439 1.0349
15 H 0.1241 - 0.1146 -
16 C -0.0395 1.0180 -0.0279 1.0293
17 H 0.1495 - 0.1431 -
18 C -0.1400 1.0564 -0.1964 1.0506
19 N 0.4261 0.2371 0.4298 0.1031
20 C -0.1371 1.0320 -0.0763 1.0314
21 H 0.1209 - 0.1078 -
22 C 0.0424 0.1963 -0.1268 1.0330
23 H 0.1469 - 0.1263 -
24 C -0.2176 1.0519 -0.2053 1.049
34 N 0.1211 0.2975 0.1188 0.2933
46 C -0.1710 1.0629 -0.1747 1.0559
57 H 0.0506 - 0.0475 -
58 H 0.0537 - 0.0530 -
59 H 0.0726 - 0.0581 -
60 H 0.0449 - 0.0447 -
61 Cl -0.4051 - -0.3864 -
62 Cl -0.3561 - -0.3926 -
63 O -0.2001 0.0557 - -
64 O -0.1806 0.0826 - -
95
Moreover, calculation shows that molecular oxygen can be adsorbed

on the PANI surface only when both oxygen atoms form bonds to surface
atoms, i.e. a bridge model of adsorption is most probably (Figure 4).
The noticeable increase in the negative charge on O2 (ads.) from 0 to
– 0.3…-0.42 e (Table 2) indicates that the PANI surface in the PANI-O2 AC
is an electron density donor and O2 (ads.) an electron density acceptor.
Table 2. Heats of formation, H (kcal), magnitude of electron density, Q (O2), displace from
the cluster to adsorbed oxygen molecules, equilibrium interatomic distances R (A-B) and
bond orders P (A-B), corresponding to them, in PANI-O2 adsorption complexes.
Bond H, Q (O2) R (O63-O64), A0 P (O63-O64)
kɫal
C2-C5 312.41 -0.3581 1.5059 0.9926
C3-C9 312.89 -0.3699 1.5015 0.9939
C5-C7 323.86 -0.3555 1.5110 0.9877
C7-C9 322.97 -0.3729 1.5052 0.9906
C14-C16 308.58 -0.4025 1.5045 0.9933
C20-C22 312.93 -0.4212 1.4982 0.9822
C13-C14 326.98 -0.3674 1.5036 0.9887
C13-C22 327.05 -0.3807 1.4969 0.9896
O64
O63
C20 C22
H57 H59
N19 C18 C13 N12 C1
Cl61
C16 C14
Figure 4. Bridge model of oxygen adsorption on PANI.
Electron density is transferred from PANI to O2 mainly through the

loss of electron density by adsorption sites. The latter (e.g. C13 and C22)
become electropositive (Table 1). The electron density displaced from the
surface to O2 during adsorption occupies the antibonding MO’s of adsorbed
oxygen molecule which is responsible for the destabilization of the O-O
bond in O2(ads.) during the chemisorptions of O2 on PANI, S -bonds in
O2(ads.) and surface AS’s undergo rupture. Instead of two rupture S - bonds,
two donor-acceptor C-O ı-bonds are formed between PANI AS`s and the
oxygen atoms that lie outside the phenyl ring plane.
96
Calculation of the heats of formation (H) of the bonds between

oxygen atoms and PANI (62) AC atoms (Table 2) shows the bonds in which
one of the AS`s (C1, C7, C13, C18) forms a bond to the nitrogen atom to be
characterized by the highest H values. In this case, the stability of adsorption
complexes is 10 - 15 kcal higher than that of the complexes formed on the
carbon atoms not bonded to nitrogen. Thus, oxygen adsorption on pairs of
AS`s which contain one of the carbon atoms C1, C7, C13, C18 is more
probable than that on pairs of side chain sites.
Calculations show (Table 2) that the 0 - 0 bond orders in
chemisorbed oxygen molecules are almost 30 % lower than those in free O2
molecule. Besides, a noticeable increase (of more than 20 %) in O-O bond
length takes place during adsorption. Thus, chemisorbed O2 molecules have
a fairly high degree of activation and can be easily reduced, which accounts
for the catalytic activity of PANI.
3.3. PANI/Graphite (Carbon) Composites

The next important stage in the development of porous gas-diffusion
electrode is an investigation of influence of thickness of PANI layer (or more
easy controlled parameters like PANI mass or electrochemical capacity) on
the local currents of O2 electroreduction (table 3).
Analysis of data presented clearly suggests that a sharp increase of
catalytic activity takes place in relatively thin layer of PANI, which
corresponds to the mass of about 0.6...0.9 mg (or Q=108...164 mC) per
geometric surface of electrode (0.5 cm2). Further increase of PANI thickness
(or mass) cannot increase enough the catalytic activity of porous electrode.
That is the reason why the developed by us PANI/graphite composites have
relatively high content of graphite (Table 4).
Table 3. Local currents of O2 electroreduction at Graphite / PANI electrode (S=0.5 cm2) with
different mass (and electrochemical capacity) of PANI in 1M HCl solution.
Mass of PANI, Capacity of I1, I2, I3,
M PANI, µA µA µA
mg mC
0.3 52 13 72 108
0.6 108 27 89 125
0.9 164 31 97 141
1.2 193 33 103 154
1.5 240 37 106 159
2.0 341 39 107 162
Notes for the Table:
The conditions of oxygen supply to the solution: I1 - with low O2 content (saturated by N2 );
I2 – with middle O2 content (saturated by O2); I3 – with maximal O2 content (during O2
bubbling on the electrode surface).
97
Our experiments, as well as analysis of the proposed theoretical

model for a generalized system of porous electrode "active material – carbon
additive" proved that thermally exfoliated graphite (TEG) can be one of the
most effective conductive additive and structural support for the different
new and existing active materials of primary and secondary batteries,
including "Air/PANI-Zn" batteries. The reason for such wide application of
TEG is a following unique complex of TEG properties: low density,
relatively high conductivity and stability to electrochemical oxidation during
battery serviceability. Also it is interesting to note that particles of TEG are
relatively hydrophobic. Thus it should be of a great benefit to use composite
based on TEG for development of gas-diffusion electrodes.
The compositions of composites are different for primary and
rechargeable variants of batteries. In rechargeable variant it is necessary to
have considerable mass of PANI for charging of battery without oxygen
evolution. In primary variant it is enough to have a relatively small amount
of PANI for creation of its quite thin layer at the conductive and stable
graphite collector. Thus we have developed two main types of composite
materials based on TEG and PANI (Table 4).
Table 4. Composition of PANI/TEG composite electrodes.

Type of battery mockup Composition of electrode, %
PANI* TEG**
Primary battery 30 70
Rechargeable battery 90 10
*Chemically synthesized Polyaniline, KNUTD (Ukraine)
**Purified, thermally exfoliated natural flake graphite, Superior Graphite Co. (USA)
3.4. New Concept for Development of Rechargeable

Metal-Air Batteries
The main concept for development of metal-air batteries with new
low-cost composite polymeric catalysts is to use catalytic activity of
PANI/TEG composition towards the oxygen reduction during the discharge
process of battery side by side with non-Faradaic process of anion doping
during the charge process (please, see schemes below).
Schemes of Discharge & Charge Processes
1. Traditional Air-Met Battery (for instance, with carbon catalysts)

Discharge Process:
Met0 – ne Met n+ (1)
O2 + 4e + 2H2O 2OH- (2)
98
Charge Process:
Metn+ + ne Met0 (3)
4OH- - 4e 2H2O + O2 (4)
Side reaction: C + O2 CO2 (5) - destruction of a catalyst
2. Rechargeable Air-Met Battery with PANI-based Catalysts

Discharge Process:
Met0 – ne Metn+ (1a)
2O2 + 4e + 4H+ 2H2O2 (electrochemical)
2H2O2 O2 + 2H2O (chemical stage)
O2 + 4e + 4H+ 2H2O (2a)
Charge Process:
Met n+ + ne Met0 (3a)
PANI – ne + nA- [PANI+nxnA-] (4a)
(intercalation)
Total: Metn+ + PANI + nA- Met0 + [PANI+nxnA-]
Thus, using PANI-type catalysis gives a possibility to exclude

basically the typical side reaction of oxygen evolution during charging of
battery, which usually conducts to destruction of catalytic active Air
electrode.
3.5. Coin Cell Mockups of Rechargeable Metal-Air Batteries
In order to demonstrate the above concept in practicality, we have

developed the coin cell mockups (active electrode area of 1 cm2) of
rechargeable metal-air battery. The composite containing 90% of PANI and
10% of TEG was used in rechargeable Zn-Air batteries. The aim of this
investigation was to check the electrochemical behavior of PANI/TEG gas-
diffusion electrode during the charging of battery. Hence, we have used thin
Zn electrodes, which limited cell capacity (~70 mAh). The galvanostatic
charge/discharge curves of such battery are presented by Figure 5.
It is easy to conclude from the shape of discharge curves that the
first stage of battery discharge is mainly related with the discharge of PANI
component. On the second stage, a region of stable potential appears (U ~
0.95 V). This stage actually corresponds to the electroreduction of oxygen at
the PANI/TEG electrode. When approximately 65% Zn supply has been
depleted, the voltage started to go down rapidly. After that, we switched
battery to the charge mode of operation. The amount of capacity that could
be recharged is limited by the amount of PANI. We have found that it is
99
possible to utilize about 45% of initial capacity obtained on the first

discharge at least during the 50-70 cycles. The limitations on cycling life of
battery in a whole are related with well known problems of Zn electrode
cycling (such as Zn dendrites formation) and could be overcome in the future
by optimal development of separator and Zn electrode construction.
Figure 5. Galvanostatic charge/discharge curves of the coin cell mockups of Zn-Air battery
(Icharge = Idischarge = 3 mA).
In this paper we would like to propose also an additional concept of

utilization of rechargeable metal-air battery with PANI/ TEG catalysts. It is
necessary to note the ability of PANI based electrode to self-charge in the
presence of oxygen. A deep discharge of PANI electrode leads to the
transformation of emeraldine (EM) to lecoemeraldine (LEM). It is well
known that LEM is chemically non-stable at the presence of oxygen due to
occurance of the following reaction:
100
We have confirmed the above reaction by visual observation of EM

to LEM transformation at the ITO glass. Thin film of PANI was
electrochemically deposited on ITO glass. Then the glass electrode was
placed in a standard three-electrode cell as a working electrode. After
potentiostatic polarization at E= -0.2 V (vs. NHE) the green EM was reduced
to LEM, which has a yellow color. The glass electrodes with yellow LEM
were placed in the argon saturated solution. We observed that yellow color
did not change during the few hours. On the other hand, the yellow color was
changed to green at few minutes when electrode was placed in the oxygen-
saturated solution. Thus the LEM can easy transformed to EM by reaction
with oxygen.
For purposes of verifying of the concept of a self-discharge due to
the LEM oxidation by air, we have designed a coin cell with a zinc electrode
and a thin PANI/TEG cathode. The typical curves of voltage change for such
electrochemical device are given by Figure 6.
Figure 6. The galvanostatic discharge curves and self-charge curve for Zn-Air coin battery
with PANI/TEG gas-diffusion electrode.
After the first galvanostatic discharge of battery up to Umin=0.4V a

circuit was disconnected and computer system has registered an open circuit
voltage. After reaching the Umax = 1,2V a circuit was connected again (the
second discharge). It is interesting to note that during the second discharge a
cell has demonstrated about 80 % of its initial capacity.
It was shown by our investigations that such cell can be cycled with
self-charging at least during 10-15 cycles. After that, the internal resistance
of such cell is incremented considerably due to the accumulation of
101
discharge products on a zinc electrode. Thus, such battery can represent

sufficient interest for some devices with a pulse mode of power consumption
(for example, in signalling devices). It is necessary to note that by
optimization of battery construction rather high value of a current in a pulse
mode could be reached. Such battery represents some kind of the hybrid
supercapacitor. PANI/TEG electrodes can be discharged with enough high
rates and it gives possibility to use such electrodes also in supercapacitors.
3.6. The Full Size Cylindrical Metal-Air Batteries

The experimental samples of Air-Zn and Air-Mg batteries with
PANI-TEG catalysts were realized in the standard AAA size of MnO2-Zn
primary batteries. The composite material based on 70% of TEG was used
for construction of cylindrical cells of metal- air battery.
The discharge curves for all batteries (1, 2) have horizontal sections
which correspond to the properties of air PANI/TEG composite electrode
(Figure 7).
Figure 7. Discharge curves of AAA size Air/Metal cells (1, 2) with composite PANI catalysts,
as well as discharge curve of separate Air PANI/TEG composite electrode (3). Discharge
current I =5mA, temperature 18°C. Electrochemical Systems: (PANI) AirµMgBr2 µMg (1);
(PANI) AirµZnCl2 µ Zn (2); (PANI) Airµ ZnCl2 µµAgClµAg (3).
The discharge curve for Air electrode, which was measured using
additional third AgµAgCl reference electrode, is ideally horizontal since
it is determined by the oxygen reduction potential in according medium
102
(Fig. 7, curve 3). The capacities of batteries are limited by the amount of Zn
(or Mg) in batteries (Fig. 7, curves 1 and 2).
The main characteristics of cylindrical AAA size metal-air batteries
with PANI/TEG catalysts, as well as standard Zn-MnO2 battery have been
gathered in the Table 5. Realization of all types of batteries in the same AAA
size gives the possibilities for comparison of above electrochemical systems
for some applications.
Table 5. Characteristics of cylindrical AAA size metal-air batteries with PANI/TEG catalysts,
as well as MnO2-Zn battery.
Performance Characteristics of cells (PAN) O2|Zn (PAN) O2|Mg MnO2|Zn
Discharge capacity, mAh 420 340 380

Specific energy, Wh/kg 140 200 45
Weight of the cells, g 3 2.7 11
Dimensions, cm
4.5 / 1.0 4.5 /1.0 4.5/1.0
(H / d )
Open circuit voltage, V 1.2 1.8 1.5
Average discharge voltage, V (at 1.0 1.6 1.1-1.3
RDISCH = 200 Ohm)
Short - circuit current, A 1.0 1.2 1.4
Self - discharge, % within a month 1.5-2.0 2.0-4.0 1.0-3.0
Service life, years a5 <1 <2
The average weight for the experimental Air-Zn and Air-Mg

batteries of AAA size is 3 g and 2.7 g correspondingly, vs that of 11 g for
traditional MnO2-Zn battery. That is the main reason of sharp increase of the
specific energy of Air-Zn (140 Wh/kg) and Air-Mg (200 Wh/kg) batteries
vs the traditional chemistry (45 Wh/kg).
The specific energy of Air-Zn battery based on slightly acidic
electrolyte (ZnCl2 with pH 5) is not very high, as compared with an ordinary
Air-Zn battery with alkaline electrolyte, due to the relatively low average
discharge voltage (1.0 V during discharge on the resistor of 200 Ohm).
Nevertheless, the service life of such systems can be greater than that of
ordinary alkaline systems owing to the absence of electrolyte carbonization.
For example, some experimental samples of Air-Zn batteries have
effectively demonstrated the service life of about 5 years.
Air-Mg batteries have larger average discharge voltage (1.6 V) and
specific energy, but did not have a great service life and abilities for cycling
due to the high activity of Mg electrode.
In general, we believe that new types of batteries with composite
PANI/TEG catalysts can find their own diapason of practical applications
depending on concrete conditions of application.
103
4. CONCLUSIONS
Polyaniline (PANI) was investigated as electrocatalyst for the
oxygen reduction reaction in the acidic and neutral solutions. Galvanostatic
discharge tests and cyclic voltammetry of catalytic electrodes based on
polyaniline in oxygen-saturated electrolytes indicate that polyaniline
catalyzes two-electron reduction of molecular oxygen to H2O2 and HO2-.
In order to study the possible reasons and mechanisms of the
catalytic activity of conducting polymers (CPs), we have modeled the
electronic structure of some molecular CPs clusters and its adsorption
complexes with oxygen.
A specialized “MOPAC” computer software package and, in
particular, its “PM3” quantum-chemical program has been successfully
applied in calculations. The results of calculations have shown that both
oxygen atoms form bonds with two more active carbon atoms of CP
molecular cluster (so-called “bridge” model of adsorption). The total energy
of system after a chemical adsorption at such active atoms is minimal.
In the CP-O2* complex, the CP surface is an electron density donor.
For example, in the case of PANI, the bond orders in adsorbed O2* molecules
decrease by about 30%, and the bond lengths L increase by about 24%.
Thus, the adsorbed O2* molecules have a fairly high degree of activation and
can readily interact with the protons. Hence, quantum-chemical analysis
confirms the mechanism of O2 electroreduction and gives possibility to
understand the reasons of PANI catalytic activity.
The above phenomenon has found a practical application for
development of metal-air battery mockups with low costs PANI/Expanded
Graphite composite catalysts.
Specific energy, which could be attained as primary battery is of
about 140 Wh/kg for Air/PANI-Zn and 200 Wh/kg for Air/PANI-Mg
batteries. The discharge curves of such batteries is practically horizontal
since there are determined by the oxygen reduction potential.
The effective service life of Air/PANI-Zn battery can be greater by
several folds than that of conventional alkaline air-zinc batteries owing to the
absence in principle of electrolyte carbonization (pH d 7).
In addition, the batteries can be rechargeable, the cyclability attained
with them being limited now by the properties of Zn (Mg) electrodes and a
separator.
A battery can be easy realized in any shape (prismatic, coin or
cylindrical) and any size, including micro-size.
We believe that new type of conducting polymer / expanded graphite
composite electrodes as gas-diffusion cathodes will find in perspective a
practical application for some types of batteries and fuel cells.
104
ACKNOWLEDGEMENT
Authors would like to acknowledge Superior Graphite Co., Chicago,
IL, USA for manufacturing and submission of various customized graphite
samples to do the development work described in the paper. We like to also
acknowledge NATO Science for Peace Programme for financial support of
this work in framework of “Carbon” SfP 973849 project.
REFERENCES
1. Alan G. MacDiamid and Heeger, J. Semiconduction and Metallic Polimer: The Fourth
Generation of Polymeric Materials. J. Phys. Chem. 2001; B105:8475-8491.
2. Bazong Zhao, Neon K.G., Liu F.T., at.al., Interactions Between Polianiline and
Viologens. Synthetic Metals 2001;123:263-266
3. Alan G.MacDiamid, "Syntethetic <etals": A Novel role for Organic Polymers. Current
Applied Phys 2001; 1: 269-279.
MECHANICALLY RECHARGEABLE
MAGNESIUM-AIR CELLS WITH
NaCl-ELECTROLYTE
Anastassia Kaisheva* and Ilia Iliev
Central Laboratory of Electrochemical Power Sources,

Bulgarian Academy of Sciences, 1113 Sofia, Bulgaria
Abstract
Metal-air cells are developed with air gas-diffusion cathodes and

Mg-anodes. Non-aggressive NaCl-solution is used as electrolyte. Carbon
based catalysts for the oxygen reduction are selected and tested in the air
gas-diffusion electrodes. Various Mg-alloys are tested as anodes. The V-A,
power and discharge characteristics of the Mg-air cells are investigated.
Keywords
Electrochemical energy sources; Magnesium anode; Air gas-

diffusion electrode; Neutral electrolyte.
1. INTRODUCTION
Metal-air batteries are characterised by a high energy density, a flat
discharge voltage, long shelf life and relatively low price. A great advantage
of the metal-air batteries is the use of a cheap and inexhaustible reactant –
oxygen from the atmospheric air, which is reduced on an invariable air
cathode [1].
Various metals can be used as an anode in metal-air cells, the
amperhour capacity of the cell being determined by the type of metal anode.
Zinc is commonly used as an anode in metal-air batteries. Zinc-air
batteries are commercially available and used as power supply for navigation
buoys, communication systems, hearing aids, and etc. [2]. Mechanically
rechargeable Zn-air batteries are developed recently for traction [3].
*
105
106
One of the main inconveniences of the zinc-air batteries is the use of

concentrated KOH-electrolyte. Strong alkaline solutions are chemically
aggressive and their contact with the human body can cause serious injuries.
The metal-air cell is an “open” system in which the electrolyte in the cell is
in contact with the surrounding air through the pores of the air gas-diffusion
electrode. There is a risk that drops of electrolyte may come out of the cell
and cause accidents. The mechanical recharge of the zinc air cells poses also
problems: the removal from the cell of the electrolyte and of the porous,
flooded with electrolyte consumed zinc anodes is an operation with high
chemical danger, which can be accomplished by trained people or robots
only. In this respect the development of metal-air cells operating in non-
aggressive electrolyte is of great interest. Promising results have been
obtained with magnesium-air cells.
2. EXPERIMENTAL
Magnesium is a reactive metal with relatively high electrochemical

equivalent of 2.2 Ah/g and low specific gravity 1.74 g/cm3; its price is not
high. This makes magnesium suitable for use as anode in metal-air cells.
A passive film is formed on the Mg electrode surface in alkaline
solutions, which creates problems when the Mg-air cell is to be discharged.
However Mg and especially some Mg alloys are suitable to be used in
magnesium-air cells with NaCl-electrolyte [4], where the formation of the
passive film on the magnesium surface is not observed.
The magnesium-air cells under investigation are electrochemical
cells containing two air gas-diffusion electrodes (cathodes) and a magnesium
anode. NaCl-solution is used as an electrolyte.
3.1. Air Gas-Diffusion Electrode
The air gas-diffusion electrode developed in this laboratory [5] is a

double-layer tablet (thickness ca.1.5 mm), which separates the electrolyte in
the cell from the surrounding air. The electrode comprises two layers: a
porous, from highly hydrophobic, electrically conductive gas layer (from the
side of the air) and a catalytic layer (from the side of the electrolyte). The gas
layer consists of a carbon-based hydrophobic material produced from
acetylene black and PTFE by a special technology [6]. The high porosity of
the gas layer ensures effective oxygen supply into the reaction zone of the
electrode; simultaneously the leakage of the electrolyte through the electrode
107
is completely avoided because of high hydrophobicity of the pores-surface of

the gas layer. The catalytic layer of the electrode comprises uniformly
distributed zones of the same hydrophobic material and carbon-based
catalyst. An effective supply of oxygen in gas phase is achieved through the
zones made from the hydrophobic material.
Various carbon-based catalysts for the electrochemical oxygen
reduction have been tested in the air gas-diffusion electrodes [7]. The
polarization curves of the air electrodes were measured when operating
against an inert electrode in 2 N NaCl-solution. The potential of the air
electrodes was measured versus saturated calomel electrode (SCE).
It is found that some types of active carbons possess enough
catalytic activity to be used as catalysts in air electrodes operating at low c.d.
Catalysts from active carbon promoted with mixed Ni-Co- oxides
are investigated [8]. These catalysts are produced by heat-treatment of active
carbon impregnated with solution containing Co-and Ni-acetates. The
concentration of the used Co and Ni- solutions is varied in order to optimise
the content of the mixed Co-Ni oxides in the catalyst.
Figure 1. Polarisation curves of air gas-diffusion electrodes with various catalysts.

108
Catalysts from active carbon additionally activated with cobalt- or

iron- phthalocyanines are also studied [7]. The results show that at current
densities up to 50 mA/cm2, the polarization of the air electrodes with catalyst
from active carbon promoted with FePc is lower than that of the electrode
with catalyst from active carbon promoted with CoPc. At higher current
density the polarization of the electrode with catalyst from active carbon
promoted with CoPc is lower, which is probably connected to the lower
transport hindrances, due to the more favorable structure of this catalyst.
Pyrolyzed catalysts obtained by heat treatment in Argon of active
carbon impregnated with solution of the compound Co-
tetramethoxyphenylporphirine (CoTMPP) are studied [9]. Air gas-diffusion
electrodes with this catalyst show low polarisation in a wide interval of
current densities (up to 100 mA/cm2) and stable long-term performance.
These catalysts are more suitable for use in magnesium-air cells operating at
high current drains, but unfortunately their price is comparatively high.
In Figure 1 we compare the polarization characteristics of the air
gas-diffusion electrodes with the described catalysts.
One of the main difficulties in the design of air gas-diffusion
electrodes operating in NaCl-solutions is to ensure the efficient current
collection. Usually the current collection is achieved by means of a metal
grid pressed on the catalytic layer of the electrode. However, it seems
impossible to find a metal, chemically stable in contact with the saline
solutions (at the working potentials of the air electrode). That’s why the
current collection in the air electrodes developed in this laboratory is
achieved through a metal grid pressed into its hydrophobic, electrically
conductive gas layer. In this case the current collector grid (a cheap Pb-
plated iron grid is used) is not in contact with the NaCl-electrolyte and the
corrosion problems are eliminated.
3.2. Magnesium Anodes
Sheets of magnesium or magnesium alloys were used as anodes in

the developed magnesium-air cells.
Several types of magnesium alloys are tested: pure Mg, MgAM60,
MgAM50 and MgAZ91 (Norsk Hydro Magnesium GmbH) and MA8M06
(Mg/TU190116/74) (Russia). The sheet of magnesium type MA8M06 is
covered by the producer with a thin chromium protective layer, which was
being dissolved immediately after its contact with the NaCl-solution.
The elemental compositions of the Mg-alloys used in this work
(according the producers’ product information data sheets) is presented in
Table 1.
The current collection of the Mg anodes is achieved through the
magnesium itself, without using additional current collectors. This type of
109
current collection seems to be very easy but it shows some inconveniences.

At the end of the discharge, the Mg anode is broken out and some pieces of
magnesium are dropping into the electrolyte, so they cannot be discharged.
This results in decrease of the cell capacity. In order to overcome the
problem a special protective electrically conductive sheet (based on carbon)
is used pressed between two magnesium sheets to form a “sandwich”. The
use of such “sandwich”-type magnesium anodes results in an improvement
of the current collection of the magnesium anode, especially at the end of the
discharge and in a more complete utilization of the magnesium.
Table 1. Composition of the Mg-alloys used as anodes.
Metal, % Mg Fe Zn Mn Si Al
pure Mg 99.99 - - - - -
MgAM60 94 - - - - 6
MgAM50 95 - - - - 5
MgAZ91 90 - 1 - - 9
MA8M06 ~ 97 0.02 0.12 1.3 0.01 0.12
2,0
Mg - air cell Mg: MA8M06
2 air electrodes S = 80 cm2 23oC
1,8
1,6
1 N NaCl
Voltage, V
4 N NaCl
1,4
2 N NaCl
1,2
1,0
0,8
0 1.0 2.0 3.0 4.0 5.0

Current, A
Figure 2. Typical current-voltage curves of Mg-air cell (Sair = 80 cm2;

Mg – MA8M06) operating in 1 N, 2 N and 4 N NaCl-electrolyte.
110
3.3. Experimental Mg-Air Cells
Several types of experimental magnesium-air cells were tested.

These cells varied in their size (the working area of the air electrodes used)
[10]. The current-voltage curves of an experimental Mg-air cell with two air
electrodes (Sair = 80 cm2) with pyrolyzed CoTMPP catalyst and sandwich-
type Mg anode (MA8M06) operating in NaCl-electrolytes with different
concentrations are presented by Figure 2.
The difference in the obtained curves is probably due mainly to the
difference in the conductivity of the NaCl solutions with different
concentrations. As expected, the performance in 4N NaCl is superior –
currents up to 5 A can be reached at voltage higher than 1 V.
The discharge curve of the same cell at constant load ca. 0,3 : is
presented in Figure 3.
1,4 4,0
voltage 3,5
1,2
3,0
Current, A
1,0
Voltage, V
2,5
0,8
current 2,0
0,6
1,5
0,4 Mg-air cell 1,0
Sair = 80 cm2 4 N NaCl
0,2 0,5
Mg: MA8M06 24oC
0,0 0,0
0 50 100 150 200
Time, min
Figure 3. Discharge curve of the Mg-air cell from Fig. 1 at constant load 3 :.
The testing of the different magnesium anodes is performed in

magnesium-air cells with two air electrodes with pyrolyzed CoTMPP
catalyst (S = 65 cm2). One-sheet Mg anodes were used in 4N NaCl media.
In Figure 4 we compare the current-voltage characteristics of the
investigated Mg-air cell with different Mg-anodes.
The comparison of the data from Figure 4 shows that the pure
magnesium (99.99%) is the most suitable material to be used as anode in
magnesium-air cells with NaCl-electrolyte. The magnesium alloy type
111
MA8M06 (Russia) which is produced specially for use in water activated

Mg-AgCl batteries is also suitable for use in Mg-air cells.
1,8 Mg - air
2 air elecrodes S = 65 cm2
4 N NaCl
1,6 T = 21oC
99,99% Mg
1,4 Mg:MA8M
Voltage, V
Mg:AZ91
1,2
Mg: AM60
1,0
0,8 Mg:AM50
0,6
0 1 2 3 4 5 6
Current, A
Figure 4. Current-voltage characteristics of Mg-air cells with different Mg- anodes.
Figure 5. Discharge curves of Mg-air cells with anodes from Mg 99.99%

and from Mg-alloy MA8M06.
112
In Figure 5 we compare the discharge curves at a constant load of

0,6 : of similar magnesium-air cells with anodes from Mg 99.99% and from
Mg-alloy MA8M06.
3.4. Mechanically Rechargeable Mg-Air Cells

Mechanically rechargeable magnesium-air cells were designed and
investigated. The most important feature of these cells is that the cell-case
with the air electrodes can be used many times.
Figure 6. Experimental mechanically rechargeable Mg-air cell.
1,5
initial curve
Votage, V
1,0
after 12th mechanical recharge
0,5 Mg - air cell

2 air electrodes (S=223 cm2)
4N NaCl
t = 26oC
0,0
0 5 10 15 20
Current, A
Figure 7. Current-voltage characteristics of one and the same Mg-air cell:

initial and after 12th mechanical recharge.
113
Once the Mg-anode has been discharged, it was removed out of the
cell together with the electrolyte and new Mg electrode and fresh NaCl-
electrolyte were introduced, by which the cell is operational again. The
electrolyte in the cell was also changed.
The initial current-voltage curve and the current-voltage curve after
12th mechanical recharge of one and the same Mg-air cell are presented in
Fig. 7. The results show that the Mg-air cell can be recharged mechanically
several times without any significant loose in its performance.
Figure 8. Mechanically rechargeable Mg-air cell (Sair = 223 cm2).
1,8 60
50
1,6
Power, W
40
Voltage
Voltage, V
Mg - air cell
1,4 Sair electrode= 420 cm2
30
4 N NaCl
1,2 20
10
1,0 Power
0
0 10 20 30 40 50 60
Current, A
Figure 9. Current-voltage and power characteristics of the Mg-air cell from Fig. 8.
114
Two types of mechanically rechargeable Mg-air cells for operation

at high nominal currents are designed. The air electrodes in these cells are
with pyrolyied CoTMPP catalyst. Mg anodes from the alloy MA8M06 were
used in this design.
In Figure 8 we have presented a picture of the mechanically
rechargeable Mg-air cell with total working area of the air electrodes 420
cm2. It is known [4] that a large amount of heat is released during the
discharge of Mg-air cells. This heat promotes an evaporation of the
electrolyte from the cell, which can be overcome by providing an adequate
electrolyte-reservoir. That’s why the upper part of the Mg-air cell from
Fig. 8 is designed as a reservoir in which an additional amount of electrolyte
is available.
In Figure 9 we presented the current-voltage and the power
characteristics of the Mg-air cell form Fig. 8. It is seen that currents up to
50A could be obtained from this cell at voltage higher than 1 V. The specific
power of this cell is ca. 50 W/kg.
Even more powerful Mg-air cell was constructed with total working
area of the air electrodes S = 660 cm2. Current loads up to 100 A could be
reached with this cell at a voltage higher than 1 V.
The current-voltage and the power characteristics of this cell are
presented in Figure 10.
120
1,5 Voltage 100
80
Voltage, V
Power, W
1,0
60
40
0,5 Mg - air cell
Sair electrode=660 cm2
Power 20
4 N NaCl
0,0 0
0 20 40 60 80 100
Current, A
Figure 10. Current-voltage and power characteristics

of the mechanically rechargeable Mg-air cell (Sair = 660 cm2).
115
4. CONCLUSIONS
Magnesium-air air cells with NaCl-electrolyte were developed and

investigated. The current-voltage and the discharge characteristics of the
cells with were studied. Air gas-diffusion electrodes suitable for operation in
NaCl-electrolytes were designed. Various carbon-based catalysts for the
electrochemical reduction were tested in these air electrodes. Magnesium
alloys suitable for use as anodes in Mg-air cells were found.
Mechanically rechargeable Mg-air cells for operation in middle and
high nominal current were designed.
The main advantage of the magnesium-air cells is the use of NaCl-
electrolyte. This makes the magnesium-air cells environmentally friendly.
They are clean, no toxic emission; are produced in a way that is safe for user,
even for children because the NaCl-electrolyte is not chemically aggressive
and magnesium is not toxic. The mechanical recharging of these cells can be
accomplished without any chemical danger.
Something more: the magnesium-air cells possess an infinite shelf
life because the electrolyte can be removed easily before storage. When
power is needed, the electrolyte is poured back into the cell. No other
electrochemical energy sources do exist in which the electrolyte can be
removed, stored and reused by the consumer.
From the presented results it can be concluded that the magnesium-
air cells feature substantially increased power output. They can be used as
primary, mechanically rechargeable and emergency energy sources for the
residential, commercial and industrial markets.
ACKNOWLEDGEMENTS
This work has been partially supported by ZOXY Energy Systems
AG, Germany.
REFERENCES
1. Kim Kinoshita, Electrochemical oxygen technology, John Willey & Sons, Inc. New
York, Chichester, Brisbane, Toronto, Singapore.
2. Mantell C.L. Batteries and Energy Systems, New York, Mc-Graw-Kill, 1970.
3. Iliev I., Kaisheva A., Stoynov Z. Pauling H.J, Mechanically rechargeable Zinc-air cells,
Proceedings of the 3th International Battery Recycling Congress, July 2 – 4, 1997,
Noordwijk Aan Zee, the Netherlands.
4. Gregory D.P., Metal-air batteries, Mills & Boon Limited, 1972.
5. Iliev I. Air electrodes for primary metal-air batteries, 160th Meeting of the
Electrochemical Society, Denver, Colorado, Oct. 1981, Extended Abstracts, p. 268-269.
6. US Patent No. 4031 033, 1972.
116
7. Iliev I., Air electrodes for aluminium- air batteries, Bull.Soc.Chim.Beograd, 1983; 47
(Supplement): S317-S338.
8. Kaisheva A., Gamburzev S., Iliev I. Carbon air gas-diffusion electrodes for operation in
neutral electrolytes containing mixed metal oxides, Elektrokhimija 1981; 17:1362-66
9. Kaisheva A., Gamburzev S., Iliev I., Characteristics of air gas-diffusion electrodes from
active carbon and pyrolyzed CoTMPP. Acid and neutral electrolytes, Electrochimija
1982; 18:139-43.
10. Kaisheva A., Iliev I., Milusheva J. Mechanically rechargeable magnesium-air cells with
non-aggressive electrolyte. International Congress for Battery Recycling, July 3 – 5,
2002, Vienna, Austria.
APPLICATION OF CARBON-BASED
MATERIALS IN METAL-AIR BATTERIES:
RESEARCH, DEVELOPMENT,
COMMERSIALIZATION
Anastassia Kaisheva* and Ilia Iliev
Central Laboratory of Electrochemical Power Sources,

Bulgarian Academy of Sciences, 1113 Sofia, Bulgaria
Abstract
Metal-air cells have been developed, comprising a carbon air gas-

diffusion electrode providing low polarization and stable long-term operation
and metal anode with low self-discharge, operational at high current
densities. The mechanism of gas-transport in the hydrophobic gas layer of
the air electrode is investigated. Various carbon-based catalysts for the
electrochemical oxygen reduction were tested. Methods for diagnostic of the
activity and transport properties of the catalysts are proposed. A range of
metal air cells developed in this laboratory is presented.
Keywords
Gas-diffusion electrode; metal-air cells; gas-transport in porous

media; carbon-based catalysts.
1. INTRODUCTION
Metal-air batteries combine a metal anode (similar to that used in the
conventional primary batteries) and an air gas-diffusion electrode (cathode)
similar to that used in the fuel cells. During operation the metal anode is
electrochemically oxidized for the expense of the oxygen from the air, which
is reduced on the air gas-diffusion electrode.
The air gas-diffusion electrode possesses two advantages over the
metal-oxide cathode in the conventional primary batteries: infinite charge
*
117
118
capacity and low weight independent of the capacity. The amper-hour

capacity of the metal-air battery is determined by the amount of the active
metal in the anode, but not by the air electrode.
The specific characteristics (such as Ah/kg and Ah/l) of the metal-air
batteries is significantly higher than that of the classical electrochemical
systems with the same metal anode.
The theoretical data [1] shows that Li and Ca possess very high
energy density (13172 and 4560 Ah/kg respectively) but these metals are not
suitable to be used as anodes because of their instability in aqueous
electrolytes. The theoretical energy densities of Mg and Al are also high
(6846 Wh/kg and 8212 Wh/kg). It is shown that some alloys of Mg and Al
can be successfully used as anodes, especially in metal-air cells with neutral
electrolytes. The theoretical energy density of Zn is much lower than that of
Li and Ca, but the self-discharge of Zn can be effectively suppressed by the
use of suitable inhibitors. That’s why the zinc-air batteries with KOH
electrolyte are the first metal-air system brought into service.
2. EXPERIMENTAL
The cathode in all types of metal-air cells and batteries is an air gas-
diffusion electrode. The air gas-diffusion electrode is a porous thin, light
plate, which serves as a wall of the metal-air cell and separates the
electrolyte in the cell from the surrounding air. This poses some conflicting
requirements to the air electrode: it must be highly porous and permeable for
the gaseous oxygen and simultaneously the leakage of the electrolyte
through the porous electrode must be completely avoided. The electrode
must be electrically conductive and must possess enough mechanical
strength to withstand the hydrostatic pressure of the electrolyte and eventual
hydrodynamic shocks. The air electrode must contain an active catalyst for
the electrochemical reduction of oxygen in contact with the electrolyte.
Stable operation of the air electrode with time is also needed.
3.1. Air Gas-Diffusion Electrode
The air gas-diffusion electrode developed in this laboratory [2] is a

double-layer tablet comprising a porous hydrophobic gas layer and a
catalytic layer in which the catalyst is placed. The gas layer of the electrode
is prepared from hydrophobic material: carbon black modified with PTFE by
a special technology [3]. The catalytic layer contains a porous catalyst.
119
3.1.1. Hydrophobic Gas Layer
The hydrophobic gas layer of the air electrode [4] possesses high
porosity (ca. 0,9 cm2/g), such that an effective oxygen supply through this
layer is obtained. From the experimental porogrames measured by both
mercury and 7 N KOH-porometry the contact angle 4eff of the hydrophobic
material with water electrolytes is obtained (4eff =116q – 118q). Because of
its high hydrophobicity, the gas layer prevents the leakage of the electrolyte
through the electrode.
The investigation of the pore size distribution (Fig. 1) shows that
nano-size pores (radius ca. 20 nm) predominate in the gas layer from this
hydrophobic material.
0.015
0.010
dV/dr
0.005
0.000
100 101 102 103 104 105 106

Pore radius - r, nm
Figure 1. Distribution of the volume porosity as a function of the pore radius.
The mechanism of gas transport through the porous hydrophobic

layer of the developed air gas-diffusion electrodes is theoretically and
experimentally investigated [5].
The gas-permeability coefficient K of the material is obtained by
measuring the gas flow J through a flat sample with thickness L under a total
pressure difference 'P. The gas-permeability coefficient K is a two terms
sum:
K = D1K + KK /1/
viscous term KK accounting the viscous flow of the gas (accepted as a

viscous fluid) and Knudsen term D1K accounting the Knudsen diffusion
120
mechanism (which takes into account the influence of the bouncing of the
gas molecules with the walls of the porous system).
Figure 2. Permeability coefficient as a function of the mean oxygen pressure.
In Figure 2 we presented the permeability coefficient K of oxygen as

a function of the mean gas pressure experimentally obtained for a sample of
porous material from acetylene black modified with 35% PTFE. The
experimental linear dependence is obtained. The intercept with the abscissa
corresponds to the Knudsen term D1K. The value obtained is 2,89.10-2 cm2/s.
The slope of the straight line is small, so that the ratio KK/ D1K at mean gas
pressure 1 atm. is small (| 0.1) which means that the gas flow is
predominantly achieved by Knudsen diffusion and the viscous flow is quite
negligible. At normal conditions (1 atm, 25qC) the mean free path of the air
molecules (O | 100 nm) is greater than the mean pore radii in the
hydrophobic material (r | 20 nm), so that the condition (O >> r) for the
Knudsen-diffusion mechanism of gas transport is fulfilled.
Gas diffusion in the nano-porous hydrophobic material under partial
pressure gradient and at constant total pressure is theoretically and
experimentally investigated. The “dusty-gas model” is used in which the
porous media is presented as a system of hard spherical particles, uniformly
distributed in the space. These particles are accepted as gas molecules with
infinitely big mass. In the case of gas transport of two-component gas
mixture (i = 1,2) the effective diffusion coefficient (Di)eff of each of the
121
components is connected with the corresponding Knudsen diffusion

coefficient Di,K and with the effective binary diffusion coefficient (D1,2)eff by
the Bosanquet relation:
1
1 1 1 1 ª İ (1) º /2/
(Di )eff DiK (D1,2 )eff (Di )eff «( ).D1,2 » .P
¬q ¼
in which (D1,2)eff = (H/q).D1,2, where H/q is the porosity/tortuosity factor of

the porous media and D1,2 is the binary diffusion coefficient. D1,2 is
inversely proportional to the gas mixture total pressure P (D1,2 = D1,2(1)/P),
where D1,2(1) is binary diffusion coefficient at total pressure 1 atm.
Figure 3. Effective diffusion coefficient of O2 in O2-N2 mixture as a function of gas pressure.
In Figure 3, we have presented the experimentally obtained

reciprocal values of (D1)eff of oxygen in a sample of the nano-porous
hydrophobic material as a function of the total pressure P of gas mixture (O2-
N2) when the oxygen concentration in the mixture is 21%. From the intercept
of the straight line with the ordinate the value of the Knudsen diffusion
coefficient can be also determined. It must be underlined that the value of
Knudsen diffusion coefficient obtained by these diffusion measurements
(2,86.10-2 cm2/s) is in very good coincidence with the value obtained by the
gas permeability measurements.
It must be noted that the effective diffusion coefficient (D1)eff is
obtained by electrochemical measurements of air gas-diffusion electrodes
with sufficiently thick gas layer so that the limiting process is the gas
122
transport in this layer. The polarization curves of these air electrodes are
measured when operating in a close vessel with a gas cocktail consisting of
electrochemically active (oxygen with partial pressure P10) and
electrochemically non-active (nitrogen) gases. From the polarization curves
the diffusion limited current I1L is obtained which is connected with the
diffusion limited flow of oxygen J1L. The effective diffusion coefficient of
the oxygen (D1)eff is calculated according the equation:
P10
J1L (kT) 1 . (D1 ) eff . ( ) /3/
L
It can be concluded that the predominating mode of gas transport in

the investigated nano-porous hydrophobic material is Knudsen diffusion, so
that the diffusion is the main mechanism of gas transport in electrochemical
systems based on such material and operating with gaseous reactants.
3.1.2. Catalytic Layer
The catalytic layer of the air electrode is made from a mixture of the
same hydrophobic material and porous catalyst [2]. It comprises
hydrophobic zones through which the oxygen is transported in gas phase and
zones containing catalyst where the electrochemical reduction of oxygen is
taking place. It must be noted that the overall structure of the electrode is
reproducible when various kinds of carbon-based catalysts are used.
One of the main problems in the development of air gas-diffusion
electrodes for metal-air cells is to find active and stable catalysts for the
electrochemical reduction of oxygen. Carbon-based catalysts are mostly
used, because of their highly developed surface area and capability for
adsorption of O2, suitable morphology, chemical stability, good electric
conductivity and comparatively low price.
Various carbon-based catalysts were tested in the investigated air
gas-diffusion electrodes: pure active carbon [6], active carbon promoted with
silver [7] or with both silver and nickel. Catalysts prepared by pyrolysis of
active carbon impregnated with a solution of the compound Co-
tetramethoxyphenylporphyrine (CoTMPP) are also studied [8].
It is well known that the performance of the air gas-diffusion
electrode is influenced not only by the activity of the catalyst, but also by all
transport processes taking place in its porous structure. In addition, the
transport hindrances in the electrode are function not only of its overall
structure, but also of the porous structure and the surface properties of the
catalyst. Methods for diagnostic of the activity and the transport properties of
air gas-diffusion electrodes were proposed [9].
123
It was experimentally found that the polarization curves of the

investigated air gas-diffusion electrodes in a semi-logarithmic scale at low
current densities (below 10 mA/cm2) are straight lines, which can be treated
as Tafel plots. At these low current densities the transport hindrances in the
air electrode are negligible so that activation hindrances only are available.
In Figure 4 we have presented the experimental Tafel plots of air
electrodes with catalysts from pure active carbon and from active carbon
promoted with different amounts of silver. The obtained curves are straight
lines with identical slopes. It must be underlined that the investigated
electrodes possess identical gas layers and catalytic layers, which differ in
the type of catalyst used only. Therefore, the differences in the observed
Tafel plots can be attributed to differences in the activity of the catalysts
used. The current density a at potential zero (versus Hg/HgO), obtained from
the Tafel plots of the air electrodes is accepted as a measure of the activity of
the air gas-diffusion electrodes: the higher value of a corresponds to higher
activity of the air electrode.
Figure 4. Tafel portions of the polarization curves of air electrodes

with carbon-based catalysts.
From Figure 4 it is visible that the electrodes with silver in the

catalyst are more active than that with pure active carbon catalyst.
Moreover, the increase of the amount of promoting silver in the catalyst
results, as expected, in a higher activity of the electrode.
Nevertheless, the comparison (Figure 5) of the polarization curves of
the same air electrodes show that the polarization at the high current density
range of the most active electrode is much higher that that of the electrodes
124
with lower activity. This fact can be connected with the higher transport
hindrances in the catalyst containing high amount of silver.
Figure 5. Polarization curves of air electrodes with catalysts from active carbon and with
active carbon promoted with 5% and 30% of silver.
Generally, the hindrances in the transport processes, which take

place in the porous structure of the air gas-diffusion electrodes, influence
strongly the electrode performance and are of prime importance for their
practical application.
An experimental method is proposed for estimation of the transport
hindrances in air gas-diffusion electrodes. As a measure of the transport
hindrances in the air gas-diffusion electrodes is introduced the difference 'E
between the potentials of one and the same electrode when operating with air
and with pure oxygen at one and the same current density. The difference 'E
can be theoretically described by the equation:
RT Poxygen RT foxygen (I)
ǻE E O (I) E air (I) ln ln /4/
2 Į zF Pair Į zF fair (I)
which comprises two terms.
The first term /5/ takes into account the partial pressures of oxygen only
RT Poxygen
ǻE 0 ln /5/
Į zF Pair
125
and is independent of the current density. The second term depends on the
current density through the efficiency factors of the electrode when operating
with air fair(I) and with pure oxygen fox(I) [9]. The efficiency factors being
defined as the ratio of the currents generated at one and the same potential by
the real electrode and by the hypothetical electrode with the same structure
and catalyst in which there are no transport hindrances.
Figure 6. Polarization curve of an air electrode operating with air and with pure oxygen.
With the increase of the current density, the second term in the
equation /4/ is increasing, because the efficiency factor of the electrode
operating with air fair(I) decreases more rapidly with the increase of the
current density than the efficiency factor of the same electrode operating
with pure oxygen fox(I). According to equation /4/ the value of 'E will
increase with the increase of the current density.
Figure 6 presents the polarization curves of one and the same
electrode with active carbon as catalyst when operating with air and with
pure oxygen.
The value of 'E is obtained by the extraction of the potential of the
two curves at one and the same current density, so it is free from the IR drop
between the air electrode and the reference electrode.
The obtained values of 'E are presented in Fig. 7 as a function of
the current density. It is seen that the values of 'E increase with the current.
At low current densities, where the transport hindrances in the electrode are
negligible low, 'E is practically independent of the current density. This is
126
visible in Fig. 8, where are presented data from Fig. 7 for low current
densities (up to 10 mA/cm2).
Figure 7. 'E – I curve obtained from Figure 6.
Figure 8. 'E – I curve at low current densities obtained from Figure 6.

127
Experimental 'E – I curves can be used for comparison of the air

electrodes with respect to the transport hindrances. In order to illustrate this
possibility in Fig. 9 are presented the 'E – I curves for air electrodes with
identical catalytic layers and gas layers differing in their thickness only.
Apparently the hindrances in the transport of molecular oxygen will be
higher in the electrodes with thicker gas layers.
Figure 9. 'E – I curves of air electrodes with identical catalytic layer and
gas layer with different thickness.
From Figure 9 it is seen that the value of 'E increases more rapidly
with the current density for the electrodes with thicker gas layer which is due
to the higher transport hindrances.
The estimation of the transport hindrances by the use of 'E can be
performed in two ways: by comparison of the shape of the 'E – I curves, or
by comparison of the values of 'E at a constant current density.
In Figure 10, we presented the
value of 'E at current density 200
mA/cm2 as a function of the
thickness of the gas layer of the
electrodes from Figure 10.
It is seen that the transport
hindrances remain practically
constant at thickness of the gas
layer up to 1 mm. The further
increase of the gas layer thickness
results in significant increase of
Figure 10. 'E at c.d. 200 mA/cm2 the transport hindrances.
as a function of thickness.
128
The described method for the diagnostic of the activity and the
transport hindrances in air gas-diffusion electrodes is very useful in the
research of porous catalysts for air electrodes. The comparison of the
activity and the transport hindrances of air electrodes with catalysts from
various types of active carbon allow a proper selection to be accomplished.
The diagnostic method can also be successfully used in the
assessment of promoted catalysts.
As it has been previously shown (Figure 4), the activity of catalysts
from active carbon promoted with Ag increases with the increase of the
content of the promoting silver in the catalyst. But the transport hindrances
are strongly influenced by the content of silver in the catalyst.
Figure 11. 'E – I curves of air electrodes with catalysts from active carbon and active carbon
promoted with different amounts of silver (from Figure 5).
In Figure 11, we have presented the 'E – I curves of air electrodes

with catalysts from active carbon and active carbon promoted with different
amounts of silver (from Figure 5). It is seen that the transport hindrances in
the electrodes with catalysts from pure active carbon and with active carbon
promoted with 5% of silver are near to each other. The transport hindrances
in the air electrodes with catalyst containing 30% of silver are much higher.
That’s why catalysts containing large amount of silver are suitable to be used
in air electrodes operating at comparatively low current densities.
129
Using the method for diagnostic, optimal amounts of Ni in catalysts

from active carbon promoted with both Ag and Ni were found. Catalysts are
investigated from active carbon containing one and the same amount of Ag
(5%) and different amounts of Ni. It is found that the activity of these
catalysts increases with the increase of the Ni content up to 7.5%. The
further increase of the Ni-content from 7.5 to 25% does not influence the
activity of the catalyst. The investigation of the 'E values of air electrodes
with these catalysts shows that the transport hindrances remain practically
constant at Ni content up to 10%. At higher Ni contents the transport
hindrances increase significantly. It can be concluded that the optimal
content of promoting metals in the catalysts is ca. 5%Ag + 7%Ni.
The study of catalysts produced by heat treatment in argon of active
carbon impregnated with the compound C-tetramethoxyphenylporphyrine
(CoTMPP) is of great interest. The polarization curves of air electrodes with
Figure 12. Polarization curves of air electrodes with catalysts from active carbon
promoted with 5% Pt and with pyrolyzed CoTMPP catalyst.
this catalyst are comparable to that of air electrodes with catalyst containing
5% Pt (Fig. 12). Something more on electrodes with catalyst containing
pyrolyzed CoTMPP. These show stable performance: they can operate
continuously for more than 8000 hours at current density 100 mA/cm2.
The activity of air electrodes with pyrolyzed CoTMPP catalysts is
investigated as a function of the heat treatment temperature. In Figure 13 we
have presented the initial curve, and the curves obtained after long-term
operation of the electrode at c. d. 100 mA/cm2. It is seen that a maximum in
the activity is obtained at pyrolysis temperature ca. 800°C, the activity being
practically preserved with the time of operation of the air electrode.
130
Figure 13. Activity of air electrode with pyrolyzed CoTMPP catalysts as a function of the
pyrolyzis temperature.
Figure 14. Transport hindrances in air electrodes with pyrolyzed CoTMPP catalysts
as a function of pyrolysis temperature.
131
The transport hindrances (Figure 14) show a minimum at pyrolysis

temperatures 700-800qC, but the transport hindrances in the air electrode
increase significantly with the time of operation. It can be concluded that the
main cause for the aging of the investigated air electrodes is the increase of
the transport hindrances with the time of operation.
It is theoretically predicted [10] and experimentally proven [11] that
the existence of a network of very tiny hydrophobic pores in the carbon-
catalyst grains play an important role for the effective operation of the
electrodes and is of prime importance for their long exploitation life.
It was found that the H2O2 formed by the electrochemical reduction
of oxygen oxidizes the carbon surface, by which parts of the pores in the
carbon-catalyst loose their hydrophobicity and is gradually filled with
electrolyte. As a result the hindrances in the transport of oxygen through the
carbon catalyst grains increase. This seems to be the main cause for the
aging of the electrodes during long-term operation. One way to impede this
effect is the use of catalyst with high activity not only for the
electrochemical reduction of oxygen but also for the heterogeneous H2O2
decomposition (such as pyrolyzed CoTMPP catalyst).
The investigations of various types of carbon-based catalysts allow
suitable air electrodes to be developed for use in the large variety of metal-
air cells and batteries designed in this laboratory.
3.2. Metal-Air Cells
Various types of metal-air cells are developed using the described air
gas-diffusion electrodes. In the most cases zinc is used as an anode [13].
Porous zinc electrodes with highly developed surface are elaborated
in this laboratory, which show extremely high discharge rate capability in
wide temperature range (down to -40qC). The self–discharge of the zinc is
well suppressed without the use of any unacceptable mercury and a high
amper-hour capacity of these zinc electrodes is achieved.
A series of primary zinc-air cells are developed with capacity
ranging from 100 Ah to 3300 Ah and nominal currents ranging from 2 A to
50 A. Technologies for the production of the air electrodes, porous zinc
electrodes and for the assembling of the cells are elaborated. Some of the
developed zinc-air cells are in production since 1981.
The high capacity, low power zinc-air cells from the types ZV500
and ZV3000 are used as a power supply of navigation buoys in Baltic see for
more than 10 years. Batteries from these cells operate for more than 1 year
continuously. Zinc-air cells from the same types are successfully used in
Bulgarian Himalayan expeditions (Lhotze 1981, Everest 1984 and Anapurna
1986) as power supply of the telecommunication system and electric light at
the base camp. It must be underlined that in these cases the zinc-air cells
132
operate at extreme condition – altitude 5000 m above the sea level (low
oxygen concentration of in the air) and temperatures down to - 30qC.
Active carbon promoted with small amount of silver is used as
catalyst in the air electrodes of these cells. In Figure 15 we presented the
discharge curve of the zinc-air cell ZV3000 at constant current 1 A.
Figure 15. Typical discharge curve of a zinc-air cell ZV3000.
High power primary zinc-air cells are also developed with nominal
current 20A y 40A and capacity in the range 100 Ah y 320 Ah. Typical
application of these powerful zinc-air cells is as easy activated reserve power
supply. These cells also can be successfully used for traction.
In order to achieve comparatively high nominal current of the cells,
pyroyzed CoTMPP catalyst is used in the air electrodes.
Figure 16. Polarization curves of a zinc-air cell ZV320 at temperatures 20qC and -20qC
Zinc-air cells from the types ZV280, ZV300, ZV320 and others are constructed and tested.
133
Figure 17. Voltage and power of a 12 cell battery,

consisting of zinc-air cells ZV300 as a function of the current.
In Fig. 17 we have presented the current-voltage and the power

characteristics of 12-cells battery from the cells type ZV300. It must be
reminded that, when the zinc anode in the primary zinc-air cells is
discharged, the cell is out of operation. By this, the air electrodes, which in
principle can operate for a long time, are lost. In order to avoid this
inconvenience, mechanically rechargeable zinc-air cells were developed
[14]. The essential feature of the mechanically rechargeable zinc-air cell is
that the air electrodes are used many times.
Once the mechanically rechargeable zinc-air cell is discharged, the
spend (oxidized) zinc electrode is removed from the cell together with the
electrolyte and new zinc electrode and fresh electrolyte are introduced by
which the cell is operational again.
The removed from the cell discharged anode, containing the
products of the discharge of zinc is put in a separate cell, where it is
recharged electrochemically against an inert electrode. During this charging
process zinc is produced again from the discharge products. The experiments
have shown that such a procedure of recharging of the zinc anodes can be
repeated several times without significant losses of the amperhour capacity
of the zinc electrode.
Several types of mechanically rechargeable zinc-air cells are
developed in this laboratory, operating at nominal current 20A y 40 A with
capacity ca.300 Ah.
Batteries for such cells are tested in experimental electric cars and
scooters.
134
Another type of zinc-air system is designed and investigated – the

so-called zinc-air fuel cell. It consists of a cell with two air electrodes and an
inert anode. Slurry from zinc power dispersed into the electrolyte is
circulating through the cell, by which the zinc particles are discharged on an
inert electrode.
In Table 1 we have summarized data for some Zn-air cells
developed in CLEPS.
Table 1. Data for some zinc-air cells developed in CLEPS.
Capacity Current Voltage Power Specific

Type/purpose Ah A V W energy
Wh/kg
ZV-1 (Electric truck) 160 40 1,0 40 160

ZV-2 (Electric car) 250 80 1,0 80 114
Zn-Air Fuel cell - 150 1,0 150 40
Zn-O2 280 20 1,18 23 200
ZV320 320 40 1,17 46 180
ZV500 500 2 1,2 2,4 250
ZV3000 3300 2 1,2 2,4 270
ZVMR1 (Electric car)
Mechanically 500 30 1,1 33 180
rechargeable
ZVMRII (Electric car)
Mechanically 300 20 1,1 22 150
rechargeable
Air gas-diffusion electrodes with similar overall structure suitable

for operation in neutral electrolytes are developed. On the basis of these air
electrodes metal-air cells are developed using aluminum and magnesium as
an anode.
An aluminum-air battery has been designed (2500 Ah, 2A, 12V)
with sea water as an electrolyte (in cooperation with the University of
Belgrade). The battery [15] is designed for use in a small yacht. It is known
that during the discharge of aluminum a gelatinous aluminum hydroxide is
obtained. It could form an insulator layer on the air electrode surface by
which its further operation will be troubled. In order to avoid this effect, the
electrolyte in the aluminum-air battery is changed every 8 hour. After the use
of the battery during the night, the electrolyte (sea water) is pumped out of
the battery and a fresh sea water is pumped in the battery when power supply
is needed.
Several types of magnesium-air cells [16] with neutral electrolyte
(NaCl solution) are also developed and investigated.
135
4. CONCLUSIONS
Air gas-diffusion electrodes were developed, suitable for use in
metal-air cell with alkaline or with saline electrolytes. A variety of carbon-
based catalysts are used in these air electrodes. Methods for the estimation of
the activity and the transport hindrances are proposed and used successfully
for the optimization of the carbon-based catalysts.
A variety of primary zinc-air cell and batteries are designed with
capacity ranging from 100 Ah to 3300 Ah, operating at nominal currents
from 2 to 40A at temperatures in the range +40 y -40qC. Mechanically
rechargeable zinc-air cell are also developed and tested in experimental
electric cars and scooters.
Al-air and Mg-air cells with saline electrolytes have also been
designed.
REFERENCES
1. Gregory D.P., Metal-air batteries, Mills & Boon Limited, 1972.
2. Iliev I., Air electrodes for primary metal-air batteries, Prroceedings of the 160th Meeting
of the Electrochemical Society, Oct. 1981, Denver, Colorado, p. 268-269.
3. US Patent No. 4031 033, 1972.
4. Kaisheva A., Doctor Thesis, 1984.
5. Kaisheva A., Iliev I. Gas-transport in nano-porous hydrophobic media, Procedings of
the Int. Symposium Micro/nanoscale energy conversion and transport, 2002, April 14 –
19, Antalya, Turkey, p. 107.
6. Iliev I., Mrha J., Gamburzev S., Kaisheva A., On the effect of various active carbon
catalysts on the behaviour of carbon gas-diffusion air electrodes, Journal Power Sources
1976/77; 1: 35-46.
7. Iliev I., Gamburzev S, Kaisheva A., Mrha J. Influence of mass transport on the
performance of carbon gas-diffusion electrodes in alkaline solutions, J.Appl.
Electrochem., 1975; 5 :291-297.
8. Iliev I., Gamburzev S., Kaisheva A., Gas-diffusion electrodes with transition metal
macrocyclic catalysts for electrochemical reduction of oxygen, Proceedings of the 31
ISE Meeting; 1980 Sept. 22 – 26, Venice, Italy, Vol. I, p. 286-288.
9. Kaisheva A., Iliev I., Gamburzev S., Comparative methods fo estimation of the activity
and the transport hindrances of air gas-diffusion electrodes, J.Power Sources, 1984;
13:181 – 195.
10. Iliev I., Gamburzev S., Kaisheva A., Optimization of the pyrolysis temperature of active
carbon-CoTMPP catalyst for air electrodes in alkaline media, J.Power Sources, 1986;
17: 345 352.
11. Iliev I., Structure and transport processes in carbon air electrodes,Proceedings of the
Fall Meeting of Electrochemical Society, 1983 Oct. 9-14, Washington, Vol. 83-2, p. 618-
19.
12. Gamburzev S., Iliev I., Kaisheva A., Steinberg G., Mokrousov L., Behaviour of carbon
wetproofed electrodes during long-term operation in alkaline electrolyte, Elektrokhilija,
1980; 16:1069-72.
13. Budevski E., Iliev I., Varbev R., Zinc-Luft Elemente mit hohem Energievorrat –
Probleme und Eisatzmoeglichkaiten, Z. Phys. Chem. Leipzig, 1980; 261:716-21.
136
14. Iliev I., Kaisheva A., Stoynov Z., Pauling H.J., Mechanically rechargeable zinc-air cells,
Proceedings of the 3rd Int. Battery Recycling Congress, 1997 July 2 – 4, Noordweejk
Aan Zee, the Netherlands.
15. Budevski E., Iliev I., Kaisheva A., Despic A., Krsmanovic K., Investigation of a large
capacity medium power saline aluminium-air battery, J. Appl. Electrochem., 1989;
19:323-330.
16. Kaisheva A., Iliev I., Milusheva J., Mechanically rechargeable magnesium-air cells with
non-aggressive electrolyte, Proceedings of the Int. Battery Recycling Congress, 2002
July 3 – 5, Vienna, Austria.
METAL – AIR BATTERIES
WITH CARBONACEOUS AIR ELECTRODES
AND NONMETALLIC CATALYSTS
Nikolay Korovin*
Moscow Power Engineering Institute (Technical University (9)

14 Krasnokasarmennaya, Moscow, 11250, Russian Federation
Abstract
Problems of the metal-air batteries and air electrode are considered.

Results of research, design work and development of the air – Zn primary
and air-Al mechanically rechargeable batteries, as well as electrically
rechargeable electrodes are presented.
Keywords
Batteries; air electrodes; catalysts; carbon.
1. INTRODUCTION
In the batteries with air electrodes, the oxidizer is oxygen from the
atmosphere. This fact assures many advantages of these batteries, namely:
high energy density, long shelf life, environmental compatibility, relatively
low cost and flat discharge curve. Theoretical parameters of batteries with air
electrodes (please see, for reference, Table 1) are much higher than
theoretical parameters of most traditional, newly developed and emerging
battery systems.
Obviously, the practical performance of batteries with air electrodes
naturally is lower than the theoretical values indicated by Table 1, and is a
function of parameters of the reagents making up these batteries.
Sucessful application of the air electrode requires solving some key
problems: the air electrode catalyst, the alkaline electrolyte carbonization,
the oxygen reaction with anode materials, an influence of an air humidity on
an electrode behavior.
*
137
138
Table 1. Theoretical parameters of various battery chemistries based on air electrodes

(Po2=21kPa ) at 298K.
Specific Specific Energy

Anode EMF, V capacity, energy, density,
Ah / g Wh/g kWh/l
Al 2.55 2.98 7.6 22.11

Mg 3.16 2.20 6.97 12.12
Zn 1.64 0.82 1.34 9.6
Fe 1.26 0.96 1.21 9.52
LaNi5H6 1.2 0.37 0.44 2.64
Of those types mentioned in Table 1, main attention has been paid to

the development of batteries with Zn, Mg and Al anodes [1-7]. Recently
battery air-metal hydride was proposed and investigated [3, 8].
It is possible to divide the batteries with air electrodes into three
following groups: primary, mechanically and electrically rechargeable. In
these batteries alkaline, saline and seawater electrolytes are used (Table 2).
Parameters and behavior of these batteries depend on activity and
stability of air electrodes, thus it is necessary to consider processes at the air
electrode, problems of air electrode and ways of their solution, as a complex
and multifunctional task.
Table 2. Types of batteries with air (oxygen) electrodes and applicable electrolytes.
Primary Mechanically Electrically
Anode Batteries Rechargeable rechargeable
Batteries batteries
Zinc Alkaline Alkaline Alkaline
Alkaline
Aluminum Seawater Saline -
Magnesium Seawater Saline -

Iron - - Alkaline
Metal hydride - - Alkaline
2. AIR ELECTRODES
2.1. Processes
The active component of an air electrode is oxygen. The reaction of
oxygen reduction (4-electron reaction) can be written by equation (1)
139
O2 + 2H2O + 4e- = 4OH-; E01=0.401 V (1)
It is possible to attain the theoretical potentials of the reaction (1)

only at the special conditions of the electrodes, and provided the deep
purification of electrolytes. Stationary electrode potentials are usually lower
than the theoretical potentials by 100-500 mV. The reason for that effect is
occurance of parallel and side reactions in the cell.
Many scientists deeply investigated the processes of oxygen
reduction in metal-air cells [9 - 11]. The process in question is very
complicated. Products of the oxygen reduction usually contain hydrogen
peroxide. Therefore, the oxygen reduction can proceed by a two-electron
scheme:
O2 + 2H2O + 2e- = HɈ2- + ɈH-; E02 =-0.076 V (2)
Hydrogen peroxide then can be reduced or decomposed in water and

oxygen. Thus, oxygen reduction can proceed by two parallel ways: (a) the
four-electron reaction of conversion into water and (b) the two-electron
reaction of transformation into hydrogen peroxide. The four-electron
reaction can proceed, if oxygen adsorbs with a rupture of chemical bonds in
O2 :
O2 o 2Ɉɚds. (3)
Adsorbed oxygen (Ɉɚds) from process (3) then reduces, e.g. by reaction (4):
Ɉɚds + ɇ2Ɉ +2ɟ- = 2Ɉɇ-. (4)
Energy of the oxygen adsorbtion on the pyrographite, active carbon

and some other electrodes is not enough for the rupture the bonds in oxygen.
In this case the reaction proceeds by the electron addition to the adsorbed
oxygen molecule, which usually limits all process of oxygen reduction.
2.2. Catalysts
The oxygen electrode catalysts accelerate the adsorption step, or the
step of the adsorbed oxygen reduction or the hydrogen peroxide
decomposition. A catalytic activity is shown by: platinum metals, silver,
oxides (MnO2, NiO +Li2O, ɋɨ2Ɉ3), spinels (NiCo2O4, MnCo2O4, CoAl2O4),
perovskites (LaCoO3, La1-xSrxCoO3, La1-xSrxMnO3, La1-xCaxCoO3), activated
carbon, polyaniline, macrocycles, their polymers and pyropolymers [9-14].
Properties of some of them are shown in the Table 3.
140
Table 3. Catalysts for the oxygen reduction in the alkaline solution.
Catalysts Activity Stability Cost
Pt,Rh,Pd Good Very good Very high

Silver Very good Good High
MnO2 Good Middle Acceptable
Spinels (AB2O4) Middle Middle Acceptable
Perovskite (ABO3) Middle Middle Acceptable
Activated carbon Low Middle Low
Polymer Good Low Acceptable
macrocycles
Pyropolymers Good Good Acceptable
The exchange current density of Pt-metals is relatively small, but

they have high stability. Very high cost does not permit to use them in the
batteries of wide application. Noticeably higher activity, very good stability
and lower costs are demonstrated by silver. The most inexpensive catalyst is
activated carbon that has very high surface area. This type of catalyst is used
in some batteries. Activity of carbon electrode can be improved by additive
of oxide (e.g. MnO2) or pyropolymers.
2.3. Porous Electrodes

Oxygen reduction can be accelerated by an application of electrodes
with high surface area, e.g. the porous electrodes [9, 13]. The porous
electrodes usually consist of catalysts, hydrophobic agent
(polytetrafluoroethylene-PTFE) and conductive additive. Electrode kinetics
on the porous electrodes is complicated by the mass and charge transfer in
the pores and is called “the macrokinetics of electrode processes”.
An analysis of the equations of macrokinetics of oxygen reduction
on the porous hydrophobic electrodes gave some conclusions [15, 16]:
1. The dependences of current density on weight percent of PTFE
and on porosity go through a maximum;
2. Current density increases with decrease of the catalysts and PTFE
particles size, and with the temperature increase;
3. The dependence of current density on potential is very
complicated;
4. Effective Tafel coefficient can changed from b to 4b.
5. The characteristic length of processes L is determined by the
relationship between the constants of transfer and reaction velocity.
Partial pressure of oxygen is 5 times lower than that of the air
pressure. Inert nitrogen changes the mechanism transfer of oxygen in the
gaseous phase [17]. In pure oxygen, the Stephanov flux takes place, because
141
pressure in the reaction zone is lower, than pressure of air. Therefore, in

gaseous phase the oxygen diffusion does not limit process. Inert component
(nitrogen) suppresses Stephanov flux, because difference between pressure
in the air and in the reaction zone is very small. Concentration of oxygen in
the reaction zone is lower than that in air. Therefore, it is necessary to have
special thin diffusion layer for the oxygen of air transfer.
3. PRIMARY AIR-ZINC BATTERIES
3.1. Processes and Problems

The overall reaction in zinc-air battery with alkaline electrolyte at
the beginning of discharge may by written as follows:
2Zn + O2 + 2H2O + 2OH- = 2Zn[(OH)4]2-. (5)
Zincate-ions can dissociate and after the solubility exceeds the equilibrium
level precipitation of zinc oxide proceeds:
Zn[(OH)4]2- = ZnO + H2O + 2OH-. (6)
Thus, after partial discharge, the reaction (5) can be interchanged with
reaction (7):
2Zn + O2 = 2ZnO. (7)
The cathode can of an air –zinc cell has a hole for the air access (the
respiratory hole). The substitution of the cathode oxides, which are used in
traditional batteries, with air increases capacity and energy density of such a
battery. The battery voltage depends first of all on the air potential, which at
constant current density and temperature is also constant; hence the voltage
of an air-zinc battery at the discharge practically does not change.
There are several recognized problems in air-zinc batteries. Among
these are: relatively high cathode polarization, the anode passivation, the
zinc corrosion, the electrolyte aging, the carbonization of electrolyte and the
air humidity influence on battery performance parameters. The cathode
polarization is decreased by the porous electrode with catalyst (activated
carbon, pyropolymers and other) use. The anodic passivation arises owing to
the zinc oxide formation at high current density. For its prevention it is
necessary to decrease current density, for example by the zinc powder
electrode application. Some additives, e.g. Li+ or SiO32- slow down the aging
of electrolyte. The zinc corrosion proceeds as with the hydrogen evolution.
142
Zn + 2H2O = Zn2+ + H2 + 2OH-, (8)
and as with the oxygen reaction: (5) and (7). For purposes of decreasing of
the corrosion rate, zinc of high purity is applied, and also the additives in
zinc (e.g. Pb) and in electrolyte (e.g. amine) are used.
If battery is open to the air, carbonization of alkali takes place in
accordance with process (9). For purposes of prevention of electrolyte
carbonization, the respiratory holes should be sealed until after the beginning
of battery use.
ɋɈ2 + 2Ɉɇ- l ɋɈ32- + ɇ2Ɉ. (9)
The air humidity influences the performance parameters of the air-

zinc batteries. At high air humidity the electrode flooding is possible. On the
other hand, at low air humidity a concentration of the electrolyte increases,
the air electrodes get too dry.
3.2. Portable Batteries

These types of batteries are available in button and prismatic forms.
Their main application is as power sources for hearing aids. Other
applications include various specialty uses in the notebook computers,
electronic pagers, portable battery chargers, various medical devices, the
wireless crew communicator systems [18, 19].
Figure 1. Discharge curve of a primary button air – zinc cell

PR-44 at the resistance 625: and at 293 K.
143
Button cells consist of cathode and anode cans (used as the

terminals), powdered zinc anode, containing gelled electrolyte and the
corrosion inhibitor, separator with electrolyte, thin (0.5 mm) carbon cathode
with catalyst and PTFE, waterproof gas-permeable (teflon) layer and air
distribution layer for the even air assess over the cathode surface.
Parameters of battery depend on the air–transfer rate, which is determined by
quantity and diameters of air access holes or porosity of the gas-diffusion
membrane. Air-zinc batteries at low rate (J=0,002-0,01C at the idle drain and
J= 0,02-0,04C at the peak continuous current) have flat discharge curves
(typical curve is shown by Figure 1).
With the air-transfer acceleration, current density increases (or the
cathode polarization decreases), but at very high air velocity, electrode can
be flooded (at high air humidity of more than 60% at 293K), or get too dry
and fail (at low air humidity). Some batteries have the devices for the air-
transfer control. Capacity of batteries decreases with temperature decrease
(e.g. by a factor of two with corresponding temperature decrease from 294 to
273K), as well as with current density increase. Now the button batteries
with capacity from 50 to 6300 mAh are being produced. Some of them were
developed at MPEI(TU) [2, 20] and are manufactured at the battery plant
‘Energiya”, c. Eletz, Russia. Typical useful service-life of these batteries are
2-3 months, specific energy –200-400 Wh /kg and 400-800 Wh /L. Air-zinc
batteries are the best power sources for hearing aids. Constant voltage of air-
zinc batteries provides also best compatibility of these batteries with hearing
aids from the medical point of view.
4. AIR–ALUMINIUM MECHANICALLY RECHARGE-

ABLE BATTERIES WITH SALINE ELECTROLYTE
Air-Al batteries with saline electrolyte (NaCl solution) are simple in
maintenance and low cost [4, 7, 19, 20].
4.1. Processes and Problems

Cathodic reaction in this battery chemistry is described by
equation (1). Anodic reaction may be written in a form of reaction (10):
Al - 3e- oAl3+. (10)
In saline solutions, Al is also exposed to corrosion, which can be

decreased (down to 10-6 mm/h) by additions of small quantities of Ga, Sn
and Pb to alumimum. Real processes in batteries are very complicated. The
Al3+ ion is exposed to hydrolysis with decrease of pH (reaction (11)).
144
Al3+ + H2O l AlOH2+ + H+. (11)
Chloride-ions participate in the anodic processes. Activation of Al

by Cl- and pH decrease accelerate anodic dissolution of Al. At the cathodic
department concentration of OH- increases, and as a result the complex ions,
e.g. [Al(OH)4]-, are formed. In solution more complicated ions can be
formed, for example: [Al(OH)xCly]. With the chloride concentration
increase, share of hydroxide ions in the complex ions decreases and pH
solution grows, as it was shown in our experiments [7, 20]. The dependence
of voltage of air-Al cell on NaCl concentration has a maximum at
concentration of 15wt%. That type of dependence is connected mainly with
influence of NaCl concentration on the anodic polarization. For our batteries
we choose 15wt% solution of NaCl. Reaction products are most of all are in
the colloidal condition. Experiments show that in non-stirred solution it is
possible to receive an energy density of up to 100Ah/liter. Intensive stirring
increases said performance parameter.
4.2. Battery Performance Parameters
In Moscow Power Engineering Institute (TU) portable air–aluminum

batteries with saline electrolyte were developed [7, 18, and 20]. In our
devices, the air electrodes consist of two layers. Diffusion layer contains
PTFE, carbon black and metal screen; active layer consists of activated
carbon and PTFE. At 293 K and the range of current density 2-25 mA/ cm2
dependence of cathode potential E (in H-scale) upon current density J
(Figure 2) may by written by the Tafel equation (12).
Figure 2. Volt - amperic curves of an air-Al cell with NaCl solution at 20°C.
145
E =0.03- 0.18lg J. (12)
Dependence of cell voltage U on current density (Figure 2) also is

expressed by equation: U = 1.5-0.25 lgJ. The discharge voltage at a current
density of 20 mA/cm2 increases in the beginning and after that changes only
a little (Figure 3).
Figure 3. Discharge curves of air-Al cell with NaCl solution at 20 mA/cm2 and 20oC.
Coulombic efficiency of Al is 80-90 %. Batteries with capacities in

the range of 50-1000 Ah (at repeated exchanges electrolyte after 6-8 hours of
work and anodes after 50-100 hours of work) have been developed. Specific
energy is 80-150 Wh/ kg without anode exchange and 200-400 Wh/kg with
anode exchange. Batteries, having thermal insulation, can work in the low
surrounding temperatures (down to 236K). Specific costs of these batteries is
in the range of 100-120 $/kWh.
4.3. Application
The main fields of application for this battery chemistry include:
portable TVs, radio receivers, lamps, flashlights, electric shavers, barrier
lightning, instruments, batteries for portable rechargers, emergency power
supplies, small refrigerators, power sources for tourists, hunters, geologists,
shepherds and so on. Parameters of some batteries, assembled from the
modules, are summarized in Table 4.
146
Table 4. Parameters of some air-Al batteries with saline electrolyte.

Voltage, V Dimensions, Weight without Application
Uocv, U(disch) ɫm electrolyte, kg
3.7 (2.4) 14ɯ8ɯ13 0,5(with insulation) Lamp, 1.9W

6.1 (4.0) 10ɯ10ɯ10 0,4 Flashlight, 4W
9.1 (7.0) 16ɯ12ɯ10 0.9 Lamp, 7W
12.2 (8.0) 20ɯ10ɯ10 0.8 Tape-recorder, 12 W
18.3 (12.0) 30ɯ10ɯ10 1.2 Portable TV, 12W
5. ELECTRICALLY RECHARGEABLE
AIR ELECTRODE
The most attractive batteries with air electrodes are the so-called
electrically rechargeable batteries, because they have very high specific
energy, simple maintenance, environmental compatibility and relatively low
cost. But there are many problems of rechargeable air electrodes. Aluminum
and magnesium can not recharge in water solution. Therefore zinc, iron and
metal hydride can be used as rechargeable anode. Main attention is paid to
the air-zinc batteries. Very complicated problem is development of the
rechargeable air electrode.
Reversible reaction on air electrode may be written by equation (13):
O2 + 2ɇ2Ɉ + 4e- l 4OH-; E0 = 0.401V (13)
Limiting steps of the oxygen reduction and the oxygen evolution are
different. The catalysts of direct reaction are usually not the best catalysts of
reverse reaction. Process of the oxygen evolution is accelerated by nickel
and some oxides (spinels, RuO2-TiO2, IrO2, RhO3). That process proceeds at
high anodic potentials, at which many catalysts and carriers are not stable.
The graphitized cloth, the carbon black, treated at high temperature and
nickel can be used as the catalyst carriers and current collectors.
Three approaches of solution of the reaction (13) catalyst problem
are used: application of bifunctional catalysts, two catalysts in one layer and
two catalysts in two layers.
The bifunctional catalysts, which were proposed are: silver, rhodium
and iridium oxides [3], NaxPt3O4 [21], pyrochlores A2B2O7 (A=Pb, Bi;
B=Ru, Ir) [22], perovskites –ȺȼɈ3 (Ⱥ=La, Ca; B=Co, Ni, MnɈ2) [3],
pyropolymers [24, 25], NiOxCoOx.
Comparative activity parameters of catalysts and electrodes based on
these catalysts are summarized in the Table 5.
147
Table 5. Differences in the potentials of oxygen evolution and reduction from air
at the current density of 100 mA/cm2 at 298K for some catalysts [3, 21-25].
Catalyst 'ȿ, mV Catalyst 'ȿ, mV
Pb2Ru1.42Pb0.58O7 392 LaNiO3 620
NaxPt3O4 532 La0.6Ca0.4CoO3 637
Pb2Ir2-xPbxO7 557 Electrode with Ni-PP* 850
RhO3 587 Electrode with PP* 950
NiO-CoO 1200 Electrode with Ag 910

Note for the Table: PP*-pyropolymer.
Figure 4. Polarization curves of the rechargeable air electrode at 293 K.

Electrode with (1) and without (2) pyropolymer catalyst.
As it follows from Table 5, many catalysts contain metallic

platinum. We have developed bi-layer porous hydrophobic air electrodes,
which do not contain platinum metals, are active and can be cycled [24, 25]
(Figures 4-6). These bifunctional catalysts are pyrolized Co – macrocyclic
compounds. Said catalyst has high catalytic activity for the oxygen reduction
and also features acceptable stability, however its activity for the oxygen
evolution is not high enough.
Our recent our works show that even higher activity and stability can
be demonstrated by the three-layer electrodes with nickel layer, active in the
oxygen evolution, middle layer with catalyst, active in the oxygen reduction
(MnO2, pyropolymer or a perovskite), and a diffusion (waterproof) layer,
148
needed for the air transfer. The potential difference of oxygen evolution and
reduction on such electrodes was 800-900mV (Table 5). Haas with co-
workers developed rechargeable electrode with a bifunctional catalyst
La0.6Ca0.4CoO3 [23]. The potential differences of oxygen evolution and
reduction at current densities of 100 mA/cm2 were 900-955 mV. Electrodes
are believed to be stable and permit to perform more than 500 charge-
discharge cycles.
Figure 5. Charging and discharging curves of an air electrode

with (1) and without (2) pyroplymer catalyst at 293 K.
The current density (mA/cm2) is 50 (charging) and 100 (discharging), respectively.
Figure 6. Change of potentials of the electrode with pyropolymer upon cycling at the
discharge (1) and at the discharge (2) at 293K, with the current density of 50mA/cm2.
6. CONCLUSIONS
Air-Zn primary and air–Al mechanically rechargeable batteries with

carbonaceous air electrodes were designed and tested. Air–Zn batteries are
mostly used in the hearing aids, while air-Al batteries find wider range of
applications in the portable devices. Electrically rechargeable air electrodes
with pyropolymer catalyst have been developed.
149
REFERENCES
1. Bagotzky V.S., Skundin A.M., Batteries. Moscow: Energoatomizdat. 1981, (Russian).
2. Korovin N.V. “Advanced half-gas systems for rechargeable batteries”. In New Promising
Electrochemical Systems for Rechargeable Batteries (NPESRB), V.Z. Barsukov, F.
Beck, ed. Dordrecht: Kluver Academic Publisher. 1996; 171-180.
3. Jorissen L., Garche J. “Application of bifunctional air electrodes”, In (NPESRB), V.Z.
Barsukov, F. Beck, ed. Dordrecht: Kluver Academic Publisher. 1996; 159-169.
4. Korovin N.V., New Batteries. Moscow: Energiya.1978. (Russian).
5. Blurton K.F., Sammells A.F. Metal/air batteries: their status and potential – a review. J.
Power Sources. 1979; 4: 263.
6. Budevsky E., Iliev I., Varbev R., Zinc – Luft Elemente mit hohen Energievorrat-
Probleme und Eisatzmoeglichkaiten, Z. Phys. Chem. Leipzig, 1980; 261: 716-21.
7. Korovin N.V., Kleimenov B.V., Dyachkov E.V., Voligova I.V. Aluminium-air batteries
for portable devices.Ext Abstr.1997 Joint. Intern. Meeting ES and ISE. Paris.1997.P. 88.
8. Gamburtzev S., Velev O.A., Danin R., Srinivasan S., Appleby A.J. “Performance of an
improved design of metal hydride/air rechargeable cell”. In Batteries for portable
application and electric vehicles. C.Holmes, A.Landgrebe ed. Pennington: Electroch.
Soc, 1997, 726-33.
9. Korovin N.V. Electrochemical Power Sources. Moscow: Energoatomizdat.
1991.(Russian).
10. Wiesener K., Garche J., Schneider W. Electrochemische Stromquellen. Berlin: Akademie
Verlag. 1981.
11. ɇɨɚr J.P. The electrochemistry of oxygen. N.Y: Interscience Publ. 1968.
12. Ʉɨrɨvin N.V., Ʉɚsɚtɤin E.V. Electrocatalysts of electrochemical devices. Russian
Electrochemistry. 1993; 39: 448.
13. Iliev I., Gamburzev S., Kaisheva A., Mrha J. Influence of mass transport on the
peformance of carbon gas-diffusion electrode in alkaline solutions. J. Appl.
Electrochemistry. 1975; 5: 291-97.
14. Barsukov V., Khomenko V., Chivikov Santonenko P. Physico-chemical base of
development of the air-meral batteries with the catalysts on anyline base.
Electrochemical Power Sources. (Russian). 2001; 1: 24 – 30.
15. Chismadjev Yu.A., Markin V.S., Tarasevich M.R., Chirkov Yu.G. Macrokinetics of
processes in porous electrodes. Moscow: Nauka, 1971 (Russian).
16. Korovin N.V., Luzhin V.K. Knudsen diffusion in oxygen and air electrodes. Russian
Electrochemistry. 1979; 15: 360 - 62.
17. Korovin N.V., Kicheev A.G. Luzhin V.K. Influence of the inert components on mass-
transfer in the gas phase of porous electrode. Russian Electrochemistry. 1972;.8: 146 –49.
18. Korovin N., Kleimenov B., Agaphonov N., Guryanov M. Air-metal batteries Ext. Abstr.
46th ISE Meeting; 1995; Xiamen, China. 1995. 1-5-12.
19. Hamlen R.P. “Metal/air batteries.” In Handbook of batteries, D. Linden, ed. N.Y:
McGrow-Hill Inc., 1995, 38.1-38.45.
20. Voligova I.V., Korovin N.V., Kleimenov B.V., Dyachkov E.V. “ Air-Al batteries with
saline electrolyte.” In Fundamental problems of electrochemical power sources
(FPEPS), I Kasarinov, ed. Saratov: SGU, 1999, 169.
21. Swette L., Kackley N., McCatty S.A. Oxygen electrodes for rechargeable alkaline fuel
cells, J. Power Sources. 1991; 36: 323 - 39.
22. Horowitz H.S., Longo J.M., Horowitz H.H. Oxygen electrocatalysis on some oxide
pyrochlores. J. Electrochem. Soc. 1983; 130: 1851- 59.
23. Muller S., Striebel K., Haas O. Electrochimica Acta. 1994; 39: 1661.
24. Korovin N.V., Voligova I.V., Kleimenov B.V. Bifunctional catalysts for the air
electrodes. Ext. Abstr. 50th ISE Meeting.; 1999; Pavia: 1999, 1.19.
25. Korovin N.V. Batteries with air electrodes. Electrochemical power sources. 2001; 1:
16-23. (Russian).
CHAPTER 3:
CARBON ANODES FOR

LITHIUM-ION BATTERIES
Currently, lithium-ion and lithium-ion polymer cells use graphitic

carbons as the active material of the negative electrode. The phenomenal rate
of growth of the lithium-ion battery chemistry has resulted in a steadily
increasing demand for anode grade carbons. Thus, for reference, the actual
worldwide sales volumes of graphitic carbon powders into this market
amounted to approximately 3,000 tons in 2001, propelled up to 6,000 tons in
2003, and well exceeded 11,000 tons in 2004. Due to fierce competition
among carbon manufacturers and also the lithium-ion battery becoming
commodity chemistry, the prices for anode grade carbons declined from the
levels of $50/kg to approximate levels of $20/kg within the same time frame.
Still, one may estimate that as of the end of 2004, the market of anode
materials for lithium-ion batteries amounted to a minimum of USD 220
million per year. A collection of 11 papers in this chapter seeks to address
latest issues associated with development, synthesis, characterization and use
of new advanced carbonaceous materials for the rapidly growing segment of
electrochemical energy storage market in question.
This chapter begins with an overview paper by T. Takamura and R.
Brodd. The authors broadly discuss various uses of carbon materials in the
main types of power sources. Besides heavy focus on carbon materials for
lithium-ion batteries, authors also review a recent work by U.S.
manufacturers aimed at incorporation of expanded graphite into the cathodes
of alkaline primary batteries. The role of irreversible capacity loss in the
design of lithium-ion grade carbons was also discussed. Placing surface
coatings and control of morphology of the graphite assists in minimizing the
irreversible capacity loss. In addition, the editors wish to comment that
spherodizing of lithium-ion grade graphite is a necessary, but not sufficient
step, when designing a carbon with low irreversible capacity loss.
The second paper by H.J. Santner et al. is a derivative of an
overview presentation given during the NATO-CARWC by Professor M.
Winter of Graz Technical University, Austria. This interesting work
summarizes over 10 years of research and development investigations by the
University in the search for the most efficient electrolyte additives and
solvents, which are the working environments for graphite anodes when in
lithium-ion batteries. This fundamental paper could be quite useful, in
particular, to those battery developers, who seek to understand the
phenomena of formation of protective layers on graphite. The authors clearly
differentiate between the types of layers, which are being created on graphite
during the process of formation of lithium-ion batteries.
In the third paper, M. Walkowiak et al. report on findings of Central
laboratory of batteries and Cells (CLAiO) in Poland, as related to the
electrochemical performance of spherodized purified natural graphite and
boron-doped carbons in lithium-ion batteries. While it is noteworthy that
153
154
these authors are still working on debugging of their electrochemical

devices, the fact that full cells are being used to qualify new materials at
CLAiO deserves a complement, as it is less common that full cells are used
for such purposes, with most academia scientists doing individual electrode
studies, often being unaware of the real requirements of the battery industry.
The article by Professor D. Aurbach et al. offers a comprehensive
analysis of the reasons for why the synthetic and natural graphite, and
mesocarbon microbeads fail in PC-containing organic electrolytes (PC is a
highly attractive solvent for lithium-ion battery technology due to its low
cost and inherent ability to work at temperatures below –40 qC). These
authors (from the Bar-Ilan University in Israel) suggest that the failure
mechanism of graphite/carbon electrodes in PC-based solutions is due to
particle cracking, followed by their electrical isolation by surface films,
rather than due to complete graphite exfoliation due to co-intercalation of PC
molecules with Li+ into graphite lattices. Though not widely accepted by the
scientific community, the concept appears to be very innovative. Authors
report to have collected lots of evidence, proving the above theory to be
valid. Among other advanced analytical methods used in their work, readers
are encouraged to review the presented data utilizing atomic force
microscopy (AFM).
One of the obvious values of paper by F. Henry et al. is the fact that
it comes from a U.S. commercial manufacturer of the industrial carbon and
graphite. The reader will get an unusual insight into the evolution in
understanding of what were the properties, which needed to be tailored in
order to create a graphite line suitable for lithium-ion battery application. It
is noteworthy that reversible and irreversible capacities, which are being
usually looked at in the academia as key qualifying properties, carry a
secondary role, when compared to the practical needs of having excellent
adhesion to the copper substrates, increased packing densities and acceptable
abuse tolerance in various lithium-ion battery environments.
A fundamental work by Professor F. Beguin et al. addresses the
mechanisms of reversible and irreversible intercalation and BET adsorption
in lithium-ion grade carbons. The authors discuss, among other phenomena,
an interesting concept that values of the active surface area (ASA) on carbon
correlate with irreversible capacity loss. This is a departure from
traditionally accepted belief that regular BET single point values of carbon
are in proportional relationship with the irreversible capacity loss. It is
noteworthy that editors also have seen the later theory to be highly
inaccurate, and thus, consider the approach taken by authors as valuable.
Readers should be aware that the method of determining the ASA currently
is being actively promoted by Micromeritics. So far, it has received limited
acceptance in the electrochemical labs dealing with fuel cell research, but the
method remains relatively unknown to the battery research community.
155
The seventh paper in this chapter (R. Yazami et al.) has focused on
fundamental aspects of understanding the transition between graphite
intercalation stages #1 and #2. Authors note peculiar “anomalies”,
apparently not described in the literature up until now (even though the
subject of investigation of graphite staging upon its lithiation has been
around for decades). The paper offers a fundamentally new look into the
stage transition phenomena.
In the article by the renowned Accumulator Research and Design
Institute “Istochnik” of St. Petersburg, Russia, basic performance of various
types of carbon materials in the coin and pouch-type lithium-ion cells is
described. On a side note, detailed descriptions of various component
suppliers furnished by the authors, suggest that Russia is slowly but surely
establishing an internal supply chain of anode, cathode, separator, current
collector materials and components for the needs of the lithium-ion
technology, which is being actively developed internally as well. Authors
reported performance of the following carbon materials: pyrolytic carbon; a
version of nanotubes; expanded graphite; “spectrally pure” graphite; special
“RIECP graphite” (derived from natural flake) – all are Russian-made
grades. Authors also tested thermally purified spherodized flake graphite
SLA1020 and surface coated spheroidal flake graphite SLC1115 from
Superior Graphite (USA). According to conclusions by the authors, the best
for Li+ intercalation are: SLA, SLC and “Spectral Pure”carbon materials.
Editors, however, have no doubts, that many inquiries will be made to
“Istochnik” after publishing their comments from testing of Russia’s Astrin
Co. nanotubes. What could not be effectively achieved by other “classic”
nanotube developers thus far, apparently could be achieved with a particular
grade of Russia-made nanotubes. The authors report observing good stability
of these nanomaterials under conditions of prolonged cycling in lithium-ion
cells. The editors wish to comment that the authors did not discuss the nature
of these nanotubes, but it may be speculated that such nanotubes might be
rather coarse, and partially-to-well-graphitized nanoparticles of untraditional
nature.
In the next paper by R. Yazami, I. Goncharova and N. Plakhotnik, a
joint work between the U.S./France and Ukraine researchers is presented.
Authors investigate mechanisms of BF4- and PF6- anion intercalation into the
single walled carbon nanotubes (SWCNT), produced by pulsed laser ablation
technique. Authors conclude that BF4- and PF6- anions can intercalate into
SWCNT structures, and that this process has some reversibility. It is
noteworthy that the only electrochemical method used by the authors to
arrive with the above conclusion was a method of cyclic voltametry (CV). It
is known that CV is a highly sensitive method when it comes to studying the
surface processes occurring on working electrodes, while the limitation of
the method is its inability to detect what is happening in the bulk electrode.
In the mean time, intercalation is a process, which is happening in the bulk
156
of carbon. Intercalation processes are also well studied by the technique of

galvanostatic cycling, a constant current technique that was not used in the
present paper.
A fundamental review paper by Professor A. Churikov et al. of
Saratov State University, Russia focuses on cons and pros of lithium
intercalation into hard and soft carbons, as related to application of these
materials as lithium-ion battery anodes. Besides offering what the editors
consider a good tutorial on various forms of carbon used for intercalation,
the authors arrive with the conclusion that an optimum carbon for lithium-
ion battery should be a physical blend of soft and hard carbons. While being
an interesting concept, editors wish to note that North American and Pacific
Rim lithium-ion battery industries already use an enhanced version of this
concept, e.g. soft or hard carbon coated graphitic carbons. Authors also
report observing a synergistic capacity increase effect from combining two
types of carbons. As shown by Figure 2 in the subject paper, the resultant
capacity, surprisingly goes above 372 mAh/g, which is the theoretical limit
for graphite formulated as LiC6.
A paper by Dr. J. Liu et al. of Argonne National Laboratory in the
U.S. describes some of the analytical procedures used to qualify candidate
carbon materials for their use as anodes of emerging high power lithium-ion
batteries for hybrid electric vehicle (HEV) application. This work in the
United States is being supported by the U.S. Department of Energy’s
FreedomCar program, which, besides conducting the development work,
enables cooperation between all members of the battery production chain
with three major car manufacturers. The particular paper describes
performance of thermally purified, spheroidal natural graphite, SL-20, and
also, that of surface coated by amorphous carbon, thermally purified,
spherodized natural graphite SLC1015 (both grades are the products of
Superior Graphite Co., a Chicago, IL, USA graphite/carbon manufacturer).
Carbon coating and reduction of particle size were seen to improve high rate,
pulse performance and abuse tolerance of natural graphite to the extent of
making it a suitable candidate material for the HEV battery application.
We hope you will find these papers interesting and useful in your
work.
CARBONACEOUS MATERIALS
FOR BATTERIES
Tsutomu Takamura1 and Ralph J. Brodd2*

1
Department of Applied Chemistry, Harbin Institute of Technology, Harbin, 150001, China
2
Broddarp of Nevada, Inc., 2161 Fountain Springs Dr., Henderson NV USA
1. INTRODUCTION
Practically every battery system uses carbon in one form or another.

The purity, morphology and physical form are very important factors in its
effective use in all these applications. Its use in lithium-ion batteries (Li-
Ion), fuel cells and other battery systems has been reviewed previously [1 –
8]. Two recent applications in alkaline cells and Li-Ion cells will be
discussed in more detail. Table 1 contains a partial listing of the use of
carbon materials in batteries that stretch across a wide spectrum of battery
technologies and materials. Materials stretch from bituminous materials used
to seal carbon-zinc and lead acid batteries to synthetic graphites used as
active materials in lithium ion cells.
Table 1. Partial Listing of the Uses of Carbon and carbon materials in Batteries.
Battery System Use/Function
Carbon-Zinc Sealant, cathode conductive carbon black matrix

Alkaline Zn-MnO2 Sealant, cathode conductive graphite matrix
Zinc-Air Sealant, cathode catalyst support and current collector
Li- CFx Cathode active mass
Li-SO2 Cathode polymer bonded conductive matrix
Li-Ion Anode active mass, cathode conductive diluent
Lead Acid Anode Conductive diluent
Ultracapacitors Anode and cathode active matrix
Fuel Cells Cathode and anode catalyst support
Intercell connectors
*
157
158
It is estimated that over 14 billion alkaline Zn-MnO2 cells and 30

billion carbon-zinc cells were made worldwide in 2003. This compares to the
about 500 million lead acid automotive batteries that were made in the same
period.
2. ALKALINE CELLS
Today, the alkaline Zn-MnO2 cells set the standard for performance
for primary cells. The need for higher performance alkaline cells was met in
part by the use of “expanded” graphite from Superior Graphite. This type of
graphite is produced from natural graphite by treatment in nitric acid
followed by heating to cause exfoliation of the graphite. Photomicrographs
of the expanded graphite and natural graphite are shown in Figure 1. These
expanded particles mixed with MnO2 form a matrix with higher conductivity
than that produced by from the same formulation using natural graphite.
a b
Figure 1. Purified expanded (a) and natural (b) graphite. (Courtesy of Superior Graphite).
1.7 AAA-size battery

1.5
Cell Voltage
SGC Expanded Graphite

1.3
1.1
0.9
Synthetic Graphite
0.7
0 0.1 0.2 0.3 0.4 0.5 0.6
Discharge Capacity, Ah
Figure 2. Continuous Discharge of AAA-Size Alkaline Cells at 500 mA.

(Courtesy of Superior Graphite).
159
This permits a reformulation of the cathode to include more active

material and a lower internal resistance, as noted in Figure 2. The resultant
lower internal resistance and higher active mass content produced a 30%
increase in cell performance at high rate discharges.
The proper mixing of the graphite with the MnO2 is critical for high
rate performance. It is necessary to “put work into the mix” to uniformly
coat the manganese dioxide particles with a thin layer of graphite. This
provides a low resistance path from the manganese to the cell terminals to
minimize the internal resistance of the cell. A similar effect can be expected
if used in cathode formulations for Li-Ion cells.
2. LI-ION CELLS
The mobile electronic devices, including the 3rd generation of
portable phones, portable personal computers and personal digital assistants
(PDAs) have an insatiable appetite for “evermore” power and energy
(capacity). The power sources of choice for these devices are Li-Ion
batteries. There are several excellent reviews of recent work in this area [9 –
11]. The working voltage of Li-Ion batteries is about 3.5 volts, which is over
two times that of conventional portable batteries, and yields batteries that are
capable of delivering very high specific energy density. Li-ion has become
the system of choice for portable electronic devices and shows promise to
become the next generation universal rechargeable battery system.
2.5
2
Am pere Hours
1.5
Treated Graphite
1 and Fibers
Electrolyte Additives
Osaka Mesophase
0.5 Hard Carbon Pitch Beads
0
1990 1992 1994 1996 1998 2000 2002 2004
Ye a r
Figure 3. Capacity improvement in the ICR 18650 cell and the anode composition.
The capacity of the system has more than doubled since it was
introduced in 1991. The improvements in performance were due in large part
160
from the use of new carbon anode materials, as noted in Figure 3. Over 850
million Li-Ion cells were manufactured worldwide in 2002.
The Li-Ion system was developed to eliminate problems of lithium
metal deposition. On charge, lithium metal electrodes deposit moss-like or
dendrite-like metallic lithium on the surface of the metal anode. Once such
metallic lithium is deposited, the battery is vulnerable to internal shorting,
which may cause dangerous thermal run away. The use of carbonaceous
material as the anode active material can completely prevent such dangerous
phenomenon. Carbon materials can intercalate lithium into their structure (up
to LiC6). The intercalation reaction is very reversible and the intercalated
carbons have a potential about 50mV from the lithium metal potential. As a
result, no lithium metal is found in the Li-Ion cell. The electrochemical
reactions at the surface insert the lithium atoms formed at the electrode
surface directly into the carbon anode matrix (Li insertion). There is no
lithium metal, only lithium ions in the cell (this is the reason why Li-Ion
batteries are named). Therefore, carbonaceous material is the key material
for Li-Ion batteries. Carbonaceous anode materials are the key to their ever-
increasing capacity. No other proposed anode material has proven to perform
as well. The carbon materials have demonstrated lower initial irreversible
capacities, higher cycle-ability and faster mobility of Li in the solid phase.
Figure 4. Depiction of lithium intercalated into the carbon/graphite lattice.
The heart of the carbon anode utility is its structure, shown in

Figure 4. Lithium ions in the organic electrolyte approach the surface of the
graphite where they exchange an electron and enter the carbon structure
(intercalate) along the basal plane. The lithium atoms settle into the center of
the hexagonal carbon structure in the amount of up to one lithium to six
carbon atoms (LiC6). As the lithium content of the graphite increases, the
graphite approaches the lithium potential. With graphite, there is a small
(5%) volume change as the electrode reaches full charge. As the electrode
161
approaches the lithium potential a thin film forms on the surface from the
reaction with the organic solvent to form a protective passive layer called the
Solid Electrolyte Interphase (SEI) layer. This reaction uses lithium from the
cathode and is called “first cycle loss.” It is irreversible and related to the
surface and type of carbon/graphite. Electrolyte additives can control the
losses related to the formation of the SEI. In a well formed film, there is no
capacity loss associated with solvent reduction and SEI layer formation after
the first cycle.
The reaction at the anode in Li-Ion cells is given in Equation 1.
During charge the lithium ions approach the surface of the carbon where
they accept an electron and enter the lattice. On discharge, the opposite
reaction occurs. The electrochemical reaction is thought to occur on the edge
planes and not the basal plane of the carbon/graphite particles.
Li+ + e- = Li (in C) (1)
3. KINDS OF CARBONACEOUS MATERIALS,

WHICH CAN BE USED AS ANODES IN LI-ION CELLS
These kinds of carbon materials include:
1) Highly Oriented Pyrolytic Graphite (HOPG)

2) Natural graphite
3) Artificial graphites
4) Hard carbons
5) Mesophase low temperature fibers and MCMB carbons
Figure 5. TEM photograph of natural graphite [12].
While it is possible to prepare square centimeter, and larger, pieces

of HOPG. the performance of HOPG is very similar to that of natural
162
graphite but much more expensive and it will not be considered further here.
A TEM photograph of natural graphite is shown in Figure 5 in which the
movement of lithium in the carbon skeleton is very rapid along the
horizontal (basal plane) direction. The discharge curve is very flat and the
charge/discharge reversibility after the first cycle is essentially 100 %
(Figure 6). The first cycle loss for typical 5 m2/g natural graphite is about
10% to 12%. The high power rate, however, is restricted because the
graphite tends to form leaf-like platelets with very little edge plane exposed
and a long diffusion path for the lithium reaction.
Figure 6. Low rate charge discharge performance of natural graphite electrode [13].
Figure 7. Photomicrograph of treated graphite “potatoes” (Courtesy of Superior Graphite).

163
When the natural graphite powders are coated on the current

collector, they tend to be aligned to be parallel to the collector surface, which
raises the ohmic resistance perpendicular to the electrode surface, which is
important for the high rate charge/discharge. Surface treated natural
graphites, called potatoes, shown in Figure 7 can avoid these surface
orientation related problems to yield a high capacity material with a low first
cycle loss of about 5%. Since the lithium transfer is through the edge plane
to the electrolyte, the electrode reactions may be hindered by the structure of
the SEI layer at the electrolyte-graphite interface. If the graphite particles can
align with the basal plane perpendicular to the current collector, the reaction
is facilitated. Graphite materials presently used in the practical commercial
batteries include mesophase carbon beads and fibers (MCF), artificial
graphites having the potato-like shape and surface treated natural graphite.
3.1. Mesophase Carbon Fibers

Petoca, Ltd. prepared mesophase carbon fibers by melting the
mesophase pitch. The fiber is formed by blowing molten pitch through tiny
holes in a die into a heated chamber. The fibers are then heated at about
300ºC in an atmosphere of low-pressure oxygen and water vapor, followed
by heating at 500ºC in nitrogen to produce a thin hard protecting skin.
Finally, heating up 3000ºC produces a well-graphitized bead or fiber.
The formed graphite fiber has a radial structure of the crystal as shown in
Figure 8 (a) and (b). The fiber is cut into short lengths before the
graphitization.
(a) (b)
Figure 8. (a) Schematic structure of carbon fibers and (b) an SEM of a Melblon 3100 fiber
showing the fiber and internal structure of a fracture.
164
This procedure produces fibers, such as Melblon 3100, which have

1) a radial crystal orientation which can accept Li ions from all the side wall
areas and 2) the side wall is covered with a porous thin graphen skin which
protects the fiber from the shape change during the Li intercalation and de-
intercalation (swelling and shrinkage). Fibers as thin as 7 Pm can be
prepared to facilitate their use in electrode coating operations. The anode
discharge curves are shown in Figure 9, where (a) is for the normal product
and (b) is for the boron-doped product. [14]. In addition, surface coating of
the fibers with silver, zinc, etc., improves the transfer across the interphase
boundary to yield very good high rate performance [15].
Volts vs. Li
Volts vs. Li
0 0.1 0.2 0.3 0.4 0 0.1 0.2 0.3 0.4

Ah/g Ah/g
(a) (b)
Figure 9. Discharge curves of (a) normal and (b) boron doped Melblon carbon fiber
at 1, 2, and 6 mA/cm2.
Toshiba, among others, have succeeded to produce a very thin (as

thin as 1 mm thick for the 7 mm batteries that have superior performance, as
compared to conventional Li-Ion batteries. The success is based on the use
of the Petoca boron doped MCF (BMCF), new electrolytes containing LiBF4
dissolved in EC/MEC/J-BL, and an Al laminate plastic sheet container. The
low temperature performance is superior to that of the conventional Li-ion
battery. The carbon fiber is electrically highly conductive but it is still
strongly recommended to coat the milled carbon fiber with an appropriate
amount of conductive additives, such as natural graphite powder or Timcal
synthetic graphite powder. The particle size of the powder should be 5Pm, or
less, to maintain good contact with the fibers.
Acetylene/carbon black is also quite effective but has an initial
irreversible capacity that cannot be ignored. The amount of irreversible loss
for acetylene black component ranges up to 20%. The particle size of
conductive additives is recommended to be less than 5 microns. The addition
is very effective to improve to improve: 1) cycle life, 2) high power
capability, and 3) the initial charge efficiency (reduce the initial irreversible
165
capacity loss). The use of the expanded graphite can serve the same purpose
without the high irreversible loss associated with carbon blacks. The coating
know-how is very important in producing high performance electrode
structures. The know-how is not restricted to the slurry formulation but the
control of coating machine operations are equally, or more, important for
thickness control and particle orientation.
3.2. MCMB (Mesophase Carbon Micro Beads)

When the mesophase pitch is heated, tiny spherical beads having a radial
orientation of the graphite embryo form and grows in the mother pitch.
When the size of the beads in the pitch reaches an appropriate size (10-30
Pm), the pitch is cooled down and the beads can be collected by dissolving
the pitch matrix with an appropriate solvent. The beads are then heat-treated
in an oxygen/water gas mixture to produce the protective skin, as for the
MCF, and then graphitized, as depicted in Figure 10.
Figure 10. Depiction of the (a) Onion structure, the (b) Radial structure,
and (c) Actual product MCMB, [16].
MCMB produced by Osaka Gas Co. has very good performance and
is easily coated on the Li-Ion anode current collector (Cu). These materials
are used widely throughout the world. The price is expensive, and cannot be
reduced. This production process is inherently expensive due to the large
volume of solvent required to be wash out and recover the beads from the
pitch matrix.
3.3. Hard Carbons

When thermosetting resins, such as, phenol resin, epoxy resin,
cellulose, sugars or proteins are heated in an inert gas or vacuum, the
166
materials decompose to release water, and other gases to form carbon. The
carbon particles are inherently joined together in a complicated structure and
cannot reorient to form graphite. The resulting carbon is very hard and has
no preferred crystallite orientation but has a randomly oriented micro-
structure of graphite and voids or cavities. Therefore, the hard carbons have
a higher capacity than that of graphite.
a b
Figure 11. (a) Photomicrograph of Kureha Carbotron P,
(b) TEM of 2800 C heat-treated Sugar carbon.
These materials have an advantage that they do not decompose

propylene carbonate (PC) so that the low cost PC electrolytes, having very
good performance, can be used. In contrast, PC tends to decompose
vigorously when it is in contact with lithiated graphite. These are two
advantageous points of hard carbon as compared with graphite. In addition,
hard carbons have very little change in volume during the charge/discharge
reaction. However, these carbons have several demerits: 1) they have very
poor high rate performance related to the slow diffusion of lithium in the
internal carbon structure; 2) the discharge curves is not flat but gradually
slopes down to the end-of-charge voltage; 3) large initial irreversible
capacity, as noted in Figure 11, of 20% or more. A comparison of the three
materials is given in Table 2.
Ah/kg
Figure 12. Charge and discharge curves of Carbotron P.
167
Table 2. Comparison of the anode performances of three carbonaceous materials.
Materials
Natural Artificial
Hard carbons
graphite graphite
Performance
Capacity ** ** ***
High rate capability ʊ *** *
High working
*** *** *
voltage
Flatness of discharge *** ***
Price *** * **
Low temp. perform. ** *** **
Use PC based
ʊʊ ʊʊ **
electrolytes
Electric conductivity *** *** **
Sony has stuck to their hard carbon for the anode because of its long
cycle life and still makes a line of hard carbon anode cells. Sony has moved
to graphite because they have recognized that hard carbon cannot be
improved to deliver the high power and high capacity requirements of
current portable electronic devices. Some hard carbons have demonstrated
very high capacities but are considered unusable in their present form
because of their high initial irreversible capacity and slow rates of Li doping
and undoping (intercalation). It should be noted that recently electrolyte
additives have been developed that allow the use of natural graphite in low
cost propylene carbonate (PC) electrolytes (17, 18). These materials control
the initial SEI layer formation, improve storage and cycle life.
Li-Ion batteries are poised to enter a new era of high performance.
Figure 13 depicts the segmentation into two market segments. One path will
continue the drive for higher performance but will require new, more
expensive materials [19]. The other path will emphasize low cost materials
with essentially equivalent performance to serve new developing markets
such as power tools and hybrid vehicles. This path puts the Li-Ion system in
a position to challenge the Ni-Cd and Ni-MH for power tools and the hybrid
automotive applications. It also raises the option to compete with primary
alkaline Zn-MnO2 cell for low-end portable electronic applications. One key
to meeting these challenges is the availability of low cost electrolyte and
electrode materials. Solution additives have lowered the first cycle loss and
opened the possibility to use lower cost electrolyte compositions and
eliminate the use of LiPF6 solutes. The other key is the increased safety
realized by the use of low cost, stable cathode materials.
168
High
Period of Rebirth Performance
Doped Nickel Cathode
3 Ah
Materials 2.2 Ah Alloy Anode
Understanding
Low Cost
2003 Mn or Phosphate Cathode
1 Ah New – anode,
Engineering Treated Graphite Anode
Improvements cathode. electrolyte,
1991 binders, packaging, LiBOB - PC Electrolyte
etc.
Figure 13. Depiction of the development of the Li-Ion battery system.
4. ULTRACAPACITORS
The carbon-carbon ultracapacitor has become the main memory

protection device in appliance applications. These devices store charge in the
electrical double layer that forms at all electrode-solution interfaces. High
surface area carbon blacks (1000 m2/g) are the material of choice. Since no
electrochemical reactions are involved, these devices have very long life (105
cycles or more). This field has been very active as it is possible that the
ultracapacitor will find application in hybrid vehicles. Recently, Okamura
reported that devices with 45 Wh/kg and 60 Wh/l are possible with carbons
where development of a high internal surface area results from intercalation
and removal of potassium [20]. The potassium opens channels of
nanoporosity with high surface area. Asymmetric capacitors, with a carbon
anode and a battery electrode (nickel, lead, etc.), have been developed with
very long cycle life. In another development, Naoi deposited S-stacked
supramolecule redox couples onto carbon nanotubes. He reported long cycle
life and reasonable energy density [21].
5. CLOSING REMARKS
The future remains bright for the use of carbon materials in batteries.
In the past several years, several new carbon materials have appeared:
mesophase pitch fibers, expanded graphite and carbon nanotubes. New
electrolyte additives for Li-Ion permit the use of low cost PC based
electrolytes with natural graphite anodes. Carbon nanotubes are attractive
new materials and it appears that they will be available in quantity in the
near future. They have a high ratio of the base plane to edge plain found in
HOPG. The ultracapacitor application to deposit an electronically conductive
polymer on the surface of a carbon nanotube may be the wave of the future.
169
REFERENCES
1. The Electrochemistry of Carbon, S. Sarangapani, J. Akridge and B Schuum, Eds., The
Electrochemical Society, Pennington, NJ, 1984, PV84-5.
2. K. Kinoshita, Carbon Electrochemical and Physicochemical Properties, J. Wiley, New
York, 1988.
3. M. Winter and J. Besenhard in Handbook of Battery Materials, J. O. Besenhard (Ed.)
Wiley-VCH, Weinheim (1999).
4. Mitio Inagaki, Carbon Material Engineering, Nikkan Kogyo Publ. (Japanese) (1985).
5. M. Inagaki and Hishiyama, New Carbon Material, Gohodo, (1994) (Japanese).
6. Denchigijutu, (No 6) 1994-(No. 10) 1999 Battery Committee Japan (Japanese).
7. Handbook of Batteries and Fuel Cells, D. Linden and T. B. Reddy, Eds., McGraw-Hill,
New York, 3rd Edition, 2001.
8. Handbook of Battery Materials, J.O. Besenhard, Ed., Wiley-VCH, Weinheim, 1999.
9. M. Yoshio and A. Kozawa, Lithium ion Batteries, Nikkan Kogyo, Tokyo, 2nd edition
(2000).
10. Lithium Ion Batteries, M. Wakihara and O. Yamamoto, Wiley VCH Weinheim, 1998.
11. Advances in Lithium-Ion Batteries, W.A. van Schalkwijk and B. Scrosati, Klewer
Academic/Plenum Publishers, 2002.
12. X.Y. Song, Xi Chu and K. Kinoshita, in New Sealed Rechargeable Batteries and
Supercapacitors, MRS Symposium Proceedings Volume 393, Materials Research
Society, Warrendale, 1995, p321.
13. S. Mori, H. Asahina, H. Suzuki, A. Yonei and K. Yokoto, J. Power Sources, 68, 59
(1997).
14. T. Tamaki, T. Kawamura, and Y. Yamazaki, in Materials for Electrochemical Energy
Storage, MRS Symposium Volume 496, D.S. Ginley, D.H. Doughty, B. Scrosati, T.
Takamura and Z. Zhang, Eds., Materials Research Society, Warrendale, 1998, p.569.
15. J. Suzuki, O. Omae, K.Sekine and T. Takamura, Solid State Ionics, 152-153, 111(2002).
16. M.Nishizawa R. Hashitani, T. Ito, T. Matsue and I. Uchida, Electrochem. and Solid
State Letters, 1, 10 (1998).
17. D. Aurbach In Proceedings 1st PBFC at Jeju Island, Korea, June 1 – 6, 2003, J. Power
Sources, in press.
18. M. Broussely, S. Herreyre, P. Biensan, P. Kaztejna, K. Nechev and R. J. Stanowitz, J
Power Sources 97-98, 13 (2001).
19. T. Takamura, Solid State Ionics, 152-153, 19 (2002).
20. M. Takeuch, T. Maruyama, K. Koike, A. Morgami and M. Okamura, Electrochemistry,
66, No. 12 (1998).
21. K. Naoi, In Proceedings 1st PBFC at Jeju Island, Korea, June 1 – 6, 2003, J. Power
Sources, in press.
ANODE-ELECTROLYTE REACTIONS IN
Li BATTERIES: THE DIFFERENCES BETWEEN
GRAPHITIC AND METALLIC ANODES
H. J. Santner, K. C. Möller, W. Kohs, C. Veit, E. Lanzer, A. Trifonova,

M. R. Wagner, P. Raimann, C. Korepp, J. O. Besenhard, and M. Winter**
Institute for Chemical Technology of Inorganic Materials, Graz University of Technology,
A-8010 Graz, Austria,
Abstract
An overview about more than 10 years of R&D activities on solid

electrolyte interphase (SEI) film forming electrolyte additives and solvents at
Graz University of Technology is presented. The different requirements on
the electrolyte and on the SEI formation process in the presence of various
anode materials (metallic lithium, graphitic carbons, and lithium storage
metals/alloys are particularly highlighted.
Keywords
Solid electrolyte interphase (SEI), electrolyte additive, lithium ion

battery, Li metal, graphite, lithium alloy.
1. METALLIC LITHIUM AS ANODE MATERIAL IN

RECHARGEABLE LITHIUM BATTERIES
Considering thermodynamic reasons for the selection of a metallic
anode material, light metals like Li, Na, or Mg are favored as they combine
outstanding negative standard potentials with low equivalent weights, with
lithium showing the lowest weight and most negative potential. The strong
reducing power of metallic lithium leads to its thermodynamic instability in
most known battery electrolytes. Fortunately, in suitable electrolytes, films
of electrolyte decomposition products spontaneously form after the
*
171
172
immersion of Li into the electrolyte solution. In the ideal (but so far not fully
realized) case, the films still act as a membrane for the active charge carrier,
the lithium cation, but are electronically insulating and are impermeable to
other electrolyte components and thus hinder further electrolyte reduction
[1]. In other words, these films behave as a solid electrolyte interphase (SEI)
[2]. The observation of the kinetic stability of lithium in a number of
nonaqueous electrolytes induced a rapid proliferation of research in the field
of lithium batteries beginning in the 1950s, and the commercialization of
primary lithium batteries followed quickly. The commercial breakthrough of
rechargeable batteries using metallic lithium, however, failed, as upon
recharge in liquid electrolytes lithium plating occurs simultaneously with Li
corrosion and passivation (formation of the SEI). The SEI is inhomogeneous,
both with regard to composition and thickness, and lithium deposition
through the SEI therefore does not proceed uniform and smooth but locally
very different, that dendritic highly reactive metal depositions can be
observed (cf. Fig. 1). The dendrites are substrates for further inhomogeneous
SEI formations, which results in further dendritic Li deposition, etc. In other
words, the dynamic lithium surface during cycling results in continuous SEI
formation, which again promotes the dynamics of the surface. As a
consequence of this "vicious cycle", Li and electrolyte losses occur in each
cycle and thus the cycle life of the cell is drastically reduced. Even worse,
the dendrites may grow to filaments which locally short-circuit the cell. Due
to the low melting point of metallic Li (about 180qC) the local overheat can
trigger a disastrous thermal runaway [3].
10.0 µm
Figure 1. Lithium deposition on Ni performed in 1 M LiClO4 in propylene carbonate (PC)

as electrolyte, Q = 2.7 C, I = 1 mA/cm2, magnification: x 1000 [4].
173
2. CARBON REPLACES METALLIC LITHIUM IN

RECHARGEABLE LITHIUM BATTERIES
The market breakthrough of rechargeable lithium batteries was
achieved when the performance and safety limited metallic lithium electrode
had been replaced by layered carbons (favorably graphitic carbons), the
graphite being an intercalation host for lithium cations. The main difference
to metallic lithium is that the positions for charge storage and protective SEI
formation are different. The SEI is formed on the surface of graphite, and the
lithium cations are transported through the SEI into the graphite, where they
are accommodated into the host structure. Thus, the surface of graphite in
contact with the electrolyte remains stable and so does the SEI (Figure 2). As
a consequence and at contrast to metallic lithium, SEI formation on graphites
and other carbons takes place in a confined period of time, i.e., mainly in the
first cycle. Therefore, SEI formation on graphite can be (but also has to be)
controlled during the 1st charge process [3, 5, 6].
Anode SEI Electrolyte
Li+
LiCn +
Li Li+(solv) y
Carbon surface
Figure 2. SEI film formation on the carbon surface (for details, cf. text).
3. FILM FORMING ELECTROLYTE COMPONENTS

AND ADDITIVES
SEI formation control is the key to good performance and the safety
of the whole lithium ion battery, as not only anode operation but also
cathode properties are strongly affected by the SEI formation process (the
cathode is the lithium cation source of lithium ion cells). Apart from control
of the graphite (surface) properties, an appropriate composition of the
electrolyte is usually helpful for creation of an effective SEI.
Currently, graphite-based lithium ion batteries use mixed solvent
electrolytes containing highly viscous ethylene carbonate (EC) and low
viscosity dilutants such as dimethyl carbonate (DMC) or diethyl carbonate
(DEC) as main solvents. EC is indispensable because of its excellent filming
characteristics. DMC and/or DEC are required to get the low temperature
174
performance of the electrolyte at least reasonable. The performance

limitations of EC/DMC or EC/DEC based electrolytes, e.g., at lower and
higher temperatures, under abuse conditions, or with regard to the 1st cycle
irreversible charge losses, stimulated various R&D efforts for improved
electrolytes, either by addition of certain electrolyte additives or even by
replacement of certain solvent components, e.g., either EC or DMC/DEC by
other solvents and additives [3, 5, 6].
Figure 3. Charge recoveries vs. wet stand time of LiCn in PC electrolytes without and with
electrolyte additives. Electrolyte: 0.5 M LiClO4 in PC, Carbon: Highly graphitic carbon fiber
P100 (Amoco), 1st charge with -50 µA mg-1. [7].
S : without additive
z : CO2 (saturated)
: N2O (saturated)
¡ : 3 x 10-4 M Sx2-
Figure 4. Charge recoveries vs. wet stand time of LiCn in PC electrolytes without and with
electrolyte additives (CO2, N2O, Sx2-). Electrolyte: 0.5 M LiClO4 in PC,
Carbon: Highly graphitic carbon fiber P100 (Amoco), 1st charge with -50 µA mg-1 [4].
The first promising results have been achieved with inorganic

additives such as CO2 [4, 7, 8] (Figs. 3-5), and N2O [4, 8] or Sx2- [4, 8] (Figs.
4-5) added to organic solvent-based electrolytes, which in comparison to the
175
additive-free electrolyte showed some improvements of the SEI formation

process. However, most surprisingly also inorganic component based
electrolytes, e.g., SO2-based electrolytes [4, 9] (Fig. 6) and (with some
limitations) also SOCl2-based electrolytes (Fig. 7) [4, 9] showed some
compatibility with graphitic carbons.
Figure 5. Integrated capacity losses of LiCn in PC electrolytes without and with electrolyte
additives (CO2, N2O, Sx2-). Electrolyte: 0.5 M LiClO4 in PC, Carbon: Highly graphitic carbon
fiber P100 (Amoco), i = 50 µA mg-1, cut-off: 0-2 V vs. Li/Li+ [4].
The beneficial effect of CO2 on the SEI formation process and the
cycling performance of graphitic carbons have been reported by many
groups [4, 7, 8, 10, 11]. It is most nicely demonstrated in J-butyrolactone
based electrolytes obviously, CO2 suppresses the decomposition reaction of
J-butyrolactone at approx. 0.8 V vs. Li/Li+ (Figs. 8, 9). Independent of the
used electrolyte salt (LiClO4 (Fig. 10, top, left), LiBF4 (Fig. 10, top, right),
LiPF6 (Fig. 10, bottom, left), and LiN(SO2CF3)2 (Fig. 10, bottom right), the
cycling performance of the J-butyrolactone based electrolytes improves
when CO2 is present. Nevertheless, the significant differences in the
electrochemical performance indicate that the electrolyte salt anion plays a
strong role in the SEI formation process, i.e., that the filming behaviour of
the anion may interfere with that of the additive or with that of J-
butyrolactone [12, 13].
176
Capacity
Losses,
%
Cycle number
Figure 6. Integrated irreversible capacities of LiCn in 1 M LiBr in SO2/acetonitrile (AN) and
0.5 M LiClO4 in PC as electrolytes. Carbon: Highly graphitic carbon fiber P100 (Amoco),
i = 50 µA mg-1, cut-off: 0-2 V vs. Li/Li+ [4].
C
H
A
R
G
E
R
E
C
O
V
E
R
Y
Cycle number
Figure 7. Constant current charge/discharge cycling of highly graphitic carbon fiber P100
(Amoco) in 1.8 M LiAlCl4 in SOCl2 electrolyte, i = 300 µA mg-1, cut-off: 0-2V vs. Li/Li+ [4].
177
Figure 8. Constant current charge/discharge cycling (1.-3. cycles) of graphite (Lonza KS44
synthetic graphite) in 1 M LiClO4 in γ-butyrolactone as electrolyte without and with CO2
(saturated in electrolyte) as electrolyte additive, i~10µA mg-1, cut-off: 0-1.5V vs. Li/Li+
[12,13].
Figure 9. Excerpts of the peaks due to "electrolyte decomposition" recorded by voltammetry

in 1 M LiClO4 in J-butyrolactone as electrolyte without and with CO2 (saturated in the
electrolyte) as electrolyte additive, scan rate: 0.01 mV s-1 [12, 13].
178
Integrated Capacity Losses / mAh g-1
Cycle number Cycle number

Integrated Capacity Losses / mAh g-1
Cycle number Cycle number
Figure 10. Integrated irreversible capacities of LiCn in J-butyrolactone based electrolytes

without (full symbols) and with (open symbols) CO2 as electrolyte additive using various
electrolyte salts: LiClO4 (top, left), LiBF4 (top, right), LiPF6 (bottom, left), LiN(SO2CF3)2
(bottom, right). Carbon: Lonza KS44 synthetic graphite,
i = ~10 µA mg-1, cut-off: 0-1.5 V vs. Li/Li+ [12].
2.0
1.5
E / V vs. Li/Li+
1.0
0.5
0
250 500
C / mAh g-1
Figure 11. First cycle constant current charge/discharge curves of synthetic graphite
TIMREX£ SFG 44 using 1 M LiClO4 in PC:PS (propylene sulfite) (95:5 by volume)
as electrolyte. i = r 20 mA g-1, cut-off = 1.8/0.025 V vs. Li/Li+.
179
Due to easier handling, liquid electrolyte additives and solvents have

found more interest than gaseous additives. Among these components,
organic sulfites [14-19] (Fig. 11), partially fluorinated aprotic solvents [12,
20-25], and vinylene additives [24, 26-32] have found particular attention.
Partially fluorinated components can be used either as electrolyte
solvents (Fig. 12) or as electrolyte additives (Fig. 13). In many cases they
show much superior SEI forming capabilities compared to their non-
fluorinated counterparts. Moreover, fluorinated solvents are in general much
less flammable as less hydrogen is available, which might contribute to cell
safety [12, 23, 25].
E / V vs. Li/Li
1.4
Solvated intercalation
1.2
followed by solvent
EC/DME
E / V vs. Li/Li+
1.0 decomposition
0.8
0.6
0.4
EC/F-DME
0.2
0
0 0.5 1.0 x in LixC6
(186) (372) (C/mAh.g-1)
Figure 12. 1st cycle constant current charge curves of synthetic graphite LONZA KS 44 (i)
using 1 M LiN(SO2CF3)2 in EC:DME (dimethoxy ethane, CH3OCH2CH2OCH3) (3:2) as
electrolyte. The measurement was stopped when the graphite was exfoliated, (ii) using 1 M
LiN(SO2CF3)2 in EC:F-DME (partially fluorinated dimethoxy ethane, CH3OCF2CF2OCH3)
(3:2) as electrolyte. i = r 20 mA g-1, cut-off = 0.0 V vs. Li/Li+ (adapted from [12]).
0.3
0.2
0.1
-1
0.0
i / mA mg
-0.1
-0.2
-0.3
-0.4
0 500 1000 1500 2000 2500 3000
E / mV
Figure 13. Cyclic voltammogram of graphite (Lonza KS6, synthetic graphite) in 1 M LiClO4
in PC/novel fluorinated additive (90:10, v:v); scan rate 30 µV s-1, Potentials vs. Li/Li+.
180
Vinylene additives are very attractive at the moment both from an

academic and industrial point of view [24, 26-32]. Usually, vinylene
additives can be very effective even when they are used in very low amounts
(d 1 vol.-%), in particular when the first charging step is properly adjusted
(Fig. 14).
(a) (b)
0.5
0.4
0.4
-1
-1
capacity / mAh mg
capacity / mAh mg
0.3
0.3
0.2
Charge 0.2 Charge
Discharge Discharge
0.1 0.1
0.0 0.0
1 2 3 4 5 6 7 5 10 15
cycle number cycle number
Figure 14. (a) Constant current charge cycling of graphite (TIMREX SFG 44) in 1M LiClO4
in PC/AAN (99:1, v:v) with a constant current of r20 mA g-1, cut-off: 0.02/1.5 V vs. Li/Li+.
The efficiency was 70% in the 1st cycle, and <85% in the following cycles. (b) Same graphite
and electrolyte, but different charging process: Potentiodynamic ramp from 3.0 to 0.5 V vs.
Li/Li+ in the 1st charge, then constant current cycling at r20 mA g-1, cut-off: 0.02/1.5 V vs.
Li/Li+ in the rest of the 1st cycle and subsequent cycles. The efficiency was >80% in the 1st
cycle and >>90% in the following cycles. The applied combined potentiodynamic/
galvanostatic 1st cycle charge process might is not the ultimately best one, but illustrates the
effect of the charge conditions on the performance.
4. WHAT IS SPECIAL ABOUT THE SEI FORMATION

ON GRAPHITE?
With regard to SEI formation, a main difference to lithium metal
(above) and lithium storage metals/alloys (below) is that the structure of
layered graphite gives rise to basically two kinds of surfaces, prismatic
(edge) surfaces and basal plane surfaces. Ideal (= defect and contaminant
free) basal plane surfaces are homogeneous and "smooth" and consist only of
carbon atoms. In contrast the prismatic surfaces are heterogeneous and
"rough" and apart from carbon may contain various, mostly oxygen-
containing surface groups [33]. It is well known that the prismatic and basal
plane surface areas of graphite show a different SEI formation behaviour
[33, 24]. For example, in propylene carbonate (PC) based electrolytes
ternary solvated graphite- intercalation compounds take part via the
prismatic surfaces in the electrolyte reduction and SEI formation mechanism.
181
In fact, crystalline graphites usually cannot be operated in PC

electrolytes, unless effective film forming electrolyte additives are used (see
above) as propane gas evolution [35], creation of solvated graphite
intercalation compounds (sGICs) [36], and graphite exfoliation take place.
Recently [37, 38], it was found that propylene evolution is observed at
graphite, while absent at lithium active metallic anodes, e.g., Sn and SnSb.
The compared to other anode materials unique propane gas
evolution process at graphite, can be explained by decomposition of the only
kinetically stable sGICs, as only graphite, but not metallic Li and lithium
storage metals/alloys are vulnerable to solvent co-intercalation (cf.
experiment, in Fig. 15). Apparently, the decomposition products of the
sGICs, which are created apart from propane, do not form an effective SEI
hindering further co-intercalation. Thus, even when only a limited graphite
area may be subject to solvent co-intercalation (which may be not visible by
XRD), gas evolution takes place inside graphite and the formed gas could act
as strong lever enabling exfoliation and mechanical destruction of graphite
[39, 40] (Fig. 16).
0.0 TIMREX T200-2000

+
TIMREX T200-2000
V vs. Li/Li
2.0
-0.5
mA
-1.0 1.5
potential
-1.5 1.0 plateau
-1000 0 1000 2000
Mass Signal
Mass Signal
-14
-15 m/e = 41 -14
5 * 10 A m/e = 41
5 * 10 A
propene gas
evolution
-1000 0 1000 2000 200
200
200
100
100
100
00
0
-100
-100 -200
-200
+
mV vs. Li/Li time / s
Figure 15. In order to monitor gas evolution in parallel to the electrochemical experiment we
took use of on-line mass spectrometry. The experiment was performed as follows: Following a
fast potentiodynamic step with 10 mV s-1 from ca. 3.0 to 0.65 V vs. Li/Li+, the electrode
(TIMREX T200-2000) has immediately been kept at open circuit. At open circuit, the
potential and the mass signal of propylene have been recorded vs. time. At these fast scan
rates, only small areas of the graphite react. Thus only small, but still significant mass signals
can be recorded: No strong propylene evolution is observed during the voltammetric scan
(Fig. 15 left). At open circuit the graphite potential jumps from 0.65 V to >1 V vs. Li/Li+,
where a potential plateau develops (Fig. 15, right). This plateau potential is typical for
solvated graphite intercalation compounds (sGICs). The formed sGICs are only kinetically
stable and gradually decompose during time, which is visible by a gradual shift of the open
circuit potential to more positive values. This is accompanied by propene gas evolution (m/e
= 41), the gas being a decomposition product of the solvated intercalates [39].
At the prismatic surfaces, the primary reduction reaction is the

reduction of those solvent(s) co-intercalated inside graphite, whereas on the
basal plane graphite sites, reduction reactions may proceed in an excess of
182
all electrolyte species (including electrolyte salt anions and solvents that do
not co-intercalate). Therefore, the electrolyte decomposition products formed
on the basal plane surface should have a different composition to those
formed at the prismatic surfaces [38]. In addition, it should be noted, that due
to the fact, that lithium intercalation and solvent reduction occur in parallel
at the prismatic surface sites, the SEI at the prismatic surfaces contains a
higher fraction of lithium compounds than the SEI at the basal plane sites.
Polycrystalline graphite
Crevice
(grain boundary)
Graphene layer Gas evolution

Solvated Li+ SEI products
Solvated anion Crevice formation and/or opening
Further co-intercalation
further gassing
further crevice formation
further electrolyte penetration into
crevices by capillary forces, etc.
Graphite exfoliation and

macroscopic destruction
Figure 16. Model: In PC based electrolytes, solvent co-intercalation, gas formation and
crevice formation in polycrystalline graphite materials are inter-related reactions. In fact,
there is a subsequence of reactions: (1) PC co-intercalation, (2) gas formation, (3) crevice
formation ultimately resulting in exfoliation and macroscopic destruction of graphite [40].
It should be noted here, that not only the (chemical and

morphological) composition of the protective layers at the basal plane
surfaces and prismatic surfaces is different, but that these layers also have
completely different functions. At the prismatic surfaces, lithium ion
transport into/from the graphite structure takes place by intercalation/de-
intercalation. Here the formed protective layers of electrolyte decomposition
products have to act as SEI, i.e., as transport medium for lithium cations.
Those protective layers, which have been formed on/at the basal plane
surfaces, where no lithium ion transport into/from the graphite structure
takes place, have no SEI function. However, these “non-SEI layers” still
protect these anode sites from further reduction reactions with the
electrolyte.
183
Though there is a distinct difference between the functions of the

SEI and non-SEI layers, it is accustomed in the literature to designate the
protective layers as SEI, independent of their function. However, a most
important impact of the differentiation between SEI- and non-SEI layers on
graphite is that during processes such as SEI aging, formation, or dissolution,
any measurable changes of the non-SEI layers at the basal plane surface sites
can not be directly related to the electrochemical data recorded in parallel,
such as charge, discharge or self-discharge as these electrochemical
processes take place only at the prismatic sites of graphite (Fig. 17).
Figure 17. The basal plane and prismatic surfaces of graphite have different functions with
respect to lithium intercalation and de-intercalation (= charge, discharge, self-discharge,
etc.). As a consequence, only the electrolyte decomposition product layers at the prismatic
surfaces have SEI function. Any processes related with electrolyte decomposition product
layers at the basal plane surfaces (= non-SEI layers) therefore can not be directly related to
electrochemical data such as charge, discharge, self-discharge, etc. The situation is even
more complex as the SEI composition and morphology at the basal and prismatic surface
planes varies [40].
5. SEI FORMATION AND SEI STABILITY ON LITHIUM

STORAGE METALS AND ALLOYS
Electrolyte effects on the cycling stability of lithium storage metals
and alloys indicate the importance of SEI formation in this case, too. Very
early measurements suggest that additives such as CO2 do not only improve
the cycling stability of metallic lithium [41] and graphitic carbons (see
above), but also that of lithium storage metals (Fig. 18), which may be
related with the electrical properties of the SEI (Fig. 19) [13].
184
Figure 18. Cycling of metallic Sn (2.5 µm) on a Cu-substrate in 1 M LiClO4/PC in absence

and presence of CO2 (saturated at ambient pressure). ic = id = 0.025 mA cm-2, capacity
limited charge of 1.7 Li/Sn, discharge cut-off: 0.8 V vs. Li/Li+ [13].
Figure 19. Impedance spectra of "Li1.5Sn0.72Sb0.28" after storage for 40 days in 1 M LiClO4/PC
in absence (fapex = 341 Hz) and presence (fapex = 674 Hz) of CO2
(saturated at ambient pressure) [13].
Recent studies on the electrolyte decomposition processes on

graphite and lithium storage metals however revealed that the reduction
mechanisms and SEI compositions as well as the requirements on the SEI
properties are different (see above). Because of the strong volume changes of
the metallic hosts during lithium uptake and removal in lithium alloy
electrodes, the SEI is subject to continuous mechanical stress. This may
be relieved by the formation of a more flexible SEI, whereas graphites
require a dense SEI suppressing solvent co-intercalation. From the point of
view of surface dynamics, the SEI stability on lithium storage metals/alloys
may be rather compared with that on metallic lithium than that on graphites
(Fig. 20) [42].
185
Figure 20. SEI formation on different anodes for rechargeable Li batteries: (A) lithium
metal, (B) graphitic carbon, and (C) metals and intermetallics. Different colors of the SEI
indicate SEI products formed at different stages of charge and discharge
(and do not indicate different composition) [42].
Figure 21. Effect of adapted electrolyte on the cycling performance of electroplated thin film
Sn2Co anodes with regard to: a.) discharge capacity and b.) efficiencies. (z) 1 M LiClO4 in
EC/PC (1:1), () 1 M LiClO4 in EC/PC (1:1) + 20 vol.-% new electrolyte component.
Charge procedure: potentiodynamic charging step: 2.5 - 0.8 V vs. Li/Li+, followed by
constant current charge/discharge cycling at 250 µA cm-2, cut-off: 20-1200 mV vs. Li/Li+.
186
Figure 1 presents results on the beneficial effect of a new electrolyte

solvent on the cycling behavior of the lithium storage alloy Sn2Co. The
influence of the adapted electrolyte may become even more clearly by taking
a look at the cumulated charge losses (Fig. 22). It should be noted that this
new electrolyte component is not able to suppress solvent co-intercalation
into graphite [37], emphasizing the above-mentioned different requirements
on the electrolyte.
As long as lithium storage metal and alloys do not show a
dimensional and structural stability similar to that of graphite, the graphite
anode in a lithium ion cell may not be simply replaced by the lithium storage
metal/alloy without adapting the electrolyte solution. Moreover, the fact, that
graphitic carbons and lithium storage metals and alloys may require different
electrolyte compositions may be an obstacle for the application of
graphite/lithium storage metal composites.
Figure 22. Effect of adapted electrolyte on the cumulative charge loss of thin film Sn2Co
anodes. (z) 1 M LiClO4 in EC/PC (1:1), () 1 M LiClO4 in EC/PC (1:1) + 20 vol.-% new
electrolyte component.
ACKNOWLEDGEMENTS
Support by the Austrian Science Funds through the special research
program "Electroactive Materials" is gratefully acknowledged.
REFERENCES
1. A. N. Dey, Electrochem. Soc. Fall Meeting 1970, Abstr. 62.
2. E. Peled, in: Lithium Batteries (Ed.: J.-P. Gabano), Academic Press, London, 1983, Ch.
3.
187
3. M. Winter, J. O. Besenhard, M. E. Spahr, P. Novak, Adv. Mat., 1998, 10, 725.

4. M. Winter, Diploma Thesis, University of Münster, Germany, 1993.
5. M.Winter, J. O. Besenhard: in: Handbook of Battery Materials (Ed.: J. O. Besenhard),
Wiley-VCH, Weinheim, 1999, 383.
6. M. Winter, K. C. Moeller, J. O. Besenhard (2003)Science and technology of advanced
lithium batteries (Eds: G. A. Nazri, G. Pistoia), Kluwer Academic Publishers, New York,
2003, Ch. 5.
7. J. O. Besenhard, P. Castella, M. W. Wagner, Mater. Sci. Forum, 1992, 91-93, 647.
8. J. O. Besenhard, M. W. Wagner, M. Winter, A. D. Jannakoudakis, P. D. Jannakoudakis,
E. Theodoridou, J. Power Sources, 1993, 43-44, 413.
9. J. O. Besenhard, M. Winter, J. Yang, W. Biberacher, J. Power Sources 1995, 54, 228.
10. O. Chusid, Y. Ein-Ely, D. Aurbach, J. Power Sources, 1993, 43-44, 47.
11. B. Simon, J. P. Boeuve, M. Broussely, J. Power Sources, 1993, 43-44, 65.
12. M. Winter, Doctoral Thesis, University of Münster, Germany, 1995.
13. J. O. Besenhard, M. Winter, J. Yang, International Workshop on Advanced Batteries
(Lithium Batteries), Osaka 1995, 129.
14. G. H. Wrodnigg, J. O. Besenhard, M. Winter, J. Electrochem. Soc., 1999, 146, 470.
15. G. H. Wrodnigg, T. M. Wrodnigg, J. O. Besenhard, M. Winter, Electrochem. Comm.,
1999, 1, 148.
16. G. H. Wrodnigg, C. Reisinger, J. O. Besenhard, M. Winter, ITE Batt. Lett., 1999, 1(1)
111.
17. H. Ota, T. Sato, H. Suzuki, T. Usami, J. Power Sources, 2001, 97-98, 107.
18. H. Gan, E. S. Takeuchi, Eur. Pat. Appl. No. 513428, 2000.
19. G. H. Wrodnigg, J. O. Besenhard, M. Winter J. Power Sources, 2001, 97-98, 592.
20. J. O. Besenhard, K. v. Werner, M. Winter, US Pat. No. 5916708, 1997.
21. W. K. Appel, J. O. Besenhard, S. Pasenok, M. Winter, G. H. Wrodnigg, Ger Pat No. DE
19 724 709, 1998.
22. W. K. Appel, J. O. Besenhard, L. H. Lie, S. Pasenok, M. Winter, Ger. Pat. No. DE 198
58 924, 1998.
23. K.-C. Möller, T. Hodal, W. K. Appel, M. Winter, J. O. Besenhard, J. Power Sources,
2001, 97-98, 595.
24. K.-C. Möller, S. C. Skrabl, M. Winter, J. O. Besenhard, in: Proc. 3rd Hawaii Battery
Conference, Hawaii, USA, 2001, 228.
25. L. H. Lie, T. Hodal, K.-C. Möller, G. H. Wrodnigg, W. K. Appel, J. O. Besenhard,
M. Winter, ITE Batt. Lett., 1999, 1(1), 105.
26. C. Jehoulet, P. Biensan, J. M. Bodet, M. Broussely, C. Moteau, C. Tessier- Lescourret, in
Batteries for Portable Applications and Electric Vehicles (Eds.: C. F. Holmes, A. R.
Landgrebe), The Electrochemical Society, Pennington NJ, 1997, PV97-18, 974.
27. M. Fujimoto, M. Takahashi, K. Nishio (Sanyo) US Pat. No. 5352548, 1994; B. Simon,
J.-P. Boeve (SAFT) US Pat. No. 5626981, 1997; J. Barker, F. Gao (Valence) US Pat. No.
5712059, 1998; Y. Naruse, S. Fujita, A. Omaru (Sony) US Pat. No. 5714281, 1998.
28. K.-C. Möller, H. J. Santner, J. O. Besenhard, M. Winter, in: Proc. 4th Hawaii Battery
Conference, Hawaii, USA, 2002, 238.
29. M. Yoshio, H. Yoshitake, K. Abe, T. Umeno, K. Fukuda, 11th International Meeting on
Lithium Batteries, Monterey CA, 2002, Abstract No. 9.
30. H.J. Santner, K.-C. Möller, J. Ivanþo, M.G. Ramsey, F.P. Netzer, S. Yamaguchi, J. O.
Besenhard, M. Winter, J. Power Sources, 2003, 119-121, 368.
31. K.-C. Möller, H.J. Santner, W. Kern, S. Yamaguchi, J.O. Besenhard, M. Winter, J.
Power Sources, 2003, 119-121, 561.
32. C. Korepp, H. Santner, J. O. Besenhard, M. Winter, K. C. Möller,
J. Anal. Bioanal. Chem., in press.
33. J. P. Olivier, M. Winter, J. Power Sources, 2001, 97-98, 151.
34. D. Bar-Tow, E. Peled, L. Burstein, J. Electrochem. Soc., 1999, 146, 824.
35. A. N. Dey, B. P. Sullivan, J. Electrochem. Soc., 1970, 117, 222.
188
36. M. Winter, G. H. Wrodnigg, J. O. Besenhard, W. Biberacher, P. Novák, J. Electrochem.

Soc., 2000, 147, 2427, and references therein.
37. M. R. Wagner, P. Raimann, K. C. Möller, J. O. Besenhard, M. Winter,
Electrochem. Solid-State Lett., 2004, in press.
38. M. R. Wagner, P. Raimann, H. J. Santner, A, Trifonova, C. Korepp, K.-C. Möller, J. O.
Besenhard, M. Winter, Lithium-Battery-Discussion 2003, September 14-19, 2003,
Bordeaux-Arcachon (France), Abstract No. Inv. 7.
39. H. J. Santner, M. R. Wagner, G. Fauler, P. Raimann, C. Veit, K. C. Möller, J. O.
Besenhard, M. Winter (2003). An Overview on SEI Formation Processes of Lithium
Battery Anodes in Organic Solvent Based Electrolytes, Taipei Power Forum and
Exhibition (TPF2003), December 1-3, 2003, Taipei (Taiwan) Invited lecture.
40. H. J. Santner, M. R. Wagner, G. Fauler, P. Raimann, C. Veit, K. C. Möller, J. O.
Besenhard, M. Winter (2003), Taipei Power Forum and Exhibition (TPF2003),
December 1-3, 2003, Taipei (Taiwan), Proceedings Volume.
41. S. B. Brummer, V. R. Koch, R. D. Rauh, in: Materials for Advanced Batteries (eds.:
D. W. Murphy, J. Broadnead, B. C. H. Steele, Plenum, New York, 1980.
42. M. R. Wagner, J. O. Besenhard, K.-C. Möller, M. Winter, Proceedings of the 7th
International Symposium on Advances in Electrochemical Science and Technology
(ISAEST-VII), November 27-29, Chennai (India), in press.
43. M. Winter, W. K. Appel, B. Evers, T. Hodal, K.-C. Möller, I. Schneider, M. Wachtler,
M. R. Wagner, G. H. Wrodnigg, J. O. Besenhard, Chem. Monthly, 2001, 132, 473.
PERFORMANCE OF NOVEL TYPES OF
CARBONACEOUS MATERIALS
IN THE ANODES OF
CLAiO'S LITHIUM-ION BATTERY SYSTEMS
Mariusz Walkowiak1*, Krzysztof Knofczynski1, Daniel Waszak1,

Maciej Kopczyk1, M. Rusinek1, and Jacek Machnikowski2
1
Central Laboratory of Batteries and Cells, Forteczna 12, 61-362 Poznan, Poland
2
Wroclaw University of Technology, Institute of Chemistry and Technology of Petroleum and
Coal, GdaĔska 7/9, 50-344 Wrocáaw, Poland
Abstract
The electrochemical performance of two anode materials has been

investigated in CLAiO's fully functional lithium-ion battery mockups. Our
findings show that the spherodized purified natural graphite SL20 possesses
sufficient reversible capacity and low irreversible capacity loss, and thus has
been qualified as a suitable candidate material for application in lithium-ion
batteries. Another carbon material investigated in this work was a Boron-
doped carbon (25B2) synthesized by WUT via co-pyrolysis of coal tar pitch
with pyridine borane complex. Its irreversible capacity was seen to be higher
than that of SL20, while reversible capacity is promising. The unique
electron acceptor properties, which are typical in boron-doped graphites did
not manifest themselves in the investigated sample. Ball milling of boron-
doped carbon may be an effective way of improving the WUT precursor and
thus designing a suitable lithium-ion grade anode material.
1. INTRODUCTION
Lithium-ion (shuttle-cock, rocking-chair, swing) battery is widely
considered as the most advanced power source for consumer electronics and
is regarded as the most promising battery technology for a variety of other
applications, such as electric vehicles, medicine and space exploration. One
of the most critical factors in designing successful Li-ion cell is the choice of
*
189
190
proper active anode materials. For many years carbonaceous materials have
been attracting particular attention of lithium-ion battery scientists and
engineers due to the possibility of reversible hosting lithium ions in their
structures [1-3]. The reaction of lithium insertion/deinsertion is the basic
process taking place in the anodes of Li-ion cells. Among carbons tested as
potential anode materials for Li-ion cells were natural and synthetic
graphites, non-graphitised carbons (both soft and hard), carbon fibers, carbon
composites, nanotubes and others. Recently doped carbons (especially
boron-doped) have been considered as promising new anode materials due to
their unique properties. Since boron atoms introduce electron acceptor levels
the materials have been expected to have enhanced capacities.
As the end-user in the NATO SfP project “Carbons as materials for
the electrochemical storage of energy” Central Laboratory of Batteries and
Cells does research and development works on the application of novel
carbonaceous materials to the Li-ion technology. The general idea of these
works is to build prototypes of cylindrical Li-ion cells on the basis of
materials produced in the cooperating laboratories. The aim of this paper is
to examine the applicability of selected commercial and non-commercial
carbon materials (with special attention devoted to boron-doped carbons) to
the construction of a practical cylindrical Li-ion cells.
2. EXPERIMENTAL
Two different natural graphites manufactured by Superior Graphite
Co. (SL-20 and LBG-73) were tested as received on the possibility of using
as anode of a cylindrical Li-ion battery. For comparison, typical synthetic
graphite KS-15 from Lonza was examined.
Innovatory boronated carbons (manufactured in the Institute of
Chemistry and Technology of Petroleum and Coal, Wroclaw University of
Technology, Poland) were obtained by co-pyrolysis of coal-tar pitch with a
pyridine-borane complex. In the first stage of pyrolysis (520ºC) the so-called
semi-coke is obtained. Further carbonization at 2500ºC leads to obtaining
boron-doped carbonaceous material (sample labeled 25B2).
Grinding was performed using Retsch ball-mill. Particle size
distribution was measured using Malvern (Mastersizer) laser particle sizer.
BET surface area measurements were carried out using ASAP
(Micromeritics) surface area analyzer.
Basic electrochemical parameters of the tested materials were
obtained using coin-type half-cells with metallic lithium as the counter
electrode and 1M LiPF6 in EC/DEC (Merck) as electrolyte. The cells were
galvanostatically charged/discharged at the current density 20 mA per gram
of active mass. Full cylindrical Li-ion cells were manufactured in a
laboratory scale using AA-size cases. Anode active masses after mixing with
191
a binder (PVdF, Fluka) were coated in the form of a viscous slurry onto a 20
µm thick copper foil using a hand-operated laboratory coating tool.
Similarly, cathode active mass (manganese spinel, Merck) mixed with a
binder and a conductive additive was coated onto a 20µm thick aluminum
foil. After sandwiching with a 25µm thick polypropylene separator (Celgard)
the electrode set was wound and inserted in the case. Finally the cell was
dozed with electrolyte and closed. The most critical operations were carried
out in a glove box.

Construction of lithium-ion battery is a complex problem involving
many critical stages. Among these the choice of proper active anode material
is one of decisive importance. The ideal candidate for anode should be
characterized not only by excellent electrochemical parameters but also low
cost and good processability.
500
Irreversible capacity
400 Reversible capacity 124

Capacity [mAh/g]
31
26
300
200
337 346
315
100
0
SL-20 LBG-73 KS-15
Figure 1. Comparison of the reversible and irreversible capacities of commercial graphites.
Three commercial graphite materials were evaluated as possible

candidates for anodes of a Li-ion cell: SL-20 and LBG-73 (Superior
Graphite Co.) and KS-15 (Lonza). The first two materials represent a wider
range of graphites specially designed by the SGC researchers to meet the
stringent requirements of Li-ion technologies. Prior to the actual construction
of a cylindrical Li-ion battery all materials were subjected to electrochemical
tests in half-cells in order to select the best candidate for anode material
in terms of reversible and irreversible capacities. These values are also
necessary in the actual battery designing process, which involves the stage of
192
establishing the anode/cathode mass balance. The bar chart on the Figure 1
shows the respective capacities for all the commercial graphites tested.
From the data presented it appears that all the tested materials exhibit
reversible capacities distinctly exceeding 300 mAh/g, which is typical for
graphites. Graphite grades SL-20 (SGC) and KS-15 (Lonza) are close to the
maximum theoretical reversible capacity accessible for graphitic materials
(372 mAh/g). However the irreversible capacity for KS-15 is unacceptably
high (124 mAh/g, which means 26 % of the total electrical charge delivered
to the electrode upon first charging). The corresponding values in the case of
the SGC graphites are 31 mAh/g (8 %) and 26 mAh/g (8 %) for SL-20 and
LBG-73 respectively. Such low values of irreversible losses together with
high reversible capacities make the SGC materials good candidates for the
anodes of practical Li-ion batteries.
The model cylindrical Li-ion battery (AA-size) was manufactured
using SL-20 graphite as anode active material. The general appearance of the
cells is shown by Figure 2; for more detailed description of the cells see the
experimental part of the paper.
Figure 2. The general view of the coin and cylindrical cells.
One important conclusion from the works on the construction of the

cell is that the material (SL-20) can be described as one having excellent
process properties. By this, one should among others understand perfect
adhesion to the copper foil and the ability to form smooth and uniform
layers. The discharge capacities for the first ten cycles are presented in the
Figure 3, together with the discharge profile for the first cycle.
193
Disch. cap. / mAh 550
500
450 4,5 [V]
4,0
400 3,5
3,0
0 200 400 [mAh]
350
0 2 4 6 8 10
Cycle number
Figure 3. Discharge capacities for he first ten cycles for the Li-ion cell SL-20 anode. The
inset: discharge curve for the first cycle.
25B2
25B2m2
25B2m6
Figure 4. Particle size distributions for the original and ground B-doped carbon from WUT.
194
The capacities obtained initially exceed 500 mAh (which is ca. 70 %

of those exhibited by commercial products), but decline in the subsequent
cycles to reach a fairly stable value of ca. 450 mAh in the tenth cycle.
Boron-containing carbons synthesized by co-pyrolysis of coal-tar
pitch with pyridine-borane complex (series 25Bn) have already been
considered as hosts for lithium insertion [4]. Unlike the commercial
graphites described above, the boron-doped carbon 25B2 (WUT) as received
was not suitable for direct use in the cylindrical cell due to very large and
hard particles. This feature makes the coating process very difficult.
To improve the processing properties, the material was ground in a
ball-mill for two hours (sample designated as 25B2m2) and separately for 6
hours (sample designated as 25B2m6). In order to check the effect of milling
the particle distribution of all the samples was determined (see Fig. 4).
The curves give information about the distribution (in weight
percentage terms) of particle sizes in a given sample. While the majority of
the original material (25B2) concentrates around 500 µm (see the peak of the
25B2 curve, note that the particle size axis is a logarithmic one), the particles
of the milled samples are markedly reduced in size (the peaks are located
around 150 µm and 80 µm for 25B2m2 and 25B2m6 respectively).
Prolonged grinding makes the WUT material applicable for the Li-ion
technology. Together with the reduction of particle sizes an increase in the
BET surface areas is observed (0.4 m2/g, 5.8 m2/g and 8.5 m2/g for 25B2,
25B2m2 and 25B2m6 respectively). What is interesting, the XRD pattern
analysis proves that the crystal structure of the material was not affected by
grinding.
500
Reversible capacity
Irreversible capacity
400
94 82
75
Capacity [mAh/g]
300
200
331 324 337
100
0
25B2 25B2m2 25B2m6
Figure 5. Comparison of the reversible and irreversible capacities of the original

and ground B-doped carbon from WUT.
195
Electrochemical tests in half-cells allow the preliminary assessment

of the WUT carbon as well as of the impact of grinding on the
electrochemical performance. The data from the chart on the Figure 5
indicate that the material has high reversible capacity (similar to the
capacities of the commercial graphites described earlier).
This capacity is practically not affected by grinding. This finding is
in contradiction to many literature reports stating that excessive ball milling
can have negative impact on the reversible capacities due to introduction of
turbostratic disorder. Also irreversible capacities do not change significantly
upon grinding. However the ratios of irreversible charge consumptions are
higher than in the case of SGC graphites (e.g. 20% for 25B2m6). To sum up,
ball milling of the original B-doped carbon (WUT) creates material with
satisfactory electrochemical parameters and having greatly improved
processing properties. Figure 6 shows the discharge capacities for the first
ten cycles for the AA-size Li-ion battery with the 25B2m6 anode. The insert
on this picture shows the first discharge profile for this cell.
350
Disch. cap. / mAh
300
4,5 [V]
4,0
250 3,5
3,0
2,5
0 100 200 300 [mAh]
200
0 2 4 6 8 10
Cycle number
Figure 6. Discharge capacities for he first ten cycles for the Li-ion cell SL-20 anode.
The insert: discharge curve for the first cycle.
4. CONCLUSIONS
Commercial and non-commercial carbons were tested for their

applicability as anode of lithium-ion battery. It was found that Superior
Graphite Co’s materials are characterized both by high reversible capacities
and low irreversible capacities and thus can be regarded as good candidates
for practical full cells. Cylindrical AA-size Li-ion cells manufactured using
laboratory techniques on the basis of SL-20 anode had initial capacities over
500 mAh (volumetric energy density ca. 240 Wh/dm3). Boron-doped carbon
196
25B2 (WUT) exhibits high reversible capacity but its irreversible capacity is
slightly greater than that of SGC graphites. Moreover, very large particles
hinder direct application of this material. It was found that ball milling could
be considered as an effective method of improving the usability of this
carbon as anode of Li-batteries.
ACKNOWLEDGEMENTS
The authors would like to thank Superior Graphite Co. for delivering
graphite samples. Special thanks for NATO SfP “Carbon” 973849 project
consortium members for valuable material and intellectual assistance, in
particular to Prof. François Beguin, Prof. ElĪbieta Frąckowiak, Prof. Jacek
Machnikowski, Prof. Andrzej Lewandowski and Prof. Vyacheslav Barsukov.
REFERENCES
1. S. Flandrois, B. Simon, Carbon, 37 (1999) 165-180.
2. M. Endo, C. Kim, K. Nishimura, T. Fujino, K. Miyashita, Carbon, 38 (2000) 183-197.
3. F. Cao, I. V. Barsukov, H. J. Bang, P. Zaleski, J. Prakash, J. Electrochem. Soc., 147
(2000) 3579-3583.
4. J. Machnikowski, E. Frąckowiak, K. Kierzek, D. Waszak, R. Benoit, F. Beguin, J. Phys.
Chem. Solids, 65 (2004)
WHY GRAPHITE ELECTRODES FAIL
IN PC SOLUTIONS: AN INSIGHT FROM
MORPHOLOGICAL STUDIES
Doron Aurbach*, Maxim Koltypin, Hnnan Teller and Yaron S. Cohen

Department of Chemistry, Bar-Ilan University, Ramat-Gan 52900, Israel
Abstract
This paper deals with capacity fading mechanisms of graphite

electrodes. Three types of graphite particles were used: synthetic flakes,
natural flakes and mesocarbon microbeads (MCMB). We used a probe
solution, EC-PC/LiClO4, in which electrodes comprising different types of
graphite particles behave very differently from each other. The tools for this
study included in-situ AFM imaging, scanning electron microscopy, FTIR
spectroscopy, XRD and standard electrochemical techniques. The
morphology of the edge planes of the graphite particles plays an important
role in their ability to develop passivating surface films when graphite
electrodes are polarized cathodically in the electrolyte solutions. Another
critical factor that determines the passivation of the graphite electrodes is the
cohesion and adhesion of the solution reduction products. A major failure
mechanism of graphite particles in PC solutions seems to be cracking of the
particles because of reactions in crevices, which lead to a build-up of internal
pressure. Such cracking processes can be clearly observed by AFM imaging.
The unexpected dependence of the irreversible capacity of some types of
graphite electrodes on the particle size (increases with larger particle size) is
also evident for the above failure mechanism. Factors that determine
stabilization and failure of graphite electrodes are demonstrated and
discussed.
Keywords
Graphite electrodes, ethylene carbonate, propylene carbonate,

capacity fading, surface films, morphology, AFM, passivation.
*
197
198
1. INTRODUCTION
There is no question that the development and commercialization of
lithium ion batteries in recent years is one of the most important successes of
modern electrochemistry. Recent commercial systems for power sources
show high energy density, improved rate capabilities and extended cycle life.
The major components in most of the commercial Li-ion batteries are
graphite electrodes, LiCoO2 cathodes and electrolyte solutions based on
mixtures of alkyl carbonate solvents, and LiPF6 as the salt.1 The electrodes
for these batteries always have a composite structure that includes a metallic
current collector (usually copper or aluminum foil/grid for the anode and
cathode, respectively), the active mass comprises micrometric size particles
and a polymeric binder.
Graphite intercalates electrochemically and reversibly with lithium
via phase transitions in four stages up to a stochiometry of LiC6
(corresponding to the capacity of 372 mAh/gr).2 Since Li insertion into
graphite electrodes occurs at very low potentials, very close to that of Li
deposition (between 0.3 and 0.01 V vs. Li/Li+), all the relevant nonaqueous
solvents and salts, in which electrochemical Li intercalation into graphite can
proceed, are reduced on the cathodically polarized graphite electrodes at
potentials that can be as high as 1.5 V above lithium insertion potentials.3
Most of these reduction processes precipitate surface films on the graphite
particles, which are comprised mostly of organic and inorganic Li salts
(similar to the Li electrodes that are covered by surface films in all
nonaqueous solutions in which lithium is apparently stable).4 These surface
films formed on graphite electrodes, which are usually ion conducting and
electronically insulating, passivate these electrodes, and were widely
explored in terms of structure, properties, etc.5-10 It appears that their
properties are critical to the stability and reversibility of graphite electrodes
in repeated Li insertion/deinsertion cycling. Hence, the behavior of lithiated
graphite electrodes depends strongly on the composition of electrolyte.
The failure mechanisms of lithiated graphite electrodes in commonly
used Li salt solutions relevant to the field of Li ion batteries are still in the
focus of study by a number of research groups.11,12 A common opinion is that
in cases where the passivation by the surface films is not sufficient, solvent
molecules may cointercalate with Li ions. Consequently, the fragile lattice
structure of graphite may be destroyed by exfoliation of the graphene planes
that simply fall apart from each other, thus causing an amorphization of the
graphite particles.11-13 Another possibility raised was that solvent molecules
cointercalate with the Li ions even at potentials as high as 1 V vs. Li/Li+, and
are reduced within the graphite, close to the surface of the particles, thus
blocking further Li-ion insertion into the lattice.14
There is an ongoing interest in the study of failure mechanisms of
graphite electrodes. It is very important to understand the source of the
199
irreversible capacity of graphite electrodes15,16 (related to the irreversible

reduction of solution species) and how to reduce it.17,18 In these studies,
issues such as morphological aspects, crystal structure (phase composition),
surface area and solution composition are considered, and their possible
impacts on the irreversible behavior of graphite electrodes are explored.15-20
Figure 1 provides several electrochemical windows of important,
relevant processes, including the reduction of alkyl carbonates, ethers, Li
insertion into graphite, and Li metal deposition. Recent studies revealed two
major failure mechanisms of graphite electrodes in repeated Li insertion/
deinsertion processes:21
Figure 1. Schematic representation of the electrochemical windows of the processes,

occurring during cathodic polarization of graphite electrodes in nonaqueous solutions.
1) When the electrolyte solutions are not too reactive, as in the case of
ethereal solutions, there is no massive formation of protective
surface films at potentials above Li intercalation potential, and most
of the solvent reduction processes may occur at potentials lower than
0.3 V vs. Li/Li+. Hence, the passivation of the electrodes is not
sufficient to prevent cointercalation of solvent molecules. This leads
to an exfoliation of the graphite particles into amorphous dust
(expholiated graphene planes). This scenario is demonstrated in
Figure 2a as the reduction of the 002 diffraction peak21 of the
graphite electrode, polarized cathodically in an ethereal solution.
2) In the case of solutions based on a solvent such as propylene
carbonate (PC), the failure of graphite electrodes is attributed by
some researches to the exfoliation of the graphite particles due to
cointercalation of PC molecules with the Li ions.14,22 The difference
between ethylene carbonate (EC) and PC in this respect may,
according to this approach, be attributed to the higher ability of PC
molecules to solvate Li ions.23 Hence, cointercalation of PC
molecules takes place because their desolvation from Li ions, which
migrate from solution phase to the intercalation sites in the graphite,
200
is more difficult than in the case of EC molecules.14,22,23 As

demonstrated in Figure 2b, the failure of graphite electrodes in PC
solutions does not necessarily involve a complete exfoliation of the
graphite particles and destruction of their 3D structure.21 Cumulative
studies, based on scanning electron microscopy (SEM), X-ray
diffraction (XRD) and standard electrochemical techniques,
converged to the conclusion that the failure mechanism of graphite
electrodes in PC solutions in more likely due to cracking of the
particles and their electrical isolation by surface films rather than
due to the complete exfoliation (on the nanometer scale).24,25
Figure 2. XRD patterns measured ex-situ from graphite particles and chronopotentiograms
of graphite electrodes polarized galvanostatically from OCV to low potentials in
CH3-OCH2CH2-OCH2CH2-OCH3 (diglyme)/LiClO4 solution (a) and in PC/LiAsF6 solution
(b). The letters near the V vs. capacity curves correspond to the relevant XRD patterns,
which have been measured after different periods of galvanostatic polarizations. Notice that
in the diglyme solution (a) the XRD patterns clearly indicate progressive destruction of the
active mass, while in the PC solutions (b) the XRD patterns show that
the active mass remains pure graphite.
This article discusses the origin of the failure mechanisms of

graphite electrodes in PC solutions, focusing on the electrodes’ irreversible
capacity, mainly from a morphological point of view. The EC-PC/LiClO4
electrolyte solution was chosen as the major probe solution, because it was
found that graphite electrodes of different types and morphologies behave
very differently in these solutions in terms of the irreversible capacity
measured during a first cathodic polarization of the electrodes and the
potentials at which the major reduction of solution species takes place, which
are lower by more than 0.5 V compared with those in the EC-DMC/LiAsF6
solution. Nevertheless, after stabilization, the electrodes behave reversibly in
this probe solution, which means that after the first few cycles, the behavior
of these systems stabilizes.25
201
2. EXPERIMENTAL DESCRIPTION
The experimental tools for this research were chronopotetiometry
(galvanostatic cycling),25 atomic force microscopy (AFM),26,27 scanning
electron microscopy (SEM), and X-ray diffraction (XRD).21,25 It should be
mentioned that the AFM imaging was conducted in-situ under potential
control and in a special homemade glove box filled with highly pure argon
atmosphere. This system has been already described in detail in the
literature.28
This article concentrates on three types of graphite particles, which
differ in their morphology, based on a rigorous examination of SEM
micrographs:25
1) Synthetic graphite flakes, obtained from Timrex Inc., whose

morphology has been characterized by a high level of crevices in the
facets perpendicular to the basal planes, through which lithium ions
are inserted into the graphite lattice (edge planes).
2) Mesocarbon microbeads (MCMB) particles, obtained from Osaka
Gas Co., which are round graphite particles. They are also far from
having a smooth morphology and have crevices on their surface.
3) Natural graphite (NG) particles, which have smoother facets and a
lesser amount of crevices in their edge planes.
3.1. On the Proposed Scenarios of Stabilization and Failure

of Graphite Electrodes
The surface reactions of graphite electrodes in many nonaqueous
solutions have been investigated intensively,29,30 and the major reaction paths
in a variety of alkyl carbonate solutions seem to be quite clear. Both EC and
PC decompose on graphite electrodes, polarized cathodically, to form solid
surface films with ROCO2Li species as major components,31 and ethylene or
propylene gases, respectively, as co-products.
Figure 3 illustrates schematically the processes of graphite
electrodes in standard solutions (such as EC-DMC/LiAsF6). During the first
cathodic polarization, highly passivating surface films are formed, which
stop the surface reactions. Since the reduction of EC is quick enough and
forms sufficiently passivating films, they protect the fragile graphite
structure from cointercalation of solvent molecules and exfoliation. As Li is
inserted, there is a volume expansion of the graphite, and hence, the surface
films on the edge planes (through which the Li insertion takes place) are
202
stretched. As Li is de-intercalated, the graphite’s volume contracts back to its

initial size.
Formation of
passivating
Surface reactions of surface films
EC, PF6- and DMC Li intercalation
Li+, e-
Li+
Li Li+
+
Li+ Li+
Li deinsertion
Graphite flakes in EC-DMC solutions

Figure 3. An illustration of morphology, surface processes and changes during a Li insertion-
deinsertion cycle of an electrode comprising synthetic graphite flakes in EC-DMC solutions,
in which the electrodes behave reversibly.
The methyl group in

the ROCO2Li species
interferes with a good
cohesion
CH=CH22
CH33CH=CH
Li22CO33
e- PC CH3CH=CH22
e- PC Li+
+ ROCO22Li
CH3CH=CH22
-
e PC Li
CH3CH=CH22
Li+ etc.
Worse passivation!
Gas bubbles are
formed insides the
Graphite particles
PC gas
particles are gas
splitted due to PC
gas pressure More reactive
surface exposed,
further surface
Surface films reactions
Electrochemically
isolated graphite
Current collector flakes
Figure 4. A scheme of the proposed mechanism of surface reactions of PC on graphite flakes,
which contain crevices.
203
The passivation of the graphite particles depends on the

accommodation of the surface films to these volume changes, and to the
extent that their stability and integrity are maintained (when the graphite is
fully lithiated). It is assumed that upon the repeated volume changes of the
graphite particles during the Li insertion-deinsertion cycling, passivation by
the surface films on the edge planes may be damaged, and thus reactive Li-
graphite compounds are exposed to the solution, thus allowing a further
small-scale reduction of solution species.32 These minor surface reactions
thicken the surface films during prolonged, repeated cycling, and hence,
explain the increase of impedance of these electrodes observed during
prolonged charge-discharge cycling.33
Figure 4 presents schematically the surface reactions in PC based
solutions, especially when the particles’ morphology is extremely rough. At
potentials below 1.5 V, PC is reduced on any metallic or carbon electrode, in
the presence of Li ion.34 The major products are propylene gas and
CH3CH(OCO2Li)CH2OCO2Li.35 The former solid product is insoluble in PC,
and hence precipitates on the active mass of the electrode as a surface film.
When the carbon particles have deep crevices in which the surface reactions
take place, pressure may be developed inside the crevices by gaseous
products such as propylene, as the solid products that precipitate may block
the outward gas diffusion. The developed pressure inside the crevices may
then mechanically crack the particles very easily, thus exposing more
reactive surfaces of the particles to the solution and allowing further surface
reactions to take place. In the following sections we present experimental
evidences that support the above failure mechanism.
3.2. Experimental Evidence:

Irreversible Capacity vs. Surface Area and Particle Size
Figure 5 shows the dependence of the irreversible capacity on the
size and the specific surface area of synthetic graphite particles in EC-PC
solutions compared to EC-DMC solutions. It is expected that as the average
size of the graphite particles decreases, and their surface area is higher, the
irreversible capacity should also be higher, as was indeed reported
previously for EC-DMC solutions.10,11,14,31 Nevertheless, in the EC-PC
solutions used herein, as the particle size is higher, the opposite trend is
observed and the irreversible capacity is higher as well. Hence, surprisingly,
the irreversible capacity is lower in these systems as the surface area of the
particles is higher.
These results are in contrast to the approach that assumes that as the
graphite flakes are bigger, the average energy needed to separate the
graphene layers by solvated ions is higher, and thus the level of exfoliation
may be lower. However, these results are in agreement with the mechanism
204
presented above in which graphite particles which contain deep crevices and
have rough morphology are cleaved because of development of internal
pressure in the crevices. A fresh surface of active mass is then exposed to the
solution species, thus allowing further surface reactions, and hence,
increasing irreversible capacity. As the particles are larger, the freshly
exposed surface of active mass due to cracking (along the basal planes) is
larger (compared to the surface area of the original facets), and thus the
irreversible capacity is higher. Thereby, the irreversible capacity behaves
inversely with the specific surface area of the particles (the specific surface
area is higher as the particles are smaller).
Synthetic graphite flakes

(a) Irreversible capacity vs. particle size
EC-PC
EC-DMC
(b) Irreversible capacity vs. specific surface area
EC-PC
EC-DMC
Figure 5.Irreversible capacity vs. particle size (a) and vs. specific surface area (b) of graphite
electrodes comprising of synthetic flakes in EC-PC and EC-DMC solutions, as indicated.
205
In the case of EC-DMC solutions, since the surface species are

deposited quickly and form very compact passivating films, the passivation
of the active mass is obtained before products such as ethylene gas have the
chance to be accumulated in crevices and an internal pressure to grow.
Indeed, in these solutions the irreversible capacity depends inversely on the
size of the particles (as expected).
3.3. Experimental Evidence:

Irreversible Capacity vs. Discharge Rate
It was found that in the case of EC-PC solutions the irreversible
capacity depends on the discharge rate of the graphite electrodes. As the
discharge rate is higher, the irreversible capacity is higher, as presented in
Figure 6. This result reflects changes in the balance between the kinetics of
particles cracking and the development of a proper passivating films, at
different polarization rates: as the discharge rate is higher, the surface
reactions occur more intensively and hence products such as propylene gas
accumulate more quickly, while the solid products precipitate fast as well
and block its way out (thereby, a fast build-up of internal pressure cracks the
graphite particles). As a consequence, fresh surfaces of active mass are
exposed to the solution, and hence, surface reactions occur intensively.
2.5
C/15
2 C/40
C/53
Voltage, V
1.5
0.5
0
0 500 1000 1500 2000 2500
Capacity, mAh/g
Figure 6. Potential vs. capacity curves obtained from cycling tests of synthetic graphite flakes
in EC-PC/LiClO4 solutions in different discharge rates. Notice that as the discharge rate
decreases – the irreversible capacity decreases accordingly.
206
As the discharge rate is low enough, gaseous products are formed

slowly, and have the chance to escape from the crevices before internal
pressure builds up and cracks the particles.
3.4. Experimental evidence:

AFM Imaging and Surface Chemistry
The main issue of this paper relates to the morphological aspects of
the graphite particles. As described above, the three types of graphite
particles, which are the subject of this study, differ in their pristine
morphology, mainly in the roughness and the level of crevices in their edge
planes (perpendicular to the basal planes).25
Figure 7 shows typical images obtained by in situ AFM
measurements, which were conducted with electrodes comprising different
types of graphite particles.26 These measurements nicely demonstrate
cracking of the graphite particles upon cathodic polarization of the
electrodes, as was described above. This figure presents images of synthetic
graphite, MCMB and natural graphite (NG) particles in an EC-PC solution at
OCV, before polarization and deposition of surface films, and after the
formation of surface films, during the first cathodic polarization of the
electrodes.
Images “a” and “b” in Figure 7 belong to synthetic graphite flakes in
the probe solution. As the electrodes are polarized from OCV (3V vs. Li/Li+,
Figure 7a.) to potentials below 1V, pronounced morphological changes are
observed: the surfaces of the particles at their boundaries are increasingly
lifted upwards. This is clearly seen when comparing images 7a and 7b while
focusing on a gap between two particles. It is clear that the electrodes retain
their integrity, and after the irreversible morphological changes observed by
the AFM imaging, they can function electrochemically and reversibly in
repeated Li insertion-deinsertion processes.
Pictures 7c and d are AFM images measured in situ during a first
galvanostatic polarization of an electrode comprising MCMB (round)
particles. The images obtained at OCV (not seen in the figure) reflect
initially a rough surface. Image 7c relates to the carbon surface covered by
surface films, whose morphology differs considerably from that of the
pristine material. The surface films cover the entire surface and, in fact, hide
the initial rough structure. As polarization proceeds, the imaging shows the
clear formation of cracks that deepen and widen during polarization, during
the irreversible surface processes (corresponding to the potential plateau
around 0.8 V vs. Li/Li+ that is seen in the galvanostatic curves). Upon further
polarization, the cracks are filled by solution reduction products, which are
formed within the cracks. Imaging electrodes comprised of MCMB particles
provide a clearer insight into the failure mechanisms, compared to imaging
207
of synthetic graphite electrodes, because MCMB particles have many

upward-faced edge planes in the composite electrodes, on which the particles
are cracked. Hence, cracking can be easily imaged during the cathodic
polarization of the electrodes while surface reactions occur.34 In the case of
the synthetic graphite flakes, mostly the basal planes (the wider dimension of
the particles) face upward, and hence cracking can be followed only
indirectly, appearing as a lifting up of the particles’ edges.
LiClO4/EC-PC
493.9 nm
922.8 nm
0 nm
5 µm
0 nm
5 µm
5 µm 2.5 µm
2.5 µm 5 µm
2.5 µm
(A) KS6, OCV 0 µm 2.5 µm

(B) 0.531 V
0 µm
42.7 nm 425.6 nm
0 nm
5 µm
5 µm 2.5 µm 0 nm
5 µm
2.5 µm
5 µm 2.5 µm
0 µm
(C) MCMB, 0.792 V 2.5 µm
(D) 0.777 V 0 µm
481 nm
0 nm 816 nm
5 µm
0 nm
5 µm
5 µm 2.5 µm
2.5 µm 5 µm 2.5 µm
(E) NG 15, OCV 0 µm

2.5 µm
(F) 0.592 V 0 µm
Figure 7. In-situ AFM imaging of synthetic graphite flakes (a, b), MCMB particles (c, d) and
natural graphite particles (e, f) during the first cathodic polarization of the electrodes in the
+
probe solution (LiClO4/EC-PC), measured at the indicated potentials vs. Li/Li . The arrows
and circles point to the relevant morphological processes, as detailed in the text (see ref. 26).
208
Finally, Figures 7e, 7f show AFM images of an electrode comprised

of natural graphite flakes, measured in situ in the probe solution (EC-
PC/LiClO4) during the first galvanostatic polarization from OCV. The
images reflect a relatively stable morphology with only minor changes,
which should be attributed to the formation of surface films on top of the
particles. While imaging natural graphite electrodes, one could hardly
observe cracking of particles, as was observed with the synthetic flakes and
the MCMB electrodes. It is suggested that electrodes comprising natural
graphite flakes behave differently in PC-containing solutions because their
edge planes are much smoother and contain fewer crevices, as compared to
synthetic flakes or MCMB particles.
The possible detrimental build-up of pressure inside the graphite
particles should depend critically on the quality of the electrode’s
passivation.
In cases where the solid reduction products of the solvent molecules
form highly cohesive and adhesive surface films, the surface reactions are
quickly blocked before further massive reduction of solution species (which
also form the gas molecules) takes place. When passivation of the graphite is
not reached quickly enough (as in the case of PC solutions), intensive
surface reactions build up the internal pressure that cracks the particles and
leads to their deactivation.
a. O Li
+ +
O
d.
O C C O e. O Li+ O R
+ + +
+
Li O Li O Li O C C O
++
Li
++ Li R O Li+ O
Li 2CO 3 ROCO2Li
b.
f. +
Li+
O
O O
LLii
O Li O
++
O O C
HH 444 LLii +++
O
O CC 222
c. O C
O C O
C O
C O
O
LLii
H 44
++
C
C H
CC22 + LLii ++
O
O O
O O
O Li+
Li O
O
(CH2OCO2 Li)2 (EC reduction product)
Figure 8. Typical FTIR spectra measured from graphite electrodes cycled in EC solutions at
25 ºC (a, b, c) taken from ref. 21 and the relevant structures to which these spectra belong
according to ref. 36 (d, e, f).
a. A graphite electrode after a first Li insertion-deinsertion in EC - LiAsF6 1 M.
b. Same as a, EC-DEC 3:1, 1 M LiAsF6 solution, 6 atm CO2.
c. FTIR spectrum of (CH2OCO2Li)2 in KBr pellet.
d. The expected structure of Li2CO3 films.
e. The expected structure of Li2CO3 films.
f. The expected structure of (CH2OCO2Li)2 films.
209
Figure 8 shows typical FTIR spectra measured from graphite

electrodes cycled in different EC-based electrolyte solutions. The basic
features belong to species such as (CH2OCO2Li)2, as previously discussed in
detail.21 However, as was found by rigorous calculations,36 these spectra
belong to dimmers of these species, as also demonstrated in Figure 8.
The adhesion of the ROCO2Li species, formed by reduction of the
alkyl carbonate solvents, to the carbon surfaces, as well as the cohesion of
the ROCO2Li species, goes through the carbonate groups and the Li ions that
span the negatively charged oxygen atoms36 and the negatively charged
carbon atoms (see Figure 8). EC and PC reduction forms (CH2OCO2Li)2, and
CH3CH(OCO2Li)CH2OCO2Li, respectively, which differ from each other in
the methyl group. It is suggested that the (CH2OCO2Li)2 species formed by
EC reduction precipitate as adhesive and cohesive surface films more
quickly and efficiently than PC reduction products, which have an extra
methyl group. It is assumed that such a methyl group should interfere badly
with both adhesion and cohesion of the Li dicarbonate species, due to steric
hindrance (which does not exist with (CH2OCO2Li)2).
From the above discussion, it is clear that the stabilization or failure
of graphite electrodes depends on a delicate balance between passivation
phenomena (due to the formation of highly cohesive and adhesive surface
films) and a buildup of internal pressure due to the reduction of solution
species inside crevices in the graphite particles. This delicate balance can be
attenuated by both solution composition (EC-DMC vs. EC-PC or PC, etc.)
and the morphology of the graphite particles (i.e. the structure of the edge
planes and the presence of crevices).
4. SUMMARY
Understanding the failure mechanisms of graphite electrodes in Li
insertion-deinsertion processes is an important issue in the field of Li
batteries. In this respect, the failure of Li-graphite electrodes in PC solutions
poses an interesting question, in light of the fact that in EC-based solutions
lithiated graphite electrodes behave highly reversibly, even though EC and
PC are so similar in their structure and properties.
In addition to the known cointercalation and exfoliation mechanism
which provides a good explanation of the failure of graphite electrodes in
ethereal solutions, this article suggests a mechanism for the failure of these
electrodes in PC solutions, which is based on cracking of the particles, due to
the build-up of internal pressure, that can be caused for instance, by the
accumulation of gas inside the crevices in the edge planes of the particles
during cathodic polarization of the electrodes. This mechanism is supported
by electrochemical and morphological results. This paper compares three
types of graphite electrodes; synthetic flakes, natural flakes and MCMB,
which show different behavior in EC-PC/LiClO4 (used as a probe electrolyte
210
solution) in terms of irreversible capacity and stability. This difference is

connected to the different morphologies of the graphite particles.
The different behavior of graphite electrodes in EC and PC solutions
can be explained in light of their surface chemistry and the fact that in
addition to surface species of the ROCO2Li (Li alkylene dicarbonates) type,
reduction of both EC and PC form alkylene gas. When EC reduction
dominates the surface chemistry, the (CH2OCO2Li)2 thus formed can
precipitate with good adhesion and cohesion to form highly passivating
surface films. When the surface chemistry is dominated by PC reduction, the
surface species that contain the methyl group are less cohesive and adhesive;
thus, formation of passivating surface films requires an intensive reduction
of solution species, and is accompanied by a massive propylene gas
formation. Reduction of PC in crevices in the absence of a rapid formation of
passivating surface films, may lead to an internal pressure buildup, which
leads to the cracking and splitting of the particles and to an increase of the
surface area, and hence, massive surface reactions take place and form
surface films that electrically isolate most of the active mass. In cases of EC-
PC mixtures, where EC reduction products also contribute to the formation
of passivating surface films, electrodes comprised of natural graphite flakes
whose edge planes are relatively “smooth” can behave reversibly with
relatively low irreversible capacity loss. However, when the electrodes are
comprised of synthetic flakes or MCMB whose edge planes have a high
concentration of crevices, the passivation in an EC-PC based solution is not
sufficient to avoid the buildup of internal pressure, which cracks the
particles. Nevertheless, cracking of these particles in EC-PC based solutions
is not massive enough to interfere adversely with the electrode integrity, nor
to changes in the particles’ orientation that lead to an electrical isolation of
the active mass, as happens in pure PC solutions. Therefore, the electrodes
can reach passivation in EC-PC solutions after pronounced consumption of
irreversible charge and then lithium can be inserted and deinserted
reversibly.
ACKNOWLEDGMENT
Partial support for this work was obtained by the BSF, Israel-USA
binational foundation.
REFERENCES
1. Li-ion Batteries, Fundamentals and Performance; Wakihara, M.; Yamamoto, O., Eds.;
Wiley-VCH: Weinheim (1998).
2. Dahn, J. R.; Sleight, A. K.; Shi, H.; Reimers, J. N.; Zhong, Q.; Way, B. M. Electrochim.
Acta, 38, 1179 (1993).
211
3. Aurbach, D.; Gofer, Y. in Nonaqueous Electrochemistry, D. Aurbach, Editor, Chap. 4,

Marcel Decker, Inc., New York (1999).
4. Peled, E. in Lithium Batteries, J. P. Gabano, Editor, Chap. 3, Academic Press, Inc., New
York (1983).
5. Aurbach, D.; Youngman Chusid, O.; Carmeli, Y.; Babai, M.; Y. Ein-Eli J.
PowerSources, 43, 47 (1993).
6. Aurbach, D.; Ein-Eli, Y.; Markovsky, B.; Carmeli, Y.; Yamin, H.; Luski, S.
Electrochem. Acta, 39, 2559 (1994).
7. Fong, R.; von Sacken, U.; Dahn, J. R. J. Electrochem. Soc., 137, 2009 (1990).
8. Naji, A.; Willmann, P.; Billaud, D. Carbon, 36, 1347 (1998).
9. Yang, C. R.; Wang, Y. Y.; Wan, C. C. J. Power Sources, 72, 66 (1998).
10. Winter, M.; Nova´k, P.; Monnier, A. J. Electrochem. Soc., 145, 428 (1998).
11. Winter, M.; Besenhard, J. O. in Handbook of Battery Materials, J. O. Besenhard, Editor,
Part III, Chap. 5, Wiley-VCH Verlag GmbH, Germany (1999).
12. Rosoken, J. M.; Decker, F. J. Electrochem. Soc., 143, 2417 (1996).
13. Tran, T. D.; Feikert, J. H.; Pekala, R. W.; Kinoshita, K. J. Appl. Electrochem., 26, 1161
(1996).
14. Winter, M.; Besenhard, J. O.; Spahr, M. E.; Nova´k, P. Adv. Mater., 10, 725 (1998).
15. Dahn, C. H.; Kim, H. S.; Moon, S. I. J. Power Sources, 101, 96 (2001).
16. Lampe-Onnerud, C.; Shi, J.; Onnerud, P.; Chamberlain, R.; Barnett, B. J. PowerSources,
97-98, 133 (2001).
17. Oliver, J.; Winger, M. J. Power Sources, 97-98, 151 (2001).
18. Joho, F.; Rykart, B.; Blome, A.; Novak, P.; Hilkelm, H.; Spahr, M. E. J. PowerSources,
97-98, 78 (2001).
19. Zanem, D.; Antonni, A.; Pasquali, M. J. Power Sources, 97-98, 146 (2001).
20. Zaghib, K.; Nadeau, G.; Kinoshita, K. J. Power Sources, 97-98, 97 (2001).
21. Aurbach, D.; Markovsky, B.; Weissman, I.; Levi, E.; Ein-Eli, Y. Electrochim. Acta, 45,
67 (1999).
22. Chung, G. C.; Kim, H. J.; Jun, S. H.; Choi, J. W.; Kim, M. H. J.Electrochem. Soc., 147,
4398 (2000).
23. Wang, Y.; Balbuena, P. B. J. Phys. Chem., A105, 9972 (2001).
24. Aurbach, D.; Levi, M. D.; Levi, E.; Schechter, A. J. Phys. Chem. B, 101, 2195 (1997).
25. Aurbach, D.; Teller, H.; Levi, E. J. Electrochem. Soc., 149, A1255 (2002).
26. Aurbach, D.; Koltypin, M.; Teller, H. Langmuir, 18, 9000 (2002).
27. Aurbach, D.; Teller, H.; Koltypin M.; Levi, E. J. PowerSources, 119-121, 2 (2003).
28. Cohen, Y.; Aurbach, D. Rev. Sci. Instrum., 70, 4668 (1999).
29. Aurbach, D.; Markovsky, B.; Levi, M. D.; Levi, E.; Schechter, A.; Moshkovich, M.;
Cohen, Y. J. Power Sources, 81-82, 95 (1999).
30. Peled, E.; Golodnitsky, D.; Pencier, J. in Handbook of Battery Materials, Besenhard, J.
O. Editor, Part III, Chap. 6, Wiley-VCH Verlag GmbH, Germany (1999).
31. Aurbach, D.; Markovsky, B.; Schechter, A.; Ein-Eli Y.; Cohen, H. J. Electrochem. Soc.,
143, 3809 (1996).
32. Koltypin, M.; Cohen, Y.S.; Cohen, Y.; Markovsky, B.; Aurbach, D. Electrochem.
Commun., 4, 17 (2002).
33. Mortinent, A.; Le Gorrec, B.; Montella, C.; Yazami, R. J. PowerSources, 97-98, 83
(2001).
34. Aurbach, D.; Moshkovich, M.; Gofer, Y. J. Electrochem. Soc.,148, E155 (2001).
35. Aurbach, D.; Gottlieb, H. Electrochim. Acta, 34, 141 (1989).
36. Matsuta, S.; Asada, T.; Kitaura., K. J. Electrochem. Soc. 147, 1695-1702 (2000).
NEW DEVELOPMENTS IN THE ADVANCED
GRAPHITE FOR LITHIUM-ION BATTERIES
Francois-Xavier Henry, Igor V. Barsukov, Joseph E. Doninger,

Scott Anderson, Peter R. Booth*, Peter L. Zaleski, Richard J. Girkant,
David J. Derwin, Maritza A. Gallego, Tomás Huerta, and Gabriela Uribe
Superior Graphite, Peter R Carney Technology Center
4201 W. 36th Street, Chicago, IL 60632, USA, www.SuperiorGraphite.com
Abstract
In this paper, three generations of natural graphite are described in

order to show the evolution in understanding of which physical and
electrochemical requirements are needed to make usable graphite for
practical lithium-ion batteries. Of the three alternative types of carbonaceous
materials discussed in the paper, the optimum has been found to be natural
graphite. Its behavior has been compared with the behavior of graphitic
mesophase carbon microbeads and hard carbon materials. The comparison
allowed summarizing key guidelines to design new generations of active
materials. The capacity, irreversible capacity loss, rate capability, safety and
packing density are some of the key characteristics, we concluded, are
essential when developing new generations of high packing density and low
surface area graphitic carbons.
1. INTRODUCTION
Due to its high energy density (3,860 mAh/g) and low voltage,
lithium is the most attractive metal of the periodic table for battery
application. Unfortunately lithium metal, and most of its alloys cannot be
used in rechargeable batteries because of their poor cyclability. Therefore,
lithium intercalation compounds and reversible alloys are among today’s
materials of choice for subject application. The most common active
materials for the negative electrodes in lithium-ion battery applications are
carbonaceous materials. The ability of graphitized carbonaceous materials to
*
213
214
be intercalated makes them excellent reservoirs of lithium [1]. Even with a

maximum theoretical capacity of 372 mAh/g (stage I of the graphite
intercalation compounds with lithium, LiC6), which is an order of magnitude
lower than the theoretical capacity of lithium metal, these materials can be
efficiently cycled, when lithium metal cannot.
Synthetic carbonaceous materials are widely used in these
applications. Several types of synthetic materials (e.g. graphitized
mesophase carbon microbeads (MCMB), graphitized milled carbon fiber,
and even, initially, hard carbons) became the materials of choice at the time
of commercialization of first successful lithium-ion batteries in late 1980s.
New trends, mainly driven by cost reduction and need for improved
performance, currently shift focus towards application of natural graphite.
Besides lower cost, natural graphite has several advantages when
compared to its synthetic counterparts. One of them is higher reversible
capacity of natural graphite. The downside of natural graphite is that
unfortunately, its almost perfect lamellar structure makes processing of
graphite uneasy. Natural graphite is also more sensitive to co-intercalation of
certain solvents, which is usually being associated with increase of the
irreversible capacity and reduction of cell’s cyclability [2].
The irreversible capacity results from formation of a surface-
electrolyte interface (SEI) layer, and is believed to be caused by
decomposition of the electrolyte on the surface of active material during few
first charge cycles [3-5]. The values of irreversible capacity and the SEI are
functions of the type of active material and the electrolyte. Also, the safety
issue, which is believed to be associated with stability of SEI, has been
identified as a major parameter in the equation [6-7]. The contribution of the
negative electrode to the thermal runaway is believed to be related to the
nature and also to the surface area of the active material [8-9].
Said subjects are being analyzed in this work. Also, the authors have
attempted to show that in order to be suitable for lithium-ion applications, a
carbon-based active material has to meet a complex number of
physicochemical and electrochemical characteristics. A simple check of
galvanostatic behavior, which is often used today to conclude about carbon’s
suitability for lithium-ion battery technology, is rarely enough for making an
accurate assessment.
2. EXPERIMENTAL
2.1. Graphite Purification

Natural flake graphite has been initially upgraded by flotation
technologies to reach a carbon level in the order of 95%C. Such precursor
has been heat-treated at high temperature to bring the carbon content up to
215
99.95%C, and to eliminate potential battery poisons contained in graphite’s

ash. The sublimation of impurities is performed on an industrial scale in
Superior Graphite’s continuous thermal purification furnaces under inert gas.
During the process of purification, the temperature, which flake is exposed
to, reaches up to 2,800qC.
2.2. Graphite Processing

The purified flake is ground on various industrial mills in order to
reduce size of the flakes and to synthesize application-specific morphologies.
In order to generate proper size distribution of graphite powder, the
milling parameters are adjusted. Changes in particle size distributions for the
end-products are controlled with laser particle size analyzers (example:
Microtrac-X100, of Microtrac, Inc., Montgomeryville, PA, USA).
2.3. Physicochemical Characterization

The surface area of graphite is determined by nitrogen adsorption
using a BET single point method (equipment is available from Quatachrome
Instruments, Boynton Beach, FL, USA).
The packing density of graphite is determined by the tap density
analysis. In this test, we are determining the volume of 100g of graphite
powder after automatic tapping (Quatachrome Autotap model) of a
graduated cylinder filled with a powdered sample, for 1,500 times in
approximately 4 minutes. During tapping, the travel height of a graduated
cylinder with graphite is 3.2 mm.
The Microtrac-X100 (a laser particle size analyzer) was used to
determine the size distribution of the powder. The method is based on a
liquid mode of operation, that is before laser scanning, the powder is added
to a solution of deionized water, containing a surfactant.
The morphology of particles was assessed by synergistic analysis of
packing density values, along with Scanning Electron Microscopy imaging.
Most of the pictures of powders used in this paper were taken using JEOL’s
JSM-6320F instrument outsourced from Drexel University, Philadelphia,
PA, USA.
Two methods were used to determine the purity of the final material.
The ash content was obtained by calculation of a weight loss of 1 g size
graphite sample, after its exposure to 900qC for 4 hours (Superior Graphite
AIMS procedure FBQ0001). This test gives a good idea of the total amount
of impurities. The nature of impurities and their specific concentrations as
measured in ppm or ppb levels, was determined with an ICP inductively
coupled plasma spectrometer (Model Jobin Yvon Ultima POXX/681,
216
available from Horiba Group, Longjumeau, France) after leaching impurities

out from graphite in a mixture of nitric and sulfuric concentrated acids.
Moisture has been determined by Karl Fisher moisture titration at
300qC (model Metrohm 756 KF Coulometer with 707 KF oven, available
from Brinkmann Instruments, Inc. of Westbury, NY, USA).
2.4. Electrochemical Characterization

The electrochemical testing of active materials has been performed
on a 24-channel battery cycler (model BT2000 Basic Charge, available from
Arbin Instruments of College Station, TX, USA).
The electrodes were prepared by mixing graphite powder with the
NMP-predissolved PVDF binder (grade KF#9306, available from KUREHA
Advanced Materials Div., Tokyo, Japan) at the ratio 9/1 of C/PVDF,
respectively. The viscosity of the slurry was adjusted at 5,000 cps by adding
more solvent (NMP).
The slurry was pasted on a copper foil at slow motion (<2.54 cm/s)
using an automatic draw down coater available from Paul N. Gardner Co. of
Pompano Beach, FL, USA) and the electrode was dried for a day under
vacuum at 120qC. After this drying step, the density of the coating was
adjusted to ~1.45-1.55 g/cc through calendaring operation.
The carbon / Li metal half-cells were built inside the Ar-filled glove
box (UniLab system available from MBraun Inc. of Stratham, NH, USA).
The design of the 2-electrode electrochemical cells used in this work
is state of the art and is based on HS Test Cell available from Hohsen Corp.
of Tokyo, Japan [10]. Unless stated otherwise, these cells were tested at
32qC at different cycling rates (from C/20 to C rate) under continuous
current. The electrolyte used was EC:DMC (1:1), LiPF6 (1M) made by Cheil
Industries, South Korea.
3.1. Electrochemical Prescreening

The capacity and profile of the galvanostatic cycling curves are
among the first key characteristics of the active material.
The first galvanostatic charge/discharge at C/20 rate for a natural
graphite (grade LBG1025 of Superior Graphite, Chicago, IL, USA), hard
carbon (grade Carbotron of Kureha Co., Tokyo, Japan) and graphitized
217
mesophase material (grade MCMB2528 of Osaka Gas Co., Tokyo, Japan)

are presented in Figure 1.
The hard carbon has a totally different electrochemical behavior than
the natural graphite and the MCMB2528. Close to half of its capacity during
the intercalation and during the de-intercalation manifests itself at a potential
above 0.4V versus lithium, when it is less than 10 percent of that for the
graphitic active materials. The fact that there is no staging process, as
observed on the curves for hard carbon, is correlated to the crystalline
structure of the material. Hard carbons have an extremely limited lattice
organization and the electrochemical curve without any plateaus shows this
clearly [11-12]. In these disorganized materials, Lithium is believed to be
adsorbed on the available sites.
This, sometimes, allows hard carbons to have higher capacity than
the theoretical capacity of graphite (was not seen to be the case with a
particular sample of commercial hard carbon). During the charge and
discharge, the potential decreases or increases almost continuously. Two
other highly graphitic materials show a classic staging process during
cycling. Each plateau corresponds to a specific phase or stage [13-14].
During intercalation, the first plateau that appears at approximately
0.2V versus lithium corresponds to the stage IV. For this stage, four graphine
layers separate two layers of intercalated lithium. Stages II and III can be
observed before the first stage appears at a potential close to 90 mV. The
theoretical capacity of purified graphite (372 mAh/g) is based on
composition of stage I, (LiC6).
2
1.8
Hard
1.6
Carbon
Voltage (V vs Li/Li+)
1.4 MCMB
1.2
Natural
1 Graphite
0.8
0.6
0.4
0.2
0
0 50 100 150 200 250 300 350 400
Capacity (mAh/g)
Figure 1. Initial Galvanostatic Charge/Discharge Curves

of Three Types of Active Carbonaceous Materials at C/20 rate.
218
The electrochemical characteristics of these three materials are

summarized in Table 1. The reversible capacity, the irreversible capacity loss
and capacity below a certain voltage (0.5V in this case) were identified to be
the key important electrochemical parameters at the time.
These results show that purified natural graphite has the highest
capacity and the highest capacity below 0.5 V. The irreversible capacity is
favorable in case of MCMB2528. With its low capacity and its higher
irreversible capacity loss, the hard carbon doesn’t seem to be an interesting
candidate for conventional lithium-ion battery applications, as its application
will result in reduced cell discharge voltages.
Table 1. Electrochemical characteristics at C/20 rate of a commercial hard carbon, MCMB

2528 and a natural graphite (LBG1025) in half cells with EC-DMC (1:1), LiPF6 (1M).
Material Natural Graphite
Hard Carbon MCMB2528
Performance Property (LBG1025)
Reversible capacity 310 330 360

(mAh/g)
Capacity < 0.5 V
206 320 354
(mAh/g)
Irreversible Capacity
16.8 5 8.9
(%)
From these data, the two types of materials that can be considered as
suitable for lithium-ion applications are MCMB and natural graphite.
MCMB are widely used at the moment. Data in Table 1 shows that natural
graphite can be an attractive candidate for the application in question as well.
Five years ago, we were considering that these electrochemical data
were enough to define if an active material would be a good candidate for
lithium-ion batteries. Based on this understanding, Superior Graphite
developed its “LBG” grades.
3.2. Natural Graphite LBG1025: Electrochemical Results

Figure 2 presents the results of galvanostatic cycling of LBG1025.
The data is summarized in Table 2. Our two design guidelines were
achieving low irreversible capacity and the highest possible capacity at high
rate.
Knowing that for any pure natural graphite the capacity at low rate
should be very close to the theoretical capacity, it did not become a surprise
when the reversible capacity value of LBG1025 reached 365 mAh/g at C/20
rate; it was still as high as 349.5mAh/g at C/5 rate.
Due to issues of polarization in a specific cell design we use for
testing, the cut-off potential at C/2 is fixed at – 40mV vs Li/Li+. At this
negative potential we never observed with LBG material any lithium plating.
219
2
1.8 C/2
C/20
1.6 C/5
1.4
1.2
1
0.8
0.6
0.4
0.2
0
-0.2 0 50 100 150 200 250 300 350 400
Capacity (mAh/g)
Figure 2. Galvanostatic charge/discharge curves of natural graphite LBG1025

at C/20, C/5 and C/2 rates. Electrolyte: EC-DMC (1:1) LiPF6 (1M).
Table 2. Electrochemical characteristics of natural graphite LBG1025 in half cells

Electrolyte: EC-DMC (1:1) LiPF6 (1M).
Material Discharge Irreversible

Charge capacity at
capacity (2V vs capacity loss,
0V vs Li/Li+, mAh/g
Performance Property Li/Li+), mAh/g %
First cycle at C/20 392.8 360.2 8.3%
Second cycle at C/20 364.2 364.2 -
Third cycle at C/20 368.7 365.2 -
C/5 351 349.5 -
C/2 287 330 -

331 at (-0.04V)
Under these conditions (intercalation from 2V to -0.040V vs Li/Li+),

the capacity during de-intercalation is 330mAh/g. At C/2 rate, the capacity at
0V during intercalation is 287mAh/g. This still represents 77% of the
theoretical capacity value for graphite. LBG1025 clearly shows some good
220
high rate performance and the irreversible capacity of 8.3% is well below the
10%, which we were targeting to achieve in our development goals.
3.3. Natural Graphite LBG1025: Physical Characteristics

and Behavior in Lithium-ion Electrode Assemblies
The physical characteristics of the powder and the mechanical
properties of the electrode made from these powders were seen to be among
key important parameters. Some physical characteristics of the LBG1025
and its typical Scanning Electron Microscope image can be found in Table 3.
The SEM shows a flaky, rounded edge smooth morphology.
Table 3. Physical characteristics and typical Scanning Electron Microscope image

of purified natural graphite LBG1025.
Surface Area (m2/g) <4.9
Tap Density (g/cc) 0.47
D10 <9
Particle Size
Distribution D50 19
(Pm)
D90 <35
Mechanical properties of the electrode turned out to be a critical

factor not clearly understood at the time of designing of LBG1025 graphite.
Authors have learned that a battery cannot be cycled if the coating does not
adhere properly to the current collector.
There are some standard adhesion tests used by the paint and
coatings industry, which we have adopted for our work. The most common
is the “Scotch tape” test, which is described by an ASTM D3359, method B.
In this test, a strip of a specific grade of a Scotch tape is applied to
the coating whose surface has been cross-cut with a multi-teeth scratching
tool, and then removed. The percentage of coating that remains on a
substrate is assessed. Usually, less than 65% of coating left on a substrate
suggests poor adhesion. The manual version of this test used in our
investigations is somewhat operator dependent but it gives a good idea about
the level of adhesion of graphite coating to the copper current collector.
221
Figure 3 presents results of adhesion for LBG1025/PVDF

composites. As one can see, in the particular testing conditions, LBG1025
has a poor adhesion (100% of the coating has been removed by the Scotch
tape even without a need to make cross-cut scratches on the surface of
current collector).
Figure 3. Adhesion (Scotch tape test) of PVDF/LBG1025 coating

showing poor adhesion of graphite to copper current collector.
Analysis of results reported for natural graphite LBG1025 allowed

us to draw a preliminary conclusion that regardless of excellent (near
theoretical) electrochemical performance, this graphite, in its pure form, is
not suitable for application in the negative electrodes of lithium-ion batteries.
This is true at least in cases when LBG1025 is used as a sole active carbon
ingredient of the anodes. Noteworthy mentioning: this grade of natural
purified graphite has been found by the industry to be of efficient use as
diluents for MCMB (up to 50% dilution) with primary goal of cost reduction
of the negative electrode.
Partially disappointing results with LBG1025 helped us define new
targets needed to be achieved in order to develop a successful natural
graphite-based material for application in the lithium-ion batteries. New
product design guidelines incorporated a combination of key physical
characteristics (packing density, morphology) and the electrochemical
behavior (capacity, irreversible capacity loss). Following these guidelines,
Superior Graphite has developed the “SL” line of products, which is
described in the following section.
3.4. The “SL” / “SLA” Product Line of Thermally Purified

Natural Graphite: Electrochemical Results
Figure 4 presents typical results of galvanostatic cycling of purified
natural graphite SLA1020 (formerly known as SL-20) from C/20 to C rates.
The results have been summarized in Table 4.
222
1.8 C C /2
C /10
C /3
1.6
1.4
1.2
C /20
0.8
0.6
0.4
0.2
0
0 50 100 150 200 250 300 350 400
-0.2
C apacity (m A h/g)
Figure 4. Galvanostatic charge/discharge curves of natural SLA1020 at C/20, C/10,C/3,

C/2 and C rates. Electrolyte: EC-DMC (1:1) LiPF6 (1M).
Table 4. Electrochemical characteristics purified natural graphite SLA1020

in half cells. Electrolyte: EC-DMC (1:1) LiPF6 (1M).
Charge (intercalation) De-intercalation Irreversible
Properties
at 0V, mAh/g at 2V, mAh/g capacity, %
First cycle at C/20 398.1 362.7 8.8%
C/10 363.8 358.9 -
C/3 315 (341 at -0.04V) 338.4 -
C/2 257 (308 at -0.04V) 306.8 -
C 151 (207 at -0.04V) 206 -
The capacity of SLA1020 at C/20 is 362mAh/g and the irreversible

capacity loss is 8.8%. From C/20 to C/3 the profile of the galvanostatic
cycling curve features the plateaus, which are specific to highly graphitic
materials. At higher rates (C/2 and C), the curve is smooth; there are almost
no plateaus and the capacity drops significantly at the C rate.
We believe that this type of smooth profile results from kinetic
limitations. Still at C/2, the capacity with cutoff potential 0.0V vs Li/Li+ is
257mAh/g. It drops to about 151mAh/g at C rate. When compared with
223
LBG1025 (discussed in the previous sections of this paper), the

electrochemical performance of SLA1020 is slightly lower in terms of
material’s rate capability. This becomes noticeable at C/2 and C rates. From
what we have reported previously, the electrochemical behavior should be
considered as only one of several critical parameters.
3.5. Natural Graphite SLA1020: Physical Characteristics and

Behavior in Lithium-ion Electrode Assemblies
Natural purified graphite SLA1020 has near similar particle size
distribution than LBG1025, which has been discussed earlier. However, due
to spheroidal morphology of SLA1020, its packing density exceeds 0.8 g/cc
(this is almost a factor of 2 of the packing density of LBG1025).
Thus SLA1020 features a particular type of morphology, which we
were targeting to manufacture in order to improve the packing density, the
adhesion to the copper and the cycling behavior. The adhesion test of an
electrode prepared with SLA1020 is presented by Figure 5.
The adhesion is noticeably improved over that with LBG1025.
When doing the “Scotch tape” test, as described by an ASTM D3359,
method B (discussed in detail earlier), only a fraction of the anode coating
could be peeled off the copper substrate.
Table 5. Physical characteristics and typical Scanning Electron Microscope image

of purified natural graphite SLA1020 (formerly known as SL-20).
Surface Area (m2/g) < 5.5
Tap Density (g/cc) >0.8
D10 <11
Particle Size
Distribution D50 20
(Pm)
D90 <35
Authors believe that this shows an effect of particle morphology on

the adhesion of the coating to the current collector. Having a spheroidal
morphology has also the advantage to limit negative effects of specific
orientation of the particles of graphite parallel to the current collector. The
term of “specific orientation”, is often associated with flaky morphologies,
and is believed to be the reason for poor cyclability of flaky graphite that
224
have a tendency to expend during the intercalation in a direction

perpendicular to the current collector. This process is highly undesirable in
practical batteries. It can be associated with loss of contact between copper
current collector and active material.
The guidelines, which we have followed during the development of
SLA1020 and similar types of graphite (those identified by Superior
Graphite’s product nomenclature as “SL” and “SLA” family of products)
were based on an improved understanding of a matter that for a particular
application in lithium-ion batteries it is not enough to have a material with
good electrochemical performance; rather, material’s packing and adhesion
have to be improved accordingly.
As a consequence, a significant deal of positive response has been
received with regards to SLA1020 graphite from a number of end-users.
Application of this graphite was found to be acceptable by certain lithium-
ion battery technologies, yet its use did not become overwhelming
throughout the lithium-ion battery industry, suggesting there has been room
left for further improvement of this grade.
Figure 5. Adhesion (Scotch tape test) of PVDF/SLA1020 coating

showing improved adhesion of graphite to copper current collector.
After having developed SLA1020 and two other grades of similar

nature (one being finer, and another one being coarser), a new parameter
missing in the specification had to be addressed. That is ability of graphite to
withstand prolonged calendar life and long term cycling.
Novel design approaches to addressing these issues resulted in
introduction of the “SLC” family of products, which is at the active stage of
commercialization by Superior Graphite at the time of writing of this paper.
3.6. The SLC Family of Purified Natural Graphite

We have learned that contribution of the negative electrode of
lithium-ion battery to the on-set temperature and kinetics of the thermal
225
runaway reaction, as well as battery’s ability to demonstrate increased

calendar life and long term cycling performance depends on the type of
active carbonaceous material, its morphology, and also on material’s specific
surface area [9, 15]. In particular, we have learned that “ideal” graphite that
meets above design criteria should feature reduced surface area.
2
C
1.8 C/2
C/20
1.6 C/10
C/3
1.4
1.2
0.8
0.6
0.4
0.2
0
0 50 100 150 200 250 300 350 400
-0.2
Capacity (mAh/g)
Figure 6. Galvanostatic charge/discharge curves of natural SLA1015 at C/20, C/10,C/3,

C/2 and C rates. Electrolyte: EC-DMC (1:1) LiPF6 (1M).
SLC1015 is what Superior Graphite has developed in order to

provide lithium-ion battery market with reduced surface area graphite.
SLC1015 is a natural graphite-based product, which features
spheroidal shapes of particles and a surface area of below 1.2m2/g. Said
significant reduction of material’s specific surface area has been achieved
through special surface treatment of natural graphite surface. This treatment
is believed to be resulting in blocking cavities and cracks in precursor natural
graphite particles, many of which are thought to be responsible for increased
surface activity (increased surface area) of graphite.
3.7. Natural Graphite SLC1015: Electrochemical Results

The electrochemical galvanostatic charge-discharge curves at
different rates are presented by Figure 6. The data is also summarized in
Table 6. In our testing, the irreversible capacity loss of SLC1015 was seen to
226
be in the range of 6 to 8.5% (function of electrolyte used). The capacity of

SLC 1015 even at C/2 rate (continuous charge) is 282mAh/g at a potential of
0.0V vs Li/Li+ reference electrode.
3.8. Natural Graphite SLC1015: Physical Characteristics and

Behavior in Lithium-ion Electrode Assemblies
The proprietary surface treatment done on SLC1015 and other SLC
grades available from Superior Graphite reduces dramatically the specific
surface area of this group of carbon products.
Table 6. Electrochemical characteristics purified natural graphite SLA1020
in half cells. Electrolyte: EC-DMC (1:1) LiPF6 (1M).
Charge (intercalation) Discharge Irreversible
Properties
at 0V, mAh/g at 2V, mAh/g Capacity, %
First cycle at C/20 383.5 351.8 8.3%
C/10 354.3 350.4 -
C/3 322.7 (351 @ -0.04V) 348.3 -
C/2 282.3 (336.4 @ -0.04V) 335.1 -
C 225.5 (255.7 @ -0.04V) 254 -
Table 7. Physical characteristics and typical Scanning Electron Microscope image of

surface treated purified natural graphite SLC1015.
Surface Area (m2
Tap Density (g/cc) >0.85
D10
Particle Size
Distribution D50
(Pm)
D90
For instance, the typical surface area of SLC1015 is 1.05 and is

statistically below 1.2 m2/g. By reducing material’s surface area, we
expected to achieve a significant reduction of the thermal runaway activity
227
of carbon (the later is believed to have correlation with improved abuse

tolerance in the full cells and batteries). Also, extension of cyclability and
calendar life of battery, we thought, could be achieved.
These expectations came true. Once developed, SLC1015
demonstrated increased packing density, reduced surface area and the
preliminary results about its abuse tolerance performance are extremely
encouraging.
A set of physical characteristics for the product under consideration
is presented in Table 7. SLC1015’s adhesion result is presented by Figure 7.
Figure 7. Adhesion (Scotch tape test) of PVDF/SLC1015 coating

showing excellent adhesion of graphite to copper current collector.
The adhesion of SLC1015 is noticeably better than that with

LBG1025 and also SLA1020, described earlier. Only a minute fraction of
the SLC1015 based coating could be removed from the copper substrate
when an adhesion check in accordance with the ASTM D3359, method B
procedure has been performed. The interface “copper/coating” was seen to
be much stronger than with any previously described generation of natural
graphite.
Performance of the surface treated purified natural graphite
SLC1015 and other similar grades in Superior Graphite’s “SLC” product
family has attracted a number of lithium-ion battery manufacturers. The
number of satisfied end-users is steadily growing.
The above success of SLC product line indicates that design
parameters taken as targets for development meet expectations of the
lithium-ion battery application.
Besides “high energy density” batteries (lap top computers, cellular
phone power sources, etc.), new material looks promising for the emerging
“high power” segment of the battery market. For instance, authors would
like to make a reference to the important work by developers at the Argonne
National Laboratory, USA, in qualifying the SLC1015 graphite as a viable
228
candidate for electrode active materials for the emerging application in

hybrid electric vehicles. The SLC1015 graphite has been reported to be fully
suitable for said application, as can be seen in [16, 17].
4. CONCLUSIONS
The comparison in between natural graphite and other carbonaceous
materials has shown that natural graphite having sufficient purity and an
optimal set of surface properties can be an outstanding candidate for lithium-
ion battery applications.
In this work we have considered three generations of natural
graphite. Using examples of these three product families, the authors have
shown evolution in understanding of requirements necessary to achieve for
an active material to be used at an industrial level. The progress achieved in
the processing of natural graphite currently allow Superior Graphite to
produce very low surface area graphite having high packing density. The
packing density of regular SLC type materials is in between 0.80 to 1.1g/cc
with surface area in between 0.85 to 1.5 m2/g.
Given the complex process to produce mesophase carbon
(graphitized microbeads and fibers), natural graphite can be very competitive
in terms of its manufacturing costs [18]. The physical characteristics of
certain SLC type materials are extremely close to the characteristics of state-
of-the-art MCMB grades.
At the electrochemical performance level, these novel natural
graphite-based materials surpass mesophase carbon’s characteristics as
related to cell/battery safety performance, low irreversible capacity loss, and
good rate capability even at high current densities.
At a broader level, a battery is a complex product and the negative
electrode is just one of its many components. The electrochemical behavior
of graphite depends on a lot of parameters (such as, for instance, current
collector, binder, electrolyte [19], special additives, application, etc.).
LBG1025 is a good example to show that the electrochemical performance
data is only one piece of the puzzle. Good initial galvanostatic performance
of graphite is not enough for material to be suitable for lithium-ion battery
application. Laboratory testing of initial performance of carbon is always
necessary, but at the end, the only way to test the material is in real
application, in full lithium-ion cells and batteries.
REFERENCES
1. Hossain, Sohrab. Rechargeable Lithium Batteries (Ambient Temperature). In Handbook
of Batteries, David Linden ed., New York, N.Y., 36.1-36.77, 1995.
229
2. Winter M., Novak P. and. Monnier A. Graphites for Lithium-ion Cells: The Correlation
of the First-Cycle Loss with the Brunauer-Emmett-Teller Area. J. Electrochem. Soc.,
145, 428-435 (1998).
3. Lithium Batteries. E. Peled, J.P.Gabano, eds., Chap 3, 43-69 Academic Press, London
(1983).
4. Naji A., Willmann P., and Billaud D. Electrochemical Intercalation of Lithium into
Graphite: Influence of the Solvent Composition and the Nature of the Lithium Salt.
Carbon, 36, 1347-1352 (1998).
5. Endo E., Tanaka K. and Seika K. Initial Reaction in the Reduction Decomposition of
Electrolyte Solutions for Lithium Batteries, J. Electrochem. Soc., 147, 4029-4033 (2000).
6. Henriksen G. Advanced Process Research: Overview/Advances./ Proceedings of the
Annual Merit Review Meeting of the U.S. DOE ATD Program, Argonne, IL, (8/2001).
7. Cao F., Barsukov I.V., Bang H.J., Zaleski P. and Prakash J. Evaluation of Graphite
Materials as Anodes for Lithium-Ion Batteries. J. Electrochem Soc. 147 (10), 3579-3583
(2000).
8. Barsukov I.V. Development of low-cost, novel carbonaceous materials for anodes in
lithium-ion rechargeable batteries – Superior Graphite Co. Snapshots of CARAT
(Cooperative Automotive Research for Advanced Technology) Projects. Publication of
OAAT & OTTEE&RE, U.S. Department of Energy, 9/2001, 26-27.
9. Barsukov I.V. Development of low-cost, carbonaceous materials for anodes in lithium-
ion batteries – Superior Graphite Co. Snapshots of CARAT (Cooperative Automotive
Research for Advanced Technology) Projects. Publication of Office of FreedomCAR and
Vehicle Technologies, EERE, U.S. Department of Energy, 5/2003, 22-23.
10. Battery R&D Equipment. Hohsen Corp. product catalog. Tokyo, Japan (2002), 1.
11. Zheng, L., Zhong Q., and Dahn J.R. High-Capacity Carbons Prepared from Phenolic
Resin for Anodes of Lithium-Ion Batteries. J. Electrochem. Soc., 142, 211-214 (1995).
12. Dahn J.R., Zheng L., Liu Y. and Xue J.S. Mechanism for Lithium Insertion in
Carbonaceous Materials, Science, 270, 590-593 (1990).
13. Billaud D., Henry F.X. and Willmann P. Electrochemical Synthesis of Binary Graphite-
Lithium Intercalation Compounds. Mat. Res. Bull., 28, 477-483 (1993).
14. Billaud D., Henry F.X., Lelaurain M., and Willmann P. Revisited Structures of Dense
and Dillute Stage II Lithium-Graphite intercalate Compounds. J. Phys. Chem Solids, 57,
775-781 (1996).
15. Joho F., Novak P., and Spahr M.E. Safety Aspects of Graphite Negative Electrode
Materials for Lithium-Ion Batteries. J. Electrochem. Soc., 149, 1020-1024 (2002).
16. Amine K., Liu J., Belharouak I. and Kahaian A. Material screening from and with
suppliers. In Chapt. Overview of cost reduction and Technology Transfer Efforts. Proc.
Of the Advanced Technology Development Program, U.S. DOE FreedomCAR annual
merit review meeting, June 24-25 (2003) Argonne National Laboratory, Argonne, IL,
USA, P.8.
17. Liu J., Vissers D.R., Amine K., Barsukov I.V., Doninger J.E. Surface Treated Natural
Graphite as Anode Material for High-Power Li-Ion Battery Applications. In New
Carbon-Based Materials for Electrochemical Energy Storage Systems: Batteries, Fuel
Cells and Supercapacitors. Barsukov I., Johnson C., Doninger J., Barsukov V. eds.
NATO ARW series Volume. Springer (2005) - in this book.
18. Barsukov I., Henry F., Doninger J., Gallego M., Huerta T., Girkant R. and Derwin D. On
the electrochemical performance of lithium-ion battery anodes based on natural graphite
with various surface properties. ITE Letters on Batteries, New Technologies & Medicine,
V.4, N.2 (2003), 163-166.
19. Kohs W, Santner H.J., Hofer F., Schröttner H., Doninger J., Barsukov I., Albering J.H.,
Möller K.-C., Besenhard J.O., and Winter M. A study on electrolyte interactions with
graphite anodes exhibiting structures with various amounts of rhombohedral phase. J.
Power Sources, 199-121, 528-538 (2003).
MECHANISMS OF REVERSIBLE AND
IRREVERSIBLE INSERTION IN
NANOSTRUCTURED CARBONS
USED FOR Li-ION BATTERIES
F. Béguina*, F. Chevalliera, M. Letelliera, C. Vixc, C. Clinarda, J.N.

Rouzauda, and E. Frackowiakb
a
CRMD, CNRS-Université, 1B Rue de la Férollerie, 45071 Orléans Cedex 02, France
b
ICTE, Poznan University of Technology, ul. Piotrowo 3, 60-965 Poznan, Poland
c
ICSI, CNRS, 15, rue Jean Starcky, 68057 Mulhouse, France
1. INTRODUCTION
Taking into account the high-energy demand of electric vehicles, an
important research effort is devoted for developing a new generation of
power sources. In the case of lithium-ion batteries, the main attention is paid
to designing materials with improved electrochemical performance. The
negative electrode of commercial lithium-ion batteries is usually graphite,
which reversible lithium intercalation capacity is limited to one lithium for
six carbon atoms in standard conditions [1]. Due to their higher storage
capacity, nanostructured carbons, such as hard carbons, seem to be
promising materials in replacement of graphite [2]. However, they
demonstrate a noticeable hysteresis and irreversibility, which can preclude a
practical application [2,3]. Despite the huge number of publications on the
subject within the last years, the causes of these two important drawbacks
are not yet clearly elucidated. Hence, the understanding of the insertion-
deinsertion mechanisms in nanostructured carbons is a key for improving the
performance of this kind of materials.
In this paper, new highlights are proposed to interpret the reversible
and irreversible capacities of nanostructured carbons in lithium batteries. A
proportional dependence between the irreversible capacity and the active
surface area (ASA) of carbon materials will be demonstrated, showing the
ASA concept more universal than any of the other parameters which were
previously considered. In-situ 7Li NMR will be also presented as a means to
*
231
232
follow stepwise the sequence of reversible lithium insertion, without all the
disadvantages of the ex-situ methods.
2. RELATION BETWEEN THE IRREVERSIBLE

CAPACITY AND THE ACTIVE SURFACE AREA
2.1. Literature Review

The irreversible capacity of lithium-ion batteries is essentially
attributed to the electrolyte decomposition on the carbon electrode during the
first reduction cycle, forming a passivation layer (also called Solid
Electrolyte Interface, SEI). The main components of the SEI identified by
infrared spectroscopy are Li2CO3, ROCO2Li, etc. [4-6]. According to some
publications, the extent of SEI formation is directly related to the surface
area accessible for the electrolyte. Therefore it is often compared to the BET
specific surface area, giving in some cases a linear relationship [7-9]. Since
nanostructured carbons are generally characterised by a noticeable value of
BET specific surface area, it could be a reasonable explanation for the
important value of irreversible capacity encountered with these materials.
However, Buiel et al. showed that this contribution to the irreversible
capacity cannot be higher than 0.13 lithium for 6 carbon atoms, and some
other parameters, for example the surface oxygenated functionality might be
involved [10]. Furthermore, for carbon materials with a mesoporous
character, such as carbon nanotubes or carbons from templates, the
irreversible capacity is always very high [11], and its evolution is better
related to the mesopore volume [12].
The influence of heteroatoms involved in surface functional groups,
such as oxygen and sulfur, on the irreversible capacity has been
demonstrated by Larcher [13]; higher the surface concentrations of groups,
higher the contribution to the irreversible capacity. Upon increasing the
highest treatment temperature (HTT) of the carbon precursor, the irreversible
capacity decreases. Following by XPS the evolution of -C-Li or –CO- Li+
surface concentration vs HTT, it has been proposed that it is due to a
decrease of the amount of free radicals or surface hydroxyl groups,
respectively [14,15]. Similarly the drastic increase of irreversible capacity
which occurs after ball milling of graphite in argon atmosphere is interpreted
by a highly developed active surface, i.e. dangling bonds [16]. However, the
involvement of free radicals to form C-Li bonds is quite hypothetic [17,18].
Nevertheless, there is no doubt that a part of lithium is irreversibly trapped
by surface functional groups. The species which are usually adsorbed on a
carbon surface, i.e. water, di-oxygen, etc. are also contributing to the
irreversible capacity, by increasing the thickness of the SEI [19,20]. Finally,
considering a series of carbons, a linear relationship has been found between
233
the irreversible capacity and the micropore volume [21]; from ex-situ 7Li
NMR results, the authors suggested that metallic lithium is irreversibly
trapped in the micropores [22].
Since claims attribute the irreversible capacity, at least partly, either
to the surface functionality or to adsorbed species, it has been attempted to
cure the surface of carbon materials. When the surface oxygenated groups
are reduced by di-hydrogen, the irreversible capacity decreases and
simultaneously the reversible capacity increases due to the introduction of C-
H functionality on the carbon surface [23]. A positive effect of HTT at ca.
900-1000°C under vacuum has been observed, which could be due to surface
groups elimination [24,25]. However, surprisingly, the performance does not
improve after high temperature treatment under secondary vacuum [26].
Chemical vapour deposition of pyrocarbon allows the irreversible capacity
of non-graphitisable carbons to be strongly depressed [10]. Since the specific
surface area measured by nitrogen adsorption is not modified by this
treatment, the observed decrease could be due to the elimination of surface
groups or adsorbed species. By analogy, increasing the hydrophobic
character of carbon by copper deposition on the surface allows to get a
smaller value of irreversible capacity [27]. The irreversible capacity of
graphite is also decreased after pyrolytic carbon deposition [28] or coke
coating from the carbonisation of a pitch/resin blend [29] or coal tar pitch
[30]. It has been suggested that the pyrolytic carbon coating prevents the
solvent co-intercalation into graphite [28]. It also allows reducing the BET
specific surface area and the O/C ratio [31]. Similarly after coke coating,
there is less gas evolution during the first discharge, that it is attributed to a
decrease of the exposed surface area of graphite [30].
To summarise, mainly two parameters seem to contribute to the
irreversible capacity of nanostructured carbons: the SEI formation, which
would depend on the BET specific surface area and the presence of species
adsorbed or bonded to the surface [10]. However, our previous observation
that the irreversible capacity is unchanged after annealing under secondary
vacuum, disclaims a direct influence of oxygenated surface groups or
adsorbed species [26]. Moreover, it is not worth to use the BET specific
surface area, which is obtained from the amount of nitrogen physisorbed on
the carbon surface in order to explain the SEI formation where high energies
are involved. Therefore, we suggest relating the irreversible capacity to the
number of active sites on the carbon surface, so-called active surface area
(ASA) [26]. The active surface area of a carbon material corresponds to the
cumulated surface area of the different types of defects present on the carbon
surface (stacking faults, single and multiple vacancies, dislocations) [32];
these sites are responsible for the interactions with the adsorbed species.
Since the surface functionality and the BET specific surface area are not two
independent parameters (in general, higher the BET surface area, higher the
number of surface groups), the ASA appears as a more universal concept.
234
In this paper, the effect of surface functionality will be studied by

comparing the irreversible capacity of viscose carbon fibres outgassed at
900°C to that of the pristine material. Graphite ball-milled in different
atmospheres is selected to demonstrate the irreversible capacity vs ASA
relationship.
2.2. Experimental
The materials chosen for our investigations were natural graphite
(UF2 from Kropfmühl <5Pm) or a carbon cloth prepared by carbonisation of
viscose under neutral atmosphere for 15 minutes, successively at 400, 700
and 1000°C. The carbon cloth is used as a self-standing electrode, without
any binder or conductivity additive, which is an advantage for investigating
only the electrochemical properties of the disordered carbon. The natural
graphite specimens (1 g) were ball-milled with 7 g of stainless steel balls
under controlled atmosphere using a Spex 8000 mill. For some experiments,
the carbon samples were coated with pyrolytic carbon by chemical vapour
decomposition of propylene (2.5 ml/mn) diluted in nitrogen (100 ml/mn)
during 10 minutes at 900°C.
Nitrogen adsorption/desorption isotherms were determined at 77 K,
using a Micromeritics ASAP 2010, after outgassing the samples at 150°C
until the pressure reached 10-6 mbar. The Active Surface Area (ASA) was
measured by the method of Walker et al.[33] which is based on di-oxygen
chemisorption. The sample (a mass corresponding to a surface area of ca. 0.5
m2) was first outgassed under dynamic vacuum (10-3 Pa) at 950°C for 5
hours to release the active sites. In a second step, oxygen is chemisorbed at
300°C during 15h (initial pressure: 66.5 Pa) leading to the formation of
surface oxygenated complexes. After elimination of the remaining di-
oxygen, the amount of oxygenated complexes is determined by mass
spectrometry analysis of evolved CO and CO2 by subsequent outgassing at
950°C.
For the electrochemical investigations, two electrode cells
(Swagelok type) with a glass micro-fibre separator (Whatman GF/D cat
1823070) were assembled in an argon atmosphere glove box. The working
electrode was a disk of the carbon cloth or a plastic carbon pellet prepared
according to the Bellcore technique [34], and metallic Li was used as counter
and reference electrodes. The electrolyte was 1M LiPF6 in a mixture of
EC/DEC (1/1) (Merck). The cells were investigated by galvanostatic charge-
discharge in the potential range from 2 V to 0 V vs Li, at C/20 rate
(corresponding to a capacity of 372 mAh/g in 20 h), using a Mac Pile multi-
channel potentiostat/galvanostat (Biologic, France).
235
2.3. Results and Discussion

As detailed in the literature background, it is often attempted to
relate the irreversible capacity either with the BET specific surface area or
with the presence of surface groups. Below, we will show two experiments,
which disclaim these kinds of correlations.
Figure 1 presents the plot of the BET specific surface area vs the
irreversible capacity measured for graphite samples milled in different
atmospheres and sometimes post-treated by pyrolytic carbon deposition. The
experimental values are quite scarce and, contrarily to several claims [7-9],
there is not any linear dependence between these two parameters. It seems
that the linearity would exist only for samples from the same family with
comparable microtextures.
The effect of surface functionality has been investigated in a next
experiment. Figure 2 shows the galvanostatic charge/discharge on the carbon
cloth outgassed at 150°C under vacuum in order to eliminate the adsorbed
water. As it is often the case for hard carbons, the irreversible capacity is
quite high, xirr = 0.71 (taken as x in the formula LixC6). Taking into account
the literature data [13], such a high value could be due to a developed
oxygenated surface functionality. Therefore, for a second experiment, the
surface groups have been eliminated by heat-treating the carbon cloth at
900°C under vacuum, condensing the vapours of the vacuum line in a liquid
nitrogen trap. Figure 3 presents the galvanostatic curve for the resulting
material, which has been transferred into the swagelok cell without any
contact with the air atmosphere, in order to avoid any pollution. It is
remarkable, that the electrochemical characteristics are even worst than for
the as received specimen.
Hence, it is clear that the irreversible capacity is not directly related
with the presence of surface oxygenated groups. In the conditions used for
outgassing at 900°C, the number of active sites has not been modified [33].
Therefore, taking into account that the irreversible capacity is almost the
same for the experiments shown in Figures 2 and 3, it might mean that only
the active sites are responsible of the electrolyte decomposition, and
consequently of the SEI development.
In order to verify this assumption, the Active Surface Area (ASA) of
the various graphite samples has been expressed as a function of the
irreversible capacity, and the result is presented in Figure 4. A perfect linear
dependence is observed, which proves that the main contribution to the
irreversible capacity is the decomposition of the electrolyte on the active
sites during the first galvanostatic reduction. A similar trend has been found
for the carbon cloth treated in different conditions [26]. Figure 4 shows also
that all the samples (4, 5, 6, 7) after milling present the highest values of
ASA, whatever the atmosphere for milling, due to the production of many
edge planes [16].
236
500 (g)
BET specific surface area (m /g)
2
450
400
350 (a)
300 (c)
250
200
(e)
150
100
50 (h) (d) (b)
(f)
0
0 0.5 1 1.5 2 2.5
x in LixC6
Figure 1. Relation between the BET specific surface area and the irreversible capacity x of
graphite samples treated in different conditions. (a) 10 h in vacuum; (b) 10 h in vacuum +
pyrolytic carbon deposition; (c) 10 h under H2; (d) 10 h under H2 + pyrolytic carbon
deposition; (e) 10 h under O2; (f) 10 h under O2 + pyrolytic carbon deposition; (g) 20 h in
vacuum; (h) 20 h in vacuum + pyrolytic carbon deposition.
x in L ixC 6
Figure 3. Galvanostatic charge/discharge of the
carbon cloth outgassed 24 hours under vacuum at 900°C.
237
It is remarkable that the samples 1, 2, 3, 8 which have been coated in

the same conditions by chemical vapour deposition, present comparable
values of ASA and irreversible capacity (Figure 4). After coating, the active
sites of milled graphite are blocked and only those of the pyrolytic carbon
layers are effective in the electrolyte decomposition. Similarly, the decrease
of irreversible capacity observed by other authors after pyrolytic carbon
deposition [28] or coke coating [29,30] on graphite is due to a blockage of
active sites. Since the active sites are rather located on the edge planes, the
BET surface area, which theoretically expresses the total surface of the basal
planes, cannot be a relevant parameter. Hence, the proportionality of
irreversible capacity with the BET specific surface area will only be found
with non-microporous solids of the same family which have a comparable
density of active sites, for example a series of graphites [9].
70
(c)
Active surface area (m /g)
60
2
50 (e)
40 (g)
(a)
30
20 (f) (d)
10 (h) (b)
0
0 0.5 1 1.5 2
x in LixC6
Figure 4. Relation between the BET specific surface area and the irreversible capacity x of
graphite samples treated in different conditions. (a) 10 h in vacuum; (b) 10 h in vacuum +
pyrolytic carbon deposition; (c) 10 h under H2; (d) 10 h under H2 + pyrolytic carbon
deposition; (e) 10 h under O2; (f) 10 h under O2 + pyrolytic carbon deposition; (g) 20 h in
vacuum; (h) 20 h in vacuum + pyrolytic carbon deposition.
3. IN-SITU 7LI-NMR STUDY OF THE REVERSIBLE

LITHIUM INSERTION MECHANISM
3.1. Introduction
Although some nanostructured carbons are able to store reversibly

higher amounts of lithium than graphite [2], they generally demonstrate a
high hysteresis (see for example Figure 2) that still precludes their use in
lithium-ion batteries. In order to improve their electrochemical properties as
electrode materials, a current effort is made to better understand the
238
insertion-deinsertion mechanism. Ex-situ solid-state 7Li-Nuclear Magnetic

Resonance (7Li-NMR) is one of the most useful methods which has been
extensively applied for the determination of the states of lithium at various
steps of reduction or oxidation of nanostructured carbons [35,36].
The principle of the measurements by ex-situ 7Li NMR is as follows.
During the galvanostatic charge/discharge of a lithium/carbon cell, the
experiment is stopped at a selected potential. Then, the carbon material is
extracted from the cell, eventually washed with a clean solvent, dried and
ground before being put into a sealed tube for the NMR analysis. Using such
a methodology, the equilibrium composition of the sample might be
modified by washing or by moisture and oxygen from the atmosphere of the
glove box during the different transfers. On the other hand, each value of
potential requires the preparation of a new sample that limits the number of
possible measurements if one wants to keep a reasonable experimental time.
Recently, to overcome these drawbacks, we have performed in-situ 7Li-NMR
during the galvanostatic cycling of a composite carbon electrode in aN ultra-
thin plastic lithium/carbon cell [37,38]. In these conditions, the carbon
substrate is the same for all the spectra, eliminating possible uncertainty in
the interpretations that might always exist when comparing the results
obtained with different starting specimens and electrochemical cells.
In the following part of this paper, in-situ 7Li-NMR is used to better
know the state(s) of lithium reversibly stored in disordered carbons.
3.2. Experimental
For these experiments, we used the carbon cloth prepared by
carbonisation of viscose up to 1000°C. It was additionally coated with a thin
layer of pyrolytic carbon by CVD under propylene. The conditions for the
sample elaboration were previously given in Part 2. As seen before, the
advantage of pyrolytic carbon coating is to depress the amount of lithium
irreversibly inserted, that should make easier the interpretation of the NMR
data related with the reversible part of insertion. For the in-situ 7Li NMR
experiments, we used a supple ultra-thin plastic lithium/carbon cell built
according to the conditions described in reference [38]. The liquid
electrolyte was 1M LiPF6 in a 1:1 volume ratio of ethylene carbonate EC and
diethyl carbonate DEC (Merck). 7Li-NMR was performed using a Bruker
Avance DSX 360 spectrometer equipped with a broadband probe head. Once
the cell to be investigated was positioned in the 10 mm coil, the copper
current collectors were welded to wires connected to a Mac Pile II
galvanostat (Biologic, France). The cell was cycled at a C/20 rate in the
potential range from 2 V to 0 V vs Li. In these conditions, the acquisition
time (8 min) was short compared to the time of a complete cycle (60 hours),
therefore the cell cycling was not stopped during recording the NMR spectra.
239
High Resolution Transmission Electron Microscopy (HRTEM,

Philips CM20, 200 kV) was applied to get structural and nanotextural
information on the fibers, by imaging the profile of the aromatic carbon
layers in the 002-lattice fringe mode. A carbon fiber coated with pyrolytic
carbon was incorporated in epoxy resin and a transverse section obtained by
ultramicrotomy was deposited on a holey carbon film. An in-house made
image analysis procedure was used to get quantitative data on the composite.
3.3. Results and Discussion
The HRTEM observation of the cross section of a coated fiber

showed that the core is constituted of aromatic layers highly misoriented,
whereas they are preferentially oriented in parallel for the thin coating; pairs
of stacked layers form mainly Basic Structural Units (BSUs) in which the
average interlayer distance is smaller than between the aromatic layers in the
bulk of the fiber. Since the nanotexture is more dense for the pyrolytic
carbon than for the fiber itself, it acts as a barrier which prevents the
diffusion of the large solvated lithium ions to the core of the fiber, allowing
the passivation layer to be less developed after this treatment. Hence, the
major amount of lithium inserted is involved in the reversible contribution;
therefore this composite material is extremely interesting for the in-situ 7Li
NMR study of the reversible insertion.
Figure 5 shows the result of in-situ 7Li NMR during galvanostatic
cycling of the supple cell. Initially, the spectrum comprises one line at 0 ppm
corresponding to Li+ from the dissolved salt and another line at 263 ppm due
to metallic lithium from the counter electrode. Since, the latter line is
completely out of the shift range noted for the new lines, it is not shown in
Figure 5. During the electrochemical reduction step, two lines related with
lithium insertion are observed. The minor one progressively shifts in the first
cycles from 0 ppm to 18 ppm, where it stabilizes until the end of the
reduction process. The position of the most important one progressively
downfield shifts to reach values of ca. 110 ppm at full discharge (Figure 5).
During the oxidation step, the reverse phenomenon is observed with some
hysteresis, the two new lines shifting back to 0 ppm and vanishing
completely when the potential reaches 2 V vs Li.
Lithium peaks at constant position are also found for graphite
intercalation compounds (GICs), at 41 ppm for the 1st stage LiC6 or the 2nd
stage LiC12 compounds [39,40], and at 10 ppm for the 2nd stage LiC18 phase
[41]. Hence, the peak found at 18 ppm for the carbon composite is attributed
to lithium intercalation (with a charge transfer to carbon) in graphitic type
intervals. HRTEM image analysis shows that such kinds of intervals are
rather found in the pyrolytic carbon coating and to a lesser extent in the
core of the fiber. The peak, which shifts up to 110 ppm is attributed to
240
quasi-metallic bidimensional lithium clusters located in narrow nanopores

formed by pairs of spaced and misoriented aromatic layers in the core of the
fibers.
Figure 5. 7Li NMR spectra recorded during 3 cycles of reversible lithium insertion-
deinsertion in the composite carbon electrode. The peak at 0 ppm is due to ionic lithium
(Li+PF6- and passivation layer). The peak of lithium at 263 ppm is not shown.
45%
Pyrocarbon
40%
Fiber
35%
30%
Fringes %
25%
(
20%
15%
10%
5%
0%
0.25 1.25 2.25 3.25 4.25 5.25 6.25 7.25 8.25
Fringes length L (nm)
Figure 6. Histogram of the fringes length obtained by computer analysis of the

HRTEM image of the fiber/pyrolytic carbon composite.
241
Table 1. V alues of electron affinity for few aromatic hydrocarbons and for graphite.
Electron affinity (eV)

C6H6 0
Naphtalene 1.1
Anthracene 1.8
Graphite 4.39
The histogram of the composite fringes lengths obtained by

computer analysis of the HRTEM images is shown in Figure 6. Short fringes
are more numerous in the fiber, whereas long fringes up to 6 nm essentially
exist in the pyrolytic carbon coating. Then, it is interesting to consider the
data reported in Table 1, which shows the values of electron affinity for
some aromatic hydrocarbons and for graphite. From this table, we observe
that higher the number of aromatic rings, from benzene to graphite, higher
the electron affinity.
The high charge transfer in graphite is correlated with the lithium
intercalation between the layers of infinite size. Hence, we can confirm that
the Knight shift of 18 ppm corresponds to lithium intercalated between
layers of large size, which are essentially located in the pyrolytic carbon
coating, whereas the values up to 110 ppm are due to lithium close to layers
of smaller size. It means that lithium becomes more and more “metallic” (i.e.
less ionic) during the reduction process, occupying nanopores where it is
close to layers of decreasing diameter.
4. CONCLUSION
In this paper, we presented new information, which should help in
optimising disordered carbon materials for anodes of lithium-ion batteries.
We clearly proved that the irreversible capacity is essentially due to the
presence of active sites at the surface of carbon, which cause the electrolyte
decomposition. A perfect linear relationship was shown between the
irreversible capacity and the active surface area, i.e. the area corresponding
to the sites located at the edge planes. It definitely proves that the BET
specific surface area, which represents the surface area of the basal planes, is
not a relevant parameter to explain the irreversible capacity, even if some
papers showed some correlation with this parameter for rather low BET
surface area carbons. The electrolyte may be decomposed by surface
functional groups or by dangling bonds. Coating by a thin layer of pyrolytic
carbon allows these sites to be efficiently blocked, without reducing the
value of reversible capacity.
242
Regarding the reversible capacity of disordered carbons, we found

that lithium is mainly in a quasi-metallic state, another minor part being in a
state comparable to intercalation. If we consider that the lithium density in
the intercalated domains is comparable to that of graphite, we can conclude
that the enhanced reversible capacity of the disordered carbons is related
with the quasi-metallic state of lithium. In this paper, we postulate that the
electron density of the quasi-metallic lithium clusters is related with the size
of the neighbour carbon layers. In future work, we plan to investigate
different materials in order to determine if there is any relationship between
the reversible capacity and the fringe length.
ACKNOWLEDGMENTS
This research was supported by the NATO Science for Peace
programme (Project CARBON SfP 973849) and by a PhD grant from the
Conseil Régional du Centre (France) to Frédéric Chevallier. Thanks are due
to Luc Aymard for providing the samples of milled graphite and to Mathieu
Morcrette for building the supply cells; both are from Université de Picardie,
Amiens (France).
REFERENCES
1. Guérard D., Hérold A., Carbon 1975; 13: 337
2. Zheng T., Dahn J.R., Applications of carbon in Lithium-ion batteries, In Carbon
materials for advanced technologies, Burchell T.D. Ed., Elsevier, Oxford, 1999, pp 341-
388.
3. Frackowiak E., Béguin F., Carbon 2002; 40: 1775.
4. Naji A., Ghanbaja J., Humbert B., Willmann P., Billaud D., J. Power Sources 1996; 63:
33.
5. Aurbach D., Zaban A., Ein-Eli Y., Weissman I., Chusid O., Markovsky B., Levi M., Levi
E., Schechter A., Granot E., J. Power Sources 1997; 68: 91.
6. Kanamura K., Tamura H., Shiraishi S., Takehara Z., J. Electroanal. Chem. 1995; 394: 49
7. Fong R., Von Sacken U., Dahn J.R., J. Electrochem. Soc. 1990; 137: 2009
8. Simon B., Flandrois S., Fevrier-Bouvier A., Biensan P., Mol. Cryst. Liq. Cryst. 1998;
310: 333
9. Winter M., Novak P., Monnier J., J. Electrochem. Soc. 1998; 145: 428.
10. Buiel E., Dahn J.R., J. Electrochem. Soc. 1998; 145: 1977.
11. Duclaux L., Frackowiak E., Béguin F., J. Power Sources 1999; 81-82: 323.
12. Béguin F., Méténier K., Chevallier F., Azaïs P., Pellenq R., Bonnamy S., Duclaux L.,
Rouzaud J.N., Frackowiak E., Mol. Cryst. Liq. Cryst. 2002; 386: 151.
13. Larcher D., Mudalige C., Gharghouri M., Dahn J.R., Electrochim. Acta 1999; 44: 4069.
14. Y. Matsumura, S. Wang, J. Mondori, J. Electrochem. Soc. 1995; 142: 2914.
15. Y. Matsumura, S. Wang, J. Mondori, Carbon 1995; 33: 1457.
16. Azaïs P., Duclaux L., Faugère A.M., Béguin F., Appl. Phys. Lett. 2002; 81: 775.
17. Y. Matsumura, S. Wang, Y. Nakagawa, C. Yamaguchi, Synth. Met. 1997; 85: 1411.
18. M. Winter, H. Buqa, B. Evers, T. Hodal, K.C. Moller, C. Reisinger, M.V. Santis
Alvarez, I. Schneider, G.H. Wrodnigg, F.P. Netzer, R.I.R. Blyth, M.G. Ramsey, P.
243
Golob, F. Hofer, C. Grogger, W. Kern, R. Saf, J.O. Besenhard, Monatsh. Chem. 2001;
132: 473.
19. Xing W., Dahn J.R., J. Electrochem. Soc. 1997; 144: 1195.
20. Aurbach D., Schechter A., Electrochim. Acta 2001; 46: 2395.
21. Guérin K., Février-Bouvier A., Flandrois S., Simon B., Biensan P., Electrochim. Acta
2000; 45: 1607.
22. Guérin K., Ménétrier M., Février-Bouvier A., Flandrois S., Simon B., Biensan P., Solid
State Ionics 2000; 127:187.
23. Lu W., Chung DDL, Carbon 2001; 39: 493.
24. M. Kikuchi, Y. Ikezawa, T. Takamura, J. Electroanal. Chem. 1995; 396: 451.
25. R. Alcantara, F.J. Fernandez-Madrigal, P. Lavela, J.L. Tirado, J.M. Jimenez-Mateos, C.
Gomez de Salazar, R. Stoyanova, E. Zhecheva, Carbon 2000; 38: 1031.
26. Béguin F., Chevallier F., Vix C., Saadallah S., Rouzaud J.N., Frackowiak E., J. Phys.
Chem. Solids 2004; 65: 211.
27. Y. Wu, C. Jiang, C. Wan, E. Tsuchida, Electrochem. Comm. 2000; 2: 626.
28. M. Yoshio, H. Wang, K. Fukuda, Y. Hara, Y. Adachi, J. Electrochem. Soc. 2000; 147:
1245.
29. Kuribayashi I., Yokoyama M., Yamashita M., J. Power Sources 1995; 54: 1.
30. Yoon S., Kim H., Oh S.M., J. Power Sources 2001; 94: 68.
31. Natajaran C., Fujimoto H., Tokumitsu K., Mabuchi A., Kasuh T., Carbon 2001; 39: 409.
32. Lahaye J., Dentzer J., Soulard P., Ehrburger P. Carbon gasification : The active site
concept. In Fundamental Issues of Control of Carbon Gasification Reactivity, Lahaye J,
Ehrburger P. Ed., Academic Publishers, London, 1991, p.143-158
33. Laine N.R., Vastola F.J., Walker P.L. Jr., J. Phys. Chem. 1963; 67: 2030.
34. Tarascon J.M., Gozdz A.S., Scmutz C., Shokoohi F., Warren P.C., Solid State Ionics
1996; 86-88: 49.
35. Gautier S., Leroux F., Frackowiak E., Faugère A.M., Rouzaud J.N., Béguin F., J. Phys.
Chem. A 2001; 105: 5794.
36. Guérin K, Ménétrier M., Février-Bouvier A., Flandrois S., Simon B., Biensan P., Solid
State Ionics 2000; 127: 187.
37. Letellier M., Chevallier F., Clinard C., Frackowiak E., Rouzaud J.N., Béguin F., J.
Chem. Phys. 2003; 118: 6038.
38. Chevallier F., Letellier M., Morcrette M., Tarascon J.M., Frackowiak E., Rouzaud J.N.,
Béguin F., Electrochem. Solid State Lett.2003; 6: A225.
39. C. Conard J., Estrade H., Mat. Science Engineering 1997; 31: 173.
40. Imanishi N., Kumai K., Kokugan H., Takeda Y., Yamamoto O., Solid State Ionics 1998;
107: 135.
41. Tatsumi K., Zaghib K., Sawada Y., Abe H., Ohsaki T., in Rechargeable Lithium and
Lithium-Ion Batteries, Megahed S., Barnett B.M., Xie L. Editors, PV 94-28, p. 97, The
Electrochemical Society Proceedings Series, Pennington, NJ (1994).
SOME THERMODYNAMICS AND
KINETICS ASPECTS OF THE
GRAPHITE-LITHIUM NEGATIVE ELECTRODE
FOR LITHIUM-ION BATTERIES
Rachid Yazami1,2*, Audrey Martinent1 and Yvan Reynier1,2

1
LEPMI, UMR CNRS-INPG 5631, 38041 St Martin d’Hères, France
2
California Institute of Technology, Division of Engineering and Applied Science,
Pasadena, California 91125, USA
1. INTRODUCTION
The chemistry of lithium-ion batteries is based on the lithium-ion
shuttling between the graphite negative electrode and the transition metal(s)
oxide positive electrode. The overall reaction can be schematized as:
LixC6 + LiyMO2 ļ 6C + Liy+x MO2 (1)
Due to differences in the chemical potential of lithium in the LixC6

and Liy+xMO2 materials, the operating voltage of the battery may be as high
as 4V.
The lithiated graphite electrode has many attractive features that
have made it the best choice for negative electrode in commercial batteries.
These include high mass and volume capacity, low operating voltage, long
cycle life and good safety. From the crystal chemistry point of view,
lithiation of graphite undergoes a characteristic series of phase transitions
called stages. An intercalation stage describes the 2D stacking sequence of
the lithium layers between the graphene layers. The stage number is the
number of periodically stacked graphene layers between two adjacent
intercalated layers. Starting from pure graphite, successive stages are formed
as the lithium concentration increases. One way to intercalate lithium into
graphite is by electrochemical cathodic polarization of a graphite electrode in
a Li organic liquid or solid-state polymer electrolyte cell. Based on in situ
and ex-situ x-ray diffractometry, the stages that form as the lithium
concentration increases are: dilute stage 1, stage 4, stage 3, liquid-type stage
2, stage 2 and stage 1 [1-4]. The composition induced stage transitions occur
*
245
246
at relatively well defined ‘x’ values around 0.13, 0.18, 0.23, 0.33 and 0.5
respectively. In dilute stage 1, lithium is randomly distributed between the
graphene layers, which causes a slight shift of the (002) main peak of
hexagonal graphite to lower diffraction angles. Liquid-like stage 2 has
lithium in every other layer, with no long-range in plane ordering. This
contrasts with the stage 2 compound of nominal composition Li0.5C6 (LiC12)
in which the lithium forms an ordered array with the characteristic hexal
structure [5]. The same hexal structure occurs in stage 1 LiC6 in which
lithium occupies each Van der Waals gap between the graphene sheets.
Figure 1 shows typical lithiation/delithiation voltage profile curves
obtained with graphite under a low galvanostatic cycling regime (i.e. C/100
rate) [6]. It shows plateaus and semi-plateaus corresponding to the stage
transitions described above. A flat plateau is the signature of a two-phase
system as in the case in the composition range of 0.5<x<1. It has been well
accepted that the plateau is related to a two-phase system that is a mixture of
stage 2 LiC12 and stage 1 LiC6 phases following the equilibrium reaction:
LiC12 + yLi ļ2y LiC6 + (1-y) LiC12 (2)
According to equation 2, the amounts of each phase should vary

linearly with intercalated lithium intercalated beyond LiC12.
0.4
cellule 'Stella'
0.3
E/ V vs. Li /Li
+
0.2 lithiation
charge
stage 2 ļstage 1
régime C/100
0.1
discharge
décharge
delithiation
0
0 0.2 0.4 0.6 0.8 1
Q /Q =x
m th, m
Figure 1. Voltage vs. composition curves during the

graphite electrode lithiation and delithiation at C/100 rate.
Besides staging, the other important characteristic of the graphite

negative electrode is the formation of the solid electrolyte interphase (SEI)
during the first cycles [7, 8]. The SEI comes mainly from surface reactions
247
between the electrolyte and the graphite electrode. The SEI plays an
important role in the characteristics of the LixC6 electrode including the
irreversible capacity loss and the thermal and chemical stability that govern
the electrode self discharge and cycle life [9, 10].
In this paper we present results from independent studies on the
stage 2 to stage 1 transition area that show some unexpected features
(anomalies). The results are obtained by electrochemical impedance
spectroscopy (EIS), entropy measurements (ǻS(x)) and in situ x-ray
diffractometry (XRD). The aim is to understand the mechanism of stage
transition dealing with the observed anomalies.
2. EXPERIMENTS AND RESULTS
2.1. The Kinetics (EIS) Anomaly

Coin-type half-cells with metallic lithium as the counter (and
reference) electrode and natural graphite as the working electrode were used
for the EIS study. The electrolyte consisted of a molar solution of LiPF6 in
equivolumic mixture of ethylene carbonate (EC) and dimethyl carbonate
(DMC) [5]. The cell was cycled several times between 1.6V and 0V. Starting
from the fully delithiated state at 1.6V, EIS spectra were taken every 200mV
in the voltage window between 1.5 and 200mV and then every 50 mV
between 200mV and 0mV. For the latter, EIS spectra were taken in both
decreasing and increasing voltage sweeps.
Typical Nyquist plots of the Li /LixC6 cell contain three frequency
domains: a high frequency (HF) range covering the ~104-10 Hz, a medium
frequency (MF) range between 10 and 1 Hz, and a low frequency range in
the 10-10-2 Hz. Each frequency region may be associated with various
physical chemical phenomena that were shown and discussed in the
literature [6, 11-21]. The focus of this work is on the anomalous dimensional
changes observed in the HF semi circle during the stage 2-stage 1 transition.
Figures 2a and 2b show a series Nyquist plots during delithiation (a)
and re-lithiation (b) of graphite in the 0V -0.2V voltage window. The HF
area of the plots consists of a semi circle of composition dependant radius R.
Starting from ~0V where the electrode composition is close to LiC6, R
remains constant with increasing voltage. At a threshold voltage, R suddenly
decreases and then remains almost constant until 200mV. When the voltage
is decreased again from 200mV, the diameter does not vary much. It then
increases suddenly to reach a new mostly constant size. If the intercalation -
de-intercalation cycle is repeated again, a similar behavior is observed; that
is the HF semi circle changes suddenly in size at well-defined voltages
during lithium de-intercalation and re-intercalation. The reversible character
248
of the HF semi-circle “breathing” suggests it should be associated with either

surface’ or bulk-related reversible electrode phenomena.
(a) (b)
Figure 2. Evolution of the high frequency EIS semi-circle

during delithiation (a) and lithiation (b) of the graphite electrode.
We found an equivalent electrical circuit that fits best the LixC6

electrode behavior at high frequency. The circuit consists of a resistor R in
parallel with a constant phase element (CPE). The latter is defined with a
pseudo-capacitance Q and a parameter Į with 0< Į <1 [6]. The impedance of
the CPE is expressed as:
1
Z CPE (3)
Q( jZ) D
where j is the complex unity number (j2 = -1) and Z = 2SQ, Q = frequency.
The impedance of the equivalent circuit is therefore:
R
Z circuit (4)
1 RQ( jZ) D
where 2R is the diameter of the HF semi circle.

Figure 3 shows the voltage dependence of R during the delithiation
(3a) and re-lithiation (3b) cycle. In the lithiation step (Fig. 3a), as the voltage
decreases R remains mostly unchanged until it suddenly increases and
reaches a new plateau. The corresponding voltage onset is at about 70mV vs.
Li+/Li. During de-intercalation, R varies a little, as the voltage is more
negative, it then suddenly decreases to stabilize at a lower plateau. The onset
249
voltage here is about 100mV. The onsets of increase and decrease in the R
value occur at the same voltages of 70 and 100mV in the following cycles. It
is interesting that 70 and 100mV match respectively the voltage plateaus
during intercalation and de intercalation in the stage 2 to stage 1 phase
transition showed in Figure 1.
Figure 3. Voltage dependence of the HF semi circle radius during

Li de-intercalation (a) and re-intercalation (b) into graphite.
2.2. The Crystal Structure ‘Anomaly’

In-situ x-ray diffraction (XRD) was performed on a coin type cell
with a 4x6 mm Kapton window coated with conductive thin copper layer.
The graphite electrode was pressed against the Kapton window so as to be
reached by the x-ray beam. After several lithiation/delithiation cycles under
a C/10 rate between 1.5 and 0V, the cell was fully delithiated up to 1.5V.
The cycle capacity achieved with the graphite electrode is about 360mAh/g.
The cell was then re-lithiated under a slower rate of C/20. XRD patterns
were taken for about five minutes every hour while the cell is under
continuous discharge. As result the lithium composition x in LixC6 was
incremented by 0.05 between two successive XRD scans.
Figure 4 shows the evolution of the XRD spectra during the
discharge (lithiation, Fig. 5a) and the charge (Fig. 5b). The diffraction angle
area shown is around the strongest peaks (i.e. the 002 of graphite and stage 2
and the 001 of stage 1). Starting from fully delithiated graphite with 002
peak at around 26.6° (d002=3.35Å, Cu KD radiation) the peak shifted
continuously to lower angles until it reached about 25.2° at x=0.33. This
angle is the one of the 002 stage 2 compound.
The angular position of the peak then remained almost unchanged.
No other peak appeared in the 0.5<x<0.7 composition domain, where stage 1
is expected to form according to the stage 2 – stage 1 two-phase model. Such
a model has been widely used in the literature to explain the flat voltage
250
plateau in the 0.5<x<1 composition range. It was only after x>0.7 that a new
peak at 2T~24° could be noticeable. The latter is the 001 peak of stage 1
corresponding to d001=3.71 Å. As x increased from 0.7 to 1, the 002 peak of
stage 2 decreased in intensity while the 001 peak of stage 1 increased.
Lithiation
(a)
x=1
x=0.05
x=0
24 25 26 27
2θ (°) Cu
Delithiation
(b)
x=0
x=0.07
x=1
24 25 26 27
2θ(°) Cu
Figure 4. XRD (w/ Co K ) spectra during

a lithiation (a) and de-lithiation (b) cycle of graphite under C/20 rate.
The spectra were taken for 5 minutes after each hour charge or discharge.
During lithium de-intercalation the peaks intensities varied in

opposite direction. However the decrease rate of the 001 peak is lower at the
beginning of this process that its increase rate during intercalation. In order
to illustrate this hysteretic behavior, we calculated the average interlayer
spacing <d(x)> based on the integrated 00s peak intensities of stage s. for
pure graphite we took the 002 peak intensity. Taking D(s) the fraction of
stage s in the mixture, <d(x)> is given by:
3.71 3.36(s 1)
d( x ) ! ¦ D s (9)
s s
¦ Ds 1
s
251
Figure 5 displays the composition dependence of <d(x)> during

intercalation and de intercalation. This curve seems to feature four
distinguishable domains.
1. 0<x<0.15 where <d(x)> varies linearly with almost no

hysteresis.
2. 0.16<x<0.33: a small hysteresis appears with <d(x)> is
slightly higher during intercalation than de-intercalation.
3. 0.33<x<0.5: <d(x)> is almost constant around 3.51Å and
4. 0.5<x<1: strong hysteresis where <d(x)> is lower during
intercalation than during de-intercalation (arrows indicate
the HF semi circle ‘anomaly’ (EIS)
Figure 5. Composition dependence of the average graphene interlayer spacing during the
lithium intercalation and de-intercalation.
2.3. The Entropy ‘Anomaly’

We used the temperature dependence of the open circuit voltage
(OCV) at constant composition x in LixC6 to measure the entropy variation
'S(x) expressed by:
w (OCV( x ))
'S( x ) F x
(5)
wT
Figure 6 shows the entropy curve obtained during the lithiation process.
Starting from a high and positive value of 60J/(mol.K) at x~0, ǻS(x)
252
decreases steadily, changes in sign and reaches a local minimum of about

-15J/(mol.K) at 0.4<x<0.5. Then ǻS(x) suddenly increases at x~0.5 and
takes values close to zero then it makes two plateaus before decreasing
sharply in the x=0.9-0.95 composition range.
I' IV II
60
50
40
ǻS (J.mole-1K-1)
30
20
10
-10
-20
0 0.2 0.4 0.6 0.8 1
x of Li xC 6
Figure 6. The entropy of lithium intercalation into natural graphite, after [22].
3. DISCUSSION
3.1. Kinetics and Strains in the SEI
We found three anomalies during the stage 2 to stage 1 phase

transition. The kinetics anomaly observed by EIS in the high frequency area
where the semi circle reversibly changes in size in the 70-100 mV voltage
domain where the transition takes place. There is a slight voltage difference
during intercalation and de-intercalation (~30mV) probably due to kinetics
effects such as phase nucleation and growth.
The origin of the HF semi circle, which is present even at the initial
state of the cell before any cathodic polarization of graphite, has been
discussed in details in the literature [6, 11, 12, 21]. Several
phenomenological models were proposed to deal with the HF semi-circle
including the grains-substrate and grains/grains contact resistances within the
electrode [12]. Electrosorption was also considered in relation with the
voltage dependence of the HF region during the early stage of the cathodic
polarization [6, 21].
The formation of the solid electrolyte interphase (SEI) on the surface
of the graphite grains occurs at electrode potentials above 250mV vs. Li+/Li.
253
The characteristic reduction semi plateau associated with the SEI formation
appears in the 800-600mV region in EC-DMC/LiPF6 electrolyte under C/10
discharge rate [23].
Depending on the graphite and electrode composition, the SEI
formation typically accounts for 5 to 20% of the total electrode capacity loss.
The SEI plays a key role in the LixC6 electrode operation. It has been
described as a heterogeneous multilayers structure [23–25]. The inner layer
is believed to consist mainly of inorganic lithium compounds that result from
the reductive decomposition of the lithium salt anion and solvent. Because of
the high reticular energy of inorganic compounds, the inner layer is expected
to prevent solvent co-intercalation since it acts as a molecular sieve with
efficient lithium ion de-solvation. The upper SEI layer should consist of
organic lithium salts coming from solvent decomposition whereas the
outmost layer has more diffuse structure. It is likely made of oligomers and
polymers, which result from ring opening of cyclic carbonates solvents.
As the graphite electrode reduction proceeds, the lithium ion
transfers from the liquid electrolyte to the graphite surface through the
successive SEI layers. Reduced lithium on the surface will intercalate
between the graphene layers under a diffusion controlled mechanism. During
the lithium intercalation, the average interlayer spacing <d(x)> increases
between x = 0 and 0.33, then stabilizes between x = 0.33 and 0.5 as shown in
Figure 5. Above x = 0.5 <d(x)> increases again although smoothly. The
inner inorganic and rigid SEI layer should accompany the morphological
changes of the active surface. Such changes should induce distortions of the
SEI’ inner layer that most likely affect its lithium ion transport
characteristics in particular increasing its resistivity. We anticipate that the
increase in the interfacial resistance due to distortions of the SEI should be at
the origin of increase of the HF semi-circle size during the stage 2 to stage 1
transition. During the reverse stage 1 to stage 2 transition, <d(x)> decreases
and re-stabilizes below x = 0.5 at the same value than during intercalation.
This should result in strain relaxation and the recovery of the ion
conductivity of the SEI internal layer. Since stage 2 to stage 1 transition is
overall reversible, the HF semi circle will decrease to almost its original
value.
3.2. Possible Origins of the Entropy Anomaly

We attributed the sudden increase in entropy at x~0.5 where stage 1
should begin to form to combined contributions from configurational and
vibrational entropy [22]. Configurational entropy goes with in-plane or the
out-of-plane lithium vacancy ‘disorder’. This translates into either the
formation of lower Li density layers or the presence of Li stacking faults.
Indeed a possible mechanism of the stage 1 phase formation may involve a
254
‘dilute lithium layer’ (noted dil.-Li) in similar way than in dilute stage 1 that
forms at early stages of the graphite lithiation. Such a compound would have
alternating ‘normal’ Li layer (Li) with the hexal structure and a dilute lithium
layer following the sequence (Li)-G-(dil.-Li)-G. Figure 7 illustrates the
structure of such pseudo stage 1 compound. Another source of disorder is the
formation of “fractionary” stages such as stages 3/2 and 4/3 [26]. The latter
have the following stacking: (Li)-G-(Li)-G-G and (Li)-G-(Li)-G-G-(Li)-G
respectively. Figure 7 illustrate the fractionary stages as possible path during
the formation of stage 1, the other path involves the pseudo stage 1
formation. A combination of fractionary stages may also be considered to
deal with the lithium layers stacking disorder during the early formation of
n 1
stage 1. The most likely fractionary stages are the ; (n>1) stages,
n
which consists of a periodic stacking of n lithium layers separated by n+1
carbon layers. Is such a phase the periodicity along the c axis is:
c = 3 .35 + n 3.71 (Å) (6)
The strongest x-ray diffraction line is then the (00 n+1) line, with
corresponding reticular distance d(00 n+1) of:
3.35 3.71 n
d(00n+1)= (7)
n 1
and Bragg diffraction angle 2ș of:
(n 1)O
2T 2Arc sin (8)
2(3.35 3.71 n )
The Bragg diffraction angle 2ș is a decreasing function of n and

varies between 2șa25.2q and 2ș~24° corresponding to the pure stage 2 (n=1)
and pure stage 1 (n= f ) respectively. Since we didn’t observe any
continuous shift of the 002 peak of stage 2 towards lower angles in the
0.5<x<~0.7 composition domain and no 001 peak of stage 1 compound were
detected in that domain, we suggest the formation of superdense stage 2
compound. As a matter of fact the average interlayer spacing in superdense
stage 2 compound should be very close to that in normal stage 2. The
difference lies in the in-plane density of the lithium atoms as illustrated in
Figure 8-a.
The hysteresis in the average interlayer spacing in the 0.5<x<1
composition range during intercalation and de-intercalation in Figure 5
suggests that different crystallographic paths are followed during
intercalation and de-intercalation.
255
3 4
Stage Stage
2 3
Stage 2 Stage 1
Pseudo stage 1
Figure 7. Possible paths of intermediary phases between stage 2 and stage 1.
As in the latter the interlayer spacing is higher than in the former, it

is very likely that starting from LiC6, lithium is de-intercalated from each
equivalent layer, which preserves the stage 1 stacking sequence as sketched
in Figure 8-b. This implies that at the early stage of de-intercalation, lithium
layer have cation vacancies, which may be a mirror phase to dilute stage 1
formed at early steps of intercalation. Li vacancies would then form a dilute
stage 1 compound keeping almost constant the average interlayer spacing.
a)
Stage 2 Superdense stage 2 Stage 1

b)
Stage 1 Dilute Li layers stage 1 Stage 2
Figure 8. Schematic representation of a) the stage 2 ĺ stage 1 transition through superdense

stage 2 compound such as Li1.5C12 (or Li0.75C6) and, b) stage 1 ĺ stage 2 transition involving
dilute stage 1 compound.
256
In super dense phases, the average Li-Li distance within the layer is
lower than 4.26 Å of the hexal structure of normal stage 2 and stage 1
compounds. In bcc lithium, the Li-Li distance is 3.04 Å higher than 2.46 Å
in stage 1 superdense LiCn phases (n<6) shown in Figure 9. The latter phases
cannot be formed by electrochemical intercalation under normal pressure
and temperature conditions [27-29]. Rather they are typically obtained under
high pressure/high temperature synthesis conditions [30] or by ball milling
[31]. Stage 1 superdense LiC2 decomposes to intermediary and “kinetically”
more stable stage 1 phases such as Li7C24 and Li7C24 [32]. The ultimate phase
of decomposition is LiC6. All superdense phases described in the literature
have the lithium sub-lattice commensurate with that of graphite, with short
Li-Li distances. The latter are at origin of metastability as the electrostatic
Li-Li repulsion becomes dominant. Should superdense stage 2 form as
intermediary phase between stage 2 and stage 1, its stabilization may go
through a change in the Li in-plane ordering. This may involve a non-
commensurate Li sub-lattice having a Li-Li distance lying between 2.46 Å
and 4.26 Å of LiC2 and LiC6 respectively. More work is necessary to
confirm the formation of superdense stage 2 and eventually the type of
lithium in-plane organization.
LiC6 (hexal structure) LiC4
LiC2 LiC3
Figure 9. In-plane organization of the LiCn (n6) phases including some superdense phases.
The vibrational entropy is related to a disorder originated from the

Li vibrational motion about the equilibrium position at the center between
two adjacent carbon hexagons. Such a motion can be decomposed in two
257
terms: a vibration along the c axis and a vibration within the medium plane
between the graphene layers [22]. Each vibration mode is defined by its
Debye temperature, T and Tfor movement perpendicular and parallel to the
graphene plane respectively. The vibrational entropy during the stage 2 to
stage 1 transition is given by:
'Svib=R(ln(TTln (T1/T2)) (8)
The Debye temperatures of stages two and one were determined by

inelastic neutron scattering measurements [33]. The total entropy variation
using equation 8 is in the order of about 2 J/(mol.K). Although smaller in
value, such variation accounts for 10-15% of the total entropy and should not
be neglected. We are currently carrying on calculations of the vibrational
entropy from the phonon density of states in LixC6 phases.
4. CONCLUSION
Independent EIS, XRD and entropy measurements show some

anomalies at the stage 2 ļ stage 1 phase transition. It seems there are
converging evidences that for compositions between 0.5<x<1 in LixC6 the
compound doesn’t consist of a linear mixture of stage 2 and stage 1 as
suggested by equation 2. Rather intermediary phases may form such as
superdense stage 2 and Li vacancy containing stage 1. Fractionary stages are
also possible, which may increase the configurational entropy at transition.
We also showed that the crystallographic paths may be different during the
transition one way the other. These paths involve intermediary states with
close in energy. Phase transition kinetics should account for the differences.
We suggested a mechanism involving distortion of the inner
(inorganic) layer of the SEI to deal with the “breathing” of the high
frequency semi circle in the Nyquist plots. Such a distortion should be
induced by a large (>5%) volume change during the stage 2 ļ stage 1
transition. More studies are needed to confirm some of the models we
proposed including theoretical calculations and in-situ NMR and diffraction
studies.
ACKNOWLEDGEMENTS
The authors thank Prof. Brent Fultz from Caltech for fruitful
discussions and Prof. Claude Montella from LEPMI-Grenoble for his
important input in the EIS part of this paper. They acknowledge the financial
258
supports from Ato-Fina S.A. and CNRS both from France and the US-DOE
through Basic Energy Sciences Grant DE-FG03-00ER15035.
REFERENCES
1. R. Yazami, P. Touzain, J. Power Sources 9 (1983) 365.
2. J.R. Dahn, Physical Review B. 44 (1991) 9170.
3. D. Billaud, F.-X. Henry, P. Willman, J. Power Sources 54(1995)207.
4. D. Aurbach, B. Markovsky, I. Weissman, E. Levi, Y. Ein-Eli, Electrochim. Acta 45
(1999) 67.
5. D. Guérard and A. Hérold, Carbon 13 (1975) 337.
6. A. Martinent Ph. D. Thesis, Univ. Grenoble (2001)
7. R. Fong, U. Von Sacken, J.R. Dahn, J. Electrochem. Soc. 137 (1990) 2009.
8. R. Yazami, D. Guérard, J. Power Sources 43-44 (1992) 39.
9. R. Yazami, Electrochemica Acta 45 (1999) 87.
10. R. Yazami and Y. F. Reynier, Electrochemica Acta 47 (2002) 1217.
11. E. Barsoukov, J.H. Kim, C. Oh Yoon, H. Lee, J. Electrochem. Soc. 145/8, 1998, 2711.
12. Y.C. Chang, H.J. Sohn, J. Electrochem. Soc. 147/1, 2000, 50.
13. A. Funabiki, M. Inaba, Z. Ogumi, J. Power Sources 68, 1997, 227.
14. M.D. Levi, D. Aurbach, J. Phys. Chem. B 101, 1997, 4641.
15. E. Barsoukov, J.H. Kim, C. Oh Yoon, H. Lee, J. Electrochem. Soc. 145/8, 1998, 2711.
16. P. Liu and H. Wu, J. Power Sources 56, 1995, 81.
17. B. Markovsky, M.D. Levi, D. Aurbach, Electrochim. Acta, 43/16-17, 1998, 2287.
18. A. Funabiki, M. Inaba, Z. Ogumi, S. Yuasa, J. Otsuji, A. Tasaka, J. Electrochem. Soc.,
145/1, 1998, 172.
19. D. Aurbach, M.D. Levi, E. Levi, A. Schechter, J. Phys. Chem. B, 101, (1997) 2195.
20. T. Piao, S.M. Park, C.H. Doh, S.I. Moon, J. Electrochem. Soc. 146/8, (1999) 2794.
21. M. Holzapfel, A. Martinent, F. Alloin, B.Le Gorrec, R.Yazami, and C. Montella, J.
Electroanal.Chem. 546: 41-50 APR 10 2003
22. Y. Reynier, R. Yazami, B. Fultz, J. Power Sources 119 (2003) 850-855.
23. S. Genies, Ph. D. Thesis, Univ. Grenoble (1998).
24. R. Yazami, Proc. Int. Semin. Prim. Sec. Batt. Deerfield Beach, Florida, 1997
25. E. Peled, D. B. Tow, A. Merson, A. Gladkich, L. Burstein, D. Golodnitsky, J. Power
Sources 97-8 (2001) 52.
26. C. D. Fuerst, J. E. Fischer, J. D. Axe, J. B. Hastings and D. B. McWhan, Phys. Rev.
Lett., 50 (1983) 357.
27. R. Yazami, A. Cherigui, V. A. Nalimova and D. Guérard in Proc. Of the 182nd Meeting
of the Electrochemical Society, 1992 - Toronto, Canada, p. 17.
28. C. Bindra, V. A. Nalimova, D. E. Sklowsky, Z. Benes and J. Fischer, J. Electrochem
Soc. 145 (1998) 2377.
29. R. Tossini, R. Janot, F. Nobili, D. Guèrard and R. Marassi, Electrochem. Acta 48 (2003)
1419.
30. K. N. Semenenko, V. V. Adveev, V. Z. Mordkovich, Doklady An. SSSR 271 (1983)
1402 and V. A. Nalimova, V. V. Adveev and K. N. Semenenko, High Pressure Res. 6
(1990) 11.
31. D. Guérard and R. Janot, J. Phys. Chem. Sol. 65 (2004) 147.
32. J. Conard, V. A. Nalimova and D. Guérard, Mol. Cryst. Liq. Cryst. 245 (1994) 25.
33. R. Moreh, N. Shnieg and H. Zabel, Phys. Rev. B 44 (1991) 13111.
CHARACTERIZATION OF ANODES BASED ON
VARIOUS CARBONACEOUS MATERIALS
FOR APPLICATION IN LITHIUM-ION CELLS
Alexandr N. Kozhevnikov, Yuri A. Podalinski, Olga R. Yakovleva,

Vladimir S. Kotlyar*, Mikhail E. Petropavlovski, Victor G. Smirnov, and
Vladimir V. Dzhurzha
Accumulator Research and Design Institute "Istochnik"

10 Dahl street, St.-Petersburg 197376, Russian Federation
Abstract
Testing a number of carbon materials was carried out with the

purpose to choose one or several for application as a reagent of the anode for
Lithium-Ion cells for their experimental and batch production. The testing
consists of cycling of an electrode of the material being examined against
Lithium metal electrode, or LiCoO2 electrode. The basic control parameters
were: specific weight capacity, irreversible capacity of the first cycle,
capacity degradation degree during cycling. The best parameters among the
tested anodic samples have the materials based on natural graphite “Formula
BT SLC 1115” and “Formula BT SLA 1020” by Superior Graphite Co.
(USA), and graphite SPG (Russian Federation).
Keywords
Lithium-Ion cell, graphite, specific capacity, irreversible capacity.
1. INTRODUCTION
The characteristics of a carbon material used as active reagent of the
negative electrode (anode) of a Lithium-Ion cell considerably influence the
power characteristics of the cell as a whole. Thus, the major parameters are:
the values of specific capacity per unit weight and volume, and also the
*
259
260
value of irreversible capacity. The last parameter determines the amount of

superfluous active material of the positive electrode, necessary for formation
of Solid Electrolyte Interface (SEI) of the anode on the first cycle of life of
the cell [1]. SEI is determined as a surface film of solid electrolyte,
permeable to diffusion by the ion of Lithium (Li+). SEI is formed as a result
of interaction of the active surface of the carbon anode with the electrolyte.
Presence and the properties of SEI on the surface of the carbon anode
determines such characteristics of a Lithium-Ion cell, as retentivity of the
charge, degree of reversibility of the capacity in a “charge-discharge” cycle
and the rate of capacity fade during long cycling of the cell. It is accepted,
that ionic conductivity of SEI is caused by presence of insoluble Lithium
carbonates in the film, during formation of which a part of Li+, passing into
the electrolyte from the lithiated reagent of the cathode, is being spent [1].
In other words, SEI is formed at the expense of loss of a part of
capacity of the cathode (the ways of preliminary chemical processing of a
carbon material with the purpose of formation of SEI are not considered in
this paper). Taking into account, that the cathode reagent is the heaviest and
most expensive component of a Lithium-Ion cell, it is desirable, that anode
reagent, having the largest possible specific capacity, would require the least
quantity of Li+ for formation of SEI. In turn, the characteristics of a carbon
material depend on the set of factors (the nature of a material, the technology
of its processing, its particle size distribution, true surface area, etc.). From
the point of view of practical application, the characteristic of a carbon
material can be most completely and most precisely evaluated by carrying
out the “charge-discharge” cycles. The models of electrode of the material
under examination were cycled in the model of a "half-element" against a
Lithium electrode, or in the model of a cell against an electrode of lithiated
metal oxide, or other cathode capable to reversibly intercalate Li+. In this
work, an attempt to evaluate the characteristics of a set of carbon materials
by the approach mentioned above was carried out. These materials are: pyro-
carbon; nano-tubes; "expanded" graphite; spectrally pure graphite; the
material based on natural graphite developed by RIECP (the above are
produced in Russia); the materials based on natural graphite "Formula BT
SLA 1020" and "Formula BT SLC 1115" by Superior Graphite Co. (USA).
2. TECHNIQUE
The electrodes for examination were produced by smearing of
suspension of active materials in the solution of binding substance -
polyvinylidenefluoride (PVDF) - on the metal substrate by means of
applicator "Doctor Blade" by Hohsen Corp. N-methyl-2-pyrrolidone was
used as the solvent for the binding substance.
261
For manufacturing of negative electrodes, suspensions containing

45wt% powder of the carbon material being examined, 5wt% PVDF, and
50wt% of the solvent. Copper foil with thickness of 0,02 mm was applied as
substrate.
For manufacturing of positive electrodes, pastes with the following
ratio of the ingredients were applied: Lithium cobaltate by Merck or by
"Baltiyskaya Manufaktura" (Russia) - 42,5wt%, conductive additive
(acethylene soot) - 3,5wt%, PVDF – 4wt%, solvent - the balance.
Aluminium foil with the thickness of 0,02 mm was used as a current
collector.
Removal of the solvent was carried out by exposure of intermediate
products of electrodes at the temperature 130-150ºɋ within 30 - 40 minutes.
Then the intermediate products were rolled for giving the required density to
the electrode layer; finished electrodes were cut from the obtained tape by
means of the cliché.
In the case of using the metallic Lithium electrode, it was cut out of
Lithium foil with the thickness of 0,2 mm.
As the electrolyte, 1M solution of Lithium hexafluorophosphate in
the mixture of ethylene carbonate and diethyl carbonate (1M LiPF6
EC:DEC=1:1 - electrolyte LP-40 by Merck) was used.
As the separator, microporous polypropylene film (PORP by NPO
UFIM, Russia) with the thickness of 25 Pm was used.
Lithium
Current
electrode
collector
Carbon
Clip electrode
3-layers
material
Figure 1. Ɍest cell in "soft" pouch.
On the Figures 1 and 2, the schematical drawings of cells for

“charge-discharge” cycles are shown. On the Fig. 1, the cell in the "soft"
pouch, formed by welding of three-layer material (lavsan-aluminium-
polyethylene) on polyethylene layer, is shown.
During the cycling, the cell was compressed by a clip with the
purpose to provide dense assembly of the electrodes. By the Figure 2, the
262
cell made with application of parts of "coin" element CR-2016 is presented.

The capsule and the top of the cell are made of Nickel-plated steel, the
sealing lining is of polypropylene. The models of half-elements Li-C and the
cells LiCoO2-C for testing were made both in the first and in the second
variants of cells.
Carbon Separator Lithium
electrode electrode
Figure 2. CR2016 test cell.
All assembly operations were carried out in tight camera in the

atmosphere of dry air with the dew point of -40qɋ. Before the assembly, the
parts of the capsule and the separator were dehydrated under vacuum at 50qɋ
within 24 hours; the electrodes being tested were exposed in vacuum furnace
at 150qɋ within 24 hours.
Charge and discharge current densities were in the range of 0,15-1,0
mA/cm2, the final charge voltage of the model of the cell was 4,18-4,20 V.
During the cycling of the half-elements on the first charge
(intercalation of Li+), the minimal current density was chosen, thus the
theoretical capacity was given to the electrode (calculated from the weight of
reagent and in view on formation of LiC6). The discharge (deintercalation of
Li+) was stopped upon achieving voltage of 1V.
Irreversible capacity was evaluated as the ratio between the
difference of the values of discharge capacity on 3-4th and 1st cycles and the
fixed capacity (3-4th cycles).
3. RESULTS OF EXPERIMENTS
3.1. Nano-Tubes
The samples of nano-tubes designated as NTC and NTA were given
for tests by research-and-production firm "Astrin" (Russian Federation). Due
to limited quantity of the material, the tests had rough character. During the
tests it was revealed, that specific capacity per unit weight of nano-tubes is
comparable to those of other carbon materials. During 50 (for NTC) and 76
(for NTȺ) cycles “charge-discharge”, the tendency to good stability of
discharge capacity was revealed. The tests were stopped on spending of
metal Lithium electrode. Good retentivity of charge in charged state during
263
storage within 50 day was also revealed. The loss of capacity was no more
than 7%. The irreversible capacity was 30r10%.
3.2. Pyro-Carbon
The samples of pyro-carbon (PC) were synthesized by the Central
Research Institute of Materials (Russian Federation). [2]
1,2
Voltage, V
1 a
0,8
0,6
0,4
0,2
0
0 20 40 60
cycle 1 cycle 3 Capacity, mAh/g
1,2
Voltage, V
1
b
0,8
0,6
0,4
0,2
0
0 20 40 60 80 100 120 140 160 180 200 220
cycle 1 cycle 20
Capacity, mAh/g
Figure 3. Discharge curves of PC anodes based on ND (a) and NC (b).
The raw material for the synthesis was methane. Powder of Nickel
carbonyl (NC) or powder of nano-diamond (ND) was the catalyst. Attempts
to synthesize pyro-carbon on copper powder were not successful. Powder
with the composition 70%PC, 30%NC, and also the set of powders with
various ratios of PC and ND were tested. Anodes made of the powder
70PC30NC showed satisfactory cycle behavior and had specific capacity
180 mAh/(g of powder) (260 mAh/(g of carbon)) (Fig. 3a). The anodes made of
powder ɯPCɭND, irrespective of the components ratio, had specific capacity
264
no more than 70 mAh/(g of powder) (Fig. 3b). The results obtained show the
activating influence of Nickel on the processes of synthesis of pyro-carbon
and intercalation of Li+ on it. It should be noted also, that the anodes of
powder 70PC30NC in propylene carbonate electrolyte showed considerably
lower values of capacity, than in electrolyte LP-40 (Figure. 4).
Voltage, V
1,4
1,2
a
1
0,8
0,6
0,4
0,2
0
0 20 40 60 80 100 120 140 160 180 200
cycle 1 cycle 3 Capacity, mAh/g
Voltage, V
1,4
1,2
1
b
0,8
0,6
0,4
0,2
0
0 20 40 60 80 100 120 140 160 180 200 220
cycle 1 cycle 20
Capacity, mAh/g
Figure 4. Discharge curves of PC anodes in

a) 1 M LiPF6 in propilene carbonate + DME
b) 1 M LiPF6 in ethelene carbonate + diethylcarbonate.
265
3.3. Expanded Graphite

The samples of expanded (thermally expanded) graphite (ɌEG) were
provided by the Central Research Institute of Materials (Russian Federation).
TEG is a powder of light grey color with low bulk density and extremely
developed true surface; it can be easily pressed in a flexible plate
(“cardboard”). The anodes made of TEG, had the capacity about 200 mAh/g
(Figure 5). However, in spite of the fact that the anodes were made of
pressed powder, they had volumetric capacity in 3-4 times smaller, than
other tested materials. The discharge curve is shown on Figure 5.
Figure 5. Discharge curves of Thermally Eexpanded Gfraphite (TEG).
1,2
Voltage, V
0,8
0,6
0,4
0,2
0
0 50 100 150 200 250 300 350 400
Capacity, mAh/g
cycle 1 cycle 2 cycle 3 cycle 4
Figure 6. Discharge curves of the anodes based on RIECP carbon.

266
3.4. Carbonaceous Material RIECP

The material was developed by the Research Institute of
Electrocarbon Products (RIECP, Russian Federation) on the basis of natural
graphite specially for application in Lithium-Ion cells. The samples were
given with symbolic numbers, without certificates. The anodes, made of
RIECP powder, had the capacity about 300 mAh/g. The typical dicharge
characteristic is submitted on Fig. 6. Irreversible capacity is about 16%.
Concerning the cycle behavior, this material is not as stable as other ones
tested (Fig. 10).
3.5. SPG Graphite

SPG is a spectrally pure graphite. It is a powder containing rather
large fractions (up to 500 microns). This material was not intended for
application in Lithium-Ion cells and has no technological properties
necessary for manufacturing of a thin electrode. The fraction up to 100
microns was selected from it, and this allowed making the required
electrodes.
Voltage, V
0,9
0,8
0,7
0,6
0,5
0,4
0,3
0,2
0,1
0
0 50 100 150 200 250 300 350 400
cycle 1 cycle 2 cycle 10 Capacity, mAh/g
Figure 7. Discharge curves of C (SPG).
On Figure 7, the typical discharge curves of SPG electrode, obtained

with use of half-element, are shown, and on Figure 8 discharge curves of the
model cells are presented. As it is visible from the diagrams, on SPG
electrodes the value of specific discharge capacity reached 340 mAh/g, the
stability of discharge capacity was observed during 40 cycles (Fig. 10), the
irreversible capacity loss was not more than 20%. The tests are being
continued.
267
Voltage, V
4,3
4,1
3,9
3,7
3,5
3,3
3,1
2,9
2,7
2,5
0,00 1,00 2,00 3,00 4,00 5,00 6,00
cycle 1 cycle 3 cycle 40 Capacity, mAh
Figure 8. LiCoO2 – C(SPG) discharge curves.
1,2
Voltage, V
0,8
0,6
0,4
0,2
0
0 50 100 150 200 250 300 350 400 450
cycle 1 cycle 3 cycle 7
Capacity, mAh/g
Figure 9. Discharge curves of anode Formula BT SLA 1115.
3.6. Performance of Thermally Purified Spheroidal Natural

Flake Graphite SLA 1020 and Surface Coated Natural
Graphite SLC1115
Carbonaceous materials provided by Superior Graphite Co. (USA)
were developed on the basis of natural graphite specifically for application in
Lithium-Ion cells. Samples had designations of “FormulaBT SLA1020” and
“FormulaBT SLA1115”. The anodes made out of these materials, during the
tests in half-elements, have shown high and comparable values of specific
capacity (up to 350 mA·h/g and up to 360 mA·h/g, respectively). Figure 9
shows these results. Both grades demonstrated good stability under cycling
(Fig.10). The tests for stability under cycling will be continued on model
cells with the purpose to prevent the possible influence of the limited
reversibility of metal Lithium electrode on the results of tests. The
irreversible capacity loss on both materials was no more than 20%.
268
400
ɋɪ,mAh/g
350
300
250
200
150
100
50
0
0 5 10 15 20 25 30
RIECP Superior Graphite Co 1020
Superior Graphite Co 1115 C(SPG) N cycle
Figure 10. Capacity fade of anodes based on different materials.
4. CONCLUSIONS
In summary, we wish to note, that the results reported provide only
initial characterization of the materials tested not to the full. The further
profound and extended research and tests in scale of full-sized cell, first of
all concerning to long cycling, is necessary. Nevertheless, in our opinion,
when choosing an anode material for Lithium-Ion cells being developed, it is
necessary to prefer the carbon materials based on natural graphite - “Formula
BT SLA 1020” and “Formula BT SLA 1115” by Superior Graphite Co.
(USA). Spectrally pure graphite (Russia) in its modified form, also can be
recommended for application. Besides, the carbon material RIECP (Russia)
remains a promising candidate for application in question.
ACKNOWLEDGEMENTS
The authors would like to thank all companies, which presented the
samples for tests; Dr. Igor Barsukov (Superior Graphite Co.) for his efforts
in providing of our participation in the conference; U. S. Civilian Research
& Development Foundation (CRDF) for financial support of our trip.
REFERENCES
1. Peled E. Lithium Stability and Film Formation in Organic and Inorganic Electrolyte for
Lithium Battery Systems.// in Gabano J.-P., Ed. Lithium Battery. London etc. Academic
Press, 1983. Pp. 43-72.
2. Gordeev S. K., Grechinskaya A. V., Zhdanov V. V., and Krasnobrizhiy A. V. Patent of
Russian Federation ʋ 2122589.
A CARBON COMPOSITE FOR THE NEGATIVE
ELECTRODE OF LI-ION BATTERIES
A.V. Churikov*, N.A. Gridina, N.V. Churikova

Chemistry Department, Saratov State University,
Astrakhanskaya 83, 410012 Saratov, Russian Federation
Abstract
Composite electrodes made of two carbon components were

evaluated experimentally as anodes for Li-ion batteries. The electrochemical
activity of these electrodes in the reaction of reversible lithium intercalation
from/to a solution of LiPF6 in ethyl carbonate and diethyl carbonate was
studied. Compositions of the electrode material promising for the usage in
Li-ion batteries were found.
1. INTRODUCTION
As it is known, graphite and layered carbon materials capable of
reversible lithium intercalation, which form Li-C compounds of various
compositions are widely used as the active substance of the Li-ion battery
anodes [1-10]. When carbon is used as a matrix for lithium, the structural
characteristics of a specific material are of importance. Though the literature
devoted to correlations between the intercalation depth and various structural
parameters of a material is rather inconsistent, there is, however, a certain
common point of view. From the microscopic standpoint, the optimal carbon
material for the anode of Li-ion batteries should be neither perfect crystalline
nor completely amorphous. It may contain centers of crystalline graphite (a
mesophase) distributed over an amorphous carbon matrix. According to
Dahn et al. [10-13], the suitable carbon materials fall into one of three
classes, namely, graphitic materials, hydrogen-containing materials, and
single-layer hard carbons.
There are several classifications of carbon materials, given in
accordance with their different properties. Hard carbon is meant as a carbon
*
269
270
material whose structure is insensitive to heat treatment. Its precursors

usually contain heteroatoms, mainly oxygen. That is why the hard carbon
structure comprises crosslinking groups to prevent the formation of graphite
stacks and their coalescence at heat treatment [9-13]. The disordered
structure of soft carbons changes with high-temperature treatment towards
the perfect graphite one and approaches it at about 3000qC. These materials
are composed of more or less uniformly oriented crystallites whose sizes and
ordering degree increase with graphitization heat treatment [9].
Graphites exhibit an optimal discharge curve with a low-voltage
plateau and a good intercalation capacity (300 – 370 mAh/g or x = 0.8–1.0 in
the theoretical formula LixC6). Many of them allow long cycle life and a
weak hysteresis of the charge-discharge curves to be achieved. However,
graphites were noted to be less stable at cycling [8,14]. Their degradation is
usually caused by exfoliation of the structure by coarse penetrating particles,
which leads to electrode failure. Besides, graphites were found to be unable
to provide high discharge rates because of slow lithium diffusion [14].
Partially graphitized cokes produced by means of thermal
decomposition of organic raw materials, polymers and graphitized carbon
fibers show a good performance [1,2,6-17]. The properties of carbon
materials are often improved due to large amounts of dopants (H, O, S, N, P,
Si, etc.) [9,18].
Therefore, there is a wide spectrum of carbon materials suitable for
the usage in Li-ion batteries; the choice of a specific one determined by
many factors. According to Ref. 7, the percentage of various carbon
materials used in commercial Li-ion batteries was as follows: graphites – 43
%, hard carbons – 52 %, soft carbons – 5 %.
The interlayer distance d002 and the crystallite sizes La and LC are the
principal microstructural characteristics of carbon. d002 = 0.3354 nm for the
perfect graphite structure while disordered carbon materials have noticeably
larger d002's due to the graphene planes being incompletely parallel to each
other. At the early stage of graphitation (about 1200qC) the crystallites
achieve about 5 nm in both directions with d002 being about 0.344 nm. An
increase of the treatment temperature improves the structure and raises the
crystallinity degree of soft carbon. Simultaneously, the shape of the
discharge curve (lithium deintercalation) gets better; it becomes more and
more horizontal with a sharp potential elevation at complete lithium
extraction, the discharge plateau approaching the potential of pure lithium.
Therefore, a certain correlation between the electrochemical and
crystallographic properties exists and both depend on the perfection degree
of the carbon structure [9].
As to the intercalation capacity, its relation to the carbon crystalline
structure is ambiguous. The reversible capacity is reported to go through
a minimum at d002 = 0.343 nm. The LC dependence of capacity is reported
to be of a similar shape (a minimum near LC § 5 nm, an increase when
271
LC § 1 nm or LC>20 nm). Highly graphitic materials (d002 within 0.335 –

0.338 nm and LC>30 nm) show some increase of the capacity when the
lowest limit of d002 is achieved and LC increases. The specific reversible
capacity vs crystallite size or vs treatment temperature diagrams are often
published. They look as smooth curves with a minimum within medium
temperatures or medium crystallite sizes [7,11,17].
On the one hand, such a behavior of capacity is explained by great
importance of the perfection degree of the structure of graphitic materials at
lithium intercalation. According to numerous papers, the ability of carbon to
reversibly absorb significant amounts of lithium enhances with the
graphization degree (graphitization is meant as the carbon structure getting
more ordered and the areas with perfect graphite structure expanding). On
the other hand, the capacity of heavily disordered carbon materials prepared
by low-temperature (500-1000qC) pyrolysis of organic compounds, on the
contrary, rises with disordering due to involvement of extra mechanisms of
lithium absorption, besides interlayer insertion. The carbon residue after
primary carbonization at 500-600qC has almost the same structure. The
primary crystallite stacks are comprised only of 2–4 almost parallel to each
other graphite planes of 10–20 aromatic rings each [9]. Some of those
disordered substances exhibit supercapacity (i.e. x>1, up to 1 Ah/g) at
lithium absorption [3,7-9,11-14]. The nature of the mechanisms of excessive
lithium storage in low-temperature carbons is different. In particular, a part
of lithium atoms is assumed to be in the interlayer space while the other part
is bounded immediately to carbon atoms with hydrogen atoms involved [19-
21], or the fixed excessive lithium atoms are located on the edges and
breakages of the graphite structure [22]. The lithium immediately bounded to
a carbon atom is fixed much stronger than that in the interlayer space of
graphite, thus the bounded Li splits out at a more positive potential. The
uncertainty of the nearest environment in a disordered structure leads to an
almost linear shape of the discharge curve with no noticeable plateau
observed. Thermal treatment of such materials causes a sharp drop of their
capacity with its minimum within 1500-2000qC, and then, when approaching
3000qC, the capacity increases due to structure becoming more graphitic.
There are some demerits of such materials, namely, instability of the
structure at lithium insertion/extraction, a large irreversible capacity for the
first charging, and a strong hysteresis of the charge-discharge curves [10-13].
Though a common viewpoint has been produced as to the structure-
property relationship after intense studies of this problem, there remain many
contradicting examples.
The replacement of an individual carbon material by a mixture
(composite) in order to improve the electrode's performance seems
promising [23]. Discharging at low potentials (0–0.25 V vs Li/Li+) has been
proven to be provided by the graphite-structure component while it is the
non-graphitic one, which works above 0.25 V [4,9]. The present work deals
272
with mixtures of two sorts of carbon. Substances are rather different in

structure, e.g. non-graphitic carbon and crystalline graphite (graphite and
low-temperature coke) should be taken as source materials for making a
composite. The composite prepared may have a rather negative discharge
plateau like graphite with a long-term cycling life of the electrode to be
provided by amorphous carbon. Using such an approach, we expected the
mixed electrodes to combine the advantages of both carbon materials and to
compensate their demerits.
The goal of the present work was to examine binary carbon +
graphite mixtures in reversible electrochemical intercalation of lithium from
a nonaqueous solution. The following carbon materials different in structure
were chosen: soft carbon prepared by means of pyrolysis of bitumus oil and
thermally expanded graphite (TEG). Thermally expanded graphites have a
larger d002 distance, and lithium insertion distorts their crystal lattice weaker
in comparison with natural and artificial graphites.
2. EXPERIMENTAL
The thermally expanded graphite was used without any preliminary

treatment; the carbon was synthesized by carbonation of the raw material in
an argon flow at 500qC with subsequent thermal treatment. Carbon residues
strongly differing in the crystallinity degree were obtained as depends on the
temperature which ranged from 500qC to 1300qC. The identity of each
material was checked by X-ray spectroscopy.
The working electrodes were prepared on a nickel support with
polytetrafluoroethylene in N-methylpyrrolidone as a binder. All the
operations of assembling and hermetization of cells were carried out in a
glove box. Electrochemical measurements were made in mockups of the
elements with lithium counter electrode and reference electrode filled with a
1 M LiPF6 solution in a mixed solvent (ethylene carbonate (EC) + diethyl
carbonate (DEC), both from Merck). Charge-discharge characteristics were
registered on a series of samples at 35qC. Cycling was performed with a
current density 0.2 mA/cm2, charging/discharging were advanced to
potential E = 0 and 3 V (relative to the Li/Li+ electrode), respectively.
The source carbon materials show a significant electrochemical

activity for lithium intercalation though the reversible capacity is relatively
low and tends to reduce with cycling. For the thermally expanded graphite
273
the reversible capacity was about 220 mAh/g and 290 mAh/g at discharge to
0.5 and 3 V, respectively.
The mesocarbons prepared by bitumus oil pyrolysis at low
temperatures (500 and 600qC) turned out to be practically incapable of
reverse functioning as a lithium intercalating material. The highest values of
the specific capacity (200 mAh/g) were obtained for the carbons synthesized
within 700–1100qC. They also show a tendency of gradual reduction of
performance characteristics with cycling. Increasing the pyrolysis
temperature up to 1300qC led to a decrease in the specific capacity. A
positive effect of the treatment temperature on the shape of the discharge
curve was noted. The substances obtained by low-temperature pyrolysis have
a practically linear discharge curve. As the pyrolysis temperature rises, the
discharge curve undergoes flattening, i.e. the greater part of the discharge
capacity is released by the electrode at low potentials. These observations are
consistant with the literature data for graphitized carbon materials [7-14].
Thus, the electrochemical properties of the individual carbon
materials are not so high as to enable their commercial usage in Li-ion
batteries. In order to improve the performance, we started making composite
materials from two individual carbon ingredients. Figure 1 shows a typical
result of electrochemical tests of an electrode made of a blend of graphite
and soft carbon treated at 1100qC (C1100) in comparison with the
discharge curves of the individual constituents.
2
C1100 TEG
1.5
Voltage (V)
0.5
C1100+TEG
0
0 100 200 300 400
Capacity (mAh/g)
Figure 1. Discharge curves (10th cycle) of the electrodes made of the individual carbon
materials and the 1:1 composite. The current density 0.2 mA/cm2.
274
A positive effect of blending was observed in each case, i.e. the

cycling capacity of the mixed electrode was significantly higher than that of
the electrodes made of the constituents instead of the arithmetic mean which
one might expect. At the same time, the composite carbon+graphite
electrode preserved its discharge plateau near 0.1 V, similar to pure graphite.
The C1100+TRG electrodes have exhibited the highest and most stable
characteristics of all those under study.
Experiments aimed at determining the optimal ratio of the
components in the composite were conducted. The maximum capacity
corresponds to a 1:1 wt ratio while the 7:3 and 3:7 electrodes show a
somewhat lower capacity.
800 C1100+TEG
C800+TEG
Capacity (mAh/g)
600 C600+TEG
400
200
0
0 5 10 15 20 25
Cycle number
Figure 2. Discharge capacity of the electrodes of different composition versus the number of
cycles. The cycling is within 0 – 3 V (solid lines) and 0 – 0.5 V (dashed lines).
The current density is 0.2 mA/cm2.
Some data on the cycling stability of the composite electrodes are

presented in Figure 2 as the discharge capacity vs the number of cycles.
Practically, of the most importance is the behavior of the negative electrode
within 0 – 0.5 V where the major part of the capacity stock is realized. Thus,
we used two modes of cycling, namely, soft discharge to 0.5 V in order to
prolong the cycle life, and deep discharge to 3 V in order to evaluate the
maximum capacity released by the material. The C1100+TRG electrodes
have exhibited the best performance, the cycling efficiency approaching
100 %. The average capacity at 0.5 V discharge was 300 and 265 mAh/g for
the 1:1 and 3:7 (7:3) carbon : graphite ratio, respectively. This index was
lower for C800+TRG (240 mAh/g at a 1:1 wt ratio) and much lower for
275
C600+TRG (200 mAh/g at a 1:1 wt ratio). Thus, decreasing the treatment

temperature worsens the capacity characteristics of the composite electrode
as well as those of the individual components.
300
Capacity (mAh/g)
200
100
0
0 0.5 1 1.5
2
Current density (mA/cm )
Figure 3. Dependence of the discharge capacity of the C1100+TRG electrodes

on the charge-discharge current density.
However, deep discharge is a different matter. As Figure 2

demonstrates, values of the discharge capacity over 400 mAh/g were
registered for the C1100+TRG electrodes at discharge to 3 V, which are
more than the theoretical limit of 372 mAh/g at x=1. A supercapacity
considerably exceeding the theoretical limit was detected at deep discharge
of the composites with the low-temperature carbon. The electrodes
containing the carbon treated at 600(C show the initial values of the
discharge capacity up to 900 mAh/g. These results for the carbon not
subjected to high-temperature treatment are quite comparable to the
literature data on the supercapacity of disordered carbon structures formed
by low-temperature pyrolysis of organic substances [3,7-9,11-14]. This
supercapacity disappears at potentials above 0.5 V, and then, when E<0.5 V,
the discharge capacity, on the contrary, falls with the treatment temperature,
which is also in good agreement with the literature data on the role of
graphitic and non-graphitic components in lithium absorption.
Attention should be paid to the steadiness of the average discharge
capacity in cycling to 0.5 V and the gradual decrease of this value in the
course of deep cycling. The electrodes with low-temperature C600 exhibit
rather fast reduction of the discharge capacity while those with C1100 and
C800 reduce this value slower. These facts are in accord with some literature
data on the instability of disordered carbon structures at cycling [10-13].
The influence of the charge-discharge current density on the specific
capacity properties of the electrodes is depicted in Fig. 3. One can conclude
that even at a 0.2 mA/cm2 current density the potentialities of the material
are not developed fully. Further increasing the cycling current density up to
276
1.5 mA/cm2 led to a threefold reduction of the capacity taken up. However,
the usage of a combined mode of battery charging (the galvanostatic mode in
combination with the potentiostatic one) instead of constant current charging
allows to avoid deterioration of the electrode under high current loads.
4. CONCLUSION
The electrochemical activity of thermally expanded graphite and soft
carbon produced by means of bitum oil pyrolysis was studied in reversible
lithium intercalation from a 1 M solution of LiPF6 in the EC+DEC mixture.
The effect of mixing different-type carbon materials for improvement of
the performance is proposed to be employed. The cycling capacity of the
composite electrode was found to be higher than that of the electrodes
made of the constituents. The electrodes made of the 1:1 blend exhibit the
maximum discharge capacity. The highest discharge characteristics were
found for the thermally expanded graphite + soft carbon (treated at 1100qC)
system, namely, a capacity of 300 and 400 mAh/g at discharge to 0.5 and 3
V, respectively; the Coulombic efficiency approaching 100%.
REFERENCES
1. Scrosati B. J. Electrochem. Soc. 1992. V.139. P.2776.
2. Abraham K.M. Electrochim. Acta. 1993. V.38. P.1233.
3. Armstrong A.R., Bruce P.G. Nature. 1995. V.373. P.598.
4. Fujimoto H., Mabuchi A., Tokumitsu K., Kasuh T. Carbon. 2000. V.38. P.871.
5. Levi M.D., Aurbach D. Electrochim. Acta. 1999. V.45. P.167.
6. Noel M., Santhanam R. J. Power Sources. 1998. V.72. P.53.
7. Endo M., Kim C., Nishimura K., Fujino T., Miyashita K. Carbon. 2000. V.38. P.183.
8. Broussely M., Biensan P., Simon B. Electrochim. Acta. 1999. V.45. P.3.
9. Flandrois S., Simon B. Carbon. 1999. V.37. P.165.
10. Dahn J.R., Zheng T., Liu Y., Xue J.S. Science. 1995. V.270. P.590.
11. Buiel E., Dahn J.R. Electrochim. Acta. 1999. V.45. P.121.
12. Xing W., Xue J.S., Zheng T., Gibaud A., Dahn J.R. J. Electrochem. Soc. 1996. V.143.
P.3482.
13. Xing W., Xue J.S., Dahn J.R. J. Electrochem. Soc. 1996. V.143. P.3046.
14. Shi H. J. Power Sources. 1998. V.75. P.64.
15. Mabuchi A., Fujimoto H., Tokumitsu K., Kasuh T. J. Electochem. Soc. 1995. V.142.
P.3049.
16. Takami N., Satoh A., Ohsaki T., Kanda M. J. Electrochem Soc. 1998. V.145. P.478.
17. Takami N., Satoh A., Ohsaki T., Kanda M. Electrochim. Acta. 1997. V.42. P.2537.
18. Larcher D., Midalige C., Gharghouri M., Dahn J.R. Electrochim. Acta. 1999. V.44.
P.4069.
19. Peled E., Eshkenazi V., Rosenberg Y. J. Power Sources. 1998. V.76. P.153.
20. Wang Z, Huang X., Xue R., Chen L. Carbon. 1999. V.37. P.685.
21. Zheng T., McKinnon W.R., Dahn J.R. J. Electrochem. Soc. 1996. V.143. P.2137.
22. Wu Y.-P., Wan C.-R., Jiang C.-Y., Fang S.-B., Jiang Y.-Y. Carbon. 1999. V.37. P.1901.
23. US Patent 5,677,082. Compacted carbon for electrochemical cells (Prior. Oct.14, 1997).
ELECTROCHEMICAL INTERCALATION OF
PF6 AND BF4 INTO SINGLE-WALLED
CARBON NANOTUBES
Rachid Yazami1, Irina V. Goncharova2* and Vladimir N. Plakhotnik2

1
CNRS-Grenoble, PB 75, 38401 Martin d’Heres, France and
California Institute of Technology, Pasadena, California 91125, USA
2
Department of Chemistry and Engineering Ecology,
Dniepropetrovsk National University of Railway Transport, Dniepropetrovsk 49010, Ukraine
Abstract
Thin films single-walled carbon nanotubes (SWNTs) are used as the

positive electrode in lithium half-cells with LiPF6 and LiBF4 in liquid
organic electrolytes. Upon driving the potential to a high 4.5V, oxidation
peaks appear in the cyclic voltammograms, suggesting the anions may
intercalate into the SWNT triangular structure. By analyzing the shape of
voltammogram changes upon cycling, we found that the anion intercalation
into SWNT is partly reversible.
Keywords
Single-walled carbon nanotubes; SWNT; anion; intercalation.
1. INTRODUCTION
Since their first discovery by Iijima in 1991 [1], carbon nanotubes
have attracted a great deal of interest due to their very exciting properties.
Their structure is characterized by cylindrically shaped enclosed graphene
layers that can form co-axially stacked multi-wall nanotubes (MWNTs) or
single-walled nanotubes (SWNTs). Like in graphite, carbon atoms are
strongly bonded to each other in the curved honeycomb network but have
much weaker Van der Waals-type interaction with carbons belonging to
*
277
278
adjacent nanotubes. Such a weak interaction is believed to be at the origin of

the bundle structure of carbon nanotubes. Intercalation compounds have
been reported in MWNTs and SWNTs, with intercalated species occupying
the Van der Waals gaps. Most reported intercalation compounds were
concerned with alkali metals [2, 3]. So far, there have been no reports on
anion intercalation although covalent compounds especially with fluorine [4]
were successfully prepared. In this paper we have performed some
preliminary experiments on the anion electrochemical intercalation into
SWNTs with the following reaction equation:
nC + A- ĺ CnA + e- (A- = PF6 or BF4 ) (1)
2. EXPERIMENTAL
SWNT obtained by pulsed laser ablation of graphite [5], were shaped
into a thin film (buckypaper) of approximately 50 Pm in thickness. XRD (Co
K ҏradiation) and TEM (high resolution imaging and electron diffraction)
were used to characterize their microstructure. Discs of 6 mm in diameter
were cut from the film and dried in vacuum at 200q C for 12 hours before
their use as positive electrodes in coin-type (2016) cells. Cells contained
metallic lithium as the negative pole and a polypropylene microporous
separator soaked in EC:DMC-LiPF6 or EC:DMC-LiBF4 molar solutions as
electrolyte. The cells were then cycled between initial open-circuit voltage
and 4.5V, and then between 4.5V and 2V by cyclic voltammetry technique
under 70 PV/s voltage sweeping rate.
Gr (002)
10.
61.
11. 31.
10 20 30 40 50 60
2 Theta (degree Co KD )
Figure 1. XRD pattern of purified SWNT.

279
3.1. Structural Characterization

Characterization of the initial SWNTs included XRD and HR-TEM
imaging and diffraction. Fig. 1 shows the XRD chart and Fig. 2 the HR-TEM
image of the starting SWNTs. The low angle XRD diffraction peaks are
characteristic of the bundle structure, with characteristic hk lines. The ‘a’
crystal parameter, the triangular structure is about 1.7 nm. Graphite impurity
(<5%) is clearly evidenced by the (002) characteristic peak. The bundle
structure is visible in the HR-TEM image in Fig. 2a. Fig. 2b shows typical
diffraction rings of the triangular structure with corresponding indexation.
(a) (b)
Figure 2. HR-TEM of SWNTs. Insert show the diffraction rings.
3.2. Electrochemical Characterization

Figure 3 and Figure 4 show the first cyclic voltammogram (CV)
obtained with PF6 and BF4 , respectively. Arrows in the figure indicate the
direction of the voltage sweep. During the first voltage sweep, starting from
ca. 1.1 V vs. Li+/Li, the oxidation current remains very low until a threshold
value of about 4.15V is reached, then it increases sharply. When the sweep is
reversed after 4.5V, a strong oxidation peak appears at about 4.4V followed
by several other oxidation peaks especially around 3.5V. The current is then
reversed and a reduction peak appears in the 2.5-2.2V area.
In the following cycles, the CV pattern does not show anymore the
4.15V anodic potential threshold, indicating a lowering of the anion
intercalation overpotential.
280
st
1 cycle
-5
8 10
-
6 10
-5
PF
6
-5
4 10
-5
2 10
-5
-2 10
-5
-4 10
1 1.5 2 2.5 3 3.5 4 4.5 5
Voltage(V vs. Li)
Figure 3. First CV (70 mV/s) of Li/EC:DMC-LiPF6/SWNT cell.
st
1 cycle
-5
1.5 10
-
1 10
-5 BF
4
-6
5 10
-6
-5 10
-5
-1 10
-5
-1.5 10
1.5 2 2.5 3 3.5 4 4.5 5
Voltage(V vs. Li)
Figure 4. First CV (70 mV/s) of Li/EC:DMC-LiBF4/SWNT cell.
New anodic and cathodic peaks appear which indicates some degree
of reversibility of the anion intercalation. Figure 5 and Figure 6 show the
second cyclic voltammogram (CV) obtained with PF6 and BF4
respectively. However, the anodic part remains dominant.
281
nd
nd 2 cycle
2 cycle 4 10
-5
-5
2 10
-
- -5 BF
1.5 10
-5
PF 3 10
4
6
-5
1 10 -5
2 10
-6
5 10
-5
1 10
0
-6 0
-5 10
-5
-1 10 -1 10
-5
-5
-1.5 10 2 2.5 3 3.5 4 4.5
2 2.5 3 3.5 4 4.5
Voltage(V)
Voltage(V vs. Li)
Figure 5. Second CV (70 mV/s) Figure 6. Second CV (70 mV/s)

of Li/EC:DMC-LiPF6/SWNT cell. of Li/EC:DMC-LiBF4/SWNT cell.
The appearance of oxidation peaks in the CV is a direct evidence of

PFC 6 and BFC 4 anions incorporation within the SWNT structure. The
macromolecule Cn should then bear a positive charge as a result of
subsequent oxidation following. In the case of PFC 6 the anodic oxidation is:
nC + PF6 o ( C n , PF6 )+ e- (2)
It is very likely that PF6 occupies the Van de Waals gap between
the SWNTs in a similar way than in graphite intercalation compounds [6]
and in anion-doped fullerenes [7]. Because of lower surface area between
SWNTs, the stability of the ( C n , PF6 ) ionic compound should be lower than
in flat graphite layers. Therefore, during the electrochemical intercalation a
chemical de-intercalation (decomposition) may take place, which explains
the low faradaic yield of the anodic intercalation.
The observation of a high threshold voltage during the first
intercalation suggests high activation energy to form the C n macrocation.
Since alkali metals do intercalate into SWNTs with associated formation of
macroanions C n , SWNTs are amphoteric in character. However it seems
that the macroanion form is more stable and reversible as reported for the
lithium intercalated SWNTs.
282
4. CONCLUSION
We have shown that BFC4 and PFC6 intercalate into SWNTs in a

similar way than in graphite or in fullerenes (C60 and C70 molecules). The
intercalation (equation 1) takes place in different steps, which evolve with
the cycle number. We suppose that in the first cycle it needs to break the
tube-tube interaction (separation) and to transfer the electron. After
separation is made, it becomes easier to intercalate and deintercalate the
anions, although the reaction is not 100 % reversible. The technological
output of different cycles upon anion intercalation into SWNTs is obvious,
knowing the need to improve the energy density and the power density of
existing lithium-ion batteries. We believe that increasing the energy and
power density will make the nanotube-based lithium batteries even more
attractive.
ACKNOWLEDGEMENTS
The authors give thanks to Dr. Channing Ahn from California
Institute of Technology for providing the SWNTs sample and for allowing
us to show his TEM results. Thanks also to Dr. Dayal Meshri from Advance
Research Chemicals (Catoosa, OK, USA) for providing high purity LiPF6
and LiBF4 for this investigation.
REFERENCES
1. Iijima S. Nature 1991; 354: 56-58.
2. Thess A., Lee R., Nikolaev P., Dai H., Petit P., Robert J., Xu C., Lee Y. H., Kim S. G.,
Rinzler A. G., Colbert D. T., Scuseria G. E., Tomarek D., Fischer J. E. and Smalley R. E.
Science 1996; 273: 483-487.
3. Zhou O., Fleming R. M., Murphy D. M., Chen C. H., Haddon R. C., Ramirez A. P. and
Glarum S. H. Science 1994; 263: 1744-1747.
4. Yazami R., Gabrisch H., Fultz B. J. Chem. Phys. 2001; 115: 10585-10588.
5. Ajayan P. M. and Ebessen T. W. Rep. Prog. Phys. 1997; 60: 1025-1062.
6. Yazami R, Touzain P. Solid State Ionics 1983; 9-10: 489-494.
7. Yazami R, Cherigui N. Mol. Cryst. Liqu. Cryst. A-1994; 244: 209-214.
SURFACE TREATED NATURAL GRAPHITE AS
ANODE MATERIAL FOR HIGH-POWER LI-ION
BATTERY APPLICATIONS
J. Liu1*, D. R. Vissers1, K. Amine1, I.V. Barsukov2, J.E. Doninger3

1
Electrochemical Technology Program, Chemical Engineering Division
Argonne National Laboratory, 9700 South Cass Ave., Argonne, IL 60439, USA
2
Superior Graphite Co., 4201 West 36th Street, Chicago, IL 60632, USA
3
Dontech Global, Inc. 427 E. Deerpath Road, Lake Forest, IL. 60045, USA
Abstract
High power application of Li-ion battery in hybrid electrical vehicles

requires low cost and safe cell materials. Among the various carbon anode
materials used in lithium-ion batteries, natural graphite shows the most
promise with advantages in performance and cost. However, natural graphite
is not compatible with propylene carbonate (PC)–based electrolytes, which
have a lower melting point and improved safety characteristics. The problem
with it is that the molecules of propylene carbonate intercalate with Li+ into
graphite, and that frequently leads to the exfoliation of the graphite matrix.
In this work, we improved the performance of the natural graphite
by surface modification. In this effort, two types of natural graphite were
evaluated for high power applications in Hybrid Electrical Vehicles. One
was a round edge type natural graphite SL-20, and the second graphite
sample was surface treated with an amorphous carbon coating (version,
SLC1015), both materials are available from Superior Graphite of Chicago.
Test cells using the above graphite materials were evaluated in PC-
based electrolytes. For this work, the hybrid pulse power characterization
(HPPC) test was performed on cells with the different graphite anodes and
PC-based electrolytes to evaluate their high power capabilities. These
electrochemical experiments indicate that cells containing the surface-
modified natural graphite can meet the power requirement set by the
FreedomCar partnership for the hybrid vehicle applications.
Keywords
Natural graphite, carbons, surface modification, coating, lithium-ion

cells, high power, hybrid electric vehicles (HEV).
*
283
284
1. INTRODUCTION
Carbon materials are commonly used as the negative electrode
(anode) in lithium-ion cells. Among the various carbon materials, natural
graphite appears to be the most promising candidate because it has a high
capacity (372 Ah/kg, with a stoichiometry of LiC6), a potential near lithium,
a relatively flat potential profile, good cycle-ability, lower cost and
reasonable safety. Thus Natural graphite is among the most attractive
negative electrode materials for lithium-ion batteries [1-3]. Moreover,
natural graphite was also successfully investigated in high power lithium-ion
batteries for hybrid electrical vehicle propulsion [4].
But it is known that natural graphite has a problem in some of the
more promising electrolyte solvents, like those containing propylene
carbonate (PC). PC has several advantages over other organic carbonates as
a Li-ion battery electrolyte solvent component, e.g., lower cost, lower
volatility and higher flash point and better low temperature performance.
However, when using PC in Li-ion battery, the solvated PC lithium ion will
co-intercalated into the graphite layers and cause subsequent exfoliation and
failure of graphite electrode [1].
The performance of natural graphite in PC-containing electrolyte can
be improved by its surface modification. An amorphous carbon coating was
one method to prevent graphite exfoliation which was based on a concept
that a core-shell-structured carbon composite may be applied to the graphite
anode material in Li-ion batteries by Kuribayashi et al.[5]. In this effort, two
types of natural graphite were evaluated for high power application. One was
a round edge type natural graphite SL-20, and other sample was surface
treated (carbon coated) version, SLC1015. Both materials are the grades
available from Superior Graphite Co. of Chicago IL.
It is known that the electrolyte solvent may be decomposed during
the first lithium ions intercalation (or formation cycle) into graphite to form a
solid electrolyte interphase (SEI) film on the graphite negative electrode
surface. It is this SEI film (or layer) that largely determines the performance
of graphite as a negative electrode in Li-ion batteries. Aurbach et al.,
postulates that the initial surface reaction and SEI film controls the nature of
further reactions [6]. Clearly, the electrode performance of natural graphite
depends on the nature of solvent and the graphite surface modifications
which may inhabit the solvated lithium ions from insertion into the graphite
structure.
One of the promising techniques to modify the surface of natural
graphite is soft carbon or an amorphous carbon coating. Mitsui Mining
Company used the thermal vapor decomposition (TVD) treatment [2] to
obtain modified natural graphite and Superior Graphite Co. uses a new
technique with lower cost to form a thin film carbon coating on the natural
graphite materials. It is therefore anticipated that the modified natural
285
graphite will have stable performance in PC based electrolytes and will

reduce the cost for both the negative electrode and electrolyte in lithium ion
battery systems.
The objective of the present study is to evaluate the performance of
surface coated natural graphite in PC based electrolyte under high power test
conditions and to investigate the effect and impact of surface coating the
natural graphite in Li-ion batteries for high power applications and their
safety. Our goal is to select the best natural graphite as an advanced negative
electrode material for Li-ion battery that are safe, low cost, high power,
possess excellent calendar life and PC based electrolyte tolerant. From our
results, modified natural graphite in Li-ion batteries has shown excellent
performance and meets the requirements for high power applications in
hybrid electrical vehicles (HEV) [7].
2. EXPERIMENTAL
The graphite materials used in this study include: SL-20 which is
round edge natural graphite, and SLC-1015 which is the surface coated
natural graphite from the type of untreated precursor SL-20, and both
samples were obtained from Superior Graphite Co.
The anodes of these two graphite samples were fabricated from a
slurried mixture which contains 92 wt% of active graphite powder and 8wt%
polyvinylidene difluoride (PVDF) polymer binder (Kureha #9130) and using
1-methyl-2-pyrrolidinone (NMP) (Aldrich, >99%) as the solvent. After
getting the homogenous slurry, the electrode laminates were coated on Cu
current collector foil using a doctor blade in the laboratory-made laminate-
coater. The laminates were then dried first at 75ºC in air for 3 hrs and then
the final heat treatment was carried out in a vacuum oven at 75ºC for 10 hrs.
Finally, the laminates were calendared to about 35% porosity in a dry room.
Coin-type cells (size CR2032, Hohsen Co.) with a lithium foil
counter electrode (FMC Corporation) and a polypropylene separator
(Celgard 3501) were assembled inside a helium-filled glovebox (<5ppm,
H2O and O2). Laboratory made cell fixtures were also used for the HPPC test
at room temperature.
The electrolyte used in the lithium cell studies was typically 1.2M
LiPF6 in ethylene carbonate (EC): propylene carbonate (PC): methyl ethyl
carbonate (MEC) in a 3:3:4 mixture. The cells were cycled at room
temperature using Maccor Series 4000 control unit in a galvanostatic mode
under a constant current density of 0.1 to 1 mA/cm2.
Postoperative investigations of the anode material were carried out
using scanning electron microscope (SEM, JEOL model JSM-6400 at 20
kV) to view the morphology of graphite particles.
286
To evaluate the high power cell performance, hybrid pulse power

characteristic (HPPC) tests [8, 9] were carried out at Argonne National
Laboratory using their own designed cyclers that are operated by VAX
station computers to control the experiments. ANL designed cell fixtures
were also used for cell tests. In these cell studies, LiNi0.8Co0.15Al0.05O2 was
used as a positive electrode and 1.2M LiPF6, EC:PC:MEC (3:3:4) was used
as electrolyte.
Safety studies of the graphite anode samples were performed using a
Perkin-Elmer Differential Scanning Calorimeter (DSC, model Pyris 1)
instrument. The temperature scanning rate was 10ºC/min over a temperature
range of 50 to 375°C.
3.1. Morphology of Materials

SEM representative images of the surface treated natural graphite SLC-
1015 and its untreated precursor SL-20 are shown by Figure 1. The graphite
particles with the rounded edges having less active sites tend to limit the
reaction on its surfaces and thus improve its cycling performance and safety.
(a) (b)
Figure 1. SEM images of (a) untreated natural graphite precursor SL-20 and
(b) surface treated natural graphite with round edge shape (SLC-1015).
SEM images indicate that both graphite samples have very similar
morphologies. Negative electrode laminates with active materials of SL-20
and SLC-1015 were then prepared using similar compositions. These
laminates were then used to prepare negative electrodes that were inserted
into Li-ion cells having similar cathodes and electrolyte materials.
287
3.2. Capacity of Graphite Anodes

The initial charge/discharge capacity curves of SL-20 and SLC-1015
electrodes vs. Lithium counter electrodes are present in Fig.2 and 3 at the
C/5 rate. The carbon surface-coated material, SLC-1015 in 1.2M LiPF6,
EC/PC/MEC electrolyte has a higher discharge capacity (335mAh/g in
comparison to 310mAh/g for SL-20 uncoated material. These results indicate
that this material is more stable in PC-based electrolyte than the uncoated
material, SL-20. Since the electrolyte contains 30 wt% PC, it is interesting
that the uncoated round edged SL-20 material showed no apparent
exfoliation during the initial cycle. This is very different from the natural
graphite with flake type morphology, where the graphite exfoliation takes
place immediately with much reactivity in PC-based electrolytes.
3.0
2.5
Voltage, V
2.0
1.5
1.0
0.5
0.0
0 100 200 300 400
Capacity, mAh/g
Figure 2. Initial charge/discharge curve of graphite sample SL-20, in 1.2M LiPF6,

EC/PC/MEC (3:3:4) electrolyte. Test cell was cycled at C/5 rate.
In Figure 3, the discharge capacity of surface modified graphite

sample is about 335mAh/g in 1.2M LiPF6, EC/PC/MEC electrolyte for SLC-
1015 graphite. This is about 8% higher than un-coated graphite precursor
SL-20. The reason for this may be due to the better morphology of the
powder and less capacity loss in PC-based electrolyte during cycling. For
this reason we focus on the surface modified natural graphite SLC-1015 in
the following experiments.
To verify the rate capability of the SLC-1015 material, cell with
lithium counter electrodes were cycled at the C/10 and C/2 rates,
respectively, as shown in Figure 4. The practical discharge capacities for the
288
surface treated material, SLC1015 are superior to those obtained for the
untreated graphite SL-20. As indicated in Fig.4, SLC1015 shows reversible
capacity values of about 349mAh/g at the C/10 rate and 320mAh/g at C/2
rate.
Figure 3. Initial charge/discharge curve of graphite sample SLC-1015, in 1.2M LiPF6,

EC/PC/MEC (3:3:4) electrolyte. Test cell was cycled at C/5 rate.
Figure 4. Charge/discharge curves of the surface treated natural graphite SLC1015

vs. Li/Li+ in 1M LiPF6, EC:DMC (50:50 wt%).
289
In the above study, SLC-1015 showed a low irreversible capacity

when the cell was cycled at the C/10 rate, e.g., 7.3% of its capacity, that is a
very good performance.
3.3. High Power Studies

Characterization (HPPC) test of full cells were performed with
graphite anodes in accordance with the procedure described in the PNGV
test manual [8, 9]. Figure 5 and 6 show the typical HPPC test results. This
cell had LiNi0.8Co0.15Al0.05O2 as the positive active material and carbon-
coated natural graphite SLC1015 as the negative active material.
Figure 5 shows the area specific impedance (ASI) for the HPPC 18-s
discharge and 2-s charge pulses as a function of depth of discharge (DOD).
The ASIs for both charge and discharge are very low over a wide range of
DOD. These values are lower than the ASIs, needed to meet the FreedomCar
performance goal, which is 35 ohm-cm2 for 18-s discharge and 25 ohm-cm2
for 2-s charge pulses. The cell pulse-power capability for both charge and
discharge was thus found to meet the FreedomCar power requirements,
especially in the cell “sweet spot” of 30-70% DOD.
40
35
30
2
ASI, ohm*cm
25
20
15
10
5 ASIDischarge18s
ASIRegen2s
0
0 20 40 60 80 100
DOD, %
Figure 5. ASI vs. DOD for cell using surface treated natural graphite SLC1015 as anode in
1.2M LiPF6 , EC/PC/MEC (30:30:40 wt%). Pulse discharge at 10C rate.
290
The power capability of the cell against DOD was plotted in

Figure 6. The performance of such cells and test results thereof indicate that
cells containing the surface-modified natural graphite meet and exceed the
power requirements set by the FreedomCar partnership for the hybrid
electric vehicle applications.
In comparison to SLC-1015 graphite, one cell was tested with
92wt% SL-20 vs. LiNi0.80Co0.15Al0.0502 in 1.2M LiPF6, EC/PC/MEC (3/3/4)
by the same procedure and conditions described earlier. The ASI vs. DOD
plot is shown in Figure 7.
160
140
120
2
Power, mW/cm
100
80
60
Discharge
Charge
40
20
0 20 40 60 80 100
DOD, %
Figure 6. Power capability of cell using surface treated natural graphite SLC1015 as anode,
and,LiNi0.80Co0.15 Al0.05i cathode in 1.2M LiPF6 , EC/PC/MEC (30:30:40 wt%).
The ASI of cell using SL-20 as negative electrode is relatively

higher compare to SLC-1015 in Fig. 5. But this value also meets the power
requirements of the FreedomCar program for HEV application.
3.4. Safety Study

To evaluate the safety property of SLC-1015 graphite, DSC was
used to study the thermal behavior of fully charged anode samples in
291
electrolyte as a function of increasing temperature. Results are plotted in

Figure 8.
50
45
40
2
ASI, ohm*cm
35
30
25
20
15
10
ASI-Disch18s
5
ASI-Regen2s
0
0 20 40 60 80 100
DOD, %
Figure 7. ASI vs. DOD for cell using surface treated natural graphite SL-20 as anode and
LiNi0.80Co0.15Al0.05 cathode in 1.2M LiPF6 , EC/PC/MEC (30:30:40 wt%).
Pulse discharge at 10C rate.
3
Heat Flow, W/g
-1
0 50 100 150 200 250 300 350 400
Temperature, ºC
Figure 8. DSC of graphite sample (SLC-1015) at 100% SOC with electrolyte: 1.2M LiPF6 ,
EC/PC/MEC (30:30:40 wt%).
292
The natural graphite at 100% SOC (state of charge) was tested with
electrolyte using DSC and it was found that the heat generation is small. The
first peak at 125ºC was mainly due to SEI layer breakdown and until 250ºC
the main reaction of the lithiated carbon is the reaction with electrolyte that
results in gaseous products. This second peak indicates that the heat
generation was largely limited by using PC-based electrolyte and this may
improve the safety of Li-ion batteries for high power application, e.g., HEV.
4. CONCLUSIONS
In conclusion, the surface modified natural graphite has good
performance in PC based electrolyte and also meets the power requirements
for hybrid electrical vehicle applications. Surface carbon coated natural
graphite SLC1015 is a very promising material in high power Li-ion
batteries with lower cost, reasonable safety, and low irreversible capacity.
ACKNOWLEDGEMENT
Authors representing ANL acknowledge the financial support of the
U.S. Department of Energy, FreedomCAR & Vehicle Technologies
Program, under Contract No. W-31-109-Eng-38. Also, all authors are very
grateful for the continued support of their DOE sponsor, Dr. Tien Duong.
REFERENCES
1. A. N. Dey, B. P. Sullivan, J. Electrochem. Soc., 117, 222 (1970).
2. M. Yoshio, H. Wang, K. Fukuda, Y. Hara and Y. Adachi, J. Electrochem. Soc., 147 (4),
1245 (2000).
3. H. Wang and M. Yoshio, J. Power Sources, 93, 123 (2001).
4. J. Liu, A. Kahaian, I. Belharouak, S. Kang, S. Oliver, G. Henriksen and K. Amine.
“Screening Report on Cell Materials for High-Power Li-ion HEV Batteries”, ANL-
03/16, DOE report, April, 2003.
5. I. Kuribayashi, M. Yokoyama, and M. Yamashita, J. Power Sources, 54, 1 (1995).
6. D. Aurbach, M. D. Levi, E. Levi, A. Schechter, J. Phys. Chem. B, 101, 2195 (1997).
7. K. Amine and J. Liu, ITE Letters, Vol.1, No.1, B39 (2000).
8. “PNGV Battery Test Manual”, Revision 1, Idaho National Engineering Laboratory,
Department of Energy, DOE/ID-10597 (May, 1998).
9. PNGV Test Plan For Advanced Technology Development Gen 2 Lithium Ion Cells,
EHV-TP-121, Revision 6, October 5, 2001.
CHAPTER 4:
EMERGING METAL/CARBON COMPOSITE

ANODES FOR NEXT GENERATION
The following chapter contains a collection of six papers specifically
dedicated to the topic of metal/graphite composites as candidate active
materials for the negative electrodes of the lithium-ion batteries of the near
future. Editors believe this chapter to be a very first attempt made in the
worldwide electrochemical literature to group metal/graphite composite
lithium-ion battery developers into a stand-alone section of a book.
Recently the metal-carbon composites are being widely viewed as
the most probable short-term solution to obtain a significant improvement of
the classic carbon-based anode materials for lithium-ion batteries. There
appears to be a significant driving force to commercialize these materials. It
is based on a fact that fierce market competition already resulted in creation
of lithium-ion batteries with near theoretical performance for a given cell
size. Thus, editors estimate that it is going to be a matter of no more than one
to two years before the practical limit of approximately 2,500 mAh* for the
18650-size lithium-ion cell is reached. The ever-growing power demand will
require new electrode chemistries with higher specific capacities, and, likely,
these materials shall be metal/graphite composites for the anode portion of
the lithium-ion cells. The recent announcement by Sony of new
commercially viable ‘Nexelion’ cells that contain a Co-Sn-C anode signals a
trend and expands the types of future Li-ion cells that are envisioned.
In the opening paper of this chapter, a collaborative partnership of
authors from Ukraine and USA (Prof. V. Barsukov and Prof. J. Doninger)
formulate key theoretical principles for designing anode materials of new
generation lithium-ion batteries. The authors propose to coat graphite with
nano-dispersed Si or Sn, then imbed these composites in some highly stable
shell, which would suppress volumetric changes of Si and Sn upon reversible
formation of an alloy with lithium. The paper also goes into the theoretical
grounds of the role of expanded graphite and conducting polymers for use in
various battery and supercapacitor systems.
In the next paper by Y. Illin et al., capabilities of Sn anodes are
considered as a possible alternative to carbon. Thin films of Sn were
deposited onto current collector in vacuum, and tested in the coin cells.
Authors were able to obtain reversible alloying reaction, which stabilized at
100 mAh/g between cycle number 100 and 400. The stability of Sn and its
characteristics upon cycling was seen to be a function of the current collector
material. The best results were achieved with non-copper-based substrates.
The third paper in this chapter is a product of successful
international collaboration, which came into being through assistance of
*
This capacity limit for a given cell size has been determined as a result of computer
simulation by GP Battery Technologies, as referenced from: J. Fan and D. Magnuson, in
“Rechargeable Lithium Batteries” K.M. Abraham, E.S. Takeuchi, M. Doyle eds., PV 2000-
21, the Electrochemical Proceedings Series, Pennington, NJ (2000).
295
296
NATO Science for Peace, NATO Advanced Research Workshop, and US

Civilian Research and Development Foundation’s programs. Novel
metal/graphite composite materials proposed for development by Ukraine’s
Kiev National University of Technologies and Design were produced and
further upgraded by Superior Graphite in the U.S. In the next stage,
development effort linked more organizations in the U.S., one of which is an
industry leading specialty lithium-ion battery manufacturer, Lithion, Inc.
Even though the project has not yet finished, materials being developed look
promising, as judged by their testing in the full prismatic lithium-ion cells.
The paper is on carbon coated Si, and coated Si- and Al-modified graphite.
A paper by L. Reiter et al. is a joint contribution from two academic
institutions and a battery technology development organization in Ukraine. It
introduces carbon/graphite materials modified by d-metal complexes via
reactions of pyrolysis of precursor carbons with various organics. The
approach taken appears to be promising for both gas-diffusion electrodes and
anodes of lithium-ion batteries. While non-graphitic materials used by the
authors as precursors, did not appear to have benefited from grafting,
graphites were seen to show improved high rate performance in the half-cells
with metallic lithium as a counter electrode. Editors would like to comment
that potential issues of high cost and stability of performance characteristics
of graphites grafted by d-metals in the practical lithium-ion battery
environments may need to be further investigated by the authors.
In the paper by Dr. L. Matzui et al., authors seek to develop
advanced composite materials based on natural graphite and nano-scale
metals, which are known to reversibly form alloys with lithium. Authors
submitted interesting results on plating of nano-sized particles of metal onto
graphite surface. Of a little reservation to editors is the choice of carbon
precursor, which is bulky expanded graphite. Expanded graphite’s packing
density properties typically make it unsuitable for application as anode active
material in the lithium-ion batteries. The paper is lacking electrochemical
performance data; hence it is difficult to assess how stable the composites
are in practical lithium-ion cell environments.
The last paper of this chapter (C. Johnson et al.) summarizes
important investigative work by researchers at the Argonne National
Laboratory in the U.S. on the subject of improvements of graphitic anodes
for high power lithium-ion batteries. The anode carbons were modified by
Ni, Cu or carbon coatings, and then tested, among others, in PC-containing
electrolytes. Authors conclude that Cu coating of graphite improves capacity
of electrodes to a greater extent in PC than Ni coating; CVD carbon coating
on graphite was also seen to also improve anode performance versus that
based on uncoated graphite. The authors hypothesize that Cu may act as a
better de-solvation catalyst than Ni when stripping the PC coordination
sphere around the Li cation, thus leading to improved electrochemical
performance.
ON THE THEORETICAL PREREQUISITES FOR
APPLICATION OF NOVEL MATERIALS IN
PROMISING ENERGY SYSTEMS
Viacheslav Z. Barsukov1* and Joseph E. Doninger2

1
Technology and Design, 2 Nemirovich-Danchenko str., Kiev 02001, Ukraine;
2
Superior Graphite Co, 10 South Riverside Plaza, Chicago, 60606,IL / Dontech Global, Inc.
427 E. Deerpath Road, Lake Forest, IL 60045, USA
Abstract
Some theoretical prerequisites for application of modified and

expanded graphites, Si- and Sn-based composites and alloys, electro-
conducting polymers as active materials, catalysts and electro-conductive
additives for lithium - ion batteries, metal-air batteries and electrochemical
capacitors are considered. The models and the main concepts of battery-
related use for such materials are proposed.
Keywords
Modified graphite; expanded graphite; Si- and Sn- composites;

alloys; conducting polymers; lithium-ion batteries; metal-air batteries;
electrochemical capacitors.
1. INTRODUCTION
Lithium-ion batteries, high-energy lithium and metal-air batteries,
electrochemical capacitors have been changing radically, in recent years, the
structure of the market of power sources.
Constant technological improvements dictated by ever-increasing
demands from the market, require the development of new materials. This
process involves extensive basic electrochemical research into the nature and
*
297
298
mechanisms of electrode processes, quantum-chemical and physicochemical

modeling, a new look at the development and testing of battery materials.
The aim of this challenging task is to be always coming up with
technological solutions that allow reaching and surpassing theoretically
calculated energy density output limits for a particular size of
electrochemical devices. This paper seeks to introduce theoretical concepts,
which became founding stones for some of the new materials, which our
groups have jointly proposed and developed for the markets of anode active
materials in lithium-ion batteries, conductivity enhancement applications in
various types of power sources, and conducting polymer matrixes for the
emerging electrochemical supercapacitors.
2.1. “Classic” Graphite and Modified Graphite Materials

Special types of natural and/or synthetic graphite are commonly
used as negative active materials (AM) for lithium-ion batteries (LIB). After
replacing their more expensive predecessors, mesophase carbon micro beads
(MCMB), they have become state-of-the art active materials in recent years,
and, therefore, we refer to them as “classic” lithium-ion grade graphite.
These AM feature stabile capacity of Q c=365-370 (mAh/g) under
prolonged cycling conditions due to their high purity and special structure of
particles; also they demonstrate rather low irreversible capacity loss during
the first SEI formation cycles ( ' d 10%). An illustrative result showing
stable cycling of purified round-edge natural graphite is shown by Figure 1.
Figure 1. Typical galvanostatic charge (1) – discharge (2) curves of the lithium-ion battery
grade graphite, SL-20 (Superior Graphite Co., USA), as tested at C/20 rate in 2016 coin cells
having Li metal foil as counter electrode and electrolyte: EC:DMC + 1M LiPF6.
299
The specific capacity obtained in such Ams, actually corresponds to

near theoretical limit of 372 mAh/g, as calculated on a basis of classical
LiC6 stoichiometry). Further increase of capacity is possible only via
switching to new or modified materials, composites or alloys, which are
capable for reversible and stable intercalation of lithium.
It is often proposed in the literature to use SnO2, Si, carbon materials
with disordered structure, and many similar materials as an alternative to
graphite (see, for example [1-4]). Our studies show that indeed it is possible
to achieve very high capacity with such active materials in the first cycle
(Q =700-800 mAh/g and higher), however this capacity decreases quite
rapidly during cycling. Moreover, the irreversible loss of capacity in the first
cycle is very high and is often over 50%, which brings to naught their
advantages.
Different investigations of the mechanisms of capacity degradation
during cycling show that one of the main reasons of this degradation is
significant change (by a factor of 3-4) in AM volume on the intercalation-
deintercalation of lithium, followed by AM destruction [2-3].
Therefore, a possible theoretical concept of using such materials in
LIB’s is based on the development of various composites (alloys), in which
the volume variations of the electroactive constituents of AM during cycling
can be compensated by the elastic properties of the electrically conductive
matrix (Figure 2).
a) b)
1
2 4
3
Figure 2. The concept of using additives of Si (or Sn) in the negative electrode: a) the
beginning of charge; b) the end of charge. 1 and 3 – elastic porous graphite (or alloy) matrix
with Li+ ions; the particle of Si (or Sn) before (2) and after Li+ ions intercalations.
300
The summary of our many try-n-error experiments suggests that for

realization of concept proposed by Figure 2, it is necessary to enable the
following conditions:
1) Furnish uniform & highly dispersed nano-sized particles of Si or Sn;
2) The amount of Si or Sn has to be small (usually on the order of 3-
5wt%);
3) The bulk electrode has to have very good elastic properties, which
can be created by proper optimization of conductive matrix.
In accordance with this concept, the experimental modified rounded-
edge graphite grade was synthesized with 3wt% of the Si coating on LBG-25
graphite (the later being a commercial product of Superior Graphite Co.
(SGC), a U.S. industrial carbon manufacturer). While reversible capacity of
LBG-25 is approximately 370 mAh/g, its Si-coated version (grade SO # 3-
43-18), demonstrated a noticeably improved discharge capacity of 412
mAh/g. The above value of capacity is higher than theoretical capacity for
pure graphite (372 mAh/g), which leads to the average utilization coefficient
P # 111% (Figure 3, Table 1) for this product.
Figure 3. Galvanostatic charge (1) – discharge (2) curves of SO# 3-43-18 (LBG-25 + 3% Si)
graphite at C/20 rate in 2016 coin cell. Electrolyte: EC:DMC + 1M LiPF6.
Counter electrode – Li foil.
It is interesting to understand the reasons for exceeding the

theoretically possible capacity limitations with graphite SO # 3-43-18.
It is known about the existence of lithium silicide, Li6Si2, which is
close to intermetallic compounds, and also that silicon is capable to form
with lithium different alloys. We have calculated the theoretical specific
capacity of such possible compounds, as well as pure silicon (Table 2). It is
possible to explain from the Table 2 the noticeable increase of capacity for
graphite electrodes (11%) even at the small content of Si (3wt%). We can
suppose that some of such compounds (LiXSiY) with high capacity may form
301
during electrochemical cycling, and in doing so will give a noticeable boost

to the total capacity of graphite. Creation of uniform highly dispersed
network of particles of Si in an elastic graphite matrix leads to a relatively
stable dependence of capacity vs. Si loading on graphite surface.
Table 1. Results of initial galvanostatic cycling of experimental 2016 coin cells with different
types of graphite from SGC at C/20 rate ('U = 0.01 - 1.0V). Electrolyte – LP-30 from Merck
(EC:DMC + 1M LiPF6). Counter electrode – Li foil.
Average
Graphite grade Utilization 1st cycle
discharge
and its main coefficient loss
capacity*
properties P (%) (%)
(mAh/g)
LBG-73
(D50 = 43 Pm; 368 98.9 7.83
S = 2.85 m2/g)**
LBG-25
(D50 = 18 Pm; 370 99.5 6.77
S = 4.9 m2/g)**
SO # 3-43-18
412 111 21.4
(LBG-25 + 3% Si)

* Results for an average sample in a series of 3-5 coin cells, during the first 4-5 cycles.
**According to the Product Information Bulletins of SGC Purified Battery Grade Graphite.
Table 2. Calculated theoretical specific capacity of some silicon compounds and pure Si.
Compounds Li6Si2 Li12Si7 Li7Si3 Li13Si4 Li22Si5 Si
qT, mAh/g 1,641 1,149 1,411 1,716 2,005 4,000
Similar effect of relatively stable cycling on the level of discharge

capacity of 500 mAh/g can be achieved, for example, using tin-based alloys
with different metals (please, see the detailed investigation of such alloys in
this book [5]). These metals perform probably the functions of elastic matrix
in such alloys and gave possibility to compensate for the volumetric changes
of Sn.
Thus, the experiments carried out suggest that using the theoretical
concept proposed by Figure 2, it is possible to create negative electrodes that
would feature reversible capacity exceeding 400 mAh/g.
302
2.2. Thermally Expanded Graphite

One of possible ways of improving the specific characteristics of
LIB electrodes is the use of thermally expanded graphites (TEG) as
electrically conductive additives of positive active mass. The efficiency of
TEG’s use may be illustrated by analyzing a simple theoretical model
proposed by us: “AM – electrically conductive additive” (Figure 4).
Figure 5. A model for composite porous electrode in a cubic close packing approximation.
The main idea of the model is that in order for the electrically
conductive additive to effectively fulfill its functions, it must form a closed
cluster (skeleton of the interconnected carbon particles, which is the
conducting pass in electrode matrix). Once the sufficient conductive network
was formed, further considerable increase of additive content is not needed,
as it leads to decrease in the percentage of the electrochemically active
constituent in the electrode.
On the basis of an analysis of this model in a cubic closed packing
approximation, it may be shown that the necessary additive content per unit
AM mass, M , can be described by the following relation:
3
M AD U AD ª6 § t · º
M « ¨¨ AD
1 ¸¸ 1» , (1)
M AM U AM «¬ S © D AM ¹ »¼
303
where M AD is a content of the electrically conductive additive; M AM is a

content of the electrochemically active material; U AD , U AM are the effective
densities of the additive and AM respectively; D AM is an average size of
AM spherical agglomerates; t AD is an average size of additive interlayer (the
average distance between the spheres).
It is clear from the equation (1) that the optimal specific capacity of
such two-component composite porous electrode would be realized when
M M min . (2)
In this case a porous electrode has a maximal possible amount of the

electrochemically active material and the minimal amount of the conductive
additive, which is necessary for formation of closed cluster (the touch of the
spheres at the Figure 5) and a good conductivity of electrode mass.
The ratio (2) can be realized when
( t AD / D AM ) SS 1. (3)
In this case
§6 · U
M min ¨ 1 ¸ u AD . (4)
©S ¹ U AM
It follows from the relation obtained that the minimum electrically

conductive additive content is directly proportional to the effective density of
the additive. By "effective density" we understand the density of the material
under real conditions of making the electrode (with allowance for the actual
molding (rolling) pressure, humidity, temperature, etc). In this respect, TEG
has unique advantages over all existing types of additives. The density of
this material in free state (bulk density) is 0.05 g/cm3, which is about one-
fourth of that for the ordinary graphite and one-fifteenth to one-twentieth of
that for the metal powders (e.g. nickel, copper powders, etc.).
Particularly, let us make comparison of densities for common and
expanded graphite in a free state:
com
U GR
exp
# 4;
U GR
For this extreme case

com com
M min U GR
exp exp
#4. (5)
M min U GR
304
Thus, using TEG under the same conditions, the content of graphite
in the different types of batteries could be reduced as far as by the factor of 4
with corresponding raise of the capacity and voltage of batteries. At this
point of time, we have few examples of experimental evidence of such
phenomena.
Application of TEG creates prerequisites for a considerable
reduction of the electrically conductive additive’s weight content in the
electrode (provided, the choice of technological operations in making
exp
electrodes, which do not reduce considerably U GR , is correct). The concept
is applicable to batteries with non-aqueous (earlier considered example of a
lithium-ion battery system), as well as to aqueous electrochemical systems.
Besides the conductive additive, TEG may sometimes be a very
effective catalyst support, for example, in the catalytic active composite with
conducting polymers for the new air-metal batteries, which we proposed [6].
When used in the aqueous electrolytes, TEG has to provide high
stability towards oxidation during the entire battery life. For achieving this
target, it is often necessary to modify of TEG surface, for example, by Boron
[7].
2.3. Conductive Polymers

The relatively high electronic conductivity of conducting polymers
is connected with the presence in polymers of internal system of poly-S-
conjugated bonds. This is easy to understand on the example of a simplest
type of conducting polymer (polyacetylene) in the main and activated states
(Figure 6).
(a)
(b)
Figure 6. The main (a) and the activated (b) state of polyacetylene.
As our quantum-chemical calculations show, similar transformation

and delocalization of bonds takes place in the conductive forms of some
other types of CPs (polyaniline, polypyrolle, etc.). Delocalization of
chemical bonds after activation leads to appearance of an electronic
conductivity in such types of conducting polymers and creates prerequisites
for their application as electrode materials of electrochemical power sources.
Such activation can be stimulated by intercalation of ions, applying the
potential, and by use of some other low energetic factors.
305
The conducting polymers show a significant non-faradaic

component of the electrochemical mechanism. The essential differences of
faradaic and non-faradaic systems in equilibrium behavior, trends of
galvanostatic charge - discharge curves and cyclic voltammograms have
been shown, and criteria for the identification of these mechanisms are
proposed [8].
For example, the investigations of the current-generating mechanism
for the polyaniline (PANI) electrode have shown that at least within the main
range of potential 'En the "capacitor" model of ion electrosorption/
desorption in well conducting emeraldine salt phase is more preferable.
Nevertheless, the possibilities of redox processes at the limits and beyond
this range of potentials 'En should be taken into account. At the same time,
these processes can lead to the fast formation of thin insulation passive
layers of new poorly conducting phases (leucoemeraldine salt,
leucoemeraldine base, etc.) near the current collector (Figure 7). The
formation of such phases even in small amounts rapidly inhibits and
discontinues the electrochemical process.
Figure 7. Schematic representation of processes of electromigration / diffusion of ions and

formation of insulation passive layer on the boundary current collector / conductive polymer.
As it logically follows from the considered mechanism and the

existing experimental data, PANI-type conductive polymers can certainly be
used to improve the characteristics of electrochemical capacitors. According
to our estimations, a specific capacity of electrodes with conductive
polymers could be increased by as much as to 3-5 times over that of the
classical carbon electrodes.
The detailed research and development on the promising hybrid
system for electrochemical capacitors with conducting polymers is described
in paper [9]. Besides the main electrochemical process (Figure 7), the side
306
reaction of oxygen reduction takes place at the surface of PANI and some
other conducting polymers. The above phenomena has found a practical
application for development of air-metal battery mockups with low costs
PANI/Expanded Graphite composite catalysts [6].
In order to study the possible reasons and mechanisms of the
catalytic activity of conducting polymers (CP), the electronic structure of
some molecular CPs clusters and its adsorption complexes with oxygen were
modeled [6]. In the CP-O2* complex, the CP surface is an electron density
donor. For example, in the case of PANI, the bond orders in adsorbed O2*
molecules decrease by about 30%, and the bond lengths L increase by about
24%. Thus, the adsorbed O2* molecules have a fairly high degree of
activation and can readily interact with the protons.
Therefore, quantum-chemical analysis confirms the mechanism of
O2 electroreduction and gives possibility to understand the reasons of PANI
catalytic activity.
3. CONCLUSION
1. We have proposed a theoretical model, and have experimentally
proven that it is possible to increase the capacity of negative graphite
electrodes for lithium-ion batteries using the additives of Si and Sn applied
as coatings on graphite. A negative active mass has to contain a composite
(or alloy) of highly dispersed (nano-sized) particles of the above additives.
To complement the efficient operation of Si- or Sn-coated graphite electrode,
it is necessary to create a very elastic conductive matrix in the electrode.
This conductive matrix will compensate the volumetric changes of Si or Sn.
Such concept of development of the modified graphites can ensure a
relatively stable cycling at high level of specific capacity of definitely more
than 400 mAh/g.
2. Thermally expanded graphite is, in principle, one of the most
efficient conductive additives to various battery active materials, first of all
exp g
due to its very low effective density U GR # 0.5 ; high enough
cm 3
conductivity and stability towards oxidation. For realization of these key
advantages, TEG has to be of an optimum particle size (function of active
material’s particle size) to ensure a ratio t AD / D AM SS 1. Under such
condition the content of graphite in the different batteries could be reduced
noticeably and changed for the corresponding amount of active material.
3. The conducting polymers show a significant non-faradaic
component of the electrochemical mechanism inside the main range of
potentials 'En. Nevertheless, the possibilities of redox processes at the limits
and beyond this range of potentials should be taken into account. At the
same time, these processes can lead to rapid formation of thin insulating
307
passive layers of new poorly conducting phases. As it follows from the

mechanism and the existing experimental data, PANI-type conductive
polymers can certainly be used in order to improve the characteristics of
electrochemical capacitors. Besides, these materials demonstrate good
perspectives for application in air-metal batteries and fuel cells due to
pronounced catalytic activity of PANI/TEG composites toward the reaction
of oxygen reduction.
ACKNOWLEDGEMENTS
Authors would like to gratefully acknowledge Superior Graphite
Co’s Dr. Igor V. Barsukov for his valuable consultations on the subject, as
well as for development of customized samples of graphite/Si used in our
research. We also wish to acknowledge KNUTD’s Ph.D. students: V.V.
Khomenko and E.A. Illin, Ph.Ds. N.A. Chmilenko, V.V. Kryukov, E.A.
Kryukova, A.S. Katashinskii and T.A. Motronyuk for various experimental
examinations and discussions, which were useful in creation of support
evidence for the above theoretical models. One of the authors, V.B. wishes
to thank the Scientific Committee of NATO Science for Peace Program for
partial financial support of this work under the umbrella of “Carbon” SfP
973849 project.
REFERENCES
1. Idota Y., Kubota T., Matsufuji A., Maekawa Y., Miyasaka T. Tin-Based Amorphous
Oxide: A High-Capacity Lithium-Ion-Storage Material. Science 1997; 276:1395-97.
2. S. Bourderau, T. Brousse, D.M. Schleich. Amorphous silicon as a possible anode
material for Li-ion batteries Journal of Power Sources 1999; 81–82:233–36.
3. Yang J., Takeda Y., Li Q., Imanishi N., Yamamoto O. Lithium insertion into Sn- and
SnSb-based composite electrodes in solid polymer electrolytes. J. Power Sources 2000;
90:64-69.
4. H. Buqa, P. Golob, M. Winter, J. O. Besenhard. Modified carbons for improved anodes
in lithium ion cells. J. Power Sources 2001; 97-98: 122-25.
5. Ye.O. Illin, V.Z. Barsukov and V.S. Tverdokhleb. Capabilities of tin thin films as
negative materials for lithium-ion batteries. This book 2004.
6. V.G. Khomenko, V.Z. Barsukov, A.S. Katashinskii and T.I. Motronyuk. New concept
for the metal-air batteries using composites: conducting polymers/ expanded graphite.
This book.
7. T. Motronyuk, I. Barsukov, V. Barsukov, E. Frackowiak and F. Béguin. New types of
graphite stable or the prolonged electrochemical oxidation. Ext. Abstract of the 4th Int.
Symp. on New Materials for Electrochemical Systems, Montreal, Canada, July 9-13,
2001; 116-17.
8. V.Z. Barsukov, V.G. Khomenko, S.V. Chivikov, I.V. Barsukov and T.I. Motronyuk. On
the faradaic and non-faradaic mechanisms of electrochemical process in conducting
polymers. Electrochim. Acta, 2001; 46: 4083 – 94.
9. V. Khomenko, E. Frackowiak, V. Barsukov, F. Béguin. Development of supercapacitors
based on conducting polymers. This book 2005.
CAPABILITIES OF THIN TIN FILMS AS
NEGATIVE ELECTRODE ACTIVE MATERIALS
FOR LITHIUM-ION BATTERIES
Yevgen O. Illin, Viacheslav Z. Barsukov*, and Viktor S. Tverdokhleb

Technology and Design, 2 NemIrovich-Danchenko str., Kiev 02001, Ukraine
Abstract
The capabilities of thin tin films and tin-based alloys with different
metals as active materials for lithium - ion battery negative electrodes are
considered. Electrochemical characteristics of such films at different
substrates and mechanisms of their functioning are discussed.
Keywords
Tin, tin-based alloys, lithium-ion batteries, negative active material
1. INTRODUCTION
In recent years lithium - ion power sources have occupied one of the
first places among other modern energy storage systems. Their functioning is
based on the possibility of reversible intercalation of lithium ions in active
materials (AM). Substances of tin are investigated often as negative
materials for lithium-ion batteries.
It is known that in the process of electrochemical cycling of tin
oxide-based thin films versus lithium metal there are at least two reactions in
standard alkylcarbonate electrolyte based on lithium salts [1]. The first of
them leads to formation of metallic tin on the first cycle:
SnO2 + 4Li o Sn + 2Li2O. (1)
*
309
310
This reaction is irreversible. In the next cycles tin metal forms

lithium-tin alloys with different stoicheometry:
mSn + nLi l LinSnm. (2)
There are several Li-Sn intermetallics [2]. Their structure and

theoretical capacities differ in the wide range (Ɍɚble 1).
Table 1. Li-Sn alloys and their theoretical capacity.
Theoretical specific
Alloy formula Li/Sn atomic ratio
capacity, mAh/g
Li2Sn5 0.4 88
Li2Sn3 0.7 145
Li3Sn2 1.5 311
Li2Sn 2.0 405
Li7Sn3 2.3 465
Li13Sn5 2.6 589
Li7Sn2 3.5 656
Li22Sn5 4.4 790
Sn 0.0 994
Thus, in the process of Li+ reaction with tin, theoretical capacity of

Li-Sn alloys can reach up to 790 mȺ·h/g. Theoretical capacity of pure Sn is
994 mȺ·h/g (7234 mȺh/ɫm3). Formation of such alloys occurs in the range
of potentials from 0 to 0.8V versus lithium (Table 2) or in the temperature
window 0-8000C. Charge-discharge curves have an inclined form.
Table 2. Formation potentials of different LiySn alloys, [3].

Potential of alloy formation, V
Stoichiometry
(versus Li)
0.40 < y < 0.70 0.660
0.70 < y < 2.33 0.530
2.33 < y < 2.63 0.485
2.63 < y < 3.50 0.420
3.50 < y < 4.40 0.380
The reactions (1) and (2) were investigated from the electrochemical
point of view (see, for example [4-7]). The particles of tin formed in the first
cycle, were embedded in the lithium-oxide matrix. Existence of such matrix
causes high cycleability of electrode. Therefore along with graphite usage of
tin oxides as negative materials was started. They take part in cycling with
some high discharge capacity loss from 600-450 mAh/g during more than
100 cycles [8].
311
Earlier it was considered, that metallic tin by itself couldn’t be used

even as thin films because of the big volumetric changes during lithium
intercalation-deintercalation [9]. In the formation of Li4.4Sn the volume of
alloy makes 360 % of initial volume of tin [10]. In Sn metal each tin atom
occupies 2,710-2 nm3. Li22Sn5 has a volume of 0,485 nm3, corresponding to
9,710-2 nm3 per Sn atom [11]. In the process of tin thermal deposition on
different metal surfaces it is mainly possible to obtain E-modification as so-
called grain tin. It has a tetragonal structure similar to diamond, but
compressed in the direction of c axis. E - modification has a high density and
conductivity [12]. Under certain deposition conditions clusters of grain tin
have rather small size. It is possible to assume that intercalation of some
lithium ions will not result in the fast destruction of electrode. If layers of tin
are alternated with layers of electrochemically inert substance with ionic
conductivity, such electrode can provide high cycleability.
Reference [13] discusses the capabilities of thin-film tin anodes for
lithium - ion batteries. Tin was electrodeposited from acid water solutions on
a copper foil. Thickness of coating did not exceed 2 mP. For studying
thermal influence, samples were annealed under vacuum at 2000ɋ during 24
h. Galvanostatic tests were carried out in standard electrolyte in the range of
potentials 0 - 2 V versus lithium foil electrode. It is interesting that all
samples have shown the theoretical capacity of tin on the first charge. But
further in the next cycles samples, which were not annealed failed discharge
capacity down to 200 mȺh/g. Annealed samples have a discharge capacity
900 mȺh/g during more than 10 first cycles. Formation of interface
consisting of Cu6Sn5 and Cu3Sn systems in the second case is rather probable
[14]. Moreover, these two phases can coexist in wide temperature window
and appear almost simultaneously. Semi-successful attempts of tin-
containing alloys deposition on various matrixes, for example on graphite,
are reported [15]. High cycleability gave materials made by electroplating
natural graphite with the systems Sn65Sb18Cu17 and Sn62Sb21Cu17. During the
cycling versus lithium-cobalt oxide in standard alkylcarbonate electrolyte,
materials saved more than 50% of discharge capacity after 34 cycles.
Interesting results were obtained on material, made by co-sputtering of tin
(45-60 mol. %) and molybdenum on a copper foil [16]. Thickness of film is
4-10 mP. Crystalline structure consists of alternating tin and molybdenum
layers with thickness of 1 mP or less. These layers were separated by Mo1-
ɯSnx film alloy. Discharge capacity in 18650 elements reaches 350 mȺh/g,
and during more than 200 cycles exceeds 200 mȺh/g.
2. EXPERIMENTAL AND DISCUSSION

Subject of our investigations became thin films of tin deposited
under high vacuum on different current collectors.
312
All samples were galvanostatically tested in coin cell 2016 semi-

elements and special Teflon T-type testing cells in the potential window 0.03
– 0.80 V versus lithium foil in different modes from C/2 to C/40. We used
standard electrolyte LP 71(Merck).
For example, we deposited E-tin on a copper current collector
similar to Sanyo Electric Co [14]. It is interesting that as-deposited samples,
obtained without annealing, have near 200 mȺh/g of specific discharge
capacity (Figure 1) in spite of smooth coating with hundred nP-size grain tin
structure. This fact directly shows necessity of bronze interface between
active material and copper current collector independently of coating
methods. Good cycleability of all samples even in semi-elements versus
lithium observed for as-deposited tin on copper current collectors without
bronze intermediate layer. Very small hysteresis between charge and
discharge capacity presents both in case of unwoven polypropylene separator
(Mogilev, Belarus) and Celgard separator. Gradual decreasing of charge and
discharge capacity changes on reversible regularity after four hundred cycles
due to electrode structure reorganization.
We made attempts to deposit 1 mP layer of E-tin film on 20 mP
tungsten foil. This foil was previously etched in nitric acid. We showed by
XRD the absence of Sn-O and Sn-Mo bonds (Figure 2). Charge-discharge
curves in all cases have sloping shape typical of tin (Figure 3).
Some hysteresis between charge and discharge capacity is caused obviously
by collateral processes and growth of dendrites in cells in the case of usual separa-
tors usage. Growth of dendrites occurs even on three-layer separators (Figure 4). It is
known that tin is active material with drastic inclination to dendrite-making.
Because of the use of special separator materials (Celgard, Teklon, etc.) we
were able to increase discharge capacity; also, decrease hysteresis between
charge and discharge capacity, and achieved certain stability upon cycling.
Application of modern lithium-ion grade separators does increase capacity
output characteristics and improves it towards having higher stability, thus
falling off of discharge capacity is not rapid (being the case for traditional
separator materials), but rather gradual (Figure 5). This discharge capacity
fade became essential after 10 cycles during cycling versus lithium foil.
The reason of slow charge/discharge capacity reduction is probably
gradual loss of contact between the active particles and current collector.
Nevertheless, in the case of copper current collector usage we observed even
smaller increase of discharge capacity after 400th cycle (Figure 1).
In spite of higher capacity of tin electrode on tungsten current
collector in comparison with tin electrode on copper current collector,
specific capacities for unit of electrode mass have the reverse correlation
(Table 3).
Nevertheless, such parameters are quite acceptable for testing in
mockups of lithium-ion batteries.
313
Specific Capacity, mAh/g Sn 250
200
150
100
50
0
0 100 200 300 400 500 600
Cycle, N
Figure 1. Specific charge (thick line) and discharge (thin line) capacity of tin electrode
without annealing. Bronze interface is absent. Current collector – copper.
Testing mode C/5. Counter electrode - lithium foil. Separators – 2 layers of unwoven
polypropylene (Mogilev, Belarus).
Ɋɚɡɧɨɫɬɧɚɹ
Differential curveɤɪɢɜɚɹ
dots - SnSn
squares -WW
Intensity
10 30 50 70 90 110
2 T , degree
Figure 2. XRD pattern of tin coated tungsten.

314
0,8
0,7
0,6
0,5
U, V
0,4
0,3
0,2
0,1
0,0
1023 1750
3 2231 00
time, h
Figure 3. Charge-discharge curves of tin on tungsten current collector.
1000
Specific Capacity, mȺh/gSn
900
800
700
600
500
400
300
200
100
0
0 2 4 6 8 10 12
Cycle, N
Figure 4. Specific charge (squares) and discharge (dots) capacity of tin electrode. Current
collector – tungsten. Testing mode C/5. Counter electrode - lithium foil. Separator – 2-3
layers of unwoven polypropylene (Mogilev, Belarus).
315
Specific Capacity, mȺh/g Sn 600
500
400
300
200
100
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Cycle, N
Figure 5. Specific discharge capacity of tin electrode without annealing. Current collector –
tungsten. Testing mode C/5. Counter electrode - lithium foil. Separators: 1 layer of unwoven
polypropylene (Mogilev, Belarus) + 1 layer of Celgard – triangles; 2 layers of unwoven
polypropylene (Mogilev, Belarus) – dots.
Table 3. Specific characteristics of active material and electrodes.

Maximum Maximum Maximum
discharge capacity discharge capacity discharge capacity
Current
on the unit of active on the unit of on the unit of
collector
material mass, electrode square, electrode mass,
mAh/g PkAh/cm2 mAh/g
W 520 758 70
Cu 200 291 95
3. CONCLUSIONS
Nanostructured tin, tin films, alloys and composites can be
considered as promising negative materials for lithium-ion energy storage
systems. Tin based materials have a good cycleability, high charge and
discharge capacity in wide range of cycling modes (C/2 - C/40). Tin-coated
tungsten electrodes have a higher capacity than tin-coated copper electrodes,
but larger hysteresis between charge and discharge curves and quicker
capacity fade. Tin-coated copper electrodes have better cycleability than tin-
coated tungsten electrodes. Tin-coated copper electrodes must be annealed
for bronze interface creation.
316
ACKNOWLEDGMENT
Authors would like to express their gratitude to Dr. Kryukov V.V.
(Kiev National University of Technologies and Design) for creation of
original electrochemical test cells; Dr. Chmilemko M.A. for the help in
assembly of standard size 2016 elements (Interdepartmental Division of
Electrochemical Power Energetic and Engineering, NAS of Ukraine), Dr.
Matzui L. U. for her assistance with XRD investigations (Kiev National T.
Schevchenko University).
Also, authors would like to acknowledge NATO Science for Peace
Program for financial support of this work through “Carbon” 973849 project.
REFERENCES
1. Y. Idota, M. Mishima, M. Miyaki, T. Kubota, T. Miyasaka, Eur.Pat.Appl., 651450 A1
94116643.1. 1997.
2. Moffatt, William G. The Handbook of binary Phase Diagrams. New York: Genium
Publisching Corp., Schnectady, 1990.
3. Elidrissi MML., Corrector JI., Tirado JL., C. Perez V. SnHP04: a promising precursor
for active material as negative electrode in Li-ion cells Electrochimica Acta
2001;47:489-93.
4. Wang J., Raistrick ID., Huggins RA., Behavior of some Binary Lithium Alloys as
Negative Electrodes in Organic Solvent-Based Electrolytes. J. Electrochem. Soc. 1986;
133:457-60.
5. Anani A., Crouch-Baker S., Huggins RA. Kinetic and Thermodynamic Properties of
Several Binary Lithium Alloy Negative Electrode Materials at Ambient Temperature. J.
Electrochem. Soc. 1987; 134:3098-101.
6. Huggins RA., Polyphase Alloys as Rechargeable Electrodes in Advanced Battery
Systems. J. Power Sources 1988; 22:341-50.
7. Huggins RA., Materials Science Principles Related to Alloys of Potential Use in
Rechargeable Lithium Cells. J. Power Sources 1989; 26:109-20.
8. Idota Y., Kubota T., Matsufuji A., Maekawa Y., Miyasaka T. Tin-Based Amorphous
Oxide: A High-Capacity Lithium-Ion-Storage Material. Science 1997; 276:1395-97.
9. Yang J., Winter M., Besenhard JO. Small particle size multiphase Li-alloy anodes for
lithium-ion batteries. Solid State Ionics 1996; 90:281-87.
10. Yang J., Takeda Y., Li Q., Imanishi N., Yamamoto O. Lithium insertion into Sn- and
SnSbx-based composite electrodes in solid polymer electrolytes. Journal of Power
Sources 2000; 90:64-69.
11. Brousse T., Retoux R., Herterich U., Schleich D.M. Thin-Film Crystalline SnO2-Lithium
Electrodes. J. of Electrochemical Society, 1998;145:1-4.
12. Remy, Heinrich, Lehrbuch Der Anorganischen Chemie. B. 1, Leipzig: Academische
Verlagsgesellschaft Geest und Portig K.-G., 1960.
13. Tamura N., Ohshita R., Fujimoto M., Fujitani S., Kamino M., Yonezu I. Study of some
Li Alloy for Lithium Secondary Battery. Proceedings of Joint (ECS & ISE) International
Meeting; 2001 2-7 September; San Francisco, 2:Abstract No. 251, 2001.
14. Hansen, Max, Constitution of Binary Alloys, New York: McGraw-Hill, 1958.
15. Ulis A., Rosenberg Y., Burstein L., Peled E. Degradation of Carbon-Supported Tin
Alloys in LiPF6 EC:DEC. Proceedings of Joint (ECS & ISE) International Meeting; 2001
2-7 September; San Francisco, 2: Abstract No. 225, 2001.
17. Turner R.L., Amik B., Krause L.J., Christensen L., Dahn J.R. Electrochemical Characte-
ristics of Sn-Mo Anodes for Lithium Batteries. Proceedings of Joint (ECS & ISE)
International Meeting; 2001 2-7 September; San Francisco, 2: Abstract No. 257, 2001.
COMPOSITE ANODE MATERIALS FOR
HIGH ENERGY DENSITY
Joseph S. Gnanaraj1, Malgorzata K. Gulbinska2, Joseph F. DiCarlo2*,

Igor V. Barsukov3, Nancy Holt3, Viacheslav Z. Barsukov4,
Joseph E. Doninger5*
1
Worcester Polytechnic Institute, Chemical Engineering Department, 100 Institute Road,
Worcester, MA 01609 USA;
2
Lithion, Inc., 82 Mechanic Street, Pawcatuck, CT 06379, USA;
3
Superior Graphite, 10 S. Riverside Plaza, Chicago, IL 60606, USA;
4
Kiev National University of Technology and Design, Electrochemical Power Engineering &
Chemistry Department, 2, Nemirovich-Danchenko street, Kiev, 252011, Ukraine;
5
Dontech Global, Inc. 427 E. Deerpath Road, Lake Forest, IL 60045, USA
Abstract
Novel composite anode materials were prepared for use in lithium-

ion batteries. The composite material formation involved coating of
powdered substrates (graphite-, or silicon-based), thus modifying their
surface properties and cycling performance. Applying an additional surface
coating resulted in the production of composites with enhanced stability in
the practical lithium-ion battery environments. Coin cell cycling, as well as
full 7Ah prismatic cell cycling, and impedance spectroscopy methods were
used to evaluate the electrochemical properties of coated materials.
Keywords
Lithium-ion batteries; high energy density; purified natural graphite;

carbon coated silicon; graphite/metal composites; chemical vapor deposition.
1. INTRODUCTION
Carbonaceous materials with varying degree of graphitic order are
the most common commercial anodes in secondary lithium-ion batteries.
Among carbon-based materials, natural graphite is the most promising anode
*
Corresponding authors. E-mails: [email protected] and [email protected]
317
318
material because of the high theoretical lithium intercalation specific

capacity of 372 mAh/g, flat voltage potential profile, low irreversible
capacity, and low cost [1]. However, graphite electrodes typically fail in a
number of commonly used electrolyte solutions, such as propylene carbonate
(PC). The use of PC based electrolyte in the Li-ion battery would improve
the rate performance, the safety of the battery, its low temperature
performance and, eventually, costs. In addition, the rapidly increasing
market for rechargeable lithium-ion batteries (portable computers, cellular
phones, a variety of military applications) calls for batteries with much
greater capacities (>500 mAh/g) than the capacity values inherent to
carbonaceous materials.
Silicon-based materials are listed among the promising novel anode
materials. One of the most important advantages of silicon-based anodes is
their exceptionally high theoretical capacity (c.a. 4000 mAh/g). This
impressive theoretical capacity value is related to the high Li/Si ratio in
lithium silicides (e.g. Li4.4Si) formed in-situ in silicon-based anodes [2,3].
Unfortunately, tremendous volume changes in silicon are associated with the
formation of lithium silicides with a high Li/Si ratio. These volume changes
cause rapid mechanical degradation of the anode that eventually leads to the
loss of electrical contact between the current collector and the silicon-based
anode material [4]. Thus, the excellent initial capacities (over 1000 mAh/g)
offered by silicon-based anodes, fade dramatically during the subsequent
cycles. The addition of carbon to silicon powder greatly improves its
cycleability and prevents capacity fading that is observed with pure silicon
anodes [5, 6, 7].
In the second part of this study, purified natural graphite substrates
(PNG) were coated with Si and other metals, followed by an additional
coating with a protective layer of disordered carbon, in order to prevent the
electrolyte reduction and to increase the cycling stability of PNG. The use of
such composite carbonaceous materials comprising a “core” of graphite, a
second layer of an alloying metal, and a protective “shell” of disordered
carbon combines the favorable lithium storage properties of graphite,
reversible alloy formation reaction with lithium, with the solvent co-
intercalation retarding properties of non-graphitic carbons. The PNG surface
modification has been purposively narrowed down to silicon and aluminum.
These coatings were deposited onto the graphite surface to provide sites for
lithium intercalation and to form the electronic conducting layer in the
composite anode.
Fluidized-bed Chemical Vapor Deposition (FBCVD) was a method
of choice in forming both carbon, and silicon coatings on graphite and
silicon powders, respectively [8, 9]. The major advantage of fluidized bed
CVD method over other coating processes is the ability to deposit uniform
coating with the controlled thickness onto large quantities of substrate
particles (silicon and graphite powders, respectively). In the typical fluidized
319
bed CVD reactor, a non-reactive gas flows through a bed of particles, at a

carefully selected velocity. There is a drag exerted on the solid particles by
the flowing gas. However, when the gas velocity increases beyond a certain
value, the total drag on the particles equals the weight of the bed, and the
particles begin to lift and fluidize. The gaseous precursor is then introduced
into the reaction chamber in order to chemically react with the substrate
particles and to form the uniform coating layer.
Initially, cycling in the coin cells, and later, in full prismatic cells
with rated capacity of 7 Ah were used in our investigations. Also, advanced
impedance spectroscopy methods were used to evaluate the electrochemical
properties of coated materials.
2. EXPERIMENTAL
The coin half-cells were assembled using metallic lithium discs as
the positive electrodes. The negative electrodes were made up of composite
anode materials films spread onto the copper foil current collectors. Coated
powders were provided by Superior Graphite Inc. (Chicago, IL, USA). The
carbon and silicon coatings on graphite particles were done by fluidized-bed
chemical vapor deposition (FBCVD) method. Anodes were prepared by
mixing approximately 1g of natural graphite (PNG), coated with: (a) carbon,
(b) silicon (5.7 wt%), and (c) aluminum, respectively. Kynar (PVdF), Super
P carbon black, and NMP solvent were also added. The anodes were made
such that there was 87% anode material mixed with 10% Kynar, and 3%
Super P carbon. The amount of NMP used to make the slurry was adjusted
such that the slurry consisted of 47 wt% solids. The slurry was then
transferred onto the flat glass surface of hand coater. Anode slurry was
coated onto copper foil. The coating gap controlled the thickness of the
slurry being applied to the foil, which determined the weight loading. The
coated foil was dried in a vacuum oven for at least 12 hours at 110oC. The
coated foil was then blanked to the desired disc size. The surface area of the
electrode was 1.25 cm2. The two-electrode coin cells were assembled in a
glove box, as cited elsewhere [9]. The following electrolytes and/or their
mixtures were used in coin half-cells: 1M LiPF6 in mixture a of either
EC:DEC:DMC (1:1:1, v/v/v) or PC:DEC:DMC (1:1:1, v/v/v). The
electrochemical measurements included galvanostatic charge-discharge tests
on the graphite electrodes. Freshly prepared natural graphite electrodes had
an open circuit potential of ca. 3.0 V (vs. Li/Li+). Galvanostatic cycling tests
were performed at 0.2, 0.4 mA/cm2 current densities.
The full lithium-ion prismatic cells with a rated capacity of 7Ah
were assembled and evaluated in the Lithion, Inc. battery test facility. The
PNG-based anode materials from Superior Graphite Co. were coated onto
copper foil and blanked into anode components for assembly of the so-called
320
Lithion’s “low rate” 7Ah cells. The cells were fabricated with Setella E20
separator, 1M LiPF6 in EC:DMC:DEC (1:1:1) electrolyte and standard
cathode components. The charge capacity of each cell is designated at 9 Ah
during the C/20 formation cycle. This value is based on a cathode specific
capacity of 190 mAh/g at the C/20 rate. The 8 Ah designation for each cell
during the initial two C/10 charge cycles is based on a specific cathode
capacity value of 170 mAh/g. After a preliminary evaluation of the three
formation cycles, the cells were sealed and cycled for additional 7 cycles to
stabilize data for the C/10 rate. A 72 h stand test was performed on cells after
completion of the ten formation cycles.
A Maccor multichannel system (Model 2000) tester and an Arbin
Instruments Inc., computerized multichannel battery tester were used for
galvanostatic cycling. For PITT measurements, a computerized Solartron
1286 electrochemical potentiostat interface driven by Corrware (Scribner
Associates, Inc. Southern Pines, NC) was used. Impedance spectra were
measured using the Solartron 1250 Frequency Response Analyzer (FRA)
driven by a Pentium-Pro PC. The amplitude of the ac voltage was 5 mV and
in the frequency range of 65 kHz – 5 mHz in automatic sweep mode from
high to low frequency with 72 points. The cells were aged by 10-15 cycles of
galvanostatic cycling at the current density of 0.2 mA/cm2 and at appropriate
equilibration potentials in open circuit conditions. Unless stated otherwise,
all the tests were performed at a constant temperature of 20oC in an
incubator.
Scanning Electron Microscopy images of powders used in this paper
were taken using JEOL’s JSM-6320F instrument at the Illinois Institute of
Technology and at Drexel University, Philadelphia, PA, USA.
3.1. Cycling of Silicon

Volume changes in silicon during lithium-ion battery cycling are
believed to be related to the formation of lithium-silicon alloys with a high
content of lithium (e.g., Li4.4Si). The volume changes lead to anode
pulverization that eventually causes the loss of electrical contact between the
current collector and the silicon-based anode material. Thus, the initial
capacities of over 1000 mAh/g offered by silicon-based anodes, fade during
the subsequent cycles (Figure 1). The data generated in our laboratory show
that the carbon coating on silicon powder prevents capacity fading that is
observed with pure (uncoated) silicon anodes.
321
Capacity (mAh/g)
Time (Hours)
Figure 1. Capacity vs. time plot for uncoated silicon anode material.
3.2. Carbon / Silicon Composites

The ball milled silicon powder was ground in an agate mortar in
order to break up any agglomerates formed during storage and/or
transportation of silicon powder. About 0.5 g of sample was loaded into the
fluidized bed CVD reactor. The temperature was raised to 950°C. Upon
reaching 950°C argon carrier gas was passed through a bubbler filled with
liquid toluene at a flow rate of 100 mL/min. After 30 min the carrier gas
flow was stopped and the reactor was allowed to cool down.
The cycling properties of selected carbon-coated silicon powders
were evaluated. The carbon-coated silicon powdered samples were cycled in
the half-cells against the lithium metal anodes. In the preliminary coin cell
test of the silicon powders, capacity cutoff was limited to 500 mAh/g in the
voltage range of 0.01 to 1.3V. Galvanostatic cycling tests were performed at
0.2, 0.4 mA/cm2 current densities. The charge-to-discharge efficiency was
nearly 100% for the 27 cycles where discharge capacity was 500 mAh/g (20
initial cycles are shown in Figure 2). The coin cell cycling results showed an
improved electrochemical performance of the carbon-coated silicon powders
(Figure 2). However, we learned that optimization of the FBCVD process is
required to achieve uniform carbon coating thickness on silicon powders in
large quantities (>100 g).
322
600
500
Capacity (mAh/g)
400
300
200
100
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Cycle Number
Figure 2. Capacity vs. time plot for cell containing carbon-coated silicon powder material.
3.3. Graphite / Metal Composites

In order to be able to work with scaled up quantities of the higher
capacity material, it was decided to continue the work described in section
3.2 and coat Si onto carbon attempting to combine high capacity of Si and
stability of performance characteristics of graphite substrates.
Two types of graphite precursors were used in this study as
substrates for deposition by the alloying metals. The origin of the precursors
was the same - purified crystalline flake graphite (commercial grade 2900G,
being manufactured by Superior Graphite Co. of Chicago, IL, USA). This
material is a 300+µm particle size flake of high purity (99.95%C), as
achieved by the process of thermal purification [10]. The high reso-
lution scanning electron micrograph (SEM) of such graphite is shown by
Figure 3(a). The 2900G graphite has been converted through steps of
grinding and classification into the 15 µm sized platelet (Figure 3b) and
spheroidal (Figure 3c) structures.
Both carbon materials were tested for their initial electrochemical
performance in the 2-electrode electrochemical cells with Li metal as a
counter electrode. Our findings have shown that with both types of carbon
materials, achieving near theoretical reversible capacity upon Li+
deintercalation was possible. Thus, in a typical half cell environment (a
CR2016 type coin cell with graphite and Li metal electrodes, a 1M LiPF6,
323
EC:DMC (1:1) electrolyte), at the formation rate of C/15, these carbons

achieved the reversible capacity in between 350 - 360 mAh/g [10].
On the next stage, the two carbon precursors were coated with
metals, which are known to form reversible alloys with lithium. It is
noteworthy mentioning that the attempts to start using the process of
reversible formation of the alloys between Li and certain metals have a long
legacy and date back to the earlier works aimed at prevention of dendrite
formation in the lithium metal – based electrochemical cells [11]. From the
many metals, which form reversible alloys with Li, we have chosen to work
with those, which are known to form Li alloys at the potentials, which
coincide with or are as close as possible to those, at which the reversible
intercalation of Li+ into crystalline flake graphite takes place. Such approach
will enable maintaining the same discharge voltage in the full lithium-ion
cells as in those, based on conventional graphitic anodes.
(a)
(b) (c)
Figure 3. Scanning Electron Microscope Images of thermally purified natural crystalline
flake graphite 2900G (a), and its ground versions displaying
platelet (b) and spheroidal (c) morphologies.
The choice of metals has been narrowed down to Si and Al for this
group of experiments. The approximate potentials of the reversible
transformation (subject of polarization) of C into LiC6 is taking place in the
range of 0.6 – 0.01 V (vs Li/Li+); while for conversion of Si into Li22C5 it is
in the range of 0.3 – 0.05 V (vs Li/Li+), and it occurs at approximately 0.27
V (vs Li/Li+) for reaction of conversion of Al into LiAl [12, 13].
324
For illustrative purposes, the process of deposition of Si onto

graphite is being used as an example. The 15 Pm natural graphite precursors
were introduced into the industrial size chemical vapor deposition reactor,
where a thermal decomposition of silane (SiH4) into the silicon and hydrogen
was taking place under inert gas in accordance with the equation (1):
Figure 4. Less uniform Si coating on plate-like graphite (a), rather uniform Si coating on
spheroidal graphite (b) and a magnified view of Si particulates on spheroidal graphite (c).
Silane decomposes to its elements at above 400qC. Process (1) is

known as direct thermal decomposition, and produces either amorphous or
polycrystalline Si (function of reaction temperature and other processing
parameters), and is commonly used, for instance, in the solar cell industry to
reduce silane to silicon.
The process was seen to be highly sensitive to impurities [14],
among which are oxygen and water groups with basic pH. Their presence
even in minute quantities may result in occurrence of the side reactions (2)
and (3), respectively:
SiH4(g) + 2O2(g) o SiO2(s) + 2H2O; (2)

325
SiH4(g) + (2+n)H2O(g) o SiO2.nH2O(s)+ 4H2(g). (3)
It was determined that plate-like morphologies are less desirable

substrates for metal deposition, as compared with spheroidal substrates in
terms of uniformity of metal deposition. This observation is consistent with
classical teachings about metal deposition and corresponding crystal growth
on foreign substrates [15]. In accordance with these teachings, metals get
deposited on the most active sites of the substrate, rather than uniformly, on
its entire surface. The most active sites are those, which feature cavities,
cracks, pores, defects of crystallites and other un-uniformities. It is evident
from the SEMs shown by Figures 3(b) and 3(c) that spheroidal graphite
precursor had many more active sites, which allowed for more uniform
distribution of the Si (please see a non-uniform and a rather uniform
distribution of Si on two different substrates, as shown by Figures 4(a), and
4(b), respectively).
It is noteworthy that while the first droplets of metal deposit are
being held to the substrate graphite, most likely, by particle-to-particle
electrostatic forces, subsequent layers of metal tend to grow where the
energy is lowest, e.g. on top and around the first layers thus creating a
thermodynamical prerequisite for making a non-uniform coating. The
smooth parts of the surface remain passive and are hard to coat (please see
uncoated particles in Figure 4(a)).
Therefore, due to considerations of uniformity of coating, in further
tests it was decided to focus on the spheroidal metal-graphite composites.
A departure from traditional approaches of depositing a significant
amount of metal onto carbon precursor differentiated our experiments from
the state of the art literature. Through several trial and error experiments, it
was possible to formulate several key principles, which should constitute a
design of the metal/carbon composites. Among those are:
(1) the amount of the alloying metal on carbon has to be kept at below
10wt%, otherwise it will be difficult to cope with large volume
changes of metal upon alloying – de-alloying reaction;
(2) the particles of alloying metal have to be as small as possible,
preferably nanosized or submicron (the reason is the same –
suppression of the volume changes, as well as enhancement of
electrostatic adhesion between the core graphite and the metal
particle, which typically goes up as metal particles become smaller
in size);
(3) the alloying metal has to be uniformly coated onto graphite substrate
(a 15 Pm base in our tests), and then it has to be additionally coated
with a stabilizing coating (in order to suppress metal’s volumetric
expansion in the direction normal to the direction of core 15 Pm
graphite base).
326
Figure 4(c) depicts an image of the Si-coated spheroidal natural

graphite substrate at higher magnification, as derived from Figure 4(b). The
size of the substrate graphite particle is estimated to be about 15 Pm, while
the size of deposited Si particles appears to be in the range of 0.1-0.35 Pm.
The comparative 2016 coin cell data of the uncoated vs. Si-coated
spheroidized natural graphite precursor vs Li metal counter electrode
(Figure 5) reveals a remarkable effect of metal addition onto the reversible
capacity of anode. Thus, the starting material’s reversible capacity increased
from approximately 350 mAh/g to over 510 mAh/g (data taken at 0.8V vs
Li/Li+), which is 1.37 times higher than the theoretical value of the reversible
capacity for graphite.
Figure 5. First charge and discharge of CR2016 coin type half-cells with anodes based on the
15 Pm, spheroidized natural graphite (bold line) and the similar graphite modified by
8wt% Si at the C/15 rate.
Unfortunately, later findings resulted in an observation that

composite materials lost their capacity upon further cycling unless the outer
surface of particles was stabilized with an additional passivating coating,
whose nature is described in a greater detail in [16] for the example of
Superior Graphite’s SLC product line. This outer stabilizing shell is made
out of rigid disordered carbon material and uniformly coats each graphite
particle. In the preliminary coin cell investigation, the outer shell was seen to
be suppressing the volumetric changes in metal, thus improving stability of
materials with cycling. It was observed that the surface shell was proven to
be more efficient for the case of Al rather than for Si coatings.
327
Two most promising types of anode materials developed in this

study were characterized by Lithion, Inc., a specialty lithium-ion battery
manufacturer of Pawcatuck, CT, USA. The work has been carried out in
standard Lithion’s 7 Ah prismatic cells. The anode materials were:
spheroidized purified natural graphite coated with Si additionally coated
with a stabilizing carbon coating (sample is referred to as PNG-Si-2C), as
well as spheroidized purified natural graphite coated with Al and also with a
stabilizing carbon coating (sample is referred to as PNG-Al-2C).
Performance of these cells has been compared with that of the control
MCMB anodes.
Figure 6 compares the charge and discharge capacities vs. cycle
number during the formation cycle. Si-2C coated graphite shows the highest
irreversible capacity while MCMB has the minimum. Al-2C coating on
graphite shows a small positive impact on the irreversible capacity but at the
cost of capacity.
8.5
MCMB
8.0
PNG-Al-2C
Discharge Capacity (Ah)
7.5
PNG-Si-2C
7.0
6.5
6.0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Cycle Number
Figure 6. Prismatic 7 Ah battery; discharge capacity vs. cycle number plot.
Figure 7 shows the charge/discharge capacities vs. cycle number at

different temperatures. It is interesting to note that the capacity at room
temperature is regained after each low temperature cycling. Although Si-2C
coating shows better capacity than that of Al-2C at low temperature, the
capacity fading is large on long term cycling. At the room temperature,
328
although Al-2C coating shows lower capacity, it was seen to be stable on

cycling while Si-2C coating lost the capacity dramatically.
Figure 7. Prismatic 7Ah battery. Typical charge-discharge capacities vs. cycle number.
Battery was cycled at different temperatures at C/4.
Figure 8 shows a family of impedance spectra for three types of

carbon in full 7 Ah batteries. Four cells with each type of carbon were
measured for impedance. The results appear to be very interesting. The
original impedance data appears to be shifted on the real axis but when it
was corrected it matches one another. Most of the times in the literature, it is
advised to take high frequency intercept and consider it for the impedance
value. This data provides valuable insight in to the real situation. The dc
resistance value calculated from current interrupt measurements (done
separately and not reported here) correlates well with the impedance’ total
resistance (extrapolation to <5mHz). The semicircle at high frequency side
represents the SEI layer and shows that the metal coating actually increases
the SEI resistance – a phenomena, which deserves a separate investigation.
Figure 9 shows the 72-hour charge stand test. The cells were charged
to 4.1 V and stored at room temperature, followed by measuring their
voltage change over the period of 72 hours. A known drawback of non-
coated natural graphites is their inability to hold voltage on this test. It was
329
pleasantly surprising to see the metal-coated natural graphite cells to hold

charge quite well (on average better than MCMB control).
Figure 8. Families of impedance spectra. Frequencies between 65 kHz and 5 MHz.
4. CONCLUSIONS
Carbon coated Si-based anode materials were successfully
synthesized by chemical vapor deposition methods. Coated materials
performed well during the coin cell testing. Future work on this section
330
involves assembly of 7Ah cells and the optimization of the FBCVD process
parameters to achieve more uniform carbon coatings.
Stable cycling was achieved in the full 7Ah cells with a composite of
spherical natural graphite coated with Al and then stabilized with a rigid
carbon coating of the disordered nature. Further investigation is needed to
fully understand the effect of rigid carbon shell on the electrochemical
performance of graphite-based composite materials.
4.1
MCMB
4.1
PNG-Al-2C
4.1
PNG-Si-2C
Cell Voltage (V)
4.1
4.1
4.1
4.1
4.1
0 10 20 30 40 50 60 70 80
Test Time (hours)
Figure 9. Prismatic 7Ah batteries. 72-hour charge stand test.
ACKNOWLEDGEMENTS
Authors representing Lithion, Inc., Superior Graphite Co. and
Worcester Polytechnic Institute would like to thank the Army Research
Laboratory for its partial financial support of this work through contract
number DAAD17-01-C-0044.
Prof. V. Barsukov acknowledges financial support by the NATO
Science for Peace project “SfP 973849”, entitled: “Carbons as materials for
the Electrochemical Storage of Energy”. Him and Prof. J. Doninger wish to
thank NATO’s Dr. F. Pedrazzini (Program Director – Advanced Research
Workshops), Dr. Chris De Wispelaere (Program Director - Science for
Peace), and Chris Hilton (Program Manager – Travel Grants Program at the
U.S. Civilian Research and Development Foundation, Arlington, VA, USA),
331
whose efforts and coordination enabled bringing together parties – authors of

this applied battery work.
REFERENCES
1. M. Wakihara, Materials Science and Engineering, R33, 109 (2001).
2. W. J. Weydanz, M. Wohlfahrt-Mehrens, R. A. Huggins, J. Power Sources, 81, 237
(1999).
3. S. Bourdeau, T. Brousse, D. M. Schleich, J. Power Sources, 81, 233 (1999).
4. H. Li, X. Huang, L. Chen, Z. Wu, Y. Liang, Electrochemical and Solid State Letters, 2,
11 (1999).
5. M. Yoshio, H. Wang, K. Fukuda, Y. Hara, Y. Adachi, J. Electrochem. Soc., 147, 1245
(2000).
6. C. S. Wang, G. T. Wu, X. B. Zhang, Z. F. Qi, W. Z. Li, J. Electrochem. Soc., 145, 2751
(1998).
7. P. C. Eklund, J. M. Holden, and R. A. Jishi, Carbon, 33, 7 (1995).
8. M. Gulbinska, F. S. Galasso, S. L. Suib, S. Iaconetti, P.G. Russell, J. F. DiCarlo, Mat.
Res. Soc. Symp. Proc., 2003, 756, EE6.14.1- EE6.14.1.6.
9. S. Iaconetti, P. Russell, J. F. DiCarlo, M. Gulbinska, S. L. Suib, J. S. Gnanaraj,
Proceedings of the 41st Power Sources Conference, 76-78, (2004).
10. I. Barsukov, Battery Power Products and Technology, 4, 9 (2000), 30.
11. Kabanov, B.N. and Chekavtsev, A.V. Electrochemistry of the Li-Al alloy and problems
of creation of new power sources; in “Itohy Nauki i Tekhniki”, Electrochimia eds.,
“VINITI” Publishing House, Moscow, Russia, 21, 140-176 (1984) – in Russian.
12. Kedrinskiy I.A., Yakovlev, V.G. Li-Ion Accumulators. Publishing House “IPK Platina”,
Krasnoyarsk, Russia, 124 (2002) – in Russian.
13. Winter, M. Besenhard, J.O. Electrochemical lithiation of Tin and tin-based intermetallics
and composites. Electrochim. Acta, 45, 1-2, 31-50 (1999).
14. Cotton, F.A., Wilkinson, G., Murillo, C.A., and Bochmann, M. Advanced Inorganic
Chemistry, Sixth Ed., Wiley, (1999).
15. Applied Electrochemistry, Kudryavtsev N.T. Ed., Second Ed., “Khimia” Publishing
House, Moscow, Russia, 334-336 (1975) – in Russian.
16. F. Henry, I. Barsukov, J. Doninger, S. Anderson, P. Booth, P. Zaleski, R. Girkant, D.
Derwin, M. Gallego, T. Huerta and G. Uribe. New Developments in the Advanced
Graphite for Lithium-Ion Batteries – in this book.
ELECTROCHEMICAL ACTIVITY OF CARBONS
MODIFIED BY d-METAL COMPLEXES
WITH ETHANOLAMINES
L. G. Reiter1, V. A. Potaskalov1, A. A. Andriiko1, V. S. Kublanovsky2,

M. A. Chmilenko3, Yu. K. Pirskiy2, V. I. Lisin3, and S. M. Chmilenko3
1
National Technical University of Ukraine “Kyiv Polytechnic Institute”,
Chemical Technology Faculty, 37 Prospect Peremogy, 03057, Kyiv, Ukraine;
2
Institute of General and Inorganic Chemistry, Ukrainian National Academy of Science,
32/34 Prospect Paladina, 04164, Kyiv, Ukraine
3
Department of Electrochemical Power Engineering, Ukrainian National Academy of
Science, 38-A Boulevard Vernadskogo, 04164, Kyiv, Ukraine
1. INTRODUCTION
The organometallic complexes with d-metals are considered as
promising electrocatalysts for oxygen electroreduction in air-metal
electrochemical cells. Obviously, the first idea was to employ the catalytic
mechanism of the oxygen reduction with porphyrin-like metal complexes [1]
found in living beings (Figure 1).
Figure1. Fragment of the structure of natural catalysts, N4-metal complexes.
Numerous N4 complexes of Co(II) [2-5], Fe(II) [6,7], Fe(III) [8], Cu

(II) [9,10], and Mn(II) [11] were studied as electrocatalysts for modifying of
the electrode surface for O2 reduction. It was found that, varying ligands and
333
334
metal atoms, as well as by applying thermal treatment, one could obtain the
catalysts with the parameters appropriate for metal-air batteries and
electrochemical generators. Unfortunately, the N4 complexes with porphyrin-
like ligands are too expensive for mass production battery technology.
It is known that some spinel-structured 3d-metal oxides are good
catalysts for many processes involving electron transfer [12]. However, their
low conductivity does not allow for the direct use in the electrode of the
battery, and grafting them onto the carbon matrix is also very difficult
technical problem. It was found recently that this problem could be solved
indirectly, creating the “spinel” catalytic centers on the surface of carbon by
means of adsorption of some 3d-metal complexes on the graphite surface
followed by subsequent pyrolysis at certain temperatures [13,14].
In this work we have studied the preparation of electrocatalysts on
the graphite matrix using tri-nuclear complexes of 3d-metals with
aminoalcohol ligands. Tri-nuclear complexes, 2[Co(Etm)3]*Me(NO3)2,
where Etm = ethanolamine, Me = Zn2+, Cu2+, Ni2+, Co2+, were investigated.
As it was shown in [15-18], such compounds, with bridging atoms
of deprotonated amino alcohol, are formed when aminoalcoholate complexes
of metal (III) react with bivalent 3d-metal ions. Many representatives of
these compounds were synthesized in crystalline state; the polynuclear
compounds were also found to form in aqueous and methanol solutions. The
structure of 2Co(III) – Ni(II) tri-nuclear complex, according to [17,18], is
shown below in Figure 2.
Figure 2. Structure of tri-nuclear complex ion [Ni(Co(Etm)3)2]2+.

335
Being thermally decomposed onto the surface of carbon, this

complex is expected to form very small catalytically active NiCo2O4 spinel
centers. Thus, we have studied the catalytic activity of the products of
pyrolysis at different temperatures toward two electrochemical reactions –
reduction of oxygen in alkaline electrolyte and intercalation of lithium into
carbons in aprotic electrolyte of Li-ion battery. To our knowledge, the
catalytic effect of the metal complexes in the second reaction was not yet
considered in the literature.
2. EXPERIMENTAL
Syntheses of the tri-nuclear complexes were performed in water-
methanol solution according to the reaction
2 fac- Co(Etm)3 + Me(NO3)2 o [Me(Co(Etm)3)2](NO3)2. (1)
The fac- isomer of trice-aminoethanol complex of cobalt (III), with 3

coplanar oxygen atoms, was prepared according to the method [16].
The electrocatalysts for oxygen reduction were prepared as follows.
These complex compounds were inoculated onto the carbon (AG-3, BET
area near 800 m2/g) by means of adsorption from dimethylformamide
solutions. The portion of complex compound weighed so as to achieve 3% of
Co content was mixed with the carbon, then 5 ml of dimethylformamide per
1 g of the carbon were added and the mixture was cured at room temperature
for 24 hours. Series of samples obtained were thermally treated (pyrolyzed),
and the resulting grafted carbons were tested as electrode materials in the
reaction of molecular oxygen reduction.
Electrochemical performance of the modified carbons was examined
by “floating” gas diffusion electrode [19] in an electrochemical cell with
separated cathode and anode chambers at room temperature with PI-50-1.1 -
potentiostat in 1N KOH aqueous solution.
The partial oxygen pressure was checked with an OP-210/3
microanalyser [20] before and after the bubbling of oxygen through the
electrolyte. The potential was measured against an Ag/AgCl reference
electrode.
Further on, the Co-Ni complexes were used for modification of
Hohsen Carbon type (10-10) and Hohsen Graphite type (10-28) anode
materials for Li-ion batteries applying similar procedure. These anode
materials were tested in 2016 size lithium coin cells with a configuration
Li/electrolyte; (LP-30)/(modified anode material). The coin cells were
assembled by standard technology in dry atmosphere of a glove box and then
336
cycled with PC-governed galvanostatic equipment in the voltage window of

10 mV to 1.5 V.
3.1. Electrochemical Properties of Modified Carbons in the

Reaction of Oxygen Reduction
It was established that carbons grafted with pyrolyzed Co(III)-Me
complexes increase its electrocatalytic activity, and the temperature of
pyrolysis is very important. This can be seen from the plots in Figure 3. The
maximum catalytic effect is achieved for the samples annealed near 600qC,
decreasing at higher temperatures and almost vanishing at 800qC (Figure 4).
Initial
-2.5 300 C
800 C
-3.0 500 C
2
log j, A/cm
600 C
-3.5
-4.0
-4.5
-5.0
-5.5
0.05 0.10 0.15 0.20

- E, V
Figure 3. Current density versus electrode polarization (Ag/AgCl ref. electrode) for oxygen
reduction in 1M KOH at 20qC on Co-Ni modified AG-3 carbon.
Such regularities indicate that some active catalytic centers are

formed at 600qC and destroyed at higher temperatures. As follows from the
formula (Figure 1), the initial structure in the course of pyrolysis on the
surface of graphite could rearrange itself into spinel-like groups bonded to
the surface via remaining nitrogen atoms, MeCo2Oy –(N)– (Cx), as it was
already presumed in ref. [13]. This hypothetic mechanism is in agreement
with DTA data reported in [14] – though the individual polynuclear complex
compounds exothermally decompose totally at 220-240qC, they become
much more thermally stable being adsorbed onto the graphite matrix.
337
-3.0
AG-3(initial)
-3.5
Co-Ni
Co-Cu
2
Log j, A/cm
-4.0 Cr-Co
-4.5
-5.0
-5.5
0 100 200 300 400 500 600 700 800

Annealing temperature, C
Figure 4. Logarithm current density at –0.1 V vs. the temperature of pyrolysis for
trinuclear complexes on the surface of AG-3 carbon.
As is can be seen from Fig. 4, the better electrocataysts for oxygen

reduction are derived from Co-Ni tri-nuclear complex, [Ni(Co(Etm)3)2]
(NO3)2.
It also consists with the “spinel” hypothesis for the structure of
active catalytic sites: the spinel compound NiCo2O4 is well known for its
catalytic activity in the reaction of oxygen reduction [21].
3.2. Electrochemical Properties of Modified Carbons in the

Reaction of Lithium Intercalation
The electrochemical intercalation of Li was studied for carbon
electrodes modified by the 2Co-Ni complex, which showed the best effect in
the reaction of oxygen electroreduction. Galvanostatic charge-discharge
technique (PC governed automatic bench) in 2016 coin type cells was used
for this purpose.
The electrode active masses were prepared by mixing the active
material with 10 wt.% of polyvynilidene fluoride slurred in 1-methyl-
pyrrolidone solvent. The actual mass was then pasted onto one side of pre-
cleaned Co foil, dried for 4 h at 100qC, pressed and the disks of 15,6 mm
diameter were cut and placed into the cell’s coins. Large excess of Li metal
(foil) was used as a counter electrode. 1M LiPF6 solution in mixture of
ethylene carbonate (50 vol.%) and methyl carbonate was used as an
electrolyte (Merck product LP30).
338
3.2.1. Carbon-Type Material
Figure 5 shows galvanostatic charge-discharge voltage profiles of

the initial non-modified Hohsen Carbon-Type material. This is a typical
behavior of non-graphite carbon, with sloped voltage-time curves, more than
20% first cycle irreversible capacity loss and good cycle retention with
moderate reversible capacity (180-200 mAh/g depending on the current).
E/V 80
1.61
1.41 33
1.21
65
1.01
0.81 1
0.61
0.41
0.21
0.01 t/min
0 100 200 300 400
Figure 5. Charge-discharge curves of Hohsen Carbon-Type Material

in the course of long-term cycling.
3.0
initial
2.5 300
500
2.0 700
U, V
1.5
Carbon - 1st charge
1.0
0.5
0.0
0 50 100 150 200 250 300 350
Q, mAh/g
Figure 6. The first charge curves of Hohsen Carbon-Type material modified with
Co-Ni complex. Annealing temperatures are shown in the figure. Rate ~ 0.1 C (10 hours).
339
The modification with Co-Ni complex pyrolysed at different

temperatures affects the shape of first charge curve as shown in Figure 6.
The sample-300 (annealed at 300qC) increases the charge capacity while
sample-500 decreases, sample-700 remaining almost unchanged. However,
no gain in the discharge capacity is observed for all but sample-500 modified
materials (Figure 7). The sample-500 increases its first discharge capacity by
5%, and, more remarkably, decreases its first cycle capacity loss from 26 to
9%. However, these advantageous features of the modified material
disappear in course of long-term cycling of the cells (Figure 8).
350
300 Q c1
Q d1
Q, mAh/g
250
200
150
100
0 100 200 300 400 500 600 700 800
Annealing tem perature, C
Figure 7. Specific capacity during the first charge (solid symbol) and discharge
of modified carbon-type materials.
500
300
300
280 700
260 initial
240
220
200
180
Q, mAh/g
160
140
120
100
80
60
40
20
0 10 20 30 40 50 60 70 80
Cycle No
Figure 8. Behavior of modified carbon-type materials

in the course of galvanostatic cycling, current load 30 mA/g.
340
As it is seen from the data of Figure 8, all modified materials have

poor cycling performance; their reversible capacities fade faster than the one
of initial non-modified material, and become lower after the first 8-10
charge-discharge cycles. Thus, we can conclude that no positive effect is
achieved by means of modification of the Carbon-Type material with
bimetal tri-nuclear complex of Co(III)-Ni(II).
3.2.2. Graphite-Type Material
Figure 9 shows voltage profiles of the initial and modified Graphite-

type materials at first charge (A) and subsequent discharge-charge (B)
processes.
3.0
G initial G raphite - 1st charge

2.5 0.4
G 300
2.0 G 500 A
E, V
1.5
G 700 0.2
1.0
0.0
50 100 150 200 250 300
0.5
0.0
0 50 1 00 150 200 250 300 350
1.4
1.2 1st discharge 2nd charge

1.0
0.8
E, V
B
0.6
0.4
0.2
0.0
0 100 200 300 400 500
Q , m A h /g
Figure 9. Voltage profiles of Hohsen Graphite-Type Material modified with

the Co-Ni complex: A – first charge, part of it is enlarged in the insert;
B – next discharge and charge half-cycles.
341
The higher first charge capacity and small plateaus at 0.16 and 0.10
V on the curves of “low-temperature” sample G300 (annealed at 300°C,
Figure 9A) are probably the indication of the reduction of remaining metal
complexes in this sample. This process is irreversible, because the plateaus
do not appear in the discharge and next charge curves. The curves for other
samples are typical for graphite electrode materials of Li-ion batteries.
The irreversible loss of capacity at first charge is an adherent
property of each carbon anode material for Li-ion battery since it provides
the charge necessary to buildup the Solid Electrolyte Interface (SEI), which
is necessary for reversible operation of the battery. As Figure 10A shows,
this quantity is larger for modified materials than for initial one (8%), being
at maximum for G300 (26%) and descending to 10% for G700, which is
close to the value of the initial material (8%).
30 A
25
K, %
20
15
10
0
0 100 200 300 400 500 600 700 800
340
320
B
Q, mAh/g
300
280
260
240
220
200 10 m A/g
180
25 m A/g
160
140
120
100
0 100 200 300 400 500 600 700 800
Annealing T em perature, C
Figure 10. Irreversible capacity losses (A) and reversible capacity at 10-th cycle (B) for two
specific current values of modified Graphite-type materials
annealed at different temperatures.
342
The reversible capacity over the first few cycles (Fig. 10B) differs
depending on the temperature of pyrolysis. In this respect the initial and
“high-temperature” G700 sample are almost the same; the capacity is lower
for “low-temperature” G300 sample and higher for the intermediate G500
sample. Thus, grafting of the Graphite-type anode materials with pyrolyzed
at 500qC Co-Ni complexes enhance their electrochemical performance. The
effect is the greater the higher is the charge-discharge current. According to
the data in Fig. 10B, the capacity gain is 13% at specific current 10 mA/g
and 21% at 25 mA/g.
The next important property of an electrode material is cycle
retention in course of long-time operation. Figure 11 provides such data for
the modified graphite materials.
350 initial
Graphite, 20 mA/g 500 C
300 700 C
300 C
250
Q, mAh/g
200
150
100
50
0
0 10 20 30 40 50
Cycle No
Figure 11. Capacity fading of initial and

modified Graphite-Type materials in course of continuous cycling.
The reversible capacity of the modified G500 material is seen to be

constantly higher than that of the initial graphite while the performance of
other modified samples is worse.
Thus, unlike in the case of Carbon-type samples, modification of the
Graphite-type anode materials with pyrolyzed at 500qC tri-nuclear
[Me(Co(Etm)3)2](NO3)2 complexes enhance their electrochemical perfor-
mance, especially at higher charge-discharge rates. The inner mechanism of
such effect still remains unclear. We suggest the formation of nanosize
cataliticaly active spinel-like groups, which promote the charge transfer
across electrolyte/graphite interface. However, why this mechanism does not
work in case of non-graphitized carbon material? We do not know the
answer yet. Additional research is necessary in order to clear out the details
343
and optimize the performance of the grafted anode materials for its
implementation in future battery technologies.
4. CONCLUSIONS
1. Modification of carbon materials by tri-nuclear complexes of 3d-
metals with ethanolamine ligands increases the catalytic activity with regard
to the electrochemical reaction of oxygen reduction. The Co-Ni complex is
most active in this reaction if pyrolyzed at 600qC.
2. Modification of Graphite-type materials by the Co-Ni complex
pyrolyzed at 500qC improves significantly the performance of these material
in the anode of Li battery. The inner mechanism of this effect is still unclear.
We suggest the formation of nanosized spinel structures which facilitate the
charge transfer across SEI.
3. No positive effect of the modification was found for Carbon-type
(non-graphite) materials.
ACKNOWLEDGEMENTS
Part of this work was funded by the State Foundation of
Fundamental Research of Ukraine, grant # F7/217-2001. The author A. A.
Andriiko gratefully acknowledges the support by Civilian Research and
Development Foundation (travel grant TGP-1056), which enabled
presentation of this work at the NATO CARWC.
REFERENCES
1. Tarasevich M.R., Radyushkina K.A. Catalysis and Electrocatalysis with Metalporphyrins
(in Russian). M.: Nauka, 1982, 168p.
2. Shi C., Anson F.C. Inorg. Chem. 1995; 34:4554.
3. D’Souza F., Hsieh Y.Y., Deviprasad R.G. J. Electroanal. Chem. 1997; 426:17.
4. Bhugun I., Anson F.C. J. Electroanal. Chem. 1997; 430:155.
5. Shamsipur M., Salimi A., Haddadzadeh H., Mousavi M.F. J. Electroanal. Chem. 2001;
517:37.
6. Forster P.A., Kuwana T. Inorg. Chem. 1983; 22:699.
7. Ozer D., Harth R., Mor U., Bettelheim A. J. Electroanal. Chem. 1989; 266:109.
8. Zhang J., Anson F.C. J. Electroanal. Chem. 1993; 353:265.
11. Bettelheim A., Ozer D., Parash R. J. Chem. Soc. Faraday Trans.1 1983; 79:1555.
12. Razina N.F. Oxide Electrodes in Aqueous Solutions (in Russian). Alma-Ata: Nauka,
1982, 160 p.
13. Kublanovsky V.S., Pirsky Yu.K. Russian J. Applied Chem. 2001; 27:1116.
344
14. Pirsky Yu.K., Levchuk Ya.N., Reiter L.G., Kublanovsky V.S. Ukrainian Chemical
Journal. 2003; 69(3-4):77.
15. Gerasenkova A.N., Udovenko V.V., Russian Journ. Inorg. Chem. 1972; 18:2446.
16. Udovenko V.V.,Stepanenko O.N., Eroshok B.G. Russian J. Inorg. Chem. 1974; 19:2455.
17. Stepanenko O.N., Rejter L.G. Ukrainian Chemical Journal. 1992; 58(12):1047.
18. Stepanenko O.N., Potaskalov V.A., Trachevskiy V.V. Coordination Chemistry (in
Russian). 2001; 27(1):57.
19. Steinberg G.V., Kukushkina I.A., Bagotskiy V.S., Tarasievich M.R., Electrochemistry
(Russian), 1979; 15:527.
20. Kublanovsky V.S., Pirsky Yu.K., Yakimenko N.G. Russ. J. Phys. Chem. 1997; 71(1):54.
21. Efremov B.N., Tarasievich M.R. “Electrocatalysis and Electrocatalytic Processes” (in
Russian). In Sbornik Nauchnykh Trudov (Collection of Sci. Transactions), Kiev:
Naukova Dumka, 1986; 44-71.
METAL-GRAPHITE COMPOSITS
AS MATERIALS FOR ELECTRODES
OF LITHIUM-ION BATTERIES
Ludmyla Matzui*, Mykhailo Semen’ko, Mykol Babich, Leonid Kapitanchuk

Kyiv Taras Shevchenko National University, Dept. of Physics,
64,Volodymyrska Street, Kiev, Ukraine.
1. INTRODUCTION
Carbonaceous materials such as graphite, hard carbon, pyrolytic
carbon and nanotubes have been extensively investigated as anode materials
for Li-ion batteries. Among many types of carbon anodes that have been
reported, several types of carbon materials exhibit superior performance
from the point of view of reversible capacity and cyclic efficiency of the
lithium-ion batteries. These carbon and graphite materials possess different
microstructure, texture, crystallinity and morphology. Two typical types of
carbon materials, namely, highly ordered graphite treated at the temperature
of about 30000C and non-graphitized carbons treated at low temperatures of
about 10000C have been used in commercial batteries. The precursor
materials include cokes, polymers, fibers etc. Besides, the behavior and the
mechanism of lithium ions penetration into different kinds of carbon and
graphite host have been extensively studied both experimentally and
theoretically. As for the highly ordered graphite, the charge/discharge
reaction is based on Li+ intercalation and deintercalation, and theoretical
lithium storage capacity of graphite anode for lithium-ion secondary battery
has been considered to be 372 mAh/g [1,2]. On the other hand, disordered
carbons, reportedly show reversible capacity from 400 to 900 mAh/g in the
voltage interval of Li/C cells ranging between 0 and 2,5V [3]. Though the
theoretical models of this phenomenon are still under extensive discussion,
there are several suggestions to explain the high capacity of disordered
carbon anodes [4]:
1. lithium occupation of the nearest neighbor sites;

2. multilayered lithium deposition at the surface of crystallites;
3. extra lithium storage at the edge of crystallites;
*
345
346
4. lithium ion clusters in nanoscaled cavities;

5. lithium binding in the vicinity of hydrogen atoms.
On the other hand, it is well known now that such metals as Al, Sn,
Pb as well as Si are able to store Li via alloy formation though
electrochemical means resulting in substantially higher capacities [5-8]. For
example, Si or metallic Sn can electrochemically alloy with Li up to about
4,4 Li per Sn, which yields a maximum theoretical capacity of 990 mAh/g
[5,7]. However, the cycling capacity of these materials is limited by the
considerable volume changes during lithium introduction and by the
brittleness of intermetallic lithium phase. In order to improve their structural
stability, the superfine or even nanoscaled intermetallic compounds have
been developed and investigated. In general, superfine or nanoscaled
materials are expected to possess lower specific volume variation and, hence,
better cyclability in electrochemical intercalation and extraction of lithium
ions. For example, it was found that ultrafine Sn-SnSb particles with
d300nm could improve the electrochemical performance of the tin based
alloys [9]. Yet superfine or nanoscaled particles are predisposed to particle
aggregation. The resultant micro-sized particles are formed during lithation
and de-lithiation reactions. One of the possible ways to solve these problems
consists in the creation of the composite materials. The performed activities
on the creation of metal-graphite composite materials by CVD deposition on
mesoscale carbon or by mechanical mixing showed this approach to be
prospective [10].
Therefore, taking into account the listed above, it is necessary to
develop metal-graphite composite materials, which could provide the
conditions for the realization of all mentioned lithium storage mechanisms
with maximal efficiency. The main goal of this work was to synthesize
graphite based composite materials containing superfine or nanoscaled Sn,
SnO2 and Si particles. To reach this goal it was necessary to work-off the
methods of metal-graphite nanocomposite materials production and to
determine the influence of the type of source graphite material and synthesis
conditions on the phase composition, dispersity and morphology of the
synthesized composite materials.
2. EXPERIMENTAL PROCEDURES
The natural crystalline flake graphite from Zavalie field in Ukraine’s
Kirovograd Province has been subjected to thermochemical treatment and
the resultant product, thermoexfoliated graphite, has been used as carbon
material for CM creation. This material was selected because of following
considerations:
347
1. Thermoexfoliated graphite (TEG) is a material obtained by deep

thermochemical treatment of natural graphite (intercalation by sulfuric acid
with subsequent heat treatment in a regime of thermal shock). Practically, it
does not differ from the source natural graphite by crystallographic
parameters, namely, by the value of d002. According to the X-ray diffraction
ɨ
data the interlayer distance d n along c axis is equal to d n (3.36 r 0.001) A ,
for TEG samples used in this work. That is in a fairly good agreement with
that for pure graphite. This result proves the absence of turbostratification of
graphite layers and allows to consider the lithium penetration due to
intercalation being occurred effectively.
2. TEG macrostructure differs from that of natural graphite: it
possesses abnormally high porosity and highly developed active surface (40-
50 m2/g) (Figure 1). The performed thermochemical treatment leads to an
essential exfoliation of graphite matrix with a formation of cellular structure.
The thickness of cell’s walls is equal to 20-25 nm. The surface of cell’s walls
contains a lot of macrocracks, outcrops of crystallites, etc. The thermo-
chemical re-treatment was applied to enhance TEG dispersivity.
Figure 1. Parts of the surface of source TEG (ɚ, b) and TEG oxidized by
sulfuric acid and subjected to thermal shock at 8000ɋ (b); part of the surface of TEG particle
oxidized by nitric acid and subjected to thermal shock at 8000ɋ ( d ).
348
a) b)
Figure 2. Part of TEG particle with Si deposited in vacuum: ɚ), b) are the different
regions of TEG particle (Si content ~ 7wt%).
TEG structure refinement has distinctly observed in electron

microscopy studies of the oxidized TEG powders subjected to the repeated
thermal shock. In this case the size of TEG macropores was equal to 1.5-2
Pm that is essentially lower that for source TEG. Figure 2 presents SEM
images of the source TEG particle (a) and TEG particle oxidized by sulfuric
acid and re-exfoliated at 8000ɋ (b).
The surface of the edges of graphite macroplanes became sheet-like
and hilly due to the intensive exhausting of the oxidant during the thermal
shock. Moreover, the morphology of TEG particle surface could be modified
by the selection of the appropriate intercalant. For example, when the nitric
acid is used instead of sulfuric, the essential refinement of TEG
macrostructure is observed and TEG particles possess more fine structure (it
is clearly seen from Figure 3): mean size of the macropores is equal to 1-3
Pm, the thickness of graphite walls (verges) – 0.1-0.2 Pm. Besides, the edges
of graphite particles contain the outcrops of dislocations etc. However,
according to the X-ray studies TEG treatment by HNO3 oxidant does not
result in (002) graphite reflection broadening as compared with the feed
graphite. This is the evidence of the possible additional refinement of
graphite macrostructure. The analysis performed, indicates that TEG
possesses all the necessary features to realize lithium storage both through
intercalation and the mechanism, which is dominant in disordered carbon
materials, i.e. lithium occupation of the nearest neighbor sites, multilayered
lithium deposit on the surface, and at the edge of crystallites, lithium ion
clusters in nanoscaled cavities, etc.
Besides, the anomalous elastic-plastic characteristics of TEG are the
important factors for application as a support. Due to these characteristics
TEG-based matrix could accumulate the possible volume changes of
metallic component at their cyclic changes during charge/discharge process.
349
Phase composition, distribution character of the modifier particles

on the surface of graphite particles has been analyzed using the X-ray
studies, the Auger spectroscopy, electron microscopy and secondary ion
mass spectrometry. The investigations have been performed using the
following equipment: DRON-4-07 X-ray diffractometer (filtered CoKD
radiation), LAS-2000 Auger spectrometer (Riber, France), JSM-840 electron
microscope (Japan).
a) b)
c) d)
Figure 3. Part of TEG particle with Sn deposited in vacuum: ɚ, b, c, d are the different
regions of particle and different image magnification (Sn content ~ 7wt%).
The following procedures of modifying coating deposition on the

surface of graphite suppor particles have been applied to produce TEG-CMs:
- the method of thermal vacuum deposition;
- the method of metal salt chemical plating from water-salt solution with
subsequent thermal or liquid-phase reduction of salt to metal or metal oxide.
The method of thermal vacuum deposition has been used to obtain
tin and silicon deposit on TEG surface. The main difficulty at metal
component deposition on graphite support is to obtain uniform metal coating
on the surface of disperse particles. The system of uninterrupted mixing of
TEG powder during the material evaporation has been created. Its principle
350
of operation is as follows: a contained holding TEG is an enclosed metal

chamber, at the bottom of which there is gauze having a mesh of 100 u100 Pm
in size. This container was suspended in electromagnetic vibrator and
placed under the evaporator. The whole system was placed into the working
chamber of VUP-5 vacuum station. Shaking frequency at deposition was
equal to 15-20Hz. As a result, the particles change their position in the flow
of metal vapor that allows relatively uniform deposition of material on
graphite particles. Here the vacuum exceeded 510-5 P.
The second method, i.e. electroless plating, has been used also to
produce graphite - Sn, SnO2 composite materials. The chemical deposition of
metals on the graphite supports is performed through two principal steps.
The first step includes the impregnation of the oxidized graphite
material by the water-salt solution of different concentration under the
certain temperature-temporal regimes. The diffusion of salt solution into the
pores of graphite material takes place at impregnation. Temperature and time
are the main factors determining this process. Temperature increase
accelerates the diffusion of salt solution, however, it should not be higher
than the temperature of salt decomposition. With respect to the microporous
structure of the set of the used graphite supports (e.g., TEG) the
impregnation should be of a long duration. UZDN-Ⱥ ultrasonic disperser has
been applied in some cases to decrease the impregnation time and to enhance
the impregnation efficiency. Thermal treatment of the carbon material
impregnated by salt in the air was used to obtain graphite-tin dioxide CM.
The second processing step consisted of salt decomposition with the
subsequent reduction to pure metal. The method of chemical deposition of
metal salts from the water salt solution with the subsequent reduction to pure
metal by liquid phase reducer has been applied to prepare graphite-tin CMs.
In this case tin chloride was used for impregnation and potassium
tetrahydroborate was used as liquid phase reducer.
This method of nanodispersed metal production has the advantages
as compared with the previous one because metal reduction is performed at
sufficiently low temperatures (20-2000ɋ). That is why the diffusion of atoms
and migration of the formed metal particles is suppressed, no metal particles
agglomeration occurred.

The following materials were synthesized using the listed methods:
1. TEG-Si, TEG-Sn and dispersed natural graphite-Sn have been
prepared by thermal vacuum deposition under continuos mixing.
SEM studies of the obtained TEG-Si powders have shown almost
uniform distribution of silicon particles on TEG surface. These metal coated
particles were found to be of the spherical shape, mostly similar in size
351
(50 nm). The absence of silicon particle segregation proves the absence of
silicon diffusion on TEG surface upon preparation process. The absence of
silicon reflections at the X-ray diffraction patterns could be the evidence of
the amorphous structure of Si.
Some other situation is realized in a case of TEG-tin CMs. Electron
microscopy studies of the obtained TEG-Sn powders revealed the uniform
coverage of TEG surface by tin particles. Tin particles are of spherical shape
and their sizes are about 40-80 nm, i.e. somewhat higher than in a case of
silicon particles. Low scatter of particle sizes is observed as in a case of
TEG-silicon system. However, as it is clearly seen from the data of the X-ray
structure analysis (Figure 4) tin particles deposited on the surface of graphite
support are in crystalline state. The distinct and narrow tin reflections at the
X-ray diffraction pattern evidence this fact.
I, arb. u. TEG-Sn, Gr-Sn (Sn content - 7 wt.%)

500
1 - TEG-Sn
400 2 - Cr-Sn
Graphite
Sn
300
2
200
1
Gr
100 Sn
0
20 30 40 50 60 70 80
2T, degree
Figure 4. X-ray diagrams for TEG and disperse graphite with Sn.
A different situation is realized when the dispersed graphite is used

as a graphite support. In this case tin particles distribution on graphite
surface is extremely non-uniform (Figure 5). An essential scatter of particle
sizes is observed. Along with small particles of 30 nm in size, which cover
separate regions of graphite surface forming almost uniform single layer big
particles of 150 – 200 nm in size are observed, which represent the
aggregations of tin particles. According to the X-ray structure analysis the
coherent domain size is almost 2 times higher in these materials as compare
to TEG-tin CMs.
Thus, the method of thermal vacuum deposition in a regime of
continuos mixing permits to obtain graphite – silicon and graphite – tin CMs
containing modifier component about 10wt%. The particles 40-80 nm in size
are almost uniformly distributed on graphite surface. However, certain
disadvantages are inherent to any vacuum deposition method. It is difficult to
control the content of the deposited component, to produce the materials of
high modifier content and the efficiency of such methods is quite low.
352
a) b)
Figure 5. Part of disperse graphite particle with Sn deposited in vacuum: ɚ and
b are the different regions and image magnification of particle
(Sn content ~ 7wt%).
2. Our studies on graphite – transition metal systems [11] have

shown the methods of chemical deposition of nano-scaled metal particles on
the surface of graphite supporter to be powerful technique for production of
CM with well dispersed, nanoscaled homogeneously distributed modifier
component.
Graphite – tin dioxide CMs have been obtained by the method of
chemical deposition of tin nitrates from the water-salt solutions with
subsequent thermal decomposition of salt. According to the X-ray phase
analysis (Figure 6) the impregnation of both TEG and graphite by water salt
solution with the subsequent salt decomposition on air leads to the formation
of tin dioxide on the surface of graphite matrix. Moreover, in spite of the
sufficiently high content (60wt%) of modifier the Bragg’s reflections
corresponding to SnO2 are wide and low intensive that could evidences the
amorphous or disperse state SnO2.
The character of the diffraction pattern remains almost invariable
upon the increase of temperature of treatment and using both TEG and
dispersed graphite as a support. Figure 7 displays the character of SnO2
distribution on the surface of graphite matrix. As it is clearly seen SnO2
forms sufficiently big segregations 0,1-0,2 Pm in size nonuniformly
distributed on the surface of graphite matrix. These data are in contrast to the
results on production of graphite – transition metal oxide CMs where the
modifier oxides were distributed on the surface of graphite supporter in a
form of continuos layer of particles 30 – 50 nm in size. That is why thermal
reduction of tin oxides in a hydrogen flow should be considered unsuitable
for graphite-tin CMs production. So, the method of liquid phase reduction of
metal cations in water medium by Ʉȼɇ4 have been used to produce such
CMs.
TEG-Sn CMs have been prepared according to the following
scheme:
353
1. The impregnation of the oxidized TEG by SnCl26H2O water salt

solution of certain concentration at 293 Ʉ for 20 hrs.
I, arb.u. TEG-SnO2 (SnO2 content ~ 55 wt.%)

500
1 - TEG+Sn nitrate+HTT at 3000C
400 0
2 - TEG+Sn nitrate+HTT at 500 C
002Gr SnO2
300
200 2
1
SnO2
100
0
10 20 30 40 50 60 70 80 90 100
2T, degree
I, arb.u. Gr-SnO2 (SnO2 content ~ 60 wt.%)

500
1 - (Gr+Sn nitrate)+HTT at 3000C
400 2 - (Gr+Sn nitrate)+HTT at 5000C
SnO2
002Gr
300
2
200 1
100 SnO2
0
10 20 30 40 50 60 70 80 90 100
2T, degree
Figure 6. X-ray diagrams for TEG with SnO2 and for dispersed
graphite with SnO2.
a) b)
Figure 7. Part of TEG (a) and disperse graphite (b) particles with deposited
SnO2 (SnO2 content ~ 55wt%).
2. Sn cations reduction by the solution of potassium tetrahydroborate.

The X-ray phase analysis of the CMs obtained by these methods
(Fig. 8) showed that the modifier coating is pure tin. Moreover, the coherent
domain size of tin particles in this case is almost the same as for the
materials obtained by thermal vacuum deposition. Besides, it is necessary to
354
note that tin content in CMs produced by this method is 6 - 7 times higher
than that for CMs produced by thermal vacuum deposition method. The
tendency to enlargement of coherent domain size is revealed when disperse
graphite is used as a supporter instead of TEG.
I, arb. u. TEG-Sn, Gr-Sn (Sn content ~44 wt.%)

500
1 - TEG-Sn
2 - Gr-Sn
400
Graphite
300 Sn
2
200
Gr 1
100 Sn
0
20 30 40 50 60 70 80
2T, degree
Figure 8. X-ray diagrams for TEG Graphite with Sn.
Figure 9 is a group of SEM images illustrating the character of metal

distribution on the surface of graphite particles. It is seen that an essential
dispersion of TEG particles upon chemical treatment is observed. The size of
graphite particles after metal deposition is equal to 200-400 Pm. The metal
layer is sufficiently uniform and consists of the particles 40 nm in size.
However, a lot of particles’ agglomerations 80-140 nm in size is also
observed. The nonuniformity of particles’ distribution and their sizes in CMs
on the base of disperse graphite is more substantial than in a case of CMs on
the base of TEG (Fig. 10). But even in this case the size of agglomeration
does not exceed 150-180 nm in spite of the high content of the deposited
component. Thus, the method of liquid phase reductions of metal cations on
TEG and dispersed graphite permits to produce graphite – Sn CMs of
uniform phase composition. The metal in these CMs is distributed in a form
of nanoscaled formations on the surface of the graphite matrix.
a) b)
Figure 9. Part of TEG particle with deposited Sn: ɚ, b are the different
regions of particle and different image magnification (Sn content ~ 44 w.t%).
355
a) b)
Figure 10. Part of disperse graphite particle with deposited Sn: ɚ, b, are the
different regions of particle and different image magnification
(Sn content ~ 44wt%).
4. CONCLUSSION
1. Graphite-tin and graphite–silicon nanocomposite materials have
been obtained by the method of thermal vacuum deposition under continuous
agitation, and also by the method of electroless plating of salts from water
solutions with the subsequent liquid phase reduction.2. The size of modifier
particles (Sn and Si) deposited on TEG was shown to be equal to 60-80nm
and was uniformly distributed on TEG surface up to 60 wt.% of modifier;
the migration of particles is essentially suppressed. This is in contrast to the
case of disperse graphite being used as a supporter. 3. The application of
TEG in the production of nanocomposite materials is preferable as compare
to the other forms of graphite. Taking into account its morphology, the
character of modifier particle distribution and their size, TEG-based CMs
could be considered as prospective materials for the construction of
graphite–anodes of lithium-ion batteries.
REFERENCES
1. Dahn J.R, Sleigh A.K., Shi H., Way B.M., Rweydans W., Reimers J.N., Zhong Q.,
Vonken U. In: Lithium batteries, new materials and perspectives, G. Pistoria, ed. New
York and Amsterdam: Elsevier, 1993.
2. Endo M., Kim C., Nishimura T., Fujino T., Miyashita K. Recent development of carbon
materials for Li ion batteries. Carbon, 2000; 38: 183-97.
3. Wang Z., Huang X., Xue R., Chen L. A new possible mechanism of lithium insertion and
extraction in low-temperature pyrolitic carbon electrode. Carbon 1999; 37: 685-92.
4. Xiang H.Q., Fang S.B., Yjiang Y. A model for lithium in high–capacity carbons with
large hysteresis. Carbon 1999; 37: 709-11.
5. Li N.C., Martin C.R. A high-rate, high–capacity, nanostructured Sn-based anode
prepared using sol-gel template synthesis. J Electrochem. Soc.2001; 148(2): A164-70.
356
6. Mao O., Dunlap R.A., Danh J.R. Mechanically alloyed Sn-Fe(-C) powder as anode
materials for lithium–ion batteries. I. Sn2-Fe-C system. J Electrochem. Soc. 1999;
146(2): 405-513.
7. Crosnier O., Brousse T., Devaux X., Fragnaud P., Schleich D.M. New anode systems for
lithium ion cells. J.Power Sourc.2001; 94: 169-74.
8. Moon H.S., Ji K.S., Ryoo K.C., Lee Y.K., Chung S., Park J.W. Characteristics of silicon-
doped tin oxide thin films as anode materials for microbatteries. J. Cer. Proc. Res. 2002;
3: 34-7.
9. Chen W.X., Lee J.Y., Liu Z. The nanocomposites of carbon nanotube with Sb and
SnSb0.5 as Li –ion battery anodes. Carbon 2003; 41: 956-66.
10. Wang C.S., Wu G.T., Zhang X.B., Qi Z.F., Li W.Z. Lithium insertion in Carbon-Silicon
composite materials produced by mechanical milling. J. Electrochem. Soc. 1998; 145:
2751-8.
11. Vovchenko L.L., Matzui L.Yu., Brusilovets A.I.S ynthesis of graphite-metal composite
materials by salt.
ELECTROCHEMICAL PERFORMANCE OF
Ni/Cu-METALLIZED & CARBON-COATED
GRAPHITES FOR LITHIUM BATTERIES
Christopher S. Johnson1 *, Kevin Lauzze1, Nick Kanakaris1, Arthur

Kahaian1, Michael M. Thackeray1, Khalil Amine1, Giselle Sandí-Tapia2,
Stephen A. Hackney3, and Robert O. Rigney4
1
Argonne National Laboratory, Chemical Engineering Division, Electrochemical Technology
Program, 9700 S. Cass Ave., Argonne, IL, 60439 USA
2
Argonne National Laboratory, Chemistry Division, 9700 S. Cass Ave., Argonne, IL, 60439
USA
3
Department of Materials and Metallurgical Engineering, Michigan Technological
University, Houghton, MI 49931 USA
4
Fluid Air Inc., 2550 White Oak Circle, Aurora, IL 60504 USA
Abstract
Synthetic and natural graphites were metallized with Cu or Ni using

a fluidized bed coating and annealing process. With this method, crystalline
nanometer-sized metal islands (~ 50 nm diameter) were deposited onto the
surface of graphite. Post-metallization, the graphite materials were cycled in
lithium coin cells to determine their electrochemical properties in a
propylene carbonate (PC) solvent based electrolyte. Better cycling
performance (316 mAh/g active graphite, 20th cycle; 26% irreversible
capacity) was obtained with the Cu-coated graphites versus Ni-coated or
uncoated types (247 mAh/g, 20th cycle; 62% irreversible capacity) in a 30%
PC blend electrolyte at 50°C. The Cu nanosized deposits help to mediate
charge transfer into the graphite particle via improved particle to particle
contacts (electrical connectivity) and may contribute to improve kinetics of
lithium ion insertion into the graphene layers by assisted Li-Cu inter-
diffusion. Further, Cu may act as a de-solvation catalyst to remove PC-
coordinated molecules from Li cations allowing Li insertion into grapheme
layers and overall improved electrochemical performance. Carbon-coated
graphite electrodes prepared by a vapor deposition process, show superior
performance versus the uncoated graphite in a PC electrolyte system. Coated
graphites are under investigation as possible new anodes for high-power
lithium-ion battery applications.
*
357
358
Keywords
copper metallization, graphite, lithium batteries, metal coatings,

fluidized bed, carbon
1. INTRODUCTION
Carbonaceous materials such as coke and graphite are the negative
electrodes or anodes of choice in various commercial lithium-ion batteries.
While intermetallic materials such as K’-Cu6Sn5 [1], InSb [2], and Cu2Sb [3]
have shown excellent promise as an alternative high volumetric capacity
anode (e.g. InSb; ~ 1710 mAh/cm3) in advanced lithium cells, they are still
exploratory and remain as a long-term research objective [4]. Graphite
electrodes are the state-of-the-art, and potential new developments to
improve graphite and other carbonaceous materials in lithium batteries may
have an immediate impact in the technology [5].
Recently there has been an interest to coat or modify the surface of
graphite materials such that these electrodes may operate more effectively in
advanced electrolyte systems containing propylene carbonate (PC) solvent
[6]. PC, a cyclic di-ester, is a desirable solvent in these electrolytes because
it has favorable low temperature properties such as increasing the solvation
of the LiPF6 salt as well as lowering the viscosity of the electrolyte.
Improvement in these properties results in better electrolyte conductivity at
lower temperatures. However, since Li cations become highly solvated in
PC, it becomes difficult to strip the solvent coordination sphere during
electrochemical reaction and as a consequence, the co-intercalation of PC
between graphite layers during the first lithiation reaction occurs [7]. Recent
studies have confirmed the production of propylene gas from a side-reaction
breakdown of the intercalated PC molecule [8]. The formation of gas is
detrimental to the local integrity of the graphite, subsequently leading to
graphene layer exfoliation, and mechanical destruction and failure of the
electrode [8, 9].
A control of the graphite electrode-electrolyte interface in a PC
solvent electrolyte is therefore a key to obtain modest electrochemical
performance in lithium cells [10]. Apart from electrolyte additives, an
approach is to modify the SEI (solid electrolyte interphase) film to create a
lithium-permeable but protective coating at the graphite edge surface to
attempt to physically block co-intercalation of PC solvent, yet allow
diffusion of Li through the coating to the graphite layers. Various approaches
for coating graphite have been previously demonstrated using carbon [11-
14], Cu [6, 15], Ni [16], Al [17], Sn [18], Ag [19-20], as well as controlled
oxidation of the graphite surface [21].
359
This research is focused on ways to improve the electrochemical

performance of graphite materials in lithium batteries. We have attempted to
modify the graphite surface via metallization in order to minimize PC
solvent co-intercalation. In a secondary effect, we may also expect to
increase the conductivity of the graphite with metallization, which, in turn,
could improve its electrochemical performance by lowering its impedance.
In this paper, two laboratory scale processes for modifying the surface of
synthetic or natural graphite are described. One new process was designed to
produce thin metal deposits or, preferably, a coating by a type of fluidized
bed process while the other was used to prepare a thin amorphous carbon
coating on the graphite surface. These methods for coating electrode
materials are general and many variations are possible. A few examples are
highlighted in this paper. These metallized or coated graphite materials were
tested as electrodes in lithium coin cells in order to determine their
electrochemical properties in PC based electrolytes.
2. EXPERIMENTAL
2.1. Materials
Synthetic graphite powder (MAG-10) was used as received from
Hitachi Chemical (Japan). Natural graphite (SL20; rounded edge type) was
provided by Superior Graphite Co. (Chicago, IL). Other chemicals used were
ACS reagent grade (99+% purity) and were purchased from Aldrich
Chemical Company. Copper formate dihyrate and (C2H2CuO4 •2 H2O;)
nickel formate dihydrate (C2H2NiO4•2 H2O), and formic acid were used in
the fluid bed coating process. Propylene gas (99.99%) was supplied by AGA
Gas Inc. Purified water (milli-Q; Barnsted Nanopore) was used to make an
aqueous solution of the formate salts in 2% formic acid that were used as the
spray solutions in the fluid bed reactor.
2.2. Characterization
Powder X-ray diffraction patterns of the samples were collected on a
Siemens D5000 powder diffractometer with Cu KD radiation between 10q
and 80q 2T at a scan rate of 0.6q 2T/min. High-resolution TEM images of the
4.8% Cu-metallized MAG-10 graphite powder were collected on a JEOL-
JEM 4000FX-1 transmission electron microscope under an accelerating
voltage of 200 keV. Samples were prepared for the electron microscope by a
standard procedure described elsewhere [22]. The level of metallization and
dispersion was checked by analyzing the powders by energy dispersive
analysis of X-rays (EDAX) and EDAX elemental mapping on a JEOL 6400
360
scanning electron microscope (SEM). Powder samples were spread onto a

double-sided carbon conductive tape. The samples were then loaded in the
SEM for EDAX analysis (Noran Voyager II System). In addition to EDAX,
the coated samples were heated in air on a microbalance to 900°C in order to
burn off the graphite. A computer-controlled simultaneous TG/DTA
EXSTAR 6300 unit (Seiko Instruments, Inc.) was used for
thermogravimetric analysis. The metal to graphite content was calculated
from the weight loss and formation of the metal-oxide, CuO or NiO.
Electrodes were fabricated by intimately mixing 92 wt% of the
appropriate graphite powder, and 8 wt% polyvinylidene difluoride (PVDF)
negative electrode binder (Kynar, Elf-Atochem), in a 1-methyl-2-
pyrrolidinone (NMP) solvent (Aldrich, 99+%). The mixed slurry was cast
onto a copper foil current collector and dried at 75qC for 3-5 h. The electrode
laminates were dried further at 70qC under vacuum overnight. Disk
electrodes with an area of 1.6 cm2 were punched from the laminates. The
electrodes were evaluated in coin-type cells (size 2032, Hohsen Corporation)
with lithium foil (FMC Corporation) as the counter electrode, a
polypropylene separator (Celgard 3501 grade), and an electrolyte of 30%
propylene carbonate (PC), 30% ethylene carbonate (EC), 40% ethyl-methyl
carbonate (EMC) (by volume), with 1.2 M LiPF6 (Tomiyama). Cells were
constructed inside a dry room (< 2 % relative humidity). Comparative
conditions of cycling was done at room temperature or 50°C at a constant
current density of 0.2 mA/cm2 between 1.5 and 0 V (Maccor Series 4000
cycler). Additional cycling experiments were done in a non-PC electrolyte
solution consisting of 1 M LiPF6 in a 1:1 EC:DEC (50:50; by volume)
solvent mixture (Merck), and a concentrated PC electrolyte of 70% PC, 20%
EC, 10% EMC (by volume) and 1.2 M LiPF6 (Tomiyama) at room
temperature or 50°C.
2.3. Fluid Bed Reactor – Cu and Ni Metallization
A benchtop fluid bed reactor (Model 2 Fluid Air fluid bed, Fluid Air
Inc., Aurora, IL, U.S.A; www.fluidairinc.com) was used to spray coat the
graphite particles with an aqueous solution of copper or nickel formate
dihydrate. The coating was done at 40-50°C using heated, filtered room air.
The weight percent aqueous solution (Ni or Cu formate salt) to graphite was
typically 10-35% going in the fluid bed reactor. Final percentages of
deposition were calculated as a weight percent after final firing of the
samples or from calculations of weight loss from thermal gravimetric
analysis (TGA) experiments used to burn away the graphite. In this paper we
present data for a range of samples that had 3.0 to 11.7wt% Cu-metallized
and another that was 4.6wt% Ni-metallized graphite. The as-deposited
graphite was subsequently annealed in 4%H2 (helium balance) gas flow at
361
325 - 400°C for 10-14 h in an alumina tray inside a Lindberg tube furnace
under controlled gas flow. On some occasions, and as a control experiment,
pure argon gas was employed during the annealing step. Resultant powders
were used in electrochemical (coin cell cycling) studies.
2.4. Quartz Reactor - Carbon Coating

A quartz reactor was used for deposition of a carbon-coating onto
the synthetic graphite MAG-10. In this experiment, the graphite sample was
loaded on an alumina tray then fed into a quartz tube reactor. After purging
the reactor with nitrogen gas for 3 hours, a propylene gas in a balance of
nitrogen was premixed to a final percentage of 75% propylene in nitrogen
then flowed into the reactor that was heated to 700°C and held for 3 h [23].
The EXSTAR 6300 TG/DTA was used to measure the carbon-coating
content in the graphite sample [24]. Samples were heated at 10°C/min to
900°C in a flow of dry air (30 cc/min). Resulting weight percent of carbon
on graphite was calculated as 22% +/- 0.5%.
3. RESULTS & DISCUSSION
3.1. Cu and Ni-metallized Graphite

The graphite particles were coated in a specially modified fluid bed
reactor that provides two distinct zones of air-suspension. A partition or tube
can be used to physically divide these zones and lessen particle-to-particle
contact resulting in agglomeration. The inner zone incorporates a high
velocity airflow that pneumatically accelerates particles past a spraying
nozzle. Spray droplets impact upon, spread out over, and evaporate from the
particles. The coated particles spout out and drop down to a bed of particles
in the outer zone. The airflow velocity is much lower in the outer zone. The
airflow is just enough to keep the particles moving back into the inner zone.
This cyclic process can be continued until the desired level of coating is
applied. Overall, this process, also known as Wurster coating, is a relatively
simple, rapid and inexpensive method for powder coating, and has been used
for some time in the pharmaceutical industry. In addition, as larger batch
runs become necessary, the process can easily be adapted to proportionally
larger-sized commercialized fluidized bed systems.
The final third step is a mild heat-treatment of the as-deposited
graphite that is necessary to form the final metallized graphite. The graphite
is heat-treated in 4%H2 (helium balance) at fairly low temperatures of
325°C. Under these conditions, the copper and nickel formates form metallic
362
Cu or Ni directly during the heat-treatment step without the formation of

unreduced side-products of CuO or NiO.
Figure 1 shows the XRD patterns of the Cu-metallized graphite
before (Fig. 1a) and after heat-treatment (Fig. 1b and 1c). The many peaks
seen in 1a. are those of the Cu-salt film present on the surface of the graphite
prior to heat-treatment. Diffraction from Cu metal crystallites in these
samples after heat-treatment were clearly observed in the patterns in 1b and
1c. The Cu diffraction lines intensities are proportional to the metal content
present in the graphite. X-ray diffraction revealed the resultant particles were
pure copper crystalline with a face-centered cubic (fcc) structure. The
graphene layers and their d-spacing is unperturbed (less than 0.2% change)
by the mild overnight heat-treatment process carried out in regeneration gas
(4% H2 in balance of helium) for metallized graphites shown in 1b and 1c.
X
= Cu metal
500 cps
X = graphite
Intensity (cps)
(d) graphite (MAG-10)
X X X X X X
(c) 11.7%-graphite (MAG-10)

X
X X
X X X
X
X X (b) 4.8%-graphite (MAG-10)
X X X X
x 0.3
(a) Cu-formate (MAG-10)
10 20 30 40 50 60 70 80
2-theta (degrees)
Figure 1. XRD of (a) Cu-formate coated graphite, (b) 4.8% Cu-graphite, (c) 11.7% Cu-
graphite and (d) control (pristine) graphite.
Similar to copper, the nickel formate is also converted to Ni metal

when heated in 4%H2 in a balance of helium gas. This is demonstrated in the
XRD pattern in Fig. 2a where a sample of nickel formate (in the absence of
graphite) was heat-treated at 400°C for 10 h. One can clearly see that Ni
metal is present. An interesting comparison may be made if the nickel
formate sample is instead heated only in argon gas flow for 10 h at 400°C. In
this case (Fig. 2b), both NiO and Ni are formed. Thus, it is important to heat-
treat in a more reducing atmosphere to generate fully reduced metal.
363
The XRD for the sample of Ni-metallized on graphite (4.6wt%) is

provided in Fig. 2c, however the strongest peaks of Ni metal are overlapping
with those peak reflections from the MAG-10 graphite. But, the X-ray
diffraction in Fig. 2a revealed the Ni particles were pure nickel crystalline
with a face-centered cubic (fcc) structure.
x = graphite
x
= NiO
= Ni
2000 cps
(c) 4.6% Ni-graphite (MAG-10)

Intensity (cps)
x x x x x
(b) Ni + NiO
(a) Ni
25 35 45 55 65 75
2-Theta (degrees)
Figure 2. XRD of (a) metallic Ni from Ni-formate heated at 400 °C in 4%H2, (b) metallic Ni
plus NiO from Ni-formate heated at 400 °C in argon, and (c) 4.6% Ni-graphite.
15 µm 15 µm
(a) (b)
Figure 3. (a) SEM micrograph and (b) EDAX Cu elemental mapping of 11.7 wt% Cu-
graphite material (synthetic graphite MAG-10). In (b) the EDAX map was taken on the image
in (a), and the lighter areas represent Cu regions.
364
The SEM micrograph and the corresponding EDAX elemental

mapping for the 11.7 wt.% Cu-graphite (MAG-10) sample are shown
respectively in Figures 3a and 3b. At the micron resolution of this
micrograph, it is evident that the Cu is homogenously distributed on the
surface of the graphite and the coverage is quite good. Thus, the fluid bed
Wurster coating method was effective in laying down in even distribution of
metal and may therefore provide a new alternative method for the coating of
powdered battery materials.
100 nm 100 nm
(a) (b)
Figure 4. HRTEM images (a) and (b) of 4.8 wt% Cu-graphite material (MAG-10); arrows
indicate deposits of nano-Cu.
To gain a deeper understanding of the morphological nature of the

Cu on graphite, the 4.8wt.% Cu-graphite (MAG-10) was examined by
HRTEM analysis. In Figs. 4a and 4b, two TEM images are shown for the
sample. In Figure 4a, 20-50 nm-sized crystallites of Cu are observed
(marked by arrows) that are roughly spherical in shape. Figure 4b also shows
another area of the sample, and in this case a slightly larger agglomeration of
100 nm in size appears along an edge of the graphite. Through an analysis of
these figures and from careful electron diffraction imaging of the basal plane
and edge plane of the graphite (not shown here) it appears that the nano-Cu
does not preferentially migrate to the graphite edge plane (i.e. intercalate
pathway) during the coating and heat-treatment step. Thus, a mechanism
where there would be direct physical blockage of co-intercalation of solvent
molecules during the first lithium insertion reaction may be limited in this
case. The differences in electrochemical performance of the Cu-metallized
graphite vs. un-metallized may include factors such as the extent and nature
of the SEI or the kinetics of de-solvation that may be affected by the
presence of the Cu.
Previous studies focusing on deposition of Cu [6, 15] or Ni [16]
coatings and deposits on natural and synthetic graphics relied on an
electroless plating method [25]. In this method many different chemical
solutions are used in multiple steps to treat the graphite surface, such as
365
chromic acid plus sulfuric acid, then tin chloride, and palladium chloride
treating baths. Finally a copper electroless plating bath is used for the
deposition. In our work, the fluid bed Wurster coating process was
implemented and it appears more suitable for metallizing graphite versus that
of the electroless method, which is not easy to scaleup and cumbersome and
time-consuming, and also as compared to the metal-carbonyl process, which
uses toxic gases such as tetracarbonyl nickel (Ni(CO)4) during operations
[26].
3.2. Carbon-Coated Graphite

Carbon-coating is an effective way to improve the performance of
electrode materials for lithium batteries, particularly with graphites [11-14].
It is also known to aid in the surface conductivity for LiFePO4 as a cathode
material [27]. There are many ways to coat powders with carbon, but in this
study, we have chosen to decompose a hydrocarbon vapor of propylene in a
nitrogen carrier gas at a moderate temperature of 700 °C. Criteria for using
this process include a material that is stable at this temperature and under a
reducing environment.
500 cps
X = graphite
(b) graphite (MAG-10)

Intensity (cps)
X X X X X X
(a) C-coated graphite (MAG-10)
* X
X X X X X
24 34 44 54 64 74
2-theta (degrees)
Figure 5. XRD of (a) 22 wt. % C-coated graphite (MAG-10)

and (b) pure (pristine) MAG-10 graphite.
In our labs, we have successfully carbon-coated Li4Ti5O12, LiFePO4

and graphite powders with this method [28]. There was no change in the
366
MAG-10 XRD pattern before and after carbon-coating (22 wt%) in Fig. 5a-b
which suggests that the carbon-coating is amorphous. All of these materials
are quite thermally stable in a reductive gas environment which allows for
propylene gas to be reactively cracked at the surface of these powders
resulting in an amorphous carbon film. As suggested by Sandí et al. [23], the
propylene gas after reaction tends to leaves a fairly homogenous and
hydrogen-free film.
3.3. Electrochemical Processes
The MAG-10 was chosen as a candidate anode for high-power

battery applications. It is flaky synthetic graphite that was provided by
Hitachi. Electrochemical performance of this graphite appears satisfactory
from high-power pulse (HPP) lab cell testing and modeling studies
conducted in our laboratory [29]. As such, it has been studied as a
“Generation -2” anode material within the Advanced Technology Program
(ATD) at Argonne under the guidance and funding of the U.S. Department
of Energy (DOE). However, eventual applications of this graphite directly in
commercial hybrid vehicle batteries will necessitate further improvement in
the material for operation at low temperatures possibly with a PC blended
electrolyte. Further, the conductivity of the graphite needs to be increased to
improve the rate performance at low temperature. It is towards these goals
that many of the experiments done herein were conducted with MAG-10
graphite material to demonstrate improvements. However, future generations
of advanced anode graphite materials, may employ using low cost natural
graphite with film-forming additives or coated and/or metallized
modifications of natural graphite without electrolyte additives.
An important physical property, which is distinctive for Cu as a
metal is that it allows for Li diffusion [30], while Ni metal does not support
Li solubility and is Li impervious. Further, the Li0.22Cu alloy is a known
phase [31] at high and low temperatures, while no Li-Ni phases are known to
exist [32]. Thus, for a lithium battery environment, if Cu coverage is
predominate at the graphite edges, then the diffusion of Li through the metal
would be allowed, unlike for Ni. For this reason, we concentrated our efforts
in electrochemical testing of Cu-metallized graphite materials. Indeed,
cycling tests conducted on Ni-metallized MAG-10 graphite did not
demonstrate any improvements over the baseline uncoated samples, unlike
that for Cu. Further, the Ni did not improve the performance of the graphite
in a 70% PC electrolyte system.
The voltage curve collected at 50°C and in 30% PC electrolyte for
the 11.7wt.% Cu-graphite (MAG-10) first full 2 cycles are shown in Figure 6
(top traces) together with the pristine (uncoated) MAG-10 graphite (bottom
traces).
367
6 3
50 °C, 3:3:4 PC:EC:EMC (1.2 M LiPF6)
5 2
11.7% Cu-graphite (MAG-10)
(b)
4 1
Voltage vs. Li (V)
Voltage vs. Li (V)

1
2
3 0
2 -1
0.6 V process uncoated graphite (MAG-10)
(a)
1 -2
1
0 -3
0 100 200 300 400 500 600 700 800
Capacity (mAh/g)
Figure 6. Voltage profiles of lithium cells with (a) uncoated MAG-10 graphite and (b)
11.7wt% Cu-graphite (MAG-10) electrodes in 30% PC blended electrolyte and at 50°C.
The vertical ticks in the voltage curves are due to current

interruptions conducted every 20 minutes during the cycling to check the cell
impedance. Note that the capacity values in Figure 6 have not been corrected
for the inactive Cu in the graphite. There is a large first cycle irreversible
capacity loss of 61.7% for the pristine sample, whereas the 11.7wt.% Cu-
graphite shows only a 26.4% loss. Most of the irreversible capacity stems
from the solvent co-intercalation plateau at about 0.6 V in the uncoated
graphite [6, 10]. Note how the 11.7 wt.% Cu-graphite does not show as
pronounced plateau. Inspection of the curve shows that approximately 30
mAh/g or ca. 10-12% total capacity originates from slight irreversible co-
intercalation of solvent at a voltage range of 0.7 to 0.55 V in this profile.
Thereafter the coulombic efficiency and cyclability of the 11.7wt.% Cu-
graphite is excellent as shown in Fig. 7 out to 20 cycles (283 mAh/g
(uncorrected for Cu weight)). This data was far better than that seen for the
un-metallized (pristine) graphite sample in 30% PC blend electrolyte. From
observation of the long plateau at ~0.6 V in Figure 6 and the cycling
instability or higher fade in Figure 7, one may compare the two materials and
see the improvement. This result implies that the performance is tied to a
partial blockage of PC solvent molecule co-intercalation into the graphite on
the first lithiation cycle in the Cu-metallized graphite electrode. As a control
experiment, and as expected, the MAG-10 graphite itself (no coating) cycles
normally in a non-PC electrolyte, such as LP-40 at 50°C (Figure 8).
368
50 °C, 3:3:4 PC:EC:EMC (1.2 M LiPF6)

400
discharge (to 0 V)
charge (to 1.5 V)
350 discharge (to 0 V)
Capacity (mAh/g)
charge (to 1.5 V)
300 11.7% Cu-graphite (MAG-10)
283
250 247
uncoated graphite (MAG-10)
200
0 5 10 15 20 25
Cycle Number
Figure 7. Capacity versus cycle number plots of the Li/graphite and Li/11.7%Cu-graphite
cells at 50 °C in 30% PC blended electrolyte.
Similar results were found with additional tests on sister cells of 11.7
wt.% Cu-graphite in 30 % PC electrolyte. However, it was not possible for
the 11.7 wt.% Cu-graphite material to cycle in the 70% PC blend electrolyte.
Other lower wt.% contents of Cu or Ni-graphite in this high percentage PC
also were unable to cycle. In general it appears to be quite difficult for
graphite and modified graphite materials to cycle whatsoever in a 70% blend
of PC (70% PC, 20% EC, 10% EMC and 1.2 M LiPF6) as an electrolyte.
More work is needed in optimizing the surface properties of natural
graphites in order to cycle more reasonably in concentrated PC-based
electrolytes.
One may expect the rate capability to be quite good in the Cu-
metallized graphite sample, particularly since the particle-particle contact
points and electrical conductivity may be improved. In Figures 8 and 10, the
cycling stability and rate capability are given. This experiment was
conducted in the non-PC based electrolyte. After 44 cycles the 11.7wt.% Cu-
graphite delivers a specific capacity of 330 mAh/g of active graphite on
discharge at a constant current C/5 rate at 50 °C in a lithium cell in LP-40
(1:1 EC:DEC, 1 M LiPF6). After a constant current (0 V to 1.5 V) rate study
up to 2C (2 mA/cm2) during the next 8 cycles, the cell was returned to its
initial rate of C/5 to see if the material’s electrochemical performance was
adversely affected. In Figure 8, the specific capacity of the cell returned back
to a steady 327 mAh/g of active graphite. In Fig. 10, the % loss in capacity is
plotted as a function of the C rate. At the 2C rate (2 mA/cm2), the material
delivered 60% of its C/5 rate baseline capacity. Further, the curve is
369
somewhat linear implying that the material may support even higher rates
such as 4C and still provide 20% capacity or about 60-70 mAh/g.
450
50 °C, 1:1 EC:DEC (1 M LiPF6)
400
11.7 wt% Cu-graphite (MAG-10) rate studies
Specific Capacity (mAh/g)
350
327
300
graphite (MAG-10)
250
200 discharge (0 V) Cu-graphite

charge (1.5 V) Cu-graphite
150 discharge (0 V) graphite
charge (1.5 V) graphite
100
0 10 20 30 40 50 60
Cycle No.
Figure 8. Specific capacity versus cycle number plots of the Li/graphite and Li/11.7% Cu-
graphite cells at 50 °C in 1:1 EC:DEC (1 M LiPF6; LP-40).
The current interruption technique was used to ascertain the

impedance of the graphite materials. In Figure 9, a typical discharge curve
(after first cycle break-in) is plotted for the 11.7%-Cu-graphite. Note the
steps in the voltage profile indicative of graphite-staging reactions. The
Fig. 9. inset is the expanded plot of a typical current interrupt (30 seconds)
for the electrode together with one from the uncoated non-metallized
(control) graphite cell. The cell’s depth-of-discharge in this analyzed region
was about 50-60% DOD and a cell voltage of ~ 0.08 V vs. Li. Note that the
greater the voltage relaxation for a fixed cell electrode area and current, the
larger the impedance of the material, and the slower the electrode kinetics.
When the applied current is off, the cell voltage of the control electrode
relaxes to a slightly greater value than the Cu-metallized graphite electrode.
The resultant ASI (at 30 seconds) of the Cu-metallized electrode is about 30
:಴ cm2 or about 60% lower.
The cycling improvement for the Cu-metallized graphite over the
pristine graphite was also observed by K. Guo et al. [15] in their study of
electroless Cu deposited on graphite cycled in a lithium cell with a 20% PC
blend electrolyte. Also, they recorded a rate capability improvement in their
Cu graphite material as well. At a current density of 1.4 mA/cm2, the cell
achieved about 60% (~ 200 mAh/g) of the charge capacity measured at 0.14
mA/cm2, compared to about 30% (~ 100 mAh/g) for the non-treated pristine
natural graphite cell [15].
370
1.5
0.089
metallized Cu-graphite
1.3 2
0.087 ASI (30 sec) = 29.7 ȍ•cm
Cell Voltage (V vs. Li)

1.1 0.085 uncoated graphite
2
Cell Voltage (V) vs. Li
ASI (30 sec) = 48.8 ȍ•cm

0.083
0.9
0.081
0.7 0.079
0.077
0.5
0.075
0.3 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95
Relative time (h)
0.1
-0.1
0 1 2 3 4 5 6 7 8 9
Time (h)
Figure 9. Typical discharge (lithiation) voltage profile of the Li/11.7%Cu-graphite cell at

50 °C in 1:1 EC:DEC (1 M LiPF6; LP-40). Inset is an expanded region showing the voltage
relaxation change during current interruption at about 0.08 V of the Li/11.7%Cu-graphite
and Li/graphite control cell.
100 50 °C, 1:1 EC:DEC (1 M LiPF6)
11.7 wt% Cu-graphite (MAG-10)

80 (this work)
% of capacity at C/5
60
3.09 wt% Cu-graphite
(ref. 15)
40
discharge to 0 V
literature [15] charge to 2.0 V

20
literature [15] charge to 2.0 V control graphite
(control- No Cu) (ref. 15)
0
0 0.5 1 1.5 2 2.5
2
Current Density (mA/cm ) or C rate
Figure 10. Rate performance plot of % capacity versus current density for Li/11.7%Cu-
graphite cell at 50 °C in 1:1 EC:DEC (1 M LiPF6; LP-40 and literature values ofLi/ 3.09%
Cu (electroless) on graphite cell and Li/graphite cell (no Cu) at room-temperature in 20% PC
blended electrolyte [15].
371
For comparative purposes, their data was interpolated and also

plotted in Fig. 10.
The room-temperature cycling data for lithium cells containing
carbon-coated MAG-10 graphite and the Superior Graphite (SGC) natural
graphite with rounded edges are shown in Figure 11. For comparison the
standard uncoated MAG-10 graphite cell is also plotted in the Fig. 11. The
electrochemical performance of the carbon-coated graphite was superior to
the uncoated but lower than that of the SGC sample. This result is consistent
with other studies that have documented that carbon coating improves the
performance of natural or synthetic graphite in PC containing electrolytes
[13]. Interestingly, despite no carbon coating, the SGC sample performed
well due to its improved morphology. The sample was mechanically milled
to a spherical morphology with rounded edges in order to reduce the number
of exposed edge planes. The irreversible capacity on the first cycle is only
14%, then 42%, and finally 75% for the SGC sample, carbon-coated MAG-
10, and uncoated MAG-10 respectively. Thus, treatment of the pristine edge
planes by milling or carbon-coating improves the first cycle performance and
suggests that less solvent co-intercalation occurred during the initial lithium
insertion into the graphene layers.
To cycle natural graphite in a PC electrolyte, another approach is to
dissolve small quantities of film-forming additives into the electrolyte prior
to cycling such as acrylic acid nitrile (AAN) [33], ethylene sulfite (ES) [34],
and vinylene or vinylethylene carbonate (VC, VEC) [35-36]. These additives
have been shown to reductively breakdown on the first cycle at the graphite
surface thereby pre-passivating it prior to electrolyte-induced SEI film
formation. This in turn can control the extent of solvent co-intercalation, gas
formation, and graphite exfoliation. The carbon-coating performed in this
study is another method of controlling the interface of the graphite electrode
material in a beneficial way. No electrolytes additives were employed in this
approach.
Interestingly, the carbon-coating which is quite high in this work at
22 wt% also provided capacity to the cell since the capacity on the 20th cycle
was 320 mAh/g compared to a theoretical of 290 mAh/g (graphite; 290
mAh/g = 372 mAh/g – 82 mAh/g or 22%) if it were, instead “inactive” dead-
weight. Future studies will concentrate on optimizing the amount of carbon
coating while still maintaining peak electrochemical performance.
The combination of rounded edge graphite and metallization results
in a much improved electrode material as presented in Figure 12. The
electrochemical cycling data in capacity versus cycle number are presented
for the Superior Graphite Co. natural graphite rounded–edge sample in the
30% PC solution at room-temperature and at 50°C under constant current
conditions between 1.5 V (de-lithiation-charge) and 0 V (lithiation-
discharge).
372
discharge SGC (1.5 V)

400 RT, 3:3:4 PC:EC:EMC (1.2 M LiPF6)
charge SGC (1.5 V)
discharge (0 V)
Specific Capacity (mAh/g)
charge (1.5 V)
SGC (rounded edge)
discharge (0 V)
350
338 charge (1.5 V)
320
carbon-coated MAG-10
300
282
uncoated MAG-10
250
200
0 5 10 15 20 25
Cycle No.
Figure 11. Specific capacity versus cycle number plots of Superior Graphite Co. natural
graphite (rounded edge type), uncoated (pristine) MAG-10 graphite, and carbon-coated
MAG-10 graphite in respective Li cells cycled at room temperature
in 30% PC blended electrolyte.
410 RT: discharge SGC (0 V)

RT: charge SGC (1.5 V)
50 °C: charge SGC (1.5 V)
390
50 °C: discharge SGC (0 V)
(mAh/g)
50 °C: charge 3.0 % Cu-SGC (1.5 V)

370 50 °C: discharge 3.0 % Cu-SGC (0 V)
RT uncoated SGC
373
Since the rounded edge morphology is designed to lessen the extent

of exposed edge planes that are reactive and susceptible to exfoliation from
co-intercalation of PC on the first lithiation cycle, an improved performance
over the MAG-10 standard (flaky edge) was noted in Figure 11. However, it
was further observed that the RT data was somewhat better than the 50°C
cycling data that showed capacity degradation after 8 or 9 cycles (Figure 12).
After Cu-metallization (3.0 wt%) using the fluidized bed (Wurster coating),
a clear improvement in cycling was seen at 50°C where the results were
similar to RT cycling and the performance was stable in 30% PC blended
electrolyte. We attribute the improvement to decreased PC co-intercalation
in the graphite electrode. The Cu may be acting as a de-solvation catalyst for
stripping the PC coordination sphere from Li cations prior to electrochemical
insertion of lithium and therefore operates more effectively at a higher
temperature.
4. CONCLUSIONS
Some new results on metallization of graphite materials using a
fluidized bed method (Wurster coating) of coating were presented. Cu and
Ni-metallization occurs homogenously over the surface of the graphite from
the formate salts. The coating method disperses the metal formate salt over
the graphite surfaces in an even manner. After a 400°C low-temperature
heat-treatment in flowing dilute hydrogen (4 vol.%) in helium gas, the Ni
and Cu metals are directly formed as spherical 10-100 nm nanodeposits.
Extensive electrochemical cycling tests were performed in lithium cells on
the 11.7 wt% Cu-graphite samples in a 30% PC blend electrolyte in order to
check the resistance to solvent co-intercalation and the improvement in rate
capability.
At 50°C, a marked improvement was seen over the baseline cell,
both in terms of more stable cycling, a higher rate capability, and less first
cycle irreversible capacity loss. Reasons for the improvement in performance
appear to be related to a lowering of the electrode material impedance and a
smaller first cycle irreversible capacity by suppression of PC solvent co-
intercalation due to a de-solvation catalyzed by Cu metal.
Also presented were data on carbon-coating of graphite powder
using a propylene gas thermal decomposition processes. High weight percent
amorphous carbon-coatings are possible with this method, and the process
appears uniquely suited to materials that are reductively stable to 700°C. The
coated materials work better in the 30% PC electrolyte solutions, thus
showing better resistance to solvent co-intercalation problems versus
uncoated types.
374
ACKNOWLEDGEMENTS
Support from the Office of Basic Energy Sciences, Division of
Chemical Sciences of the U.S. Department of Energy (DOE) for the
HRTEM studies and from the Office of FreedomCAR and Vehicle
Technologies (also of DOE) for the surface treatment of the graphite
electrodes and the electrochemical studies under Contract No. W31-109-
Eng-38, is gratefully acknowledged. We would like to especially thank Dr.
Igor V. Barsukov, and Dr. Joseph E. Doninger from Superior Graphite Co.
for supplying the rounded-edge natural graphite powders.
REFERENCES
1. Kepler K. D., Vaughey J. T., Thackeray M. M., LixCu6Sn5 (0 < x < 13): An intermetallic
insertion electrode for rechargeable lithium batteries, Electrochem. and Solid State Lett.,
(1999) 2 (7), 307-309.
2. Vaughey J. T., O'Hara J., Thackeray M. M., Intermetallic insertion electrodes with a zinc
blende-type structure for Li Batteries: A study of LixInSb (0 <= x <= 3), Electrochem.
and Solid State Lett., (2000) 3 (1), 13-16.
3. Fransson L. M. L, Vaughey J. T., Benedek R., Edström K., Thomas J. O., Thackeray M.
M., Phase transitions in lithiated Cu2Sb anodes for lithium batteries: an in situ X-ray
diffraction study, Electrochem. Commun., (2001) 3 (7), 317-323.
4. Vaughey J.T, Johnson C. S, Kropf A. J, Benedek R, Thackeray M. M, Tostmann H.,
Sarakonsri T., Hackney S., Fransson L., Edström K., Thomas J.O., Structural and
mechanistic features of intermetallic materials for lithium batteries, J. Power Sources,
(2001) 97-8, 194-197.
5. Wu, Y. P., Rahm, E., and Holze, R., Carbon anode materials for lithium-ion batteries,
J. Power Sources, (2003) 114, 228-236.
6. Lu, W., Donepudi, V. S., Prakash, J., Liu, J., and Amine, K., Electrochemical and
thermal behavior of copper coated type MAG-20 natural graphite, Electrochimica Acta
(2002) 47, 1601-1606.
7. Winter M., Wrodnigg G. H., Besenhard J. O., Biberacher W., Novák P., J. Electrochem.
Soc., 147, 2427 (2000).
8. Santner H. J., Wagner M. R., Fauler G., Raimann P., Veit C., Möller K. C., Besenhard J.
O., Winter M., Taipei Power Forum and Exhibition (TPF2003), December 1-3, 2003,
Taipei (Taiwan), Proceedings Volume.
9. Chung G. C., Jun S. H., Lee K. Y., Kim M. H., Effect of surface structure on the
irreversible capacity of various graphitic carbon electrodes, J. Electrochem. Soc., (1999)
146, (5), 1664-1671.
10. Abe, T., Kawabata, N., Mizutani, Y., Inaba, M., and Ogumi, Z., Correlation between
cointercalation of solvents and electrochemical intercalation of lithium into graphite in
propylene carbonate solution, J. Electrochem. Soc. (2003) 150 (3), A257-A261.
11. Yoshio, M., Wang, H., Fukuda, Abe, T., and Ogumi, Z., Soft carbon-coated hard carbon
beads as a lithium-ion battery anode material, Chemistry Letters (2003) Vol. 32, No. 12,
1130-1131.
12. Yoon, S., Kim, H., and Oh S. M., Surface modification of graphite by coke coating for
reduction of initial irreversible capacity in lithium secondary batteries, J. Power Sources
(2001), 94, 68-73.
375
13. Yoshio, M., Wang, H., Fukuka, K., Hara, Y., and Adachi, Y., Effect of carbon coating on
electrochemical performance of treated natural graphite as lithium-ion battery anode
material, J. of Electrochem. Soc. (2000) 147 (4) 1245-1250.
14. Lee, H-Y., Baek, J-K., Jang S-W., Lee, S-M., Hong, S-T., Lee, K-Y., and Kim, M-H.,
Characteristics of carbon-coated graphite prepared from mixture of graphite and
polyvinylchloride as anode materials for lithium ion batteries, J. Power Sources (2001),
101, 206-212.
15. Guo, K., Pan, Q., Wang, L. and Fang, S., Nano-scale copper coated graphite as anode
material for lithium-ion batteries, J. Applied Electrochemistry (2002) 32: 679-685.
16. Yu, P., Ritter, J. A., White, R. E., and Popov, B. N., Ni-composite microencapsulated
graphite as the negative electrode in lithium-ion batteries, J. of Electrochem. Soc. (2000)
147 (4), 1280-1285.
17. King, S-S., Kadoma, Y., Ikuta, H., Uchimoto, Y., and Wakihara, M., Electrochemical
performance of natural graphite by surface modification using aluminum,
Electrochemical and Solid-State Letters, (2001) 4 (8), A109-A112.
18. Veeraraghavan, B., Durairajan, A., Haran, B., Popov, B., and Guidotti, R., Study of Sn-
coated graphite as anode material for secondary lithium-ion batteries, J. Electrochem.
Society, (2002) 149, (6), A675-A681.
19. Holze, R. and Wu, Y. P., Novel composite anode materials for lithium ion batteries with
low sensitivity towards humidity, J. Solid State Electrochem. (2003) 8: 66-72.
20. Nishimura, K., Honbo, H., Takeuchi, S., Horiba, T., Oda, M., Koseki, M., Muranaka, Y.,
Kozono, and Miyadera, H., Design and performance of 10 Wh rechargeable lithium
batteries, J. Power Sources (1997) 68, 436-439.
21. Wu, Y. P. and Holze, R., Anode materials for lithium ion batteries obtained by mild and
uniformly controlled oxidation of natural graphite, J. Solid State Electrochem. (2003) 8:
73-78.
22. Shao-Horn Y., Hackney S. A., Cornilsen B. C., Structural characterization of heat-treated
electrolytic manganese dioxide and topotactic transformation of discharge products in the
Li-MnO2 cells, J. Electrochem. Society, (1997) 144, 3147-3153.
23. Sandí, G., Joachin, H., Lu, W., Prakash, J., and Tassara, G., Comparison of the
electrochemical performance of carbon produced from sepiolite with difference surface
characteristics, J. of New Materials for Electrochemical Systems, (2003) 6, 75-80.
24. Wang, H., Fukuda, K., Yoshio, M., Abe, T., and Ogumi, Z., Measurement of carbon
amount in carbon-coated graphite by thermal analysis, Chemistry Letters (2002) 238-
239.
25. Caturla F., Molina F., Molina-Sabio M., Rodriguez-Reinoso F., Esteban A., Electroless
plating of graphite with copper and nickel, J. Electrochem. Soc., (1995) 142 (12), 4084-
4090.
26. Pacchioni, G., Rosch N., Carbonylated Nickel Clusters – From Molecules to Metals,
Acc. of Chem. Res., (1995) 28 (9): 390-397.
27. Ravet N., Chouinard Y., Magnan J. F., Besner S., Gauthier M., Armand M.,
Electroactivity of natural and synthetic triphylite, J Power Sources, (2001) 97-8, 503-
507.
28. Johnson C. S. and Sandí, G, unpublished results, (2001).
29. Nelson P., Bloom I., Amine K., Henriksen G., J. of Power Sources, (2002) 110, 437-444.
30. Suzuki J., Yoshida M., Nakahara C., Sekine K., Kikuchi M., Takamura T., Li mass
transfer through a metallic copper film on a carbon fiber during the electrochemical
insertion/extraction reaction, Electrochem. and Solid State Lett., (2001), 4 (1), A1-A4.
31. Klemm W., Volavsek, B. Z. Anorg. Chem. (1958) 296, 184-187, or Bull. Alloy Phase
Diagrams, (1986) 7 (2), or Binary Alloy Phase Diagrams (1990), Vol. 2, 2nd Ed., ASM
International, Materials Park, Ohio, Editor-in-chief, Massalski T. B., pg. 1430.
32. Binary Alloy Phase Diagrams (1990), Vol. 3, 2nd Ed., ASM International, Materials Park,
Ohio, Editor-in-chief, Massalski T. B., pg. 2450.
376
33. Santner H. J., Möller K. C., Ivanco J., Ramsey M. G., Netzer F. P., Yamaguchi S.,
Besenhard J. O., Winter M., Acrylic acid nitrile, a film-forming electrolyte component
for lithium-ion batteries, which belongs to the family of additives containing vinyl
groups, J. Power Sources, (2003) 119, 368-372.
34. Wrodnigg G. H., Besenhard J. O., Winter M., Ethylene sulfite as electrolyte additive for
lithium-ion cells with graphitic anodes, J. Electrochem. Soc. (1999), 146 (2), 470-472.
35. Aurbach D., Gamolsky K. Markovsky B., Gofer Y., Schmidt M., Heider U., On the use
of vinylene carbonate (VC) electrolyte solutions for Li-ion as an additive to batteries,
Electrochim. Acta, 47 (9), 1423-1439.
36. Vollmer J. M., Curtiss L. A., Vissers D. R., Amine K., Reduction mechanisms of
ethylene, propylene, and vinylethylene carbonates - A quantum chemical study,
J Electrochem. Soc., (2004) 151 (1), A178-A183.
CHAPTER 5:
NEW NANO- THROUGH MACRO-CARBONS

FOR ENERGY SYSTEMS: SYNTHESIS,
MODELING, CHARACTERIZATION
A collection of six papers in this chapter addresses subjects of
synthesis, characterization and modeling of new promising carbon materials,
which have not yet found application in the market of power sources, but
may do so in the foreseeable future.
Thus, in the first paper on stabilization mechanisms discovered for
reaction of synthesis of graphite nitrate, M. Savoskin et al. uncovers a
number of valuable, unknown phenomena related to production of
expandable graphite (GIC), an intermediate product in manufacturing of
thermally exfoliated graphite. The later carbon type is currently being used
in many battery systems as one of the most efficient conductive additives for
electrode matrixes. Authors study intercalation of graphite via reaction of
flake with nitric acid, followed by additions of a modifier. An attempt is
made in the paper to group various (mostly organic) modifiers into classes.
Extensive experimental data is presented in order to support the suggested
mechanism of formation of stable GICs.
The paper by Kovalenko et al. introduces a method of cyclic
voltammetry as a way to estimate stability of various carbon materials
towards electrochemical oxidation. This reaction is of particular significance
in rechargeable batteries with aqueous electrolytes. Authors discuss results
of the electrochemical performance of expanded graphite, whose
surface/structure has been modified by 5wt% B2O3. Reportedly, this graphite
outperforms conventional graphite in terms of featuring higher resistance
towards electrochemical corrosion.
Dimovski et al. addresses a very interesting concept of synthesis of
various forms of carbon materials through reaction of chlorine treatment of
iron carbides, typically found in the kish graphite ore. Currently, kish is a
worthless compound, found in tailings of metallurgical plants. Recovering
graphite from it could be a very beneficial idea. The authors were able to
synthesize amorphous carbon structures at temperatures of chlorine below
1000°C, while at higher temperatures some well-graphitized synthetic
graphite has been recovered. If commercialized, the technology may have
revolutionary implications, also for the needs of energy market.
The contribution by Rouzaud et al. teaches to apply a modified
version of high resolution Transmission Electron Microscopy (TEM) as an
efficient technique of quantitative investigation of the mechanism of
irreversible capacity loss in various carbon candidates for application in
lithium-ion batteries. The authors introduce the “Corridor model”, which is
interesting and is likely to stimulate active discussion within the lithium-ion
battery community. Besides carbon fibers coated with polycarbon (a
candidate anode material for lithium-ion technology), authors study carbon
aerogels, a known material for supercapacitor application. Besides the
capability to form an efficient double electric layer in these aerogels, authors
379
380
report observing some hydrogen storage capability with this type of carbon
materials. Reportedly, about 1% of hydrogen can be stored and released by
the micropores of an aerogel synthesized at 2600°C.
The fifth paper in this chapter is by S. Kochetova and N. Tumanova
of the National Academy of Sciences of Ukraine. It addresses subjects of
electrolysis of molten carbamides. Authors report on interesting results of
their study of the mechanisms of reactions occurring in the carbamide and
carbamide-chloride melts using techniques of cyclic voltammetry in
combination with gas chromatography and IR spectroscopy.
In the last paper of this chapter, the authors M.Savoskin,
M.Mochalin et al. of L.M. Livinenko Institute of Physical Organic and Coal
Chemistry in Donetsk, Ukraine introduce simple methods of “one –step”
synthesis of carbon-carbon composites, as well as relatively new to the
world, carbon structures called “nanoscrolls”. The authors use graphite
intercalation compounds as precursors. Editors like to specifically draw
the reader’s attention to the open-ended carbon nanostructures shown by
Figure 4 in the subject paper. While being unable to qualify the yield of
these unusual nano-compounds, authors suggest the open-ended materials
could be used for storage of various foreign molecules, one of them being
hydrogen.
Editors hope that readers will find this chapter interesting and
educating.
STABILIZATION OF GRAPHITE NITRATE
VIA CO-INTERCALATION
OF ORGANIC COMPOUNDS
M.V. Savoskin*, A.P. Yaroshenko, R.D. Mysyk, G.E. Whyman

L.M.Litvinenko Institute of Physical Organic and Coal Chemistry,
National Academy of Sciences of Ukraine, 70 R.Luxemburg St., 83114 Donetsk, Ukraine
Abstract
New thermally stable graphite intercalation compounds were

obtained via additional intercalation of a series of organic compounds into
graphite nitrate. An explanation for stabilization has been proposed that
involves hydrogen bonding between the protonated modifying organic
compound and neutral surrounding molecules.
Keywords
Graphite nitrate stabilization; co-intercalation; exfoliation; proton

affinity.
1. INTRODUCTION
The structure of acceptor-type graphite intercalation compounds

(GICs) is very specific. Their quasi-two-dimensional layers of intercalant are
quite labile, being not connected to a rigid graphite matrix by covalent
bonds. These negatively charged layers alternate with positively charged
graphene layers of the carbon matrix. The mentioned features of the GICs
structure lead to their instability. For instance, even at room temperature, the
D crystal phase of graphite nitrate of a stoichiometry C24n+NO3- 3HNO3 with
an interlayer distance of 7.80 Å turns into E phase of the same n-th stage that
is characterized by a formula C24n+NO3- 2HNO3 and interlayer distance of
6.55 Å. Moisture can lead to the substitution of nitric acid by water [1, 2].
*
381
382
As we have recently shown [3], the stability of GIC with Brønsted

acids is governed by ionization potential of the intercalated anion. Solvation
of the anion substantially increases ionization potential and hence stabilizes
GIC. Such a possibility was predicted by Inagaki [4] who proposed to use
organic molecules as additional cointercalants for the purpose of
stabilization. Afterwards, we successfully used glacial acetic acid and water
to synthesize stable products [5].
The aim of the present work is to elucidate the mechanism for the
stabilization of graphite nitrate via cointercalation of different kinds of
organic compounds.
2. EXPERIMENTAL
Modified graphite nitrate was synthesized in a thermostatic reactor
at 25°C. Nitric acid with a density of 1.502g/cm3 was added to natural flake
graphite GSM-1 (Zavalie Graphite Works, Kirovograd region, Ukraine)
sample of 10g. The mixture was stirred for 10 minutes. Then an organic
compound (modifier) was added and the mixture was stirred again for 10
minutes. The consumptions of nitric acid and the modifier were 0.4 and 4
cm3 per 1g of graphite, respectively.
Further synthesis followed two ways. According to variant I, a solid
phase was separated by filtration and divided into two parts. The first one
was dried at 105°C for two hours (product A), while the second one was left
at 20°C until a sample mass became constant (product B).
Following variant II, after treating by the modifier, the solid phase
was separated by filtration and additionally washed off with 0.2 liters of
water under vacuum of the water-jet pump. Further, as in variant I, one part
of the product was dried at 105°C for two hours (product C), while the other
was dried at 20°C until a sample mass became constant (product D). The
products were investigated by the powder X-ray phase analysis (PXRD)
using nickel-filtered CoKD radiation.
The new compounds obtained are black-gray powders resembling
initial graphite. Like other GICs, they possess a unique property of
increasing their volume by a factor of tens or hundreds under heating.
Usually, the so-called expansion coefficient serves as a measure of the
mentioned property, KV=V/m, where V is the volume in cm3, and m is the
initial mass of the sample before heating. Note that KV is lower than the
specific volume of exfoliated graphite due to the mass loss resulting from the
escape of GICs-decomposition products. Values of KV were determined in
accordance with the following method. A stainless cuvette of 150cm3 was
placed into a muffle furnace preheated up to 9000C. A product sample with
the mass of 0.2-0.3g was inserted into the heated cuvette and kept in the
furnace for 60s. A deviation between parallel measurements of KV was less
383
than 5%. To estimate the thermal stability of the compounds obtained, the
stability index was introduced as Kst = KV105/KV20, where KV105 and KV20 are
the expansion coefficients for the samples with the same modifier dried at
105 0C and 20 0C, respectively. Obviously, KV105 is less than KV20, since a
product dried at a higher temperature has to contain a smaller amount of the
intercalant. Thus, the stability index Kst has to be less than unity, its value
being closer to 1 for more stable compounds.
Listed in the Table are the expansion coefficients and stability
indices for graphite nitrates cointercalated with different organic compounds,
as well as with nitric acid and water. The modified graphite nitrates
produced may be divided into three groups. Compounds of the first group
(lines 1-8 of the Table) become stable just after modifier intercalation, and,
in fact, their stability does not increase substantially after subsequent water
treatment. Moreover, the stability of some compounds (lines 5, 8) decreases
after the water treatment. Compounds in lines 9-21 belong to the 2-nd group.
They remain unstable after modifier cointercalation, but after additional
water treatment their stability rises sharply. At last, compounds of the 3-rd
group in lines 22-23 remain unstable under any condition of the synthesis, as
well as initial graphite nitrate.
It is interesting that ethers give GICs of the 1-st group, while esters
behave in different ways. Ethyl acetate and ethyl formate lead to GICs of the
1-st group, while esters 14-17 produce those of the 2-nd one, and propylene
carbonate (not included in the Table) gives the GIC of the 3-rd group. The
compounds modified by carboxylic acids belong to the 2-nd group. At the
same time, strong acids stabilize graphite nitrate to a smaller extent as
compared with weaker acids, such as acetic and propionic acids. Various
procedures of the treatment with water and nitric acid do not stabilize
graphite nitrate (see lines 22 and 23).
There are some values of Kst greater than unity in the Table. These
contradict the meaning of Kst as a stability index. This fact may be explained
by a high reactivity of the modifier (acetone, tetrahydrofuran and
dimethylacetamid), which probably causes some reactions in the interlayer
space of GICs. The high reactivity is also characteristic of some other
modifiers with Kst<1. This is particularly true for formic acid, ethers, esters
and nitriles. In this case, some products of oxidation or hydrolysis are likely
to be in the interlayer space. Structural transformations are illustrated by the
PXRD analysis of the GIC modified by acetic acid. The analysis of the X-ray
patterns of initial graphite nitrate in Fig. 1 A shows the presence of two GICs
with the basal repeat lengths Ic equal to 9.88 and 17.73 Å. According to the
well-known formula Ic = di + (n-1)de, these values correspond to the
mixture of the second stage of the E form and fourth stage of the D form, on
account of the above mentioned gallery heights, di = 6.55Å and 7.8Å.
384
Table. Expansion coefficients and stability indices of modified graphite nitrates.
No water Treated with water

No Modifier KV, cm3/g KV, cm3/g
A (dried B (dried Kst C (dried D (dried Kst
at 105 0C) at 20 0C) at 105 0C) at 20 0C)
1 Acetone 220 188 1.17 183 161 1.13
2 Ethyl acetate 172 229 0.75 216 247 0.87
3 Ethyl formate 279 283 0.99 291 303 0.96
4 Dioxane 127 175 0.72 144 172 0.84
5 Tetrahydrofurane 127 90 1.41 107 114 0.93
6 Ethylene glycol dimethyl ether 146 161 0.90 139 147 0.95
7 Diethylene glycol dimethyl ether 162 165 0.98 151 157 0.96
8 Dimethylacetamid 99 85 1.17 74 101 0.73
9 Formic acid 144 266 0.54 135 151 0.89
10 Acetic acid 45 308 0.15 317 354 0.90
11 Propionic acid 57 322 0.18 283 324 0.87
12 Trifluoroacetic acid 28 242 0.12 69 289 0.24
13 Dichloroacetic acid 22 256 0.09 85 181 0.47
14 Methyl acetate 50 242 0.21 167 252 0.66
15 Penthyl acetate 102 279 0.37 142 269 0.53
16 Dimethylcarbonate 27 266 0.10 107 248 0.43
17 Diethylcarbonate 48 336 0.14 173 308 0.56
18 Acetonitrile 31 229 0.13 81 181 0.45
19 Propionitrile 35 196 0.18 61 216 0.28
20 Nitromethane 35 242 0.15 78 198 0.40
21 Trichlorethylene 58 288 0.20 62 206 0.30
22 Water 47 169 0.28 33 178 0.19
23 Nitric acid 30 298 0.10 20 125 0.16
IIE(003)
IVE(005)
C
0,5
IVE(0010)
(Intensity)
IIE(006)
IVD(005)
B
IVD(006)
IIE(003)
IIE(006)
IVD(007) IVD(0010) IVD(0011)
IVD(005)
A
IIE(003) IVD(006)
IIE(006)
IVD(007) IVD(0010) IVD(0011)
10 20 30 40 50 60 70
24, degrees
Figure 1. PXRD patterns (00l) of the products of successive modifications: A - initial graphite
nitrate; B - graphite nitrate modified with glacial acetic acid; C - graphite nitrate
successively modified with glacial acetic acid and water.
385
Heights of the empty layers de were found to be equal to 3.33 and

3.31 Å for the E and D forms, respectively, which are lower compared to
pure graphite. Such a possibility of the contraction of an empty layer height
was discussed by Inagaki and Fuzellier [1, 4].
The PXRD pattern does not change after the treatment of graphite
nitrate with acetic acid (Fig. 1 B), indicating the conservation of Ic and the
stage number of the GIC when nitric acid is partially substituted by acetic
acid. Note that both initial graphite nitrate and that modified with acetic acid
are thermally unstable (see the Table above.)
Contrary to this, the subsequent treatment with water sharply
changes the PXRD pattern of graphite nitrate modified with acetic acid
(Fig. 1 C). Most of peaks disappear, indicating the fundamental
reconstruction of the GIC. At first sight, this pattern resembles that of pure
graphite. However, this compound produced by the successive treatment of
graphite nitrate with acetic acid and water contains about 20-26 wt% of the
intercalant and has a greater ability of expansion and higher thermal stability
compared to graphite nitrate (see line 10 in the Table.). Indeed, the detailed
analysis reveals the presence of the second and fourth stages of the E phase
only. The found patterns sequence from A to C is in good agreement with
the known regularity of the graphite nitrate staging established in Ref. [1].
For the modifiers from the first group, the transformation is more
complicated. As is shown by the corresponding X-ray patterns, the
significant change occurs in the staging along with the probable change in
the height of filled layers of GICs. Additional X-ray investigations are
required to identify the compounds obtained.
3. DISCUSSION AND CONCLUSIONS

In the quasi-two-dimensional structure of acceptor type GICs shown
schematically in Fig. 2 each anion is surrounded by six acid molecules S [7].
However, according to the stoichiometry each anion is solvated by only two
ligands. Additional stabilization might come from formation of a continuous
two-dimensional net of hydrogen bonds among the solvation ligands.
However, this is impossible in the case of pure graphite nitrate as the
molecules of nitric acid, which solvates the anion, cannot give rise to the
hydrogen bonding among them. This is because the only proton of nitric acid
has already been bound with the nitrate anion. The insertion of the second
intercalant capable of the additional hydrogen bonding could realize such a
possibility.
On modifying graphite nitrate, as follows from the material balance
of the reaction, a major part of nitric acid is replaced by the organic
intercalant. The latter, if it is protonated by the rest of nitric acid in the
interlayer space, is capable of the strong hydrogen bonding with two
386
molecules. A similar possibility for graphite bisulfate modified with acetic

acid is discussed in Ref. [8]. Noteworthy that the energy of the hydrogen
bond incorporating charged species, protonated modifiers in the present case,
is several times greater than that for neutral molecules. It can exceed
100 kJ/mol [9]. Unprotonated organic cointercalant and water can play a role
of dielectric spacers among the charged species.
A S S A
S A S
A S S A
Figure 2. Fragment of idealized structure of the GIC [7].

The intercalant layer, which comprises the anions A- and ligands S,
is placed between two positively charged graphite layers.
To verify the mechanism presented, the quantum-chemical

calculations of proton affinity, AH, were carried out for modifiers, since the
corresponding experimental data are quite rare. The calculations were
performed for isolated molecules, since the properties of species in the
interlayer space are probably closer to the gas phase rather than the liquid.
The values of AH were calculated as a difference in the total energy between
the initial and protonated forms of the modifier. Energies were calculated
using the TZV(2df, 2p) basis and MP2 electron correlation correction.
Preliminarily, geometries were fully optimized in the framework of the
MP2/6-31G(d, p) calculation. The GAMESS suite of ab initio programs was
employed [10]. Comparison between the calculated at 0 K proton affinities
for water (7.46 eV) and dioxane (8.50 eV) and the experimental data 7.50 eV
and 8.42 eV at 298 K, respectively (see [11]), demonstrates a sufficient
accuracy of the calculation.
Despite a wide range of the modifiers used, there is a clear symbasis
between the calculated proton affinities and the experimental stability
indices Kst (Figure 3). The decrease in concentration of nitric acid in the
interlayer space will give rise to a poorer stability of GICs, if the proposed
stabilization mechanism is valid. To verify this assumption, the treatment of
graphite nitrate with 25-fold excess of acetic acid, compared to that used in
the above procedure, was carried out in a glass column. Most of the
molecules of nitric acid are probably removed from the interlayer space
under such a treatment. It turns out that this procedure reduces the expansion
387
coefficients, from 317 to 147 cm3/g for product C and from 354 to 210 cm3/g
for product D. The stability indices Kst also decreases from 0.90 to 0.70. This
substantial decrease in KV and Kst additionally supports the proposed
stabilization mechanism.
1.2
Kst 1
1.0
3 6 5
10
11 2
4
0.8
14
0.6
13
18
0.4 20
12
0.2 22
23
AH (eV)
0.0
7.0 7.2 7.4 7.6 7.8 8.0 8.2 8.4 8.6 8.8 9.0
Figure 3.Comparison between the calculated proton affinities and stability indices.
Numerals correspond to the line numbers in the Table.
REFERENCES
1. Fuzellier H, Melin J, Herold A. Une nouvelle variété de nitrate de graphite. Mater. Sci.
Eng. 1977; 31:91-94.
2. Forsman W.C., Mertwoy H.E., Wessbecher D.E. Nonreductive spontaneous
deintercalation of graphite nitrate. Carbon 1988; 26:693-699.
3. Savoskin M.V., Yaroshenko A.P., Whyman G.E., Mestechkin M.M., Mysyk R.D.,
Mochalin V.N. Theoretical study of stability of graphite intercalation componds with
Brønsted acids. Carbon 2003; 41:2757-60.
4. Inagaki M. Applications of graphite intercalation compounds. J. Mater. Res. 1989; 4:
1560-68.
5. Savoskin M.V., Yaroshenko A.P., Mysyk R.D., Shologon V.I., Khripunov S.V.
Synthesis of expandable residual graphite nitrate stabilized with glacial acetic acid.
Proceedings of the International Conference on Carbon; 2003; Oviedo, Spain. P. 188-
189.
6. Inagaki M. Graphite-nitrate residue compounds with a smaller interlayer spacing than
graphite. Carbon 1967; 5:317-8.
7. Jiang J., Beck F., Krohn H. Electrochemical reversibility of graphite oxide. J. Indian
Chem. Soc. 1989; 66:603-9.
8. Kang F., Zhang T.-Y., Leng Y. Electrochemical behavior of graphite in sulfuric and
acetic acid. Carbon 1997; 35:1167-73.
9. Joesten M.D., Schaad L.I., Hydrogen bonding. New York: Marcel Dekker, 1974.
10. Schmidt M.W., Balderidge K.K., Boatz J.A., Elbert S.T., Gordon M.S., Jensen J.H.,
Koseki S., Matsunaga N., Nguyen K.A., Su S., Montgomery J.A. General atomic
molecular electronic structure system. J. Comput. Chem. 1993; 14:1347-63.
11. Gasteiger J., Hutchings M.G. Proton affinities of different gases. J. Amer. Chem. Soc.
1984; 106:6489-95.
ELECTROCHEMICAL STABILITY OF
NATURAL, THERMALLY EXFOLEATED
AND MODIFIED FORMS OF GRAPHITE
TOWARDS ELECTROCHEMICAL OXIDATION
Igor O. Kovalenko1*, Mayya L. Barsukov2, Vyacheslav Z. Barsukov3

and Y. V. Myshko
1
ChimLaborReactiv, Ltd., Remontnaya Street, 8, Kiev, 02099, Ukraine
2
Institute for Problems of Materials Science, National Academy of Science of Ukraine
3
Kiev National University of Technologies & Design, Department of Chemistry,
2, Nemirovich-Danchenko street, Kiev, 252011, Ukraine
Abstract
Acceptor-type graphite intercalation compounds (A-GIC) can

usually be obtained through electrochemical oxidation of graphite in aqueous
acidic or neutral solutions. Together with formation of A-GIC compounds,
oxygen evolution usually takes place at the anode. This may lead to
irreversible changes in the carbonaceous material, and, consequently, may
result in a loss of reversibility of the system, caused by the partial oxidation
of graphite. Thus, stability of graphite during cycling is the most important
factor of electrode durability.
A special electrochemical method of prolonged sequential cyclic
voltammetry is proposed as means of evaluation of stability of graphite
towards oxidation. Key corrosion mechanisms for different types of graphite
are cosidered. Special types of thermally expanded graphite (TEG) modified
by P2O5, B2O3 and H3BO3 are compared. A maximal cycle life (of up to 550
cycles) is obtained with TEG modified by 5% of B2O3.
1. INTRODUCTION
Graphite finds wide range of applications in the electrodes for
certain types of rechargeable batteries and supercapacitors, in electro-
sorption/desorption electrodes, as anodes in a number of processes of
*
Corresponding author. E-mail: [email protected], [email protected]
389
390
electrochemical synthesis, etc. Nevertheless, corrosion processes during

anodic polarization of such electrodes are among the main limitations for
their practical use.
For example, graphite intercalation compounds (GICs) as active
material for metal-free rechargeable batteries (MFRB) [1], are represented
by graphite anodes, that are being reversibly intercalated-deintercalated in
aqueous solutions of strong acids by the following reaction:
- -
Cx + A + yHA [CxA yHA] + e , (1)
where Cx - the carbon macromolecule, A- - the anion, y - the

stechiometric factor for solvated molecules of the acid, HA. Nevertheless, in
parallel to formation of GICs, there usually take place side corrosion
reactions [2] connected with oxygen evolution process on anode:
- the anodic evolution of oxygen

+ -
2H2O 4H + 4e + O2; (2)
- the oxidation of graphite to CO2
CX + 2H2O CO2 + CX-1 + 4H+ + 4e-; (3)
- the formation of surface groups
CX + 2H2O CX-1COOH + 3H++ 3e-; (4)
- the formation of graphite oxide
CXA + H2O CXOH +H++A-. (5)
This may lead to the irreversible changes in the material, caused by

partial oxidation of graphite, loss of reversibility of the system along with
the efficiency of reaction for intercalation-deintercalation (1).
2. EXPERIMENTAL
2.1. Electrode Preparation

We have manufactured the graphite electrodes of “plug” type. An
active mass (88wt% graphite, 10wt% carbon black and 2wt% PTFE) was
pressed into a polyethylene plug of 11 mm inside diameter. We have studied
391
the characteristics of electrodes with an equal weight of 0.280g. Before

pressing, the Pt current collector was inserted into an opening, pierced in the
plug wall.
2.2. Graphite Materials for Investigation

We have performed comparative investigation of graphite
electrodes, obtained from various types of starting natural and artificial types
of graphite. The following materials were compared:
#1. Grades of natural graphite from Zavalie (NGZ) graphite deposit,
Kirovograd Region, Ukraine;
#2. Natural graphite flakes (NGF) “Kropfmuhl Normalflocke”
£
(BASF , Munich, Germany);
#3. Thermally exfoliated graphite (TEG) and their composites from
Superior Graphite (Chicago) and TEGs prepared from Ukrainian graphite.
2.3. Synthesis of Thermally Exfoleated Graphite

Starting natural graphite from Zavalie deposit in, Ukraine was
chemically intercalated in sulfuric acid. Potassium persulphate was used as
an oxidizer. Thermal expansion of graphite was performed at 900qC [3].
Carbon content in the experimental samples was of about 99.0%.
A United States - manufactured TEG (ABG-74, a commercial grade
of Superior Graphite Co., Chicago, Illinois, USA) was manufactured in
principle in a similar manner, yet with a process that used few more
important processing steps, as follows.
First of all, the precursor natural flake graphite having a starting
purity of approximately 95%C has been heat treated is a continuous thermal
purification process (this unique process of Superior Graphite Co. has ability
to achieve and sustain a relatively large zone of uniform temperature in
excess of 2500qC). The purpose of heat treatment of the precursor (95%
carbon natural flake graphite) was reduction of the mineral ash in graphite
structure in order to achieve lowest possible resistivity (0.007 Ohmcm) and
very high pure in the final product (99.95%C), [4].
2.4. Structural and Electrochemical Investigation of

Graphite
For purposes of investigation of the micro- and macro-structural
characteristics of graphite, we have used methods of X-ray diffraction, SEM
and we also studied the pattern of particle distribution by size with the SK
laser micron sizer (Horiba, Japan).
392
2.4.1. Consecutive Galvanostatic Cycling with Control of the Electrode

Weight Loss
We have used the method of galvanostatic cycling in the 0.5-1.75V

or 0.5-2.00V potential ranges vs the Ag/ AgCl reference electrode. We have
monitored the loss of electrode active mass during cycling for purposes of
studying the possible mechanisms of electrochemical corrosion on graphite
electrodes.
2.4.2. Consecutive Voltammetric Cycling
Electrodes were manufactured of NGZ, NGF and TEG and were

tested under comparable conditions. The relative resistivity of graphite
electrode to corrosion was compared on stability of CVs during the N (stands
for “number”) of consecutive voltammetric cycles in a given potential
region. The relative resistivity to corrosion was estimated by the number of
cycles for very stable cycling (Nstable) and for saving of electrochemical
activity (Nmax).

3.1. Structural Investigations
3.1.1. Structural Investigations in the Beginning and Finish of Cycling
for NGZ and NGF
It is possible to make conclusion about the absence of noticeable

structural changes based on comparison of structural parameters of graphite
in the beginning and at the end of cycling, using the XRD method.
The parameters of an elementary crystal cell for initial and finite
state are practically identical. For example, interplanar distance C= (6.718 r
0.007)10-8 cm for NGZ (GAK-1 grade) and NGF dos not change during
cycling.
3.1.2. Comparison the Particle Size Distribution for NGF and TEG
We have studied the key physical properties of samples of TEG,

NGF and NGZ using an SK laser micron sizer and BET surface area
analyzer. Unless otherwise stated, the particle size distribution and surface
area measurements for this work have been carried out at the Institute for
Problems of Materials Science of the National Academy of Sciences of
393
Ukraine. It was observed that the particle size distribution of products does
not vary significantly during cycling.
The average particle size values for NGZ (GAK-1 type) and NGF
was determined to be, in particular, 35-40 Pm. The specific surface area of
such graphites in initial and final states is also practically identical and is of
approximately 6.8-7.0 m2/g for natural graphite and approximately 20-25
m2/g for the internally produced exfoliated graphite.
The average size of particles for TEG from Superior Graphite Co. is
sufficiently less (D50 d 10 Pm.) [4]. The specific surface area of the ABG-74
grade of exfoliated graphite is 24 m2/g.
Due to having finer particle size and higher surface area, one may
suppose that TEG is less oxidation-resistant than NGF and NGZ. However,
we have shown TEG to be much more oxidation-resistant than NGF.
The cycling data is summarized in Table 1. It becomes clear from
the data in Table 1 that the initial specific capacitive parameters of some
types of NGZ are higher than NGF by a factor of 1.5-2.0. Nevertheless, the
values of NGmax for NGF are considerably higher than those for NGZ. Thus,
NGZ are less resistant to electrochemical oxidation than NGF.
3.2. Electrochemical Investigations
3.2.1. Comparison of Cyclic Voltammograms
On the first stage of investigation, cyclic voltammograms (CVs)

were compared for the graphite electrodes during cycling in maximal
possible potential range 'Emax = 0-2.1V (vs. the Ag/ AgCl reference
electrode) at a sweep potentials rate of 10 mV/s.
As our cyclic voltammety investigations show, the minimal
resistivity to corrosion (Nstable=20-30 cycles; Nmax=50-90 cycles) is
demonstrated by the natural Zavalie graphites (group #1).
It can be clearly seen from the curves presented for TEG (Figure 1)
that there is not any evidence for the electrode surface oxidation, at least
during the first Nstable = 190 cycles, whereas the oxygen evolution rate is
quite high under these conditions. After 190 cycles a gradual decrease of
peak currents can be noticed (see curves corresponding to 200 and 250
cycles, respectively). The service life of the electrode composed of TEG
under these conditions is Nmax=250 cycles.

50% particles less than 10Pm.; 90% particles less than 24 Pm.
394
Table 1. Summary of main characteristics of NGF and different NGZ materials

(per results of galvanostatic cycling)
# Type of Graphite NGmax Dmax, qmax, Pmax

Cycles % Ah/kg %
1 Natural graphite flakes (NGF) Kropfmuhl
Normalflocke, Germany >100 67.1 5.77 20.5
Natural graphites from Zavalie (NGZ) graphite deposit, Kirovograd Region, Ukraine:
2 Graphite for Zn-MnO2 cells (Standard #7478-75):
2.1 GE-1 39 90 10.5 37.3
2.2 GE-2 35 90 9.3 33.1
3 Graphite for motor carbon brushes, EUZ-M
type (Standard #10274-79) 26 31 12.8 40
4 Graphite for electric motors, EUZ-E type
(Standard # 21-25-156-75) 16 63.5 10.8 38.4
5 Graphite for alkaline accumulators (Standard #10273-73):

5.1 GAK-1 10 35.2 11.1 39.5
5.2 GAK-2 7 27 8.29 25.9
6 GMS-1, low ash graphite for special 5 29 8.52 30.3
applications (Standard # 18191-78)
7 Special graphite, C-1 type 4 18 6.1 19.1
(Standard # 113-08-48-63-90)
In this table:
NGmax - is the maximum number of cycles before the discharge electrode capacity Qd will
reduce to the minimal level (Qdmin=4Ah/kg in our case);
Dmax = (Qd/Qch)100% - is the maximum value of current efficiency; Qch- charge electrode
capacity);
qmax - is the maximum value of specific discharge capacity (Ah/kg) during the cycling;
Pmax = (Qd/QThd)100% - is the utilization factor (%); QThd - theoretical discharge capacity.
When cycling a NGF electrode under the same conditions

(Figure 2), the deintercalation peak current values are much higher. At the
same time, under comparable conditions, a noticeable surface oxidation of
NGF starts to be observed significantly earlier, after Nstable = 50-60 cycles.
The service life of the NGF electrode is Nmax = 100 cycles.
On the second stage of investigation (Fig. 3) cyclic voltammograms
were compared for the NGF (a), TEG (b) and TEG modified by P2O5 (c),
B2O3 (d) and H3BO3 (e) during cycling in more realistic and narrower range
of potentials 'En = 0.8-2.1V at the same sweep potentials rate of 10 mV/s.
395
In such 'En, the values of Nstable and Nmax becomes higher (for instance,
Nmax = 180 and 350 cycles for NGF and TEG, accordingly).
Thus, considering the results of the above comparison, it is easy to
conclude that the TEG electrode is more resistant to the electrochemical
corrosion (oxidation). A maximal cycle life in such conditions (up to 550
cycles) is obtained with TEG modified by 5% of B2O3.
3.3 Investigation of Mechanisms of Electrochemical

Corrosion of Graphite Electrodes
3.3.1. Comparison of Electrode Capacity Loss with the Loss of Graphite
Electrodes Mass
There exists a point of view that the main mechanism of

electrochemical corrosion is assoceated with a full step by step oxidation of
graphite to CO2 during the cycling, according with reaction (3), [2].
For purposes of studying of the possible mechanisms of
electrochemical corrosion we have compared first of all the capacity loss for
graphite electrodes with the loss of their mass during prolonged
galvanostatic cycling (Table 2).
Figure 1. CVA for the TEG - electrode (m=0.28g); Emax=2.1V; v=10mV/s.

396
Figure 2. CVA for the NGF - electrode (m=0.28g). Emax=2.1V; v=10mV/s.
Table 2. The Capacity and the Electrode Weight Loss for NGZ (GAK-1 type) and NGF
(“Kropfmuhl Normalflocke”) During the Consecutive Galvanostatic Cycling up to
Umax1=1.75V and Umax2=2.00V.
Electrode capacity loss Weight loss of electrode
Graphite Umax, active mass
Type V Q0, Qmin, G, M0, Mmin G,
mAh mAh % g g %
NGZ 1.75 5.00 1.66 66.8 2.00 1.00 50.0
(GAK-1) 2.00 9.00 0.48 94.7 2.00 0.87 56.5
NGF 1.75 5.70 3.33 67.5 1.14 1.11 2.51
2.00 9.11 3.33 68.0 1.13 1.11 2.49
The comparison of electrode capacity loss with loss of their mass

during the cycling unambiguously shows, that degradation of electrical
characteristics cannot be explained only by CO2 derivation with appropriate
decrease of an active material. The main mechanism depends strongly on the
type of graphite.
For example, for NGZ, the loss of capacity is higher but comparable
with the loss of graphite active mass (66.8 and 50.0%; 94.7 and 56.5%).
Thus, corrosion mechanism (3) is a predominant one for such an unstable to
oxidation graphite class.
For NGF, the loss of capacity is much higher (approximately 68%)
than the loss of graphite active mass (approximately 2.5%). Thus, for this
type of graphite the corrosion mechanisms (4) and (5) with formation of
surface groups may be considered predominant.
Similar mechanisms (please see equations (4) and (5)) are typical
also for TEG and modified TEG. As results of additional acid-base titration
397
shows , even in the initial state of cycling, the highly developed internal
surface of TEG is covered by thin layer of -COOH and -OH groups, which
are the result of heat-treatment during preparation of expanded graphite. This
makes TEG more stable towards the electrochemical oxidation during
cycling. It is well known also that chemical activity of any carbon material is
less if this carbon material was heat treated at higher temperature.
The stabilizing influence of oxides, which cover the surface of
graphite (result of heat treatment during purification and exfoliation) is
usually associated with enhancement of its acceptor properties. This
influence in the case of TEG was seen to produce highly stable to oxidation
carbonaceous material.
Furthermore, we believe that the stabilizing influence of boron in the
structure of graphite is connected with enhancement of its acceptor
properties, which manifest themselves when Boron atoms substitute carbon
atoms in the crystalline structure (hexagon ring) of carbon. Such effects are
mentioned in the literature for some types of carbon materials [3] and the
influence of boron on TEG can be the similar.
4. CONCLUSIONS
The minimal resistance to oxidation (few dozens of cycles) was
demonstrated by the natural graphite from Zavalie deposit, Ukraine. Natural
graphite flakes «Kropfmuhl Normalflocke» proved to possess significantly
higher resistance to oxidation (up to 100 cycles). Furthermore, TEGs
demonstrated even greater resistance to oxidation (250 cycles).
The main corrosion mechanism for the natural graphite from Zavalie
deposit, as proposed by the authors, is associated predominantly with a full
step by step oxidation of graphite to CO2 upon cycling, in accordance with
reaction (3). The main mechanism of corrosion of graphite electrodes based
on NGF and TEG is believed to be the derivative accumulation of surface
groups (reactions (4), (5)), which gradually block “entrances" into the
crystalline lattice and thus hinder intercalation of anions in interlayer spaces.
Our investigations show that the most probable areas, where the redundant
acid centers are located, are the initial sites and "entrances" in the interlayer
spacing.
Application of new types of graphite, found to be more oxidation-
proof (in particular, TEG and TEG modified by boron), can largely increase
the electrochemical stability of materials used in aqueous electrolyte media.
Their high resistance to oxidation and enhanced long-term cycling stability
create realistic prerequisites for wide range of applications for such graphite

The acid-base titration was performed by a Ph.D. V. Lucenko and an engineer V. Drozdik of
Kiev National University of Technologies and Design, Kiev, Ukraine.
398
materials in rechargeable batteries, electrosorption, electrochemical

synthesis, etc. applications.
REFERENCES
1. I.V. Barsukov, T.I. Motronyuk, V.Z. Barsukov, V.I. Drozdik. Metal-Free 1.5V
Rechargeable Batteries: Steps of Optimization and Prospects for the Practical
Application// In book: Batteries for Portable Application and Electric Vehicles, C.F.
Holmes and A.R. Landgrebe eds., The Electrochemical Society, Inc., Pennington, NJ,
592-596 (1997).
2. H. Krohn, F. Beck, and H. Junge, Reversible Electrochemical Graphite Salt Formation
from Aqueous Salt Electrilytes, Ber. Bunsenges. Phys. Chem., 86, 704 - 710 (1982).
3. I.G. Chernish, I.I. Karpov, G.P. Prikhodko, and V.M. Shai, Phisical Chemical Properties
of Graphite and its Compounds, Naukova Dumka, Kiev (1990) /in Russian/.
4. Product Information Bulletin – Expanded Graphite ABG-74. Superior Graphite Co.
(1999).
LOW TEMPERATURE SYNTHESIS OF
GRAPHITE FROM IRON CARBIDE
Svetlana Dimovski, Alexei Nikitin, Haihui Ye and Yury Gogotsi*

Drexel University, Department of Materials Science and Engineering
A. J. Drexel Nanotechnology Institute, 3141 Chestnut Street, Philadelphia, PA 19104, USA
Abstract
Synthesis of graphite by extraction of iron from iron carbide by
chlorine is discussed in this work. This process is attractive because it can
produce well-ordered graphite at temperatures as low as 600°C, providing an
opportunity for low-temperature solid-state synthesis of graphitic materials
for batteries. The composition of the reaction products under equilibrium
conditions was determined by thermodynamic simulation, and the initial
process parameters such as temperature and chlorine/carbide molar ratio
were selected accordingly. Crystalline parameters of the graphitic reaction
products, such as the interlayer spacing and crystalline sizes, were calculated
from X-ray diffraction measurements. The degree of orientation of the
graphitic layers was determined by Raman spectroscopy. Three temperature
regimes have been identified. At temperatures below 500°C, amorphous or
disordered carbon is formed as shown by Raman spectroscopy and TEM
studies. Well-ordered graphite microcrystals are formed by solid-state
growth between 600°C and 1100°C. Above the eutectic temperature 1130°C
in the Fe/Fe3C system, the growth of large graphite crystals occurs from the
liquid phase, similar to the formation of kish graphite by precipitation of
carbon at high temperatures from supersaturated molten iron. Iron chlorides,
the main impurities in the material synthesized by the solid-state growth, can
be removed by using excess chlorine gas or by a separate wet chemical
purification step.
1. INTRODUCTION
It has been previously shown that selective etching of carbides is an
attractive technique for the synthesis of various carbon structures. Carbon
produced by extraction of metals from carbides is called carbide-derived
*
Corresponding author. E-mail: [email protected]; Web: http://nano.materials.drexel.edu
399
400
carbon (CDC).1 Leaching in supercritical water, high-temperature treatment

in halogens and vacuum decomposition can be used to remove metals from
their carbides, producing carbon coatings, powders or bulk shapes. For the
chlorination reaction to proceed, it is necessary that molecules of Cl2(g) are
transported to the carbide/carbon interface and molecules of metal chloride
are transported away. Linear reaction kinetics have been demonstrated for
the chlorination of SiC2, implying that the carbon film formed during the
process on the carbide surface is nanoporous3, and as such allows easy
permeation of Cl2 molecules. This phenomenon allows transformation of a
carbide particle or a bulk component to any depth until it is fully converted
to carbon. However, the microstructure of the CDC formed by selective
etching in halogens can vary significantly with experimental conditions.
Thus amorphous and nanocrystalline graphitic carbon,4 nanotube-like
structures5 and nanocrystalline diamond2 were reported. Ordered graphite
was only observed in small amounts as a thin film on the surface of SiC
crystals, because a solid state synthesis of graphite6,7 usually requires very
high temperatures (exceeding 2500ºC). This paper describes the formation
and evolution of graphite in CDC produced by extraction of iron from iron
carbide using chlorine gas. It indicates the necessity of transition metals in
the system for low temperature graphite synthesis. An increased degree of
graphitization in CDC produced by chlorination of TiC in the presence of
small amounts of Fe, Ni, and Co has been recently demonstrated.8 However,
the interaction of iron carbide with chlorine or other halogens has not been
reported in the existing literature.
2. MATERIALS AND EXPERIMENTAL
2.1. Materials
Iron carbide powder for this study was provided by Hazen Research,
Inc. Chlorination experiments were done on the as-received powder.
Scanning electron microscopy, energy-dispersive X-ray spectroscopy (EDS),
wide angle powder X-ray diffraction (XRD), and Raman spectroscopy
analyses were conducted to characterize the original material. The average
particle size was measured to be ~0.5 Pm. These particles formed grain-like
iron carbide clusters (av. size ~350 Pm). The XRD pattern corresponds to
that of synthetic Fe3C (ICDD form 35-0772). No other compounds were
detected by the XRD of raw powder. However, EDS analysis showed traces
of oxygen, probably originating in the very thin surface layer of intrinsic iron
oxide, and/or CO2 and water absorbed from the air, and aluminum, most
likely from the sample holder.
401
Chlorine (BOC Gases, 99.5%) was dried with concentrated sulfuric

acid and anhydrous calcium sulfate. Purging was done with 99.998% pure
argon (BOC Gases).
2.2. Thermodynamic Simulation

The ChemSage v4.14 advanced SOLGASMIX-based Gibbs energy
minimization program from GTT, Germany, was used to estimate
equilibrium amounts of reaction species as functions of the incoming amount
of chlorine gas (per one mole of Fe3C) and chlorination temperature for
different amounts of chlorine. The calculations were done for a closed
system with a constant total pressure. The solid phase was treated as a
mechanical mixture of compounds with unit activities, and the gas phase was
considered a mixture of ideal gases.
2.3. Synthesis
The apparatus used to treat Fe3C with chlorine at atmospheric
pressure has been described elsewhere.1 Fe3C was loaded in a quartz boat
and placed in a horizontal quartz tube reactor, which was purged with argon
gas until the desired reaction temperature was reached. The sample was then
exposed to flowing Cl2 gas for 3 h, and purged again with argon during the
cooling period, as described in literature1,9. Constant reaction temperature
and gas flow rate were maintained throughout every run. The interaction of
iron carbide with chlorine at temperatures ranging from 400qC to 1200qC
was investigated. Carbon collected from the quartz boat was purified, when
necessary, by washing away residual impurities with 10% HCl followed
double distilled water 3 times in sequence, and dried carefully at a
temperature of about 100 - 150qC.
The chlorinated samples were analyzed using a Philips XL-30
environmental field-emission (FE) SEM. SEM analysis was always
accompanied by EDS analysis to verify that the analyzed material was
carbon. TEM samples were prepared by dispersing the synthesis products in
isopropyl alcohol over a copper grid with a lacey carbon film, and
transmission electron microscopy (TEM) analysis was performed using a
JEOL JEM-2010F (200kV). A Gatan image filter for energy filtered imaging
and electron energy loss spectroscopy (EELS) was used to identify carbon
areas and impurities. Raman microspectroscopy (Renishaw 1000) with the
402
excitation wavelength of 514.5 nm (Ar ion laser) with a range of 800-2000

cm-1 and exposure time of 10 s was used to analyze the degree of carbon
ordering. The spectra were collected from at least 5 different points of each
sample with a 2-Pm spot size. Low laser power was applied to avoid
damaging and heating the sample. GRAMS 32 v 5.2 spectral analysis
software was used for peak fitting and deconvolution. The in-plane
crystallite size of graphite, La is shown to be inversely proportional to the
ratio of D band at 1350 cm-1 and the G band at 1580 cm-1 (ID/IG ratio). It can
be calculated as: La=4.4(ID/IG)-1 (nm) according to Tuinstra and Koenig’s
work10 for 2.5 nm < La < 3000 nm.
X-ray diffraction was done using a Siemens D-500 diffractometer
utilizing CuKĮ radiation (1.406 Å). The data were collected as step scans,
with a step size of 0.05° 2ș and a count time of 2 s/step between 10° and 80°
2ș.
3.1. Reaction Thermodynamics

Since iron chlorides are more thermodynamically stable than CCl4,
at high temperatures chlorine reacts selectively with the metal at the carbide
surface by the reactions:
2 Fe3C 9Cl2 6 FeCl3 2C , or (1)

Fe3C 3Cl2 3FeCl2 C (2)
The formation of chloride dimers such as Fe2Cl6 and Fe2Cl4 in the

gas phase is also possible under certain conditions (Figure 1). Thermo-
dynamic modeling can be utilized to estimate the effect of various process
parameters on CDC formation and predict optimal synthesis conditions.11
Thermodynamic analysis (Figure 1) shows that a small amount of chlorine
results in the formation of solid FeCl2 in the sample, and relatively high
temperatures (above 1000qC) are needed to remove iron salts from CDC.
Excess chlorine shifts the reaction toward the formation of Fe2Cl6 (g), while
carbon may be consumed to form CCl4 (g) at temperatures below 400qC. At
higher temperatures, CCl4 (g) is not stable, which enables the formation of a
solid carbon phase. Intermediate regimes result in the removal of iron
compounds as gaseous reaction products. However, some solid FeCl2 may
remain in the system. Further reduction of iron from its salts to metallic iron
using H2 gas may enable the design of CDC – metal catalyst systems.
403
Equilibrium amount of species/mol

3.5
3
Cl2 (g) FeCl3 (g)
2.5
2
Fe2Cl6 (g)
1.5
CCl4 (g) C
1
FeCl2 (g)
0.5
0
200 400 600 800 1000 1200
Temperature/ºC
Figure 1. Thermodynamic analysis of iron carbide chlorination: a.) Equilibrium
amount of species vs. chlorination temperature for n = 8 moles of Cl2 (g).
Thermodynamic calculations of the SOLGASMIX type, such as

shown here, provide information about closed systems in equilibrium. The
experimental system used in this work was an open system, and gaseous
reaction products were continuously removed with the gas flow, shifting the
reactions (1, 2) equilibrium to the right and favoring the formation of carbon.
Caution must be taken in the interpretation of the thermodynamic calculation
results, in which further refinement is clearly possible; however, these results
do provide some general guidelines and are generally in good agreement
with experimental results.
3.2. Raman Spectroscopy Analysis

Typical Raman spectra of Fe3C-derived carbon powders synthesized
at various temperatures are shown in Fig. 2. Raman analysis shows that Fe3C
was successfully converted into carbon with various degrees of ordering,
depending on the reaction temperatures. Three main temperature regimes can
be identified here. At temperatures below 500ºC, chlorination yields
amorphous or disordered carbon. Broad disorder-induced D (1353 cm-1) and
graphite G (1596 cm-1) Raman bands can be found in the spectrum from the
sample obtained by chlorination at 400°C (Fig. 2a). A relatively high ID/IG
ratio of §0.90, corresponding to in-plane crystalline size of graphite La<4
nm,10 indicates that amorphous carbon is produced under these conditions.
Although the D and G bands are the most prominent features in
Raman spectra of CDC, the actual Raman scattering data cannot be fitted
with 2 peaks in the case of highly amorphous structures. Such difficulties
were observed for Fe3C-derived CDCs produced below 500qC. Instead, for
an adequate representation of these spectra, three additional broad peaks of
404
relatively low intensities located in the vicinity of 1225 cm-1 (unnamed) 1620
cm-1 (D’) and 1500 cm-1 (D”) were included. Peak fitting was conducted
following the procedure described in reference.9
G 1582 cm-1
FWHM 16 cm-1
Intensity/a.u.
Graphite crystals 1200 ºC (c)
D 1353 cm -1 FWHM 35 cm-1
Nano and micro-

D’ 1620 cm-1
crystalline graphite 600 ºC (b)
1353 cm-1 1596 cm-1
Disordered carbon 400 ºC (a)
1000 1100 1200 1300 1400 1500 1600 1700 1800

1800
Raman Shift/cm-1-1
Figure 2. Raman spectra of carbon produced by chlorination of Fe3C for 3 hours at
different temperatures.
At reaction temperatures of 600°C - 1100°C, the D and the G bands

become well-defined and clearly separated (D at ~1353 cm-1, G at 1582
cm-1). In addition, a weak D’ band appears as a shoulder around 1620 cm-1
(Figure 2b). The shift of the G peak position towards that of ordered graphite
(1582 cm-1) and decrease of its full width at half maximum (FWHM) value
to ~35 cm-1 indicates the onset of graphitization. Increasing the reaction
temperature increases the intensity of the G band relative to the intensity
of the disorder-induced D band, decreasing the ID/IG ratio from 0.44
(in average) at 600°C to about 0.13 at 1000°C, which corresponds to La§30-
55 nm. For comparison, the ID/IG ratio of CDCs produced from SiC at the
same reaction temperature is close to one.9 Thus, the presence of iron in the
system facilitates graphite nucleation and growth at much lower
temperatures, which is in agreement with previous publications showing the
catalytic effect of iron on graphitization.8
Above the eutectic temperature in the iron-Fe3C system (1130qC)12,
growth of large graphite plates and flakes occurs from the liquid phase.
Carbon precipitates in the form of highly ordered graphite crystals from
molten iron supersaturated with carbon. The Raman spectrum for
chlorination at 1200qC is shown in Fig. 2c. A very strong and narrow
405
(FWHM = 16 cm-1) G band at 1582 cm-1 suggests that graphite crystals are
formed under these conditions.
3.3. Powder X-ray Diffraction

Typical X-ray diffraction patterns of three different carbon powder
samples are shown in Fig. 3. Two 00l and two hk0 diffraction peaks can be
distinguished in the patterns of samples produced at 800°C and 1000°C. The
002 (2ș § 26.9°) and 004 (2ș § 54.9°) peaks correspond to the parallel
graphene layers. The 100 (2ș § 43°) and 110 (2ș § 77.8°) diffraction peaks
are characteristics of the 2D in-plane symmetry along the graphene layers.
Based on its XRD pattern, the powder synthesized at 500°C is not
graphitized, which is in agreement with Raman analysis. This low
temperature sample also contains traces of iron chlorides.
Further analysis of these XRD profiles shows that the width of the
002 peak decreases as the reaction temperature increases. Its position is
approximately the same for both 800°C and 1000°C, which indicates that the
graphitic interplanar spacing (0.334 nm) is developed in these powders at
temperatures as low as 800ºC. This is also in agreement with the TEM and
Raman analysis. As the reaction temperature increases, the 10l band also
splits into 100 and 101 peaks.
002 + FeCl2 and FeCl3

6000
5000
Intensity/cps
4000
3000
2000
10l
100 004
1000
1000 °C 11l
+ + 800 °C
0
+ + 500 °C
20 30 40 50 60 70 80
2T/degrees
Figure 3. XRD patterns of some carbon samples synthesized at different temperatures.
XRD data for selected samples are shown in Table 1. The interplanar
spacings, d002 and d004, were evaluated from the positions of the 002 and 004
peaks respectively by applying Bragg’s equation. The crystallite size Lc
along the c-axis was calculated from the 002 peak using the Sherrer formula
406
with a value of constant K=0.9.13 The crystallite size La along the basal
planes was calculated from the 100 peak, using the Warren-Bodenstein
formula with a value of K=1.77.14 La values of 45-50 nm were obtained from
the XRD pattern analysis. They agree with the values of 30-55 nm estimated
from the Raman spectroscopy using Tuinstra and Koenig’s equation.
Table 1. XRD data for the carbon samples.

Reaction
d002 (d004) [nm] FWHM002 [º] Lc [nm] La [nm]
temperature
500 ºC - - - 28.1
800 ºC 0.331 (0.334) 0.3657 22.4 45.5
1000 ºC 0.333 (0.334) 0.2991 27.3 47.7
3.4. Scanning and Transmission Electron Microscopy

The effect of the reaction temperature on CDC microstructure is
illustrated in Figs. 4 and 5. An SEM micrograph of ‘as-received’ Fe3C is
shown in Figure 4a. At lower reaction temperatures (<500qC), overall
volume of the original Fe3C grains is converted to a fluffy, disordered carbon
by formation and evaporation of volatile iron chlorides. Chlorination at mild
regimes yields a grainy (for 600°C, Fig. 4b) or striated (for 800°C, Fig. 4c)
texture suggesting formation of graphite lamellae. Conversion of Fe3C to
carbon is possible without affecting the original Fe3C grain morphology up
to a certain temperature, between 600°C and 800°C. Samples obtained by
chlorination at 1000°C show features of expanded graphite (Fig. 4d).
A B
100 Pm 50 Pm
C D
50 Pm 2 Pm
Figure 4. SEM micrographs: a.) As-received Fe3C powder, Sample chlorinated at:
b.) 600°C, c.) 800°C, and d.) 1000°C.
407
A very different carbon microstructure was achieved after

chlorination at 1200°C, as shown in Fig. 5. Several millimeter-size graphitic
spheres (one shown in Fig. 5a) were observed. Opening the spheres revealed
that they are hollow inside, and their walls are built of very thin graphite
flakes (Fig. 5b). The sphere’s inner surface was found to be overgrown with
graphite crystals. An example is shown in Fig. 5c. Figures 5d-e show the
morphology of very thin graphite layers, which are formed during the
synthesis process due to the removal of iron salts that intrinsically intercalate
graphite. However, it is important to note that neither XRD nor TEM and
Raman analysis found intercalation15 compounds present in CDCs after their
synthesis. Above the eutectic temperature in the iron-Fe3C system, the liquid
phase is formed producing melt droplets. Upon interaction with chlorine,
iron is easily removed from the iron-carbon melt that becomes
supersaturated with carbon. Planar graphite nucleates and grows from such
melt. The surface of a sample synthesized at 1200°C is shown in Fig. 5c.
A Outer B
surface
detail
Crystals
1 mm inside the 10 Pm
sphere
10 Pm C D E
10 Pm 1 Pm
Figure 5. SEM micrographs of a graphite sample

produced by chlorination of Fe3C at the temperature of 1200°C.
TEM images of CDC powders produced at different temperatures

are shown in Figs. 6-8. Similar to the SEM image in Fig. 4b, Fig. 6a shows
the agglomerated microstructure of the CDC powders produced at low
temperatures (400 - 600qC). The HRTEM image (Fig. 6b) shows that the
particles produced at 400°C are mainly built of amorphous carbon, while
minor local ordering and formation of turbostratic graphite can be found at
the edge of the particles even at this very low temperature. For comparison,
in CDC produced from other carbides, such as Ti3SiC2, graphitic carbon was
not found below 700°C.3 The measured crystallite size along graphite basal
408
planes for 400qC is on average less then 5 nm, which is in good agreement
with the La value estimated from the Raman spectra.
A B
amorphous
carbon
turbostratic
graphite
0.344 nm
50 nm
Figure 6. TEM micrographs of amorphous carbon and turbostratic graphite in a

sample produced by chlorination of Fe3C at 400°C.
Pronounced graphitization has been observed during high

temperature CDC synthesis. CDC powders produced at 800qC consist of thin
sheets of graphite (Fig. 7). Amorphous carbon is still present in the samples,
but in significantly smaller quantities (less then 10 % vol.). For synthesis at
1200qC, TEM analysis (Fig. 8) shows that perfect graphite crystals with an
interplanar distance of 0.334 nm are produced, indicating complete
graphitization.
A B
50 nm
Figure 7. TEM micrographs of thin graphite flakes

in a sample chlorinated at 800°C.
Graphite materials produced at 600-1100°C may find applications in

lithium batteries and supercapacitors. Currently, similar flakes are produced
in a complex process including graphitization at above 2500°C,16 followed
by intercalation and exfoliation of graphite15. Here we demonstrate that
synthesis of graphite from iron carbide can be done in one step at moderate
temperatures.
409
A B
0.334nm
200 nm
Figure 8. TEM micrographs of well-ordered graphite produced at 1200°C.
4. CONCLUSIONS
Chlorination of Fe3C at temperatures of 400qC and above results in the

formation of carbon and solid or gaseous iron chlorides. Three temperature
regimes have been defined. Amorphous carbon is formed at temperatures of
400 – 500qC. Flakes and ribbons of nanocrystalline graphite form at 600 –
1100qC. Graphite ordering increases with temperature. Formation of
microcrystalline graphite and millimeter-size graphite crystals occurs above
the eutectic point in the Fe/Fe3C system.
ACKNOWLEDGEMENTS
We thank Julius Chang and Jerome P. Downey for providing the
iron carbide sample. This work was supported by DARPA via ONR contract.
The purchase of the Raman spectrometer and environmental SEM was
supported by NSF grants DMR-0116645 and BES-0216343.
REFERENCES
1 A. Nikitin and Y. Gogotsi, 'Nanostructured Carbide-Derived Carbon (CDC)', ed. H. S.
Nalwa, American Scientific Publishers, 2003.
2 Y. Gogotsi, S. Welz, D. A. Ersoy, and M. J. McNallan, Nature, 2001, 411, 283.
3 Y. Gogotsi, A. Nikitin, H. Ye, W. Zhou, J. E. Fischer, B. Yi, H. C. Foley, and M. W.
Barsoum, Nature Materials, in press, 2003.
4 W. A. Mohun, Proc. 4th Biennial Conf. Carbon, Oxford, 1959, p. 443.
5 J. Leis, A. Perkson, M. Arulepp, M. Kaarik, and G. Svensson, Carbon, 2001, 39, 2043.
6 H. O. Pierson, 'Handbook of Carbon, Graphite, Diamond and Fullerenes: Properties,
Processing and Applications', ed. R. F. Bunshah, G. E. McGuire, and S. M. Rossnagel,
Noyes Publications, 1993.
7 M. Inagaki, 'New Carbons: Control of Structure and Functions', Elsevier, 2000.
8 J. Leis, A. Perkson, M. Arulepp, P. Nigu, and G. Svensson, Carbon, 2002, 40, 1559.
410
9 D. A. Ersoy, M. J. McNallan, and Y. Gogotsi, Mat. Res. Innovat., 2001, 5, 55.

10 F. Tuinstra and J. Koenig, J. Chem. Phys., 1970, 53, 1126.
11 N. S. Jacobson, Y. G. Gogotsi, and M. Yoshimura, J. Mater. Chem., 1995, 5, 595.
12 J. M. Tanzosh, in 'Piping materials', ed. M. L. Nayyar, 2000.
13 J. Biscoe and B. Warren, J. Appl. Phys., 1942, 13, 346.
14 B. Warren and P. Bodenstein, Acta Crystallogr., 1966, 5, 602.
15 A. Yoshida, Y. Hishiyama, and M. Inagaki, Carbon, 1991, 29, 1227.
16 D. Gonzalez, M. A. Montes-Moran, and A. B. Garcia, Energy & Fuels, 2003, in press.
HIGH RESOLUTION TRANSMISSION
ELECTRON MICROSCOPY IMAGE ANALYSIS
OF DISORDERED CARBONS USED FOR
ELECTROCHEMICAL STORAGE OF ENERGY
Jean-Noël Rouzaud*, Christian Clinard, Frédéric Chevallier,

Alexandre Thery and François Béguin
CRMD, CNRS-Université, 1B Rue de la Férollerie, 45071 Orléans Cedex 02, France
1. INTRODUCTION
Numerous forms of more or less disordered carbons (soft carbons,
hard carbons, fibers, nanotubes, mesophase microbeads, etc.) are presently
tested for different energy storage applications. For instance, microporous
carbons i.e. hard carbons, are frequently studied as anodes for lithium-ion
batteries, whereas microporous/mesoporous solids such as carbon areogels
seem to be optimal for supercapacitor applications, or hydrogen storage1,2. If
the mechanism of lithium intercalation between the infinite and perfect
graphene layers of a graphite crystal is well understood3, it is not the case
with disordered carbons. These carbons are made of short and poorly stacked
polyaromatic layers with a large distribution of layer length and interlayer
spacings, and, consequently various sites are available for lithium ion
trapping.
Disordered carbons usually exhibit a multiscale organization
(structure, microtexture, texture)4. Structurally, they are made of more or less
distorted polyaromatic layers, nanometric in size. The spatial association or
the layers, from the nanometric to the micrometric scales, gives rise to
different microtextures (lamellar, porous, concentric, fibrous, etc.) forming
the carbons skeleton4. The multiscale organization is the fingerprint of the
kind of precursor and of the formation conditions (temperature, pressure,
strains, time, etc.) met either in laboratory experiments or in Nature, and is
directly related with numerous properties.
As far as the diffraction data (as well electron and X-ray diffraction)
of disordered carbons are concerned, only averaged data can be obtained
from the broad and faint bands due to the reflections on the nanometer-size
*
411
412
coherent domains. Only the 002 reflection on the aromatic layers stacks is
usually studied. The average interlayer spacing d002 and the height Lc of the
coherent domains are determined from the position and peak width,
respectively. Unfortunately, neither de-averaged structural data, nor
microtextural information can generally be obtained. By contrast, High
Resolution Transmission Electron Microscopy (HRTEM) allows the
organization to be determined by imaging directly the profile of the
polyaromatic layers5. Two particular examples of disordered carbons are
shown in this paper : (i) microporous carbon fibers (Figure 1), (ii) carbon
aerogels (see Figure 8). However, up to now, the information from such
complex images is mainly qualitative, whereas quantitative structural and
microtextural data are required to precisely describe the organization, to
propose models and to foresee the properties, for example energy storage
properties.
Figure 1. HRTEM image of a microporous carbon fiber obtained from

the pyrolysis of viscose at 1000°C.
Various models were recently proposed to interpret the lithium

storage performance of disordered carbons, especially the house of cards
model from Zheng et al6 and the porous carbon made of ‘misoriented
graphitic-like crystallites’ from Han et al7. They are usually based on X-ray
techniques (diffraction and diffusion). Some structural features of these
models disagree with the HRTEM images (see Figure 1): only small and
“single” layers in the house of card model, too ordered and large “graphitic
crystallites” in the porous model. Furthermore, the models are rather
413
qualitative, providing no data on the layer lengths and interlayer spacing for
instance. Consequently, image analysis techniques are required to obtain
more quantitative data, as it was recently successfully shown for chars or
soots8. We have undertaken this challenging task by developing an in-house
HRTEM image analysis procedure allowing useful structural data to be
extracted4. The results obtained on two types of disordered carbons used for
electrochemical energy storage, e.g. a carbon-carbon composite based on an
isotropic carbon fiber coated with pyrocarbon and carbon aerogels, will be
presented in this paper.
2. HRTEM IMAGE ANALYSIS OF

DISORDERED CARBONS
The multiscale organization (structure, microtexture, texture) was
studied by HRTEM, using a Philips CM 20 microscope, operating at 200 kV.
Adequate under focusing was selected in order to be on the transfer function
plateau. Interferences between the transmitted electron beam and hkl
diffracted beams produce fringes, which directly image the profile of the hkl
planes in the Bragg angle, i.e. quasi-parallel to the incident electron beam.
The optimal resolution of 0.14 nm allows seeing the profile of the aromatic
layers on the Bragg angle, even for weakly organized carbons (in this case,
due to the reciprocal node elongation, the angle tolerance on the Bragg angle
can reach +/- 10°). Since HRTEM images are 2D projections, objects thinner
than 10 nm are required to avoid fringe superimposition. So thin sections are
almost impossible to be obtained by ultramicrotomy or ion thinning,
therefore we used classical grinding under alcohol in an agate mortar. After
sonifying of the alcoholic suspension, a droplet is deposited on a lacey
carbon grid and dried. Using the edges of the thinnest fragments deposited
across the holes minimizes risks of fringe superimposition.
Our image analysis procedure especially developed for disordered
carbons is presented in detail in the reference4. At first the images are
skeletonized as follows. 20 x 20 nm representative areas of raw HRTEM
images obtained with a 300,000 x magnification are sampled and digitalized
on 1024 x 1024 pixels with a resolution of 4000 ppi; one pixel then
corresponds to 0.019 nm. The first step of the image analysis process is to
obtain a homogeneous image without background noise. Specific software
modules added to the Visolog (Noesis) commercial software were developed
in order to filter the image at the Fourier transform (FT) level. The ‘clean’
resulting image, always characterized by 256 gray levels reproducing the
intensity of each pixel, is then threshold and binarised. Finally, the binary
images are skeletonized giving one-pixel width lines for fringes. The
skeletonized pixel-based image is transformed into a vectorial image, where
the points coordinates determine each fringe which can be analyzed
414
individually, or/and in relation with its neighbours. Some criteria have to be

taken into account to avoid artefacts: (1) since the size of each aromatic ring
is 0.246 nm, all the fringes shorter than this value have to be eliminated; (2)
a fringe crossed by segments smaller than the C-C bond length, i.e. 0.142
nm, will be considered rectilinear; (3) the van der Waals interaction is
supposed to be negligible between graphene layers spaced by more than 0.7
nm (twice the d002 value of graphite); consequently, pairs of more spaced
fringes and/or forming an angle larger than 15° are considered non-stacked
and independent.
a b
Figure 2. HRTEM image analysis. a: raw HRTEM image; b: corresponding skeletonized

image; c: limits of a coherent domain, definition of the structural (L, La, Lc, d) and
microtextural (D̛̓ parameters; N is the number of layers stacked within a domain.
Usually, at least several hundreds of fringes are measured and

characterized. Taking the classical example of X-ray diffraction, coherent
domains are defined by the stacks of polyaromatic layers (Figure 2). The
coherent domains are distinguished from the single layers and their relative
415
abundance is determined. They are characterized by their diameter La and

their height Lc along c axis, the number N of stacked layers and the
interlayer spacings d. Whereas X-ray diffraction only gives access to
averaged data, the distribution of each structural parameter can be
represented by bar diagrams and average values can be calculated.
3.1 Application to the Study of Carbon Anodes for

Lithium-Ion Batteries
Most of the commercially available lithium-ion batteries are based
on a graphite negative electrode. In standard conditions, the reversible
capacity of graphite is limited to 372 mAh/g, that corresponds to the
formation of the saturated 1st stage intercalation compound LiC63. Higher
reversible capacities were demonstrated by disordered carbons giving up to
twice the reversible capacity of graphite6. However, these carbons are
characterised by an important irreversible capacity due to the electrolyte
decomposition on the carbon surface and a polarization between charge and
discharge, so-called hysteresis. Recently, few papers claimed that the extent
of electrolyte decomposition is efficiently depressed by a pyrolytic carbon
coating9-13.
For our experiments, a carbon fiber cloth (Figure 3a) was prepared
by carbonization of viscose under neutral atmosphere for 15 minutes,
successively at 400, 700 and 1000°C. The carbon cloth was coated with
pyrolytic carbon, using chemical vapor decomposition of propylene (2.5
ml/mn) diluted in nitrogen (100 ml/mn) during 10 minutes at 900°C. The
resulting composite carbon material exhibits a very low irreversible capacity
and 1.5 times the reversible capacity of graphite9-11.
100 µm a b
50 µm
Figure 3. Images of a cross-section of carbon fibers after propylene pyrolysis. 3a : Scanning
Electron Microscopy of a piece of the carbon cloth. 3b : optical microscopy (crossed
polarizers with a wave retarding plate).
416
Figure 3 shows the images of a cross-section obtained by

ultramicrotomy of the coated fibers. Figure 3b shows an optical microscopy
image obtained in reflexion mode, between crossed polarizers and with a
wave retarding plate; it gives textural information (micrometric scale). The
fibers, which are optically isotropic at this scale, show a characteristic
magenta tint, whereas the yellow and blue tints indicate an anisotropic
pyrocarbon coating (due to a preferential orientation of the aromatic layers
along the fiber axis).
Figure 3c. HRTEM image of a cross-section of a carbon fiber after propylene pyrolysis: the
black line represents the boarder between the lamellar pyrocarbon (at the top) and the
microporous fiber (at the bottom).
HRTEM allows to image directly the microtexture and the structure

of each component of the fiber-pyrocarbon composite (Figure 3c). The
aromatic layers of the fiber core are highly misoriented and frequently non-
stacked, as shown by the lower part of the image; this explains the isotropy
observed by optical microscopy. By contrast, the layers of the pyrocarbon
coating are more frequently stacked in Basic Structural Units (BSUs), as
shown on the upper part of the image. The BSUs are rather parallel to the
external surface of the fiber, being responsible for the lamellar microtexture
and the anisotropic rim in optical microscopy; the coating thickness is
417
estimated here to about 20 nm. Image analysis indicates that the amount of
non-stacked layers is 40% and 30% for the fiber and pyrocarbon,
respectively. Furthermore, about 90% of the BSUs in the fiber are only
formed of 2 or 3-stacked layers, whereas more than 25 % of the BSUs in the
coating contain more than 3 layers. In the histograms of interlayer spacings,
a well visible maximum is found at about 0.4 nm for pyrocarbon, whereas
the values range mainly from 0.4 to 0.6 nm for the fiber (Figure 4). Hence,
the microtexture of the pyrocarbon is denser than that of the fiber.
45
Pyrocarbon
40
Fibre
35
Percentage
30
25
20
15
10
0
0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75
Graphite Interlayer distance (nm)
Figure 4. Histogram of the interlayer spacings d obtained by computer analysis of the

HRTEM image of a cross-section of a carbon fiber coated by pyrocarbon.
45%
Pyrocarbon
40%
Fibre
35%
Fringes %
30%
25%
20%
15%
10%
5%
0%
0.25 1.25 2.25 3.25 4.25 5.25 6.25 7.25 8.25
Fringes length L(nm)
Figure 5. Histogram of the fringe length L obtained by computer analysis of the HRTEM
image of a cross-section of a carbon fiber coated by pyrocarbon.
418
The histogram of Figure 5 shows that the layers of both kinds of

carbon are short, especially in the case of the fiber where about 75 % of the
layers are smaller or equal to 0.75 nm, i.e. to 3 cycles. By contrast, this
proportion is only about 55 % for the pyrocarbon.
The tortuosity of each fringe (and consequently of the layer) was
estimated with a structural parameter called the “order degree”. The layers
forming the fiber are more distorted than the layers of the pyrocarbon
coating. It must be noted that, even if the distorted layers appear
discontinuous by TEM (some parts can be out of the Bragg angle), they are
in fact continuous, as proved by the total insolubility of these carbons.
Continuous layers are formed by small pieces of flat layers connected each
other by defects (pentagons, tetrahedral carbons, etc.), which limiting their
extent. The higher the density of defects, the smaller the order degree and the
extent of planar layers. In fact, the image analysis shows that the shortest
layers are usually characterized by the highest interlayer spacings (>0.4 nm).
It is important to underline that large interlayer spacings, generally formed
by more or less oblique pairs of layers, are not easily detected by XRD
(broad and faint corresponding 002 reflections not be easily distinguishable
from the background noise). Two extreme cases are observed in these
disordered carbons: (i) large stacks of high order degree with a noticeable
number (> 3) of long (> 2 nm) and planar layers; and (ii) pairs of short (< 1
nm) and misoriented layers, the latter forming slit-shaped pores. Both fiber
and pyrocarbon contain the two types of stacks, but large stacks are rather
present only in the pyrocarbon coating, whereas the slit-shape pores are
especially frequent in the fiber.
Based on these HRTEM data, a sketch of the multiscale organisation
of the composite is proposed13 in Figure 6. Even if this representation is very
schematic, it illustrates the two types of carbon: short and very disordered
layers mainly in the fiber, and a developed preferential orientation of longer
and better stacked layers in the pyrocarbon coating.
1 nm
____
Figure 6. Sketch of the cross-section of a carbon fiber coated by pyrocarbon, taking into
account the HRTEM image analysis data.
419
3.2. Consequences on the Electrochemical Properties

The kind of model presented in Figure 6 allows the electrochemical
characteristics of the carbon-carbon composite used as anodes of lithium-ion
battery to be better understood. First, let us try to interpret the high
irreversible capacity of the carbon fiber and its decrease after coating by
pyrocarbon. The active sites responsible for the electrolyte decomposition
during the first reduction process correspond to the edge planes and the
defects developed in the graphene planes. Such defects are numerous in the
as-received fiber, therefore the irreversible capacity is important. By
contrast, the electrochemical characterization11 of the composite shows that
the passivation layer is still formed, but is less developed. In fact, the
pyrocarbon coating, which is directly exposed to the electrolyte, contains
much less defects and, due to its lamellar microtexture, only few edge planes
are accessible for adsorbed species. The pyrocarbon coating forms a barrier
preventing the diffusion of the large solvated lithium ions to the active sites
of the fibers that results in a low value of irreversible capacity. This
observation agrees well with the CO2 adsorption measurements, which show
that the micropore surface area strongly decreases from 480 to 60 m2/g after
pyrocarbon deposition.
As far as the reversible capacity is concerned, in situ 7Li NMR
experiments12-13 reveal the presence of 2 sites for the lithium trapping in this
composite: “truly” intercalated lithium with a characteristic Knight shift of
18 ppm, and “quasi-metallic” lithium clusters characterized by a shift
increasing up to 110 ppm. The latter could correspond to lithium trapped in
the nanometer-sized voids between misoriented BSUs. However, the
HRTEM image analysis shows that the proportion of these kinds of
micropores do not represent more than 10 % of the fiber image, i.e. too
limited to explain the reversible capacity 50% higher than the graphite one.
Consequently, we assume that the “quasi-metallic” lithium is trapped in
subnanometric bidimensional clusters between the pairs of spaced graphene
layers forming slit-shaped pores.
Reminding that image analysis revealed the presence of two extreme
types of stacking: (i) large stacks of long and planar graphene layers, (ii)
pairs of short and misoriented layers forming slit-shaped pores, let us
consider the electron affinity of the aromatic layers adjacent to inserted
lithium. As reference values, the electron affinity is 0 eV for benzene, 1.8 eV
for anthracene (3 condensed rings - molecule length: about 0.75 nm) and
reaches 4.39 eV in the case of graphite. Consequently, lithium can
intercalate, with a charge transfer to carbon, in the well-ordered stacks of
long layers (interlayer spacing smaller than 0.4 nm) presenting a high
electron affinity. By contrast, the electron affinity of the “slit-shaped pores”,
with short graphene layers, is very low. Hence, lithium located in these
420
rather large interlayer spacings (> 0.4 nm) undergoes a very weak charge
transfer, being in a state close to metal which we call “quasi-metallic”.
Coupling HRTEM image analysis, in situ 7Li NMR experiments and
galvanostatic cycling of lithium/composite carbon cells, we propose the
“Corridor model” shown in Figure 712-13. During reduction, lithium
intercalates at first in the smallest intervals between the nanometer size
graphitic type layers, giving a maximum shift of 18 ppm. Then, lithium
diffuses into the slit-shaped pores where the ions are screened by delocalized
electrons through a backdonation process, forming the quasi-metallic
clusters with a shift progressively increasing up to 110 ppm. During the
galvanostatic oxidation, the reverse process occurs, with a complete
disappearance of the two NMR peaks. When oxidation is fully completed,
there is only one peak at ca. 0 ppm characteristic of the Solid Electrolyte
Interface (SEI).
Figure 7. Schematic model based on the TEM image analysis and on in situ 7Li-NMR during
galvanostatic reduction/oxidation of the carbon composite. During insertion, ionic lithium
penetrates at first in the smallest interlayer spacings, then it diffuses in the slit-shaped pores
where quasi-metallic clusters are formed.
In conclusion, the composite formed by a carbon fiber cloth coated

with pyrocarbon is a very promising material for the negative electrode of
lithium batteries, since it couples a high reversible capacity due to the
disordered fiber (1.5 time the graphite value) with a small irreversible
421
capacity due to the dense pyrocarbon coating which blocks the diffusion of
solvated lithium to the active sites of the fiber.
3.2. Characterization of Carbon Aerogels
New forms of carbon, such as carbon aerogels, were designed in

order to get materials with a controlled pore size distribution adapted for
galvanostatic energy storage in aqueous medium14. In a first step, organic
aerogels are synsethised from phenolic resins using a soft chemistry
approach15-17. Then, they are pyrolysed under an inert atmosphere to form
carbon aerogels18-19. Different organic aerogels can be obtained over an
enough large density range (from 0.05 to 0.3) by playing on the precursor
(resorcinol, phloroglucinol) and the nature and amount of catalyst (Na2CO3,
Ca(OH)2, etc.). Moreover, using the same organic aerogel, but changing the
pyrolysis temperature (between 800 and 2600°C) allows different carbon
aerogels to be obtained. The multiscale organization of the carbon aerogels
was directly imaged by HRTEM over 3 orders of magnitude (µm – nm). The
example of the carbon aerogel obtained at 1050°C from a Ca(OH)2 based
organic aerogel (Resorcinol/Catalyst = 45) is shown in Figure 8.
Figure 8. Multiscale organization of carbon aerogels imaged by TEM. Example of the

Ca(OH)2 based organic aerogel (Resorcinol/Catalyst = 45) heated at 1050°C.
At the textural scale, this carbon is constituted of the arrangement of

round-shaped nodules, about 20 nm in diameter; mesopores are formed by
the voids between the nodules. The nodules are made of more or less
distorted layers, with numerous non-stacked layers. The HRTEM images
suggest that slit-shaped pores can be formed by pairs of spaced polyaromatic
structures, whereas the misorientation of these units could be responsible for
a small mesoporosity. The final density of the carbon aerogel depends on the
antagonistic effects of weight loss and shrinkage, both increasing with the
highest temperature of treatment (HTT). The micropore and mesopore
volumes were determined from the gas adsorption data. As the carbon
aerogel density increases, the micropore volume determined by CO2
422
adsorption at 293K (giving pore size smaller than 0.7 nm) decreases,
whereas the mesopore volume (pore size larger than 20 nm) determined from
the N2 desorption isotherm at 77K increases.
After heat-treatment at 2600°C, the mesopores become clearly
visible, with walls formed of few well-staked layers (2-3 for Na2CO3 based
carbon aerogels, and 3-5 for the Ca(OH)2 based ones). As for carbon blacks,
the diameter of the closed pores obtained above 2000°C is strongly
determined by the initial size of the nodules5 in the low-temperature carbon
aerogels. The following table reports structural data determined by HRTEM
image analysis on carbon aerogels obtained at 1050°C and 2600°C from
Ca(OH)2- and Na2(CO)3-based organic aerogels.
Table. Structural data from computer analysis of the HRTEM images obtained on carbon
aerogels formed by pyrolysis at 1050 and 2600°C of Na2(CO)3
and Ca(OH)2 based aerogels heated.
Catalyst, pyrolysis L (nm) % nsl* N d (nm)
temperature
Ca(OH)2, 1050°C 0.7 49 2.4 0.47
Na2(CO)3, 1050°C 0.7 56 2.4 0.51
Ca(OH)2, 2600°C 1.3 33 2.8 0.41
Na2(CO)3, 2600°C 0.8 42 2.5 0.45
* nsl = non stacked layers
Both carbons formed at 1050°C are highly disordered: about half of

the layers are non-stacked; the layer extent is small (about 0.7 nm) and the
interlayer spacing important (about 0.5 nm). As far as the catalyst effect is
concerned, the Na2CO3 based carbon aerogels are slightly more disordered,
as shown by higher amounts of non stacked layers and larger interlayer
spacing. This confers a strong microporous character to this aerogel with
numerous and well-defined slit-shaped pores. As expected, the heat-
treatment at 2600°C leads to a noticeable structural improvement: the layer
length increases, the proportion of stacked layers increases and the interlayer
spacing decreases; consequently the micropore amount strongly decreases.
All these features are especially verified in the case of the Ca(OH)2 based
carbon aerogel.
Attempts of galvanostatic cycling have been performed on carbon
aerogels obtained at 1050°C, 1300°C, 1700°C, 2600°C in a potential range
from 0 to –1.5 V vs NHE. Two effects are observed successively upon
decreasing potential: the formation of the electrochemical double layer and
water electrolysis at potential lower than –0.9 V vs NHE with hydrogen
storage in the micropores14. The capacitive contribution decreases with the
HTT. Multi-sites occupation of hydrogen is observed for the low-
temperature carbons and a single peak in the voltammetry curve of the
2600°C treated sample seems to indicate a phenomenon close to hydrogen
423
intercalation. About 1% of hydrogen can be stored then released from these

porous materials. As expected, the values are slightly higher in the case of
the most disordered material, i.e. the Na2CO3 carbon aerogels.
4. CONCLUSIONS
Our procedure of computerized HRTEM images analysis allows
quantitative and de-averaged structural data on carbons to be obtained. Such
data give the solid experimental basis required to build reliable models of
organization. Moreover, the relationships between formation conditions -
carbon material organization and properties can thus be specified. Such
approach was successfully applied to two types of disordered carbons
presently tested for the electrochemical storage of energy: a carbon-carbon
composite material (a microporous carbon fiber-based cloth coated with a
pyrocarbon) used as anode for lithium-ion battery, and carbon aerogels with
a controlled pore size distribution (micropores and mesopores) adapted for
galvanostatic energy storage in aqueous medium. In the first case, coupling
in situ 7Li-NMR- potentiometry-HRTEM image analysis leads to a relevant
model of lithium insertion and de-insertion in these anodes (‘Corridor
model’). For the second case, different ways of energy storage in aqueous
medium are presently tested on the designed aerogels; in the future, the
storage capacity will be discussed taking to account the data on their
multiscale organization quantified by HRTEM image analysis.
REFERENCES
2. Frackowiak E., Béguin F. Electrochemical storage of energy in carbon nanotubes and
nanostructured carbons. Carbon 2002; 40:1775-87.
3. Guérard D., Hérold A. Intercalation of lithium into graphite and other carbons. Carbon
1975; 13:337-45.
4. Rouzaud J.N. and Clinard C. Quantitative high-resolution transmission electron
microscopy: a promising tool for carbon materials characterization. Fuel Proces.
Technol. 2002; 77-78:229-35.
5. Oberlin, Agnès. “High Resolution TEM studies of carbonization and graphitization”. In
Chemistry and physics of carbon, Vol 22, Thrower P.A. (Ed.), Marcel Dekker, New-
York, 1989, p. 1-143.
6. Zheng T, Dahn JR “Applications of carbon in Lithium-ion batteries”. In Carbon
materials for advanced technologies, Burchell T.D. Ed., Elsevier, Oxford, 1999, pp 341-
88
7. Han Y.S., Yu J.S., Park G.S., and Lee J.Y. Effects of synthesis temperature on the
electrochemical characteristics of pyrolytic carbon for anodes lithium-ion secondary
batteries. J. Electrochem. Soc. 1999; 146:3999-4004.
8. Shim H.S., Hurt R.H. and Yang N.Y.C. A methodology for analysis of 002 lattice fringe
images and its application to combustion-derived carbons. Carbon 2000; 38:29-45.
424
9. Chevallier F., Gautier S., Salvetat J.P., Clinard C., Frackowiack E., Rouzaud J.N. and
Béguin F. Effects of post-treatments on the performance of hard carbons in lithium cells.
Journal of Power Sources 2001; 97-98:143-5.
10. Yoon S., Kim H. and Oh S.M. J. Power Sources. 2001, 94.
11. Chevallier F., Vix C., Saadallah S., Rouzaud J.N., Frackowiak E. and Béguin F. A better
understanding of the irreversible lithium insertion mechanisms in disordered carbons, J.
Phys. Chem. Solids 2004; 65:211
12. Letellier M., Chevallier F., Clinard C., Frackowiak E., Rouzaud J.N., Béguin F. The first
in situ 7Li nuclear magnetic resonance study of lithium insertion in hard-carbon anode
materials for Li-ion batteries, J. Chem. Phys. 2003; 118:6038-45
13. Chevallier F., Letellier M., Morcrette M., Tarascon J.M., Frackowiak E., Rouzaud J.N.,
Béguin F. In situ 7Li-Nuclear Magnetic Resonance Observation of Reversible Lithium
Insertion into Disordered Carbons, Electrochem. Solid State Lett. 2003; 6:A225-8.
14. Jurewicz K., Frackowiak E., Beguin F. Towards the mechanism of electrochemical
hydrogen storage in nanostructured carbon materials. Appl. Phys. A, in press (2004).
15. Pekala RW. Organic aerogels from the polycondensation of resorcinol with
formaldehyde J. Mat. Sci. 1989; 3221-7.
16. Pekala R.W., Alviso C.T., Kong F.M., Hulsey S.S. Aerogels derived from
multifunctional organic monomers. J. Nn-Cryst Solids 1992; 145:90-8.
17. Pekala RW and F.M. Kong FM. A synthetic route to organic aerogels-mechanism,
structure, and properties. Revue de physique appliquée. ColloqueC4, tome 24, avril
1989.
18. Pekala R.W., Alviso C.T. Carbon aerogels and xerogels. Mater Res Soc Symp Proc
1992; 270-3.
19. Tamon H, Ishizaka H, Mikami M, Okazaki M. Porous structure of organic and carbone
aerogels synthesized by sol-gel polycondensation of resorcinol with formaldehyde.
Carbon 1997; 35:791-6
ELECTROLYSIS OF CARBAMIDE-CHLORIDE
MELTS AT INERT ELECTRODES
S.A. Kochetova and N.Kh. Tumanova*

Institute of General and Inorganic Chemistry, National Academy of Sciences of Ukraine,
Pr. Palladina 32/34 Kiev, 03680, Ukraine.
Abstract
Electrode reactions in carbamide and carbamide-chloride melts have

been studied by cyclic voltammetry in combination with chromatography
and IR spectroscopy of melts and gases for determining the electrolysis
products. The interest in this process is accounted for by application of these
melts as supporting electrolytes for electrochemical investigations. The
peculiarities of carbamide-NH4Cl melt electrochemical behavior have been
established. A new compound is formed during electrolyses at the potential
of 1V, which is present in melt and in anodic gases (IR Spectroscopy). The
investigation of electrochemical properties of this compound showed that it
is electrochemically active (cyclic voltammetry) and is adsorbed (C-E
method) on inert electrodes (Pt, glassy carbon). The formation of this
compound, characterized by N-Cl bond, explains the depassivation effect of
refractory metals under anodic dissolution if polarizable electrode reaches
the transpassive region.
Keywords
Electrolysis; cyclic voltammetry; molten salts; carbamide; potassium

chloride; ammonium chloride; compound; adsorbtion.
1. INTRODUCTION
Molten carbamide is known to be a good solvent for the salts of
many metals [1] and was used as a supporting electrolyte in polarographic
investigations both at a dropping mercury electrode and at a stationary
*
425
426
platinum electrode [2,3]. Because of the low electrical conductivity of

molten carbamide, mixed carbamide-nitrate [4] and carbamide-halide [5]
melts were used more often. When investigating the electrochemical
behavior of refractory metals (Nb, Ta, Ti) in a carbamide-chloride melt, it
was found that these metals are passivated in chloride-containing melts, and
that their electrochemical dissolution (Nb, Ta) is impossible if electrode
polarization does not reach 1V. It was suggested [6,7] that during melt
electrolysis at the above potential, some compound is formed, which
functions as depassivator. To confirm this, a series of investigations of the
electrochemical properties of a melt – solvent at inert electrodes has been
carried out, and the products of electrode reactions have been investigated by
the spectroscopic and chromatographic methods. This article presents the
results of the investigation carried out.
2. EXPERIMENTAL PROCEDURES
In order to determine the electrochemical properties of the solvent,
the electrode process in molten carbamide and in carbamide-MeCl (where
Me - NH4, K) mixtures on inert electrodes (platinum, glassy carbon) were
investigated using cyclic voltammetry. The electrode reaction products were
analysed by spectroscopic methods. The adsorbtion of carbamide- NH4Cl
anodic product was investigated by differential capacity method.
The electrochemical study was conducted by recording cyclic
voltamperic curves with a PI-50-1 potentiostat and C-E curves were taken
with Resistance Bridge P-5021. A platinum electrode with a working surface
area of 5-8x10-3 cm2 was used as a working electrode. The auxiliary
electrode was platinum plate and for obtaining the differential capacity
curves – the cylinder, coaxial located to cathode. The potential was
measured against a silver reference electrode (Ag/carbamide + NH4NO3 (16
mol.%) + 0.07 mol kg–1 AgNO3). Experiments were carried out in the
temperature range 100 to 1400C under argon.
The state of carbamide melts taken in “quenched” melts were
investigated by IR spectroscopy with UR-20 and Specord M-80. The validity
of extending data for “quenched” melt to the process in melt has been proved
for a number of systems in work. The investigations of the gas phase
composition were carried out on a SELMIKHROM-1 chromatograph of the
SELMI joint-stock company, which consisted of two units: analysis unit and
control-monitoring unit. Control of the chromatograph, obtaining and
processing of the incoming information were performed by means of an
IBM-486 computer. Two series of experiments have been carried out using
first argon and then helium as the carrier gas. The replacement of argon by
helium was due to the necessity to separate and identify hydrogen and
carbon monoxide.
427
Reagent grade (carbamide, MeCl) and spectral grade (NH4Cl) salts

and dried argon were used in the study. The salts were predried in vacuum at
800C (carbamide, ammonium chloride) and 200 to 3000C (alkali chlorides).

In an individual molten carbamide, the electrode processes are
feebly marked at melt decomposition potentials (Fig. 1) because of its low
electrical conductivity. Both electrode processes are accompanied by gas
evolution.
Figure 1. Cyclic voltamperogram of Pt (1) and

glassy carbon (2) electrodes in molten carbamide.
Figure 2. IR spectra of the “quenched” carbamide: anolyte (1) and catholyte (2).
428
A large amount of ammonia and a small amount of CO and CO2

have been found in the cathode gases of carbamide melt by means of IR
spectroscopy and gas chromatography. An intense CO2 band is also present
in IR spectra of anode gases; in addition, traces of CO and a number of
unidentified bands in the range 1500-1700 cm-1 have been found [8]. Using
chromatographic analysis of anode gases, the presence of N2 in them in melt
electrolysis has been shown. It has been found that the IR spectra of the
catholyte and anolyte are identical (Figure 2).
They have a signal from the carbonyl ion of carbamide, a signal that
corresponds to ammonium ion, and a signal that characterizes the C-N bond.
On the basis of the results obtained by us, only a rough scheme of cathode
and anode reactions can be proposed because a number of bands in the IR
spectra have not been identified:
2ɋɈ(NH2)2 o 2NH3 + NH2CN + CO + O2-. (1)
2CO(NH2)2 o CO2 + NH2CN + N2 + 6H+ + 6e. (2)
Then the overall reaction is:
8CO(NH2)2 o 3NH3 + 3CO + CO2 + N2 + 4NH2CN + 3NH4OH (3)
If it is assumed that electrolysis product may be cyanamide (as the

simplest substance in terms of IR spectroscopic data for the melt), then the
chemical instability of NH2CN and the possibility of the reversible reaction:
NH2CN + 2NH3 + CO2 o 2CO(NH2)2, (4)
which probably proceeds during melt electrolysis in a cell with unseparated

compartments, are worth taking into account in the analysis of the above
overall-reaction scheme.
Our investigations showed that in mixed melts of eutectic
composition carbamide-NH4(K)Cl, the oxidation and reduction of melt
constituents take place mainly independently of each other. The anodic
process at platinum electrodes in the range of potentials below 0.9V is
associated with the direct oxidation of carbamide to secondary and tertiary
amide compounds, accumulation of ammonium ions in the melt, and
evolution of the same gaseous products as in carbamide electrolysis [8]. The
cathodic process is accompanied by the formation of ammonia, CO, and
CO2, i.e. of the same products as in pure- carbamide electrolysis. In contrast
to carbamide melt, a large amount of hydrogen appears in the cathode gases
of the mixed melt, and in the anode gases of the carbamide-KCl melt, the
presence of chlorine has been established at potentials above 0.9V. In the
429
carbamide-NH4Cl melt, a new compound appears instead of chlorine, which

is characterized by an intense signal at 630 cm-1 in the IR spectra of anode
gases and, to a smaller extent, in the IR spectra of the anolyte. This signal
corresponds to the N-Cl bond. At first it was suggested that the substance
that is formed is NCl3. By analogy with aqueous solutions [9], the formation
of NCl3 may take place according to the scheme:
NH4Cl + 3Cl2 = NCl3 + 4HCl. (5)
However, the physicochemical properties of nitrogen chloride give

no grounds for the explanation of its stability under melt electrolysis
conditions. The possibility of the formation of a chlorine derivative of
carbamide according to the scheme:
2NH4Cl + Cl2 + CO o CO(NHCl)2 + 6HCl, (6)
appears to be more probable. The evolution of HCl is supported by

the IR spectroscopy of the anode gases of the carbamide-NH4Cl melt and is
observed visually after long electrolysis.
Figure 3. Cyclic voltamperogram of Pt-electrode in carbamide-NH4Cl (curve 1) and in

carbamide-KCl (curve 2) melts.
430
However, the compounds formed by the electrolysis of the

carbamide-NH4Cl melt exhibit a strong depassivating action, which is much
stronger than that of Cl- (this was proved when studying the electrochemical
behavior of titanium in this melt) [10]. Therefore, it appears not very correct
to consider that the chlorine derivatives of carbamide can have such
properties. If one bears in mind that cyanamide is formed by carbamide
electrolysis and taking into account that the formation of the compound is
only observed if the melt contains ammonium ions, then the following
reaction is possible at the anodic discharge potential:
NH2CN + 4NH4Cl + 2CO2 = 3NH2CHNCl + HCl + 4H2O. (7)
Chlorine derivative of cyanamide possesses a high reactivity and can

show depassivating properties, but it is difficult to unambiguously identify
the compound that is formed because the presence of all possible compounds
in the melt will be reflected by the same signal at 630 cm-1 [11].
Figure 4. Differential capacity of Pt electrode in carbamide-NH4Cl melt (curve 1) and

after 0.5 ( curve 2), 1 (curve 3), 1.5 (curve 4), 2 (curve 5) hours of electrolysis.
The compound formed is electroactive. The cyclic voltammogram of

a Pt electrode in the carbamide-NH4Cl melt exhibits a maximum (wave A) in
the region of the zero value of potential on the reverse anodic sweep, which
reflects the electrochemical reduction of the compound formed (Fig. 3). For
431
comparison, a voltammetric curve is shown in the same figure, which was

obtained in the carbamide-KCl melt. The absence of maxima on the reverse
sweep supports the conclusion that maximum A does not relate to the
reduction of Cl- ions formed by oxidation but reflects the reduction of a new
compound, which is formed at potential of 1V in the carbamide-NH4Cl melt.
Adsorption of the compound formed on the electrode surface has
been established by the examination of differential capacity curve [12]. The
investigation carried out is of qualitative character because the concentration
of the compound is unknown: it is formed during electrolysis. As is evident
from the examination of C-E curves (Figure 4), the compound is adsorbed on
the electrode and decreases the double-layer capacity at the zero-charge
point. The degree of coverage of the electrode surface by the adsorbed
compound (T) was calculated taking as a basis the fact that the time of
electrolysis at the compound formation potential (1V) is proportional to the
concentration of the compound. The calculated T value are 0,61; 0,76; 0,84
and 0,95, which correspond to 0.5, 1, 1.5 and 2 hours of electrolysis. The T
value are high, i.e. the electrode surface is almost completely screened by
adsorbate molecules.
Based upon the concepts of the adsorption of the anode reaction
product, the share of the anodic curve, on which the carbamide oxidation
processes is reflected as a wave, can be explained. It may be assumed that
the adsorption of the reaction product inhibits the direct oxidation of
carbamide. To verify this conclusion, the anode was polarized to the
electrolysis product formation potential, and the reverse sweep was stopped
before the electrolysis product was reduced at the electrode. Then the
carbamide oxidation process was completely inhibited on the subsequent
forward sweep, and the curve exhibited only a current increase at the
chlorine ion oxidation potential.
4. CONCLUSIONS
In an individual molten carbamide, the electrode processes are
feebly marked at melt decomposition potentials because of its low electrical
conductivity. Both electrode processes are accompanied by gas evolution
(NH3, CO, CO2, N2) and NH2CN (approximately) is formed in melt. In
eutectic carbamide-chloride melts electrode processes take place mainly
independently of each other. The chlorine must evolve at the anode during
the electrolysis of carbamide - alkali metal and ammonium chloride melts,
which were revealed in the electrolysis of the carbamide-KCl melt. But in
the case of simultaneous oxidation of carbamide and NH4Cl, however, a new
compound containing N-Cl bond has been found in anode gases instead of
chlorine. It is difficult to fully identify this compound by the experimental
methods employed in the present work, but it can be definitely stated that
432
this compound is formed in the carbamide- NH4Cl melt (IR spectroscopy of

melts and gases), is electrochemically active (cyclic voltammetry), and is
adsorbed on the electrode surface (C-E curves). This compound possesses a
high reactivity and exhibits a strong depassivation active in the time of
electrodissolution of refractory metals in carbamide- NH4Cl melt.
REFERENCES
1. Burk W. Geschmozenez Harnstoff als Losungsmi ttel fur Alkalichloride. Chemie, 1964;
3: 354-5.
2. Vecchi E., Zuliani G. Molting Carbamide Supporting Electrolyte in Polarography. Ric.
Sci., 1955; 25: 2667-69.
3. Delimarskii Y.K., Tumanova, N.Ch., Shilina G.V., Barchuk L.P. Polarography of Ionic
Melts. Kiev: Naykova Dumka, 1978 (in Russian).
4. Tkalenko D.A. Electrochemistry of nitrate melts. Kiev: Naykova Dumka, 1983 (in
Russian).
5. Tumanova N. Electrochemistry of Nb, Ta and Ti in carbamide based melts. Proceedings
of the Green Industrial Applications of Ionic Liquids. NATO Science Series. Ed. R.
Rogers, K. Seddon, S. Volkov. Kluwer Academic Publishers, 2003.
6. Tumanova N., Volkov S., Kochetova S., Triphonova T., Buryak N. Passivation and red-
ox processes of refractory metals in carbamide-chloride melt. J. Mining and Metallurgy,
2003; 39: 69-81.
7. Tumanova N., Babenkov E., Boyko O., Buryak N. The peculiarities of electrochemical
niobium dissolution in carbamide based low-temperature melts. Abstracts of the 195-th
Meeting of Electrochem. Society, 1999 May 2-6.
8. Tumanova N., Babenkov E., Chernuhin S. and others. Electrochemical properties of
carbamide-halide melts. Ukr. Chim. J., 2000; 66: 35-9 (in Russian).
9. Comprehensive inorganic chemistry. Ed. Bailar J.C. Oxford - N. Y.: Pergamon Press,
1973.
10. Volkov S., Tumanova N., Kochetova S., Buryak N. The polysurface Mechanism of Ta
and Ti Anodic Dissolution in Low-Temperature Carbamide-NH4Cl Melt. Z. Naturforsh,
2001;56a: 761-3.
11. Nakamoto K. Infrared and Raman Spectra of Inorganic and Coordination Compounds,
New York: Wiley, 1986.
12. Damaskin B., Petriy O. Adsorption of organic compounds on electrodes. M.: Science,
1968 (in Russian).
GRAPHITE INTERCALATION AS A WAY TO
CARBON-CARBON COMPOSITES AND
CARBON NANOSCROLLS
M.V. Savoskin*, A.P. Yaroshenko, V.N. Mochalin*,

N.I. Lazareva, T.E. Konstantinova
L.M. Litvinenko Institute of Physical Organic and Coal Chemistry,
National Academy of Sciences of Ukraine, 70 R.Luxemburg St., 83114 Donetsk, Ukraine
Abstract
New simple methods for synthesis of carbon-carbon, as well as

carbon-carbon-TiO2 composites and carbon nanoscrolls were proposed
making use of readily available graphite intercalation compounds.
Keywords
Carbon composites; Anatase; TiO2; Nanoscrolls; Exfoliated graphite.
1. INTRODUCTION
Exfoliated graphite – amorphous carbon (EG-AC) composites are

widely used in a broad range of applications like adsorption of gases and
liquids, separation, and high-surface catalyst supports. They also seem to be
attractive for energy- and gas- storage systems.
There are different methods of producing such materials available
today. Greinke and Howard1 have described a method of production of EG-
AC composites starting from exfoliated graphite preformed into some article,
for example a ring. Such article is impregnated by liquid polymer or polymer
solution in organics, followed by heating in order to cure polymer or
evaporate organic solvent. The semi-finished product is then carbonized to
form an amorphous carbon phase by heating in a nitrogen atmosphere at a
*
Corresponding authors. E-mails: [email protected], [email protected]
433
434
rate of 60qC/hour from 50 to 520qC with a one-hour hold time at 520qC.

Finally the article is compressed yielding the end product.
Another method of producing EG-AC composites has been proposed
recently by Mareche et al.2 These methods consists of the following
stages: 1) obtaining EG by heating the exfoliatable graphite intercalation
compound (GIC); 2) multiple impregnation of EG particles by an amorphous
carbon precursor - polyfurfuryl alcohol; 3) drying at a relatively low
temperature (of about 100qC); 4) carbonization of material obtained at the
heating rate of 5qC/min with two-hours hold time at final temperature of
520qC; 5) activation of EG-AC composite obtained by water steam at 800qC.
As could be seen from the above, the modern methods of synthesis
of EG-AC composites are multistage long-term procedures requiring as a
rule the use of some organic solvents or inert-gas atmosphere which makes
them difficult to be scaled up by the industry.
From this point of view a one-stage process of simultaneous GIC
exfoliation, impregnation of EG by amorphous carbon precursor and its
subsequent carbonization would be very attractive. But up to now this simple
one-stage process could not be realized because of high exfoliation
temperature of the ordinary GICs needed (900 – 1200qC), which leads to
almost complete burnout of AC precursor at the atmospheric air conditions.
In our laboratory we have elaborated the method of producing
residual graphite hydrosulfate, which could be exfoliated efficiently at
temperatures of about 500qC3. Having such GIC we tried to realize one-stage
procedure of producing EG-AC composites.
2. EXPERIMENTAL
We used residual graphite hydrosulfate produced from GSM-1

graphite (which is a purified natural flake graphite from Zavalye Graphite
deposit, Kirovograd region, Ukraine) and sucrose from Merck as an AC
precursor. Sucrose was ground into fine powder and mixed with residual
graphite hydrosulfate in a glass reactor for 10 min. As a function of
experiment, the content of components in the mixture varied, however, the
total weight of mixture was kept constant (25g). After being mixed, a portion
of about 3g of GIC – sucrose mixture was heated in a Fisher Scientific
Isotemp Model 650 Programmable Muffle Furnace at 550qC during 4 min.
The conditions of heat treatment, particularly temperature and heating time,
strongly influence the properties of the product. In special experiments we
have found the optimal parameters of heat treatment mentioned above.
Heating the mixture of GIC with amorphous carbon precursor results in
exfoliation of residual graphite hydrosulfate accompanied by caramelization
of sucrose and soaking of the EG honeycomb structure by melted caramel
with simultaneous sucrose carbonization, yielding EG-AC composite. The
435
particles of final product were very similar to those of individual exfoliated

graphite in their worm-like shape.
SEM studies of composites were carried out using a JEOL JSM-
T300 scanning electron microscope. Energy dispersive X-ray (EDX)
analysis was performed by energy dispersive device LINK 860-500 attached
to the SEM unit. JEOL-JEM-200 transmission electron microscope was used
for TEM investigations.
3. DISCUSSION
Typical SEM images of EG-AC composites obtained by the one-stage
method proposed are presented in Figure 1.
<–––––––––––––––––––––––> <–––––––––––––––––––––––>
52910-6 m 18710-6 m
Figure 1. Typical SEM photographs of EG-AC composites.
These images clearly show the macroporous structure of composite

with its large pore diameter inherited from the EG matrix and the amorphous
carbon phase covering the surface of exfoliated graphite worm-like particles.
During the heating process, the AC precursor has partially fused and wet its
graphite support leading to an intimate contact between the two solids. This
structure should lead to good mechanical strength and very efficient thermal
contact between the amorphous carbon and its graphite support. Note that
these images are similar to those presented by Menard, Py et al.4,5 for
composites produced from exfoliated graphite and in situ prepared activated
carbons.
Figure 2 represents specific surface area of composites obtained as
measured by low-temperature adsorption-desorption of argon in comparison
to that calculated by the additive model. Over the entire range of sucrose
content, excluding the end points corresponding to individual components
the measured specific area of the composites obtained exceeds the values
calculated by the additive model. The higher the amorphous carbon content
436
in the composite, the more pronouncing difference has been observed. Such
behavior could be attributed to the acidic catalysis of condensation reactions
during the heating and activation of AC by the acidic gaseous species
(H2SO4, SO2, SO3 etc.) and water evaporated during the GIC exfoliation
process.
250
200
150
Ssp, m2/g
100
50
0
0.0 0.2 0.4 0.6 0.8 1.0
Mass fraction of amorphous carbon in composite
Figure 2. Specific surface area of EG-AC composites produced.

Points – experiment, straight line – calculated by the additive model.
<–––––––––––––––––––––––>
9310-6 m
Figure 3. SEM image of an EG-AC-TiO2 composite (left) and corresponding EDX

mapping of titanium (right).
One of the main advantages of the method of EG-AC composite

synthesis proposed here is a possibility of using this technique for obtaining
the tri-component composites, such as EG-AC-TiO2. Composites between
EG and photoactive anatase-type TiO2 are effective buoyant photocatalysts
for degradation of water contaminants6,7. Their synthesis is typically a
437
multistage process, which has essentially the same disadvantages described

above for the known methods of EG-AC composites production.
Therefore we tried to obtain EG-AC-TiO2 composites using the
procedure similar to that elaborated for production of EG-AC composites.
The only difference here from the scheme described above is the preparation
of initial mixture for now containing three components: residual graphite
hydrosulfate, sucrose and anatase-type TiO2 (Aldrich).
Typical SEM images of EG-AC-TiO2 composites obtained are
presented in Figure 3. The rough surface sites on the left image correspond
to the TiO2 localization sites as could be seen from comparison of SEM and
EDX images obtained from the same part of composite sample. It is clear
that in this case amorphous carbon acts not only as adsorbent but also as a
binder allowing skipping the use of additional binder needed in conventional
methods. We believe that EG-AC-TiO2 composites obtained should be very
attractive and low cost buoyant photocatalysts so their photocatalytic activity
will be investigated in the near future.
Another application of GICs in modern industry is the rapid growing
field of nanomaterials. In the recent papers of Kaner et al.8 and Shioyama
and Akita9 graphite intercalation and exfoliation were used as the key
processes in the so-called “chemical” route to the new kind of carbon
nanomaterials – carbon nanoscrolls. The method they propose differs in
advance from the known “physical” methods in comparatively low energy
consumption and low-cost simple equipment, therefore it is more convenient
to realize this technology in large-scale applications. In the study of Kaner et
al. [8] and Shioyama and Akita [9] the first stage graphite intercalation
compound with potassium (KC8) was used as the initial compound. But this
compound is expensive and difficult to obtain and operate in the large-scale
scroll-shaped skew-shaped
Figure 4. TEM images of nanoscrolls prepared from graphite nitrate.

438
applications. We succeeded in producing of nanoscrolls using the graphite

nitrate which is much more attractive from the industrial point of view. Our
procedure of nanoscrolls production could be easily transferred into
commercial large-scale application. Typical TEM images of the nanoscrolls
produced are shown in Figure 4.
On these images, one could see scrolled structures formed from one
or multiple graphitic sheets as well as non scrolled transparent graphitic
sheets in the background. The diameter of nanoscrolls produced ranges from
40 to 130 nm and length comes up to 750 nm. Based on an interlayer spacing
of graphite (3.35 Å) it could be concluded that nanoscrolls typically consist
of 10 to 40 graphene layers. These values are in good agreement with data
reported in [8-9]. According to the two possible types of scrolling process
we have found two types of nanoscrolls shown on this picture: scroll- and
skew-shaped. Both types have opened tips so their internal surface could be
accessible for small molecules.
Visual estimation of nanoscrolls yield is difficult because they
concentrate near the edges of copper grid support used in TEM studies.
Nevertheless TEM images provide clear evidence of nanoscrolls presence in
large amount.
4. SUMMARY
Graphite intercalation compounds are prospective precursors of the
new carbon materials, which are in high demand by the market today.
Combined with appropriate amorphous carbon precursors graphite
intercalation compounds could be used in one-stage process of production of
carbon-carbon composites, which could possess attractive properties for such
applications as supercapacitors elements, sorbents as well as catalyst
supports and materials for energy- and gas-storage systems.
The process of producing the new kind of carbon nanomaterials
namely carbon nanoscrolls from graphite nitrate is proposed which differs in
advance from other techniques in its simplicity, availability of raw material,
low level of energy consumption.
REFERENCES
1. US Pat. 5228701 Flexible graphite particles with an amorphous carbon phase at the
surface. Greinke R.A., Howard R.A.; UCAR Carbon Technology Corporation.
20.07.1993.
2. Mareche J.F., Begin D., Furdin G., Puricelli S., Pajak J., Albiniak A., Jasienko-Halat M.,
Siemieniewska T. Monolithic activated carbons from resin impregnated expanded
graphite. Carbon 2001; 39:771-3.
3. Yaroshenko A.P., Savoskin M.V., Magasinski A.N., Savsunenko O.B. Synthesis and
thermal expanding of residual graphite bisulfates produced in Na2Cr2O7-H2SO4 system.
439
Proceedings of 12th International Symposium on Intercalation Compounds; 2003 June

1-5; Poznan, Poland. Poznan University of Technology, 2003:134.
4. Menard D., Py X., Mazet N. Development of thermally conductive packing for gas
separation. Carbon 2003; 41:1715–27.
5. Py X., Daguerre E., Menard D. Composites of expanded natural graphite and in situ
prepared activated carbons. Carbon 2002; 40:1255–65.
6. Modestov A., Glezer V., Maryasin I., Lev O. Photocatalytic degradation of chlorinated
phenoxyacetic acids by a new buoyant titania - exfoliated graphite composite
photocatalyst. J Phys Chem B 1997; 101:4623-9.
7. Tsumura T., Kojitani N., Umemura H., Toyoda M., Inagaki M. Composites between
photoactive anatase-type TiO2 and adsorptive carbon. Appl Surface Sci 2002; 196:429-
36.
8. Viculis L.M., Mack J.J., Kaner R.B. A chemical route to carbon nanoscrolls. Science
2003; 299:1361.
9. Shioyama H., Akita T. A new route to carbon nanotubes. Carbon 2003; 41:179-81.
CHAPTER 6:
CARBONS IN THE CATHODES

OF LITHIUM-ION BATTERIES;
ALTERNATIVE FORMS OF MnO2,
CATHODE / CARBON MODELING
The seven papers in Chapter 6 are focused on cathode materials for
lithium and lithium-ion batteries. Carbon is used as a conductive additive in
composite electrodes for batteries. The type of carbon and the amount can
have a large effect on the electrochemical performance of the electrode.
The first paper in this chapter is a contribution from R. Kostecki and
F. McLarnon of Lawrence Berkeley National Laboratory and consists of a
diagnostic study of the power fade issues related to high-power Li-ion
batteries that contain a Li(Ni0.8Co0.15Al0.05)O2 cathode. Li-ion batteries are
prone to poor life and experience problems with impedance rise during
elevated temperature aging. SEM, Raman and current-sensing AFM
techniques were used to address the mechanism of power fade at the cathode
with aged cells. From the results, the authors describe a process of carbon
retreat/redistribution in the composite electrode upon accelerated aging. The
carbon retreat/redistribution is said to have a major impact in the
contribution to the cathode impedance rise and loss of power. They finish
with a discussion on the loss of conductivity across the composite cathode as
a result of carbon retreat/redistribution.
In the paper from V. Matveyev of the Ukrainian State University of
Chemical Engineering, an examination of the role of conductive carbon
additives in a composite porous electrode is conducted. A model for
calculation of the local electrochemical characteristics of an electrode is
presented. A comparison on the polarization of the electrode as a function of
the redox state of the electroactive species is emphasized in the model. The
electrochemical reaction of chloranil (tetrachlorobenzoquinone) was
measured and results compare favorably to calculations derived from the
model.
S. A. Kirillov et al. describe the design, synthesis and application of
amorphous manganese dioxide in lithium batteries in the third paper of this
chapter. A synthetic route of oxidation of Mn(II) aqueous solution with
hydrogen peroxide was used to prepare the materials. The prepared MnO2
material remains amorphous up until 600°C. The amorphous MnO2 delivers
higher capacity of 247 mAh/g versus 189 mAh/g for Russian commercial
crystalline EMD when discharged to 2.0 V. Further testing has shown that
the amorphous phase demonstrates reversible cycling with lithium. These
results bode well for the development of a new class of amorphous
manganese oxides for charge storage in lithium or perhaps lithium-ion
batteries in the future.
The performance in Zn-MnO2 alkaline cells of electrolytic
manganese dioxide produced in the presence of fluoride ions is compared
with commercial EMD materials by I. Makyeyeva et al. The material
produced in the presence of fluoride ions was shown to have superior
utilization to the commercial EMD. The authors state that the improved
443
444
performance of the manganese oxide synthesized in the presence of fluoride

ions is due to an increased or larger number of defects in the crystalline
structure. The defects are associated with a higher non-stoichiometry in the
material due to the presence of more Mn(III). In addition, the authors explain
that a higher percentage of hydroxide (OH-) moieties in their material are
associated with an improved ionic conductivity.
In the fifth paper of this chapter on cathodes, an investigation of
thin-film oxide-hydroxide electrodes containing Cr, Ni, and Co compounds
was authored by N. Vlasenko et al. The thin-films were produced by
electrochemical deposition from transition metal aqueous fluorine-containing
electrolytes onto steel substrates. These thin-films were tested in Li coin
cells. Electrochemical activity appears to scale with the amount of fluoride
used in the deposition; the larger concentration of fluoride in the bath, the
greater the capacity. One Ni oxide-hydroxide film electrode showed greater
than 175 mAh/g reversible capacity on the 50th cycle with excellent
coulombic efficiency.
In a second paper from Kirillov et al., synthetic aspects of liquid
phase precipitation reactions are discussed in relation to the factors
responsible for metal-oxide formation. In particular the Pechini reaction, a
sol-gel process, was examined. A good reference section is provided to
introduce the reader to previous literature. The authors stress that more work
needs to be done to establish what factors are important for producing the
highest quality metal-oxide powders from such reactions. The conclusion
specifically lists the criteria needed in order to accomplish this task.
In the final paper of this chapter, authors S.-H. Kang and K. Amine
from Argonne National Laboratory discuss high-power lithium ion batteries
and present a new lithium nickelate material: Li(Ni0.4Co0.2Mn0.4)O2 with a
well-formed layered structure. The electrochemical performance in lithium
cells is compared for materials that are uncoated versus coated with
aluminum oxide. The capacity fade is less for coated material versus
uncoated both at room-temperature up to 100 cycles and 50°C up to 70
cycles. The improvement was found to be linked to an interfacial impedance
drop for electrodes of the coated material. The authors state that the surface
coating suppresses the degradation of the cathode/electrolyte interface and
that the mechanism may be involved with the scavenging of HF in the
electrolyte by the coating. Thus the coating would protect the underlying
layered oxide material to acid attack and possible dissolution/corrosion
effects during cycling in the LiPF6 organic electrolyte.
The editors hope that you enjoy reading the last chapter in this
NATO series volume, and sincerely thank you for your interest in all of the
papers from the NATO-Carbon Advanced Research Workshop and
Conference (NATO-CARWC).
DIAGNOSTIC EVALUATION
OF POWER FADE PHENOMENA
AND CALENDAR LIFE REDUCTION
IN HIGH-POWER LITHIUM-ION BATTERIES
Robert Kostecki* and Frank McLarnon

Environmental Energy Technologies Division
Lawrence Berkeley National Laboratory; University of California, Berkeley, CA 94720, USA
Abstract
High-power Li-ion cells with graphite anodes and

LiNi0.8Co0.15Al0.05O2 cathodes that were cycled and stored at elevated
temperatures showed a significant impedance rise and capacity fade, which
were associated primarily with the LiNi0.8Co0.15Al0.05O2 cathode. A
combination of electrochemical, physical, and chemical diagnostic
techniques, including Raman, SEM, and current-sensing AFM, was used to
characterize the cathodes from these cells in order to produce a clear picture
of the mechanism for cell degradation. Systematic Raman mapping of 50 x
80Pm areas at 0.9Pm spatial resolution produced semi-quantitative
composition maps of cathode surfaces. Raman microscopy surface
composition maps and SEM images of cathodes from tested cells revealed
that cell cycling or storage at elevated temperatures led to significant
changes in the LiNi0.8Co0.15Al0.05O2/elemental-carbon surface concentration
ratio. The loss of conductive carbon correlated with the power and capacity
fade of the tested cathodes and the loss of surface electronic conductivity.
Keywords
Li-ion battery; cathode; carbon retreat; conductivity; Raman; AFM.
1. INTRODUCTION
Lithium-ion batteries are being seriously considered for application
in all-electric vehicles (EV) and hybrid electric vehicles (HEV’s) because of
their high power and energy densities [1, 2]. The U.S. Department of
*
445
446
Energy’s Advanced Technology Development (ATD) Program supports the

development of high-power LiNi0.8Co0.15Al0.05O2/ EC-EMC, 1.2 M
LiPF6/graphite batteries for hybrid electric vehicle applications [3]. A major
goal of the ATD Program is to develop a detailed understanding of the
power and capacity fade mechanisms of a single cell chemistry designed for
the HEV application and to develop predictive models for the accelerated
life test behavior based on these mechanisms. For this purpose, this cell
chemistry has been subjected to elevated-temperature calendar life and
lifecycle testing with a pulse-type profile over a very limited range of state-
of-charge (SOC) [4].
Diagnostic evaluations of Li-ion cells that were aged and/or cycled
under various conditions were carried out to determine the mechanisms
responsible for the cell power loss that accompanies life tests at elevated
temperatures. Many cells have been examined to determine the mechanisms
for capacity and power fade, using a wide range of diagnostic techniques that
were developed and refined during studies of the ATD cells [5]. Impedance
measurements of the cell components indicated that the LiNi0.8Co0.15Al0.05O2
cathode is primarily responsible for the observed cell power loss at elevated
temperatures, similarly to the LiNi0.8Co0.2O2 cathode, which was studied in
our previous work [6]. Possible causes of the increase in cathode impedance
include formation of an electronic and/or ionic barrier at the cathode surface.
However, the observation that, in these cells, power fade is always
accompanied by a loss of cell discharge capacity suggests that the
mechanism of degradation is more complicated and may consist of multiple
processes, which may be a somewhat general phenomenon that affects
several types of cathodes. X-ray diffraction spectroscopy failed to detect
noticeable changes in the bulk structure of the tested cathodes, and no
evidence of material structural degradation or formation of new phases was
found. In this study we focused our attention on cathode surface processes,
which we believe have a dominant effect on impedance behavior. A
systematic diagnostic evaluation of LiNi0.8Co0.15Al0.05O2 cathodes removed
from high-power Li-ion cells, which were stored and/or cycled at elevated
temperatures, was carried out. The application of local probe techniques such
as current-sensing AFM and Raman microscopy to characterize physico-
chemical properties of the electrode surface at nanometer resolution can
provide unique insight into the complex mechanism of the detrimental
processes that are likely responsible for the composite cathode capacity fade
and impedance increase.
2. EXPERIMENTAL
High-power Li-ion cells with a LiNi0.8Co0.15Al0.05O2 cathode, a
synthetic graphite anode, 1.2 M LiPF6 + ethylene carbonate + ethyl-methyl
carbonate (EC/EMC) electrolyte, and a Celgard 2300 separator, were
447
manufactured, aged, cycled, and/or abused and then characterized under the
ATD Program [3, 4]. We compared a fresh cathode with cathodes taken from
cells that were aged or cycled at elevated temperatures for up to 68 weeks,
losing up to 52% of their initial power and 24% of their initial capacity.
The samples were collected from the cathodes ~ 2.5 cm away from
the current collector tab, washed in pure dimethyl carbonate (DMC), and
soaked in DMC for 30 minutes after removal from Li-ion cells inside an
argon-filled glove box. This procedure removed electrolyte salt from the
electrode to prevent its reaction with air and moisture. An integrated Raman
microscope system “Labram” made by ISA Groupe Horiba was used to
analyze and map the cathode surface structure and composition. The
excitation source was an internal He-Ne (632 nm) 10 mW laser. The power
of the laser beam was adjusted to 0.1 mW with neutral filters of various
optical densities. The size of the laser beam at the sample was ~1.2 Pm.
We used current-sensing atomic force microscopy (CSAFM) to
measure and image-map the electronic conductivity of individual grains of
the LiNi0.8Co0.15Al0.05O2 powder (Fuji Chemical) that was used to fabricate
ATD Program composite cathodes. The powder was pressed into a gold foil
to produce randomly scattered particles of LiNi0.8Co0.15Al0.05O2 in good
electronic contact with the Au substrate. The microscope consisted of a
Molecular Imaging (MI) scanning probe microscope coupled with a Park
Scientific Instruments (PSI) AutoProbe Electronic Module. The Si atomic
force microscope (AFM) tips were coated with a thin conductive layer of
W2C. All CSAFM experiments were performed in constant-force mode with
controlled oxide-tip voltage difference. A single scan of the tip over the
sample surface simultaneously produced two images: a topographic image
and a conductance image; the latter represents Au/LiNi0.8Co0.15Al0.05O2 - tip
current variations during scanning at a given sample-tip voltage difference.
CSAFM imaging was conducted in a small glove box specially designed for
scanning-probe microscopy tests under a controlled N2 atmosphere.
Scanning electron microscope (SEM) images were recorded using a
JEOL field emission microscope, model JSM 634OF. Sample preparation
was carried out in the inert-atmosphere glove box. The samples were then
sealed in a small bottle for transportation to the AC sputtering chamber
where they were coated with gold-palladium prior of being transferred to the
SEM vacuum chamber.

Cathode surface-average Raman spectra were produced with the
Raman microscope. Raman spectra were collected from multiple 52x75Pm
sections of the cathode surfaces at 0.7Pm resolution and averaged into one
representative spectrum for each cathode. MicroRaman surface-average
spectra of the cathode from the virgin cell, and cathodes removed from cells,
448
which lost 10, 24, 34 and 52 % of power are shown in Figure 1. The Raman
spectra of all cathodes are dominated by two groups of bands: a broad
maximum centered at ~510 cm-1, characteristic for LiNi0.8Co0.15Al0.05O2
oxide, and two peaks at ~1350 and ~1600 cm-1, which correspond to the D
and G bands of elemental carbon, respectively. Raman spectroscopy is a
particularly useful tool for characterizing the near-surface structure of
carbons because of its relatively large Raman scattering cross-section.
Recorded average spectra of the composite cathodes display strong carbon
bands, which are predominant at almost all locations and originate from
graphite and carbon black conductive additives.
b
a
400 800 1200 1600 200 0

cm -1
Figure 1. Average Raman microscope spectra of the fresh composite LiNi0.8Co0.15Al0.05O2
cathode (a), and the cathode from cells that exhibited 10 (b),
24 (c), 34 (d), and 52% (e) power loss.
449
The intensity ratio between the three major bands at ~510 cm-1
(LiNi0.8Co0.15Al0.05O2), and 1350, 1580 cm-1 (carbon) represent a semi-
quanti-tative comparison because of the different Raman scattering cross-
sections of LiNi0.8Co0.15Al0.05O2 and carbon. However, it is clear from
Figure 1 that the active material/elemental-carbon surface concentration ratio
increases with the increasing extent of cell and cathode degradation and is
substantially higher for the cells that lost 10, 24, 34% of power compared
with the fresh cell. The most drastic change in the oxide/carbon surface
concentration ratio was observed in the cathode from the cell that lost 52%
of its original power. Interestingly, the cathode surface state of charge varied
between and within large individual agglomerates of active material and, at
some locations, the spectra correspond to fully charged material, even
though the cell was fully discharged at the end of testing and before
disassembly.
20 Pm
Figure 2. SEM images of the fresh composite LiNi0.8Co0.15Al0.05O2 cathode (a),

and the cathode removed from a cell that lost 52% of its original power.
Figure 2 shows SEM images of a composite LiNi0.8Co0.15Al0.05O2

cathode extracted from a fresh cell and a cell that was cycled at 45qC for 68
weeks and lost 52% of its original power. Note that the surface of the fresh
cathode consists of 5-20 Pm agglomerates of LiNi0.8Co0.15Al0.05O2 (bright
particles) and a dark fine particle coating, most likely carbon black, which is
non-uniformly distributed across the surface. Interestingly, the surface
morphology of the cathode from the tested cell reveals a noticeable change.
There remain a few areas that show no or little change compared to the
original morphology, whereas in most regions the black carbon coating has
almost entirely disappeared from the surface leaving the active material fully
exposed. This observation is in full agreement with our Raman spectra,
which showed a dramatic decrease of the carbon signal in the tested
cathodes.
450
In order to better understand the potential role of carbon

retreat/redistribution on the cathode impedance rise we measured the surface
electronic conductivity of these cathodes with CSAFM. Figure 3 shows
CSAFM images of surface conductance (right-hand panel) and topography
(left-hand panel) of a representative 5 x 5 Pm region of the composite
LiNi0.8Co0.15Al0.05O2 cathode surface at 1.0 V tip-sample voltage difference
for (A) a fresh cell; and (B) a cathode from the cell that lost 34% of its
power. The surface morphology images of both cathodes (left-hand panel)
show large polycrystalline agglomerates and reveal no significant changes in
the cathode surface topography in tested cells. This result is contrary to our
earlier study of LiNi0.8Co0.2O2 cathodes, which showed that considerable
amounts of nanocrystalline deposits accumulated in the cathode inter-
granular spaces and across the crystal planes [6,7].
Topography Conductance
Fresh cathode
Cathode from cell, which lost 34% of power
Figure 3. CSAFM images of surface conductance (right-hand panel,

black areas are conductive) and topography (left-hand panel)
of a 5 x 5 Pm region of the cathode surface
at 1.0 V tip-sample voltage difference.
In the conductance image, a dark color represents high electronic

conductance, whereas a white color corresponds to areas of low or zero
electronic conductance. Taking into account that the tip is in physical contact
with the oxide, the magnitude of the current is determined by the local
electronic properties of the electrode and the tip, and the tip-sample voltage
difference. The surface conductance image of the cathode from the fresh cell
451
exhibits areas of mostly excellent electronic conductance and only few

insulating regions. Highly conductive graphite and acetylene black certainly
contribute to the observed high tip current. The LiNi0.8Co0.15Al0.05O2 may
also contribute to the tip current, but its contribution is substantially lower
than that of carbon. The insulating areas on the cathode surface are most
likely associated with the presence of PVDF binder, a solid electrolyte
interphase (SEI), and/or unusual surface morphology (e.g., deep crevices).
The conductance image of the tested cathode reveals that almost the
entire electrode surface became insulating except for a few locations, mainly
in the deep crevices and intergranular spaces, which remained conductive.
Previous results for LiNi0.8Co0.2O2 cathodes [6,7] also showed a
complete lack of conductivity of the cathode surface, consistent with loss of
electronic contact between particles at the electrode surface and the
remainder of the electrode. Another plausible explanation for the observed
phenomena could be formation of a very thin SEI layer. A relatively thick
layer of Li2CO3 and P-, O- and F- containing compounds forms upon storage
of LiNi1-yCoyO2 in LiPF6-containing electrolyte at 80-90qC, as found in [8,9].
However, our Raman and infrared measurements revealed no observable SEI
layer on the cathode surface. On the other hand, the presence of charged
material at the cathode surface may also indicate that some particles of
LiNi0.8Co0.15Al0.05O2 became electrically disconnected from the remaining
part of the cathode due to mechanical stress and/or carbon additive retreat.
A reduction in conductance of portions of the cathode can easily
explain the observed increase in cell impedance as well as loss of cathode
capacity via isolation of oxide active material. Particle isolation is also in
concert with our observations of carbon retreat or rearrangement in tested
cathodes. These results represent the most clear and obvious difference
(compared to other diagnostic studies) in cathode characteristics after
prolonged cell tests at elevated temperatures.
ACKNOWLEDGEMENTS
This work was supported by the Assistant Secretary for Energy
Efficiency and Renewable Energy, Office of FreedomCAR and Vehicle
Technologies of the U.S. Department of Energy under Contract No. DE-
AC03-76SF00098. The authors gratefully acknowledge the tested cells, help,
and advice provided by the ATD Program participants.
REFERENCES
1. G. Nagasubramanian, R.G. Jungst, D.H. Doughty, J. Power Sources, 83, (1999) 193.
2. Q. Wu, W. Lu, J. Prakash, J. Power Sources, 88 (2000) 237.
452
3. “FY 2000 Progress Report for the Advanced Technology Development Program,” U.S.
Department of Energy, Office of Advanced Automotive Technologies, Washington, D.C.
(December 2000).
4. “FY 2001 Progress Report for the Advanced Technology Development Program,” U.S.
DOE OAAT, February 2001.
5. “Handbook of Diagnostic Techniques”, Lawrence Berkeley National Laboratory Report
no. LBID-2464, April 2003.
6. A. X. Zhang, P. N. Ross, Jr., R. Kostecki, F. Kong, S. Sloop, J. B. Kerr, K. Striebel, E.
Cairns, and F. McLarnon, J. Electrochem. Soc. 148, A463 (2001).
7. R. Kostecki and F. McLarnon, Electrochem. Solid State Lett., 5, A164 (2002).
8. A. M. Andersson, D. P. Abraham, R. Haasch, S. MacLaren, J. Liu, and K. Amine, J.
Electrochem. Soc., 149, A1358 (2002).
9. D. Ostrowskii, F. Ronci, B. Scrosati, and P. Jacobsson, J. Power Sources, 94, 183 (2001).
MODELING OF ELECTROCHEMICAL
PROCESSES IN THE ELECTRODES
BASED ON SOLID ACTIVE REAGENTS AND
CONDUCTIVE CARBON ADDITIVES
Vadym V. Matveyev*
Ukrainian State University of Chemical Engineering
8, Gagarin avenue., Dniepropetrovsk, 49005, Ukraine
Abstract
The processes that occur at single crystals of an active substance

embedded in an electroconducting porous matrix are considered. Conversion
of solid reagents is assumed to proceed via a liquid-phase scheme as follows:
dissolution–electrochemical reaction–crystallization. A set of equations is
proposed. The set having been solved, an analytical expression is derived for
a local polarization characteristic as a function of polarization, oxidation
state of reagents, and structural and physicochemical parameters of the
system. In order to check the model, galvanostatic characteristics of the
electrodes made of the mixture of chloranile (C6Cl4O2) and carbon black
have been investigated. The theoretical and experimental characteristics
show very good agreements.
Keywords
Theoretical model; porous electrode; solid reagent; dissolution;

electrochemical reaction; crystallization; polarization characteristic;
chloranile; carbon black.
1. INTRODUCTION
Most types of electrodes for electrochemical power sources have a
porous structure [1]. Such electrodes are typically composed of
electrochemically active material and additional components. The basic
*
453
454
additional component ensures the essential electronic conductivity of the

electrode. It builds a permanent continuous conducting network in the bulk
of electrode matrix. Other additions are impregnated into active mass in
small amounts. They are auxiliary agents, for example, a binder. Carbon
materials such as graphite or carbon black are frequently used as the
electrically conductive additives in battery electrodes.
In recent years, the electrodes of similar structure have also found
application in the electroanalytical chemistry, for purposes of researching the
electrochemical behavior of solid substances, etc. These electrodes are
known as the modified carbon paste electrodes or carbon paste electroactive
electrodes [2].
Given a widespread and growing application of such electrodes in
electrochemical industry, a theory is required to describe the behavior of the
electrochemical cells based on them. Such a theory would have to take into
account, and to be able to distinguish between the individual contributions of
processes proceeding at micro- to macro-levels in the electrochemical cell,
as well as to furnish a multifunctional description of the whole system.
Several levels may be defined in the organization of such systems.
We shall distinguish between the following levels [3]: an electrochemical
cell as a whole; an electrode; a local section of the electrode; and an
"elementary cell". The "elementary cell" is being defined as a separate
crystal of active substance enclosed by a part of the conducting matrix. The
“local section” is defined as being part of the electrode, which is contained in
the active mass in the sufficient amount and with a set of properties, which
reflect the average properties of the entire active mass; also, we shall assume
that within this local section it may be possible to neglect spatial
macrodistribution of the processes and the reagents concentration. Other
levels of the system organization do not need to be explained due to their
obviousness.
The theory on the level of the electrode and on the electrochemical
cell is sufficiently advanced [4-7]. In this connection, it is necessary to
mention the works of J.Newman and R.White's group [8-12]. In the majority
of publications, the macroscopical approach is used. The authors take into
account the transport process and material balance within the system in a
proper way. The analysis of the flows in the porous matrix or in the cell
takes generally into consideration the diffusion, migration and convection
processes. While computing transport processes in the concentrated
electrolytes the Stefan-Maxwell equations are used. To calculate electron
transfer in a solid phase the Ohm's law in its differential form is used. The
electrochemical transformations within the electrodes are described by the
Batler-Volmer equation. The internal surface of the electrode, where
electrochemical process runs, is frequently presented as a certain function of
the porosity or as a certain state of the reagents transformation. To describe
this function, various modeling or empirical equations are offered, and they
455
are for the most part solved. However, such an approach does not permit to
adequately take into account the processes proceeding at a microlevel, while
describing the macrosystem.
There appears to be a more adequate approach when a local
polarization characteristic is obtained as a result of analysis of the processes
in the elementary cell and the local section of the electrode. This
characteristic depends on the state transformation of the solid reagents and
the concentrations of the electrolyte components. It further may be
introduced into the equations describing the macrokinetic processes in an
electrode, and may be used to model the behaviour of the system as a whole.
The model for calculation of the local electrochemical characteristics
of an electrode is presented below.
2. MODEL
Herein, we consider the case when a porous conducting matrix with
inclusion of active solid reagents represents the electrode. It is supposed, that
both the reagent and the product are nonconductive. The conversion of the
solid reagents is assumed to proceed via a liquid-phase mechanism in the
following way: dissolution – electrochemical reaction – crystallization.
Figure 1 shows the structure of the electrode and its model. The model has
been developed on the bases of several assumptions.
2.1. Assumptions for the Model
Assumption #1. The elementary cell is spherically symmetric.

Assumption #2. The reagent and product crystals are of low solubility. Their
solubility is the following
0 0
c1 , c 2 10 3 mol l 1 (1)
The following designations shall be used in the proposed model

(please see a schematic representation of the model as introduced by
Figure 1): Rc is the radius of the elementary cell, which includes a conducting
porous layer and a pore filled with crystals. R0 is the radius of the pore
containing the reagent and the product. ri (t ) is the crystal radius. It is assumed
that the radii of the crystals may vary with the time. The index i indicates
which of the reagents form, either oxidized (i =1) or reduced one (i =2), is
taken into account.
456
Figure 1. Typical structure of the porous electrode with solid reagents

and schematic representation of the proposed physical model.
Assumption # 3. The reagent’s transfer from one cell into another

electrode is not taking place. During the electrochemical process, the solid
reagent is being dissolved and diffuses into the conducting porous matrix.
The solution gets reduced (oxidizes) electrochemically and the reaction
product crystallizes within the same pore. It is supposed that the flow of the
dissolved solid reagents through the border of the elementary cell does not
take place.
dc~i
0, i 1, 2 (2)
dr r Rc
ci ci (r ) is the concentration of the dissolved solid reagents in the

pore around the crystals. The tilde (~) specifies that the parameter relates to a
porous matrix.
Assumption #4. The electrochemical reaction is taking place within
a porous electrode layer and it can be described by the Batler-Volmer
equation (3).
kj lj
ª§ c~ · § c~ j · D nF K § c~1 · § c~ j · E nF K º
iV S V i0 «¨¨ 2 ¸¸ ¨ ¸ e RT ¨ ¨
¨ c ¸¸ ¨ c
¸ e RT » (3)
«© c 20 ¨ ¸ ¸ »
¬ ¹ © c j0 ¹ © 10 ¹ © j 0 ¹ ¼
457
ɇereiV is a local polarization characteristic, referred to as a unit of electrode

volume. SV , i0 are the specific surface and the exchange current density,
ci
respectively. is the concentration of solid-phase reagents (i) in the
ci 0
solution and this concentration is normalized to its equilibrium value; ci 0 is
k j (l j )
§ cj ·
usualy equal to the standard solubility c 0
i0 . ¨¨ c ¸¸ is the product of
© j0 ¹
the concentrations of the remaining solution components divided by their
initial values. n, F , R, T ; D , E are a number of electrons transferred in
elementary electrode reaction; Faraday’s constant, universal gas constant;
temperature (°K) and transfer coefficients, respectively.
Assumption #5. Transfer processes as within the cell have been
regarded as quasistationary. The typical time of the processes in the
electrode (time of a charging or discharging being t0~104 s) is longer than the
time of the transitional diffusion process in the elementary cell tc ~ Rc2/D ~
10-1 s (radius of the cell is Rc ~10-5 m, diffusion coefficient of dissolved
reagents is D ~ 10-9 m2/s). Therefore, the quasistationary concentration
distribution is quickly stabilized in the cell. It is possible to neglect the time
derivatives in the transport equations.
Assumption #6. On the surfaces of the crystal, concentration of the
corresponding substance in the solution corresponds to a saturated solution.
c c 0, i 1, 2 (4)
ir ri (t ) i
Assumption #7. The reagent and the product crystals are considered
to be independent.
Due to tridimentionality of the elementary cell, the reagent
concentration in the solution decreases quickly (as ci~1/r) near the crystal
surface during crystal dissolution. Therefore, the obstacles, namely crystals
of a new phase, deform insignificantly the distribution of the concentration
around of the crystal discussed. Figure 2 shows this peculiarity, which has
been calculated in [13].
2.2. Basic Equations
Main processes in the elementary cell are schematically shown by

Figure 3.
458
c/c0
2 .5
1 .5
0 .5
-1 0 1 2
x
0.8
0.6
0.4
0.2
-1 -0.8 -0.6 -0.4 -0.2 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
-0.2
-0.4
-0.6
-0.8
-1
Figure 2. Allocation of the concentration in the case when the reagent diffuses into the free
volume and when the other phase crystals present.
Reagent (Ox)
c~1 ( r )
diffusion transfer
c1(r)
-
Q1 Ox + … + n e
c10
= Q2 Red + …
c2(r)
c~2 (r )
Product (Red)
diffusion transfer
c20
Figure 3. Processes in the elementary cell.

459
Taking into account the spherical symmetry and the

quasistationarity, we have the following transition equations for the
dissolved reagent and product in a pore (r<R0)
d 2 c1 2 dc 1 d 2 c 2 2 dc 2
0 0 (5)
dr 2 r dr dr 2 r dr
The equations in the porous matrix are the following:
d 2 c~2 2 d c~2 § Q2 · d 2 c~1 2 d c~1 § Q1 ·

¨¨ ~ ¸¸iV ¨¨ ~ ¸¸iV (6)
dr 2 r dr © nF D 2 ¹ dr 2 r dr © nF D1 ¹
ɇere iV is the local polarization characteristic (3). Q i is the stoichiometric

coefficient of substance i ( Q 1 ! 0, Q 2 0 ). With these coefficients, the
reagent and product are included into the equation of the electrode
reaction. Di , D i is the diffusion coefficient of the dissolved solid reagent
within the solution in the pore and within the porous matrix, respectively.
The boundary conditions on the pore surface are the following
c1 c~1 r R c2 ~
c2 r R
c c
dc 1 ~ d c~ dc 2 ~ d c~ (7)
D1 D1 1 D2 D2 2
dr dr r R dr dr r R
c c
On the boundaries of the elementary cell and the crystals, the

boundary conditions are described by equations (2) and (4), respectively.
2.3. Polarization Characteristic of the Elementary Cell
Solving the system of the equations tacking into account the boundary
conditions according to [3] we receive the following expression for the
dimensionless polarization characteristic of the elementary cell.
C 20 C k e D K C 10 C l e E K
i .
ª 1 N O º DK ª 1 N O º E K(8)
E 20 «1 C k e E 10 «1 C le
¬ [2 g 2 »¼ ¬ [1 g 1 »¼
460
ci0
Here Ci 0 is the solubility ci0 of the solid-phase reagents i normalized to
ci00
k j (l j )
§ cj ·
its standard value. C k (l ) equals ¨
¨ c ¸¸
.
© j0 ¹
ic nF r (t )
i , K K , [i i . are the dimensionless values of the
ic 0 RT R0
cD
current, the polarization, the crystal radii, respectively. ic 0 4S nFR0 0 0
Q0
is the normal (characteristic) value of the current flowing from the
Q i c0 D0 Di
elementary cell. Ei 0 , gi are the dimensionless
ci00 Di Q 0 D i
parameters; c0 , D0 , Q 0 are the normalizing values of solubility, the diffusion
coefficient and the stoichiometric coefficient ( Q 0 1 ), respectively. They are
usually accepted as equal to the appropriate parameters of the solid phase in an
oxidized (i=1) or reduced state (i=2).
2.4. Structural-Kinetic Factor
Function
1
ª [ c O 1 2 O [c 1 º
« e 1»
[cO 1
N O, [ c «1 O »
« [ c O 1 2 O [c 1 » (9)
« e 1»
¬ [cO 1 ¼
is the factor taking into account the restrictions superimposed on the

polarization characteristics (8) by the kinetics of the electrochemical
transformations within the porous matrix
Here
2
R0
O i0 S V
nFc 0 D 0
E C e
1 l
E K
E 2C k e DK (10)
is the kinetic factor. E1 ! 0, E 2 0 since Q 1 ! 0, Q 2 0 .

461
Rc
[c is the dimensionless radius ( [ c >1 ) of the elementary cell
R0
2.5. Oxidation State
The state of oxidation q t is the relation of the oxidized crystal

mass at a given moment to the crystal mass, provided that the active
substance is completely oxidized:
m1 t m t
q t 0
1 2 0 . (11)
m1 m2
In equation (11), mi t , ri t are the mass and the radius of solid

0 0
reagent at a given moment t; mi , ri are the crystal mass and radius provided
that the active substance is completely oxidized (i=1) or reduced (i=2).
Taking into account that mi ~ ri , the dimensionless radii of the crystals
3
can be expressed via the oxidation state:

1
r1 t r1 ª m1 t º 3
0 1 1
[1 t « » >G1q1 t @3 , [ 2 t >G 2 1 q t @3 (12)
R0 R0 ¬ m10 ¼
3
§ ri0 ·
Here Gi ¨¨ ¸¸ characterizes the relation between of the
© R0 ¹
completely oxidized (reduced) crystal volume and the pore volume in which
the crystal is placed.
2.6. Evolution Equation

In order to find the evolution equation, we shall consider the relation
between the current and the oxidation state of a separate crystal.
The amount of electricity used for oxidation (reduction) of the
crystal can be expressed through the change of the oxidation state in the
following way:
nF nFmi0
dQ dmi dq (13)
M iQi M i Qi
Here Mi is the molecular weight of solid reagent i.
462
The current fall-off at this crystal is equal to the charge used for the
oxidation of a single crystal within a unit of time (the anodic current is
considered as a positive one)
dQ (14)
ic
dt
Hence, from the above, one can receive the equation of evolution, as
follows:
Q0 mi0 dq
i q, ^c j `, K, K E t E 0 .
4Sc 0 D0 R0 M i Q i dt (15)
^ `
where i q, c j , K denotes the expression of the dimensionless polarization
characteristic of the elementary cell (8), where the radii of the crystals are
expressed through the oxidation state.
2.7. Polarization Characteristic Relating to a Unit of the

Solid Reagent Mass
2.7.1. Monodispersed Case
Let us assume, that the electrode includes the reagent crystals having
identical size, which are placed at regular intervals in the conducting matrix.
For such case,
3mi
N (16)

4S ri0 U i
3
would be the number of crystals in a unit of the solid reagent mass. Here mi ,
ri0 , U i are the mass, the averaged crystal radius and the solid phase density
of the component i. Then
Nic 3nFc0 D0 3 Gi
im i q, ^c j `, K. (17)
mi
Ui ri0
2
is the polarization characteristic relating to a unit of the solid reagent mass in

a monodispersed case.
2.7.2 Polydispersed Case
While performing dispersion analysis, the crystal aggregate is

divided into several fractions; the crystals of each fraction have the sizes
within a given range, and the quantity of the reagent crystals from each
463
fraction N n can be determined. The current relating to a unit of the active

component mass may be presented as
1 4Sc0 D0 R0 n
im ¦N i n cn , icn i q n , ^c j `, K. (18)
mi Q0
Here it is supposed, that the crystals of each fraction contribute to
the total current independently, i.e. the mass interchange between separate
cells is disregarded. The latter is possible in the case when the packing
density of the reagent crystals in a porous matrix is low and the distance
between the crystals is much greater than the average radius of the crystals.
Works [14, 15] consider the case when parameter G i does not
depend on the crystal radius, e.g. when G i o 1 . The first case may take
place when the initial solid reagent crystallizes from a solution within pores
of the matrix. The second one is possible when the electrode is prepared
from a mixture of the active crystal substance and the conducting additive
particles having considerably smaller size than that of the crystals. Carbon
black is usually used as such additive. The common expression for the
current, relating to a unit of active component mass in this case is the
following
f
³ f r i >q, ^c `, Kt @ r
0 0
i j i dri0
nFc0 D0 0
im f
(19)
Q 0Ui Gi
3
³ r f r dr
0 3 0 0
i i i
0
Here,
f ri0 is the size distribution function of the initial solid reagents
2.8. Peculiarities of the Polarization Characteristic

2.8.1. The Main Factor
In case the structural-kinetic factor N O, [ c o 0 the polarization

characteristic does not depend on the kinetics of the electrochemical
reaction. In this case
C 20 C k e D K C10 C l e E K
i 1 1
.
ª º ª º
E 20 «1 G 2 1 q t »C k e E10 «1 G1 q t »C l e
3
DK
3
E K (20)
¬ ¼ ¬ ¼
464
The electrode current density is depends mainly on the solid reagents

solubility and on the initial size of crystals
c0
im ~ (21)
r
0 2
During the process of charge-discharge of the electrode, polarization

varies with time due to modification in the crystal size.
2.8.2. On the Peculiarities of the Structural-Kinetic Factor
FactorN O, [ c decreases with increase of the amount of

conductive additive [ c , and the values of parameter OK increase, as
shown by Figure 4.
Figure 4. Dependence of the structural-kinetic factor on parameter O at various values of [c.
Function OK (see equation (10)) achieves its minimum
i0 S V
O min | R 0 (22)
nFc0 D0
when
465
1 § E E1 C l ·
Kmin ln¨¨ ¸~0
D E © D E 2 C k ¸¹ (23)
and increases quickly with a shift of polarization both to a positive and to a

negative side.
2.8.3. Estimations for a Case When the Kinetics Has no Effect on the
Polarization Characteristics
The kinetics of processes has no effect on the electrochemical

characteristic when factor N O, [ c 1 . As Figure 4 shows, this is possible
when the conditions O ! 1 and [ c ! 1.2 1.4 are satisfied.
Let us estimate the specific surface of the conducting additive to get
O ! 1 . From (22) we have
i0 S V 0 k 0 SV
O | R0 | ri
nFc0 D0 Di (23)
where k 0 is the electrochemical reaction rate constant

Usually k 0 ~ 10 -3 y 10 -9 m/s [27], the diffusion coefficient in
solutions is in the range of Di ~ 10 9 m 2 /s , while the average radius of the
0 5
active reagent crystals is on the order of ri ~ 10 m .
Since the specific surface area of modern graphite materials is
SV ~10 4 y 1010 m -1 ( S m ~10 2 y 10 4 m 2 g 1 ) , it is almost always
possible to make an electrode characterized by O ! 1 from a mixture of a
carbon material and a solid reagent.
In this case, the kinetics has no effect on electrochemical
performance characteristics.
3. GALVANOSTATIC CYCLING MODE
3.1. Theory
In the case when N O, [ c o 0 , it is easy to receive expression for

dependence of electrode polarization on the current, and the oxidation state
of solid reagents, using polarization characteristic (20):
466
1
ª º
C1 i E 10 «G 1 q 3 1»

Cl ¬ ¼
K ln ln 1 (24)
Ck ª º
C 2 i E 20 «G 2 1 q 3 1»

¬ ¼
The theoretical charge (discharge) curves calculated with the help

of equation (24) at the following values of E10 E 20 1 and
G1 G2 0.7 are given by Figure 5. A high value of the initial polarization
is due to small sizes of new phase crystals. The flow of the dissolved
substance from a crystal to a matrix also decreases because the size of
reagent crystals gets decreased during the dissolution process. In certain
moment, the flow cannot any more provide the specified current and
polarization sharply increases. The dependence of the limiting values and the
oxidation state on the current is described (for anode process) by the
following equation
3
1 § C · (25)
q max (anode) 1 ¨¨1 2 ¸¸
G2 © i E 20 ¹
and is shown by Figure 6.
3.2. Experimental
In order to check validity of the proposed theoretical model,
galvanostatic characteristics of thin electrodes made out of the mixture of an
organic active compound, chloranile (C6Cl4O2) and acetylene carbon black
with a weight ratio 1 : 0.75 have been assembled and their electrochemical
performance has been investigated. A three-electrode cell made of ebonite
was used to perform the electrochemical experiments. The working electrode
was prepared from active mass (mam = 9.3·10-5 g) by the way of compacting
(P=300 MPa) it on a paraffin impregnated graphite disk (S = 1 cm2). A
counter electrode was Pt wire. The saturated Ag/AgCl electrode was used as
reference electrode. A 0.5 M solution of ɇ2SO4 was used as electrolyte. The
study was conducted at 18°C. The current-generating reaction for this
process is:
C6 Cl4 O 2 2H 2e- l C6 Cl4 OH 2 (26)
At the beginning of cycling, the electrode was discharged and

charged with a current of 0.25 mA. Both charge and discharge capacities
0
were equal to Q1 = 4.43 C. Then the electrode was cycled at a constant
467
oxidation current of 0.25 mA (charge) and at different values of reduction

(discharge) current (from 0.25 mA to 10 mA). Typical discharge curves as
the function of oxidation state, which ware calculated in the following way
It
q 1
Q1
0 (27)
are shown in Figure 7.
Figure 5. Dimensionless polarization as a function of the oxidation state.
Figure 6. The limiting value of the oxidation state

as a function of the dimensionless current in a galvanostatic mode.
468
Figure 7. Comparison between the measured (solid lines) and computed (dotted lines)
discharge curves as the function of the oxidation state at different currents.
Figure 8. The experimental data and computed curve of the active material utilization factor
as the function of the current.
Figure 8 provides a comparison of theoretically computed vs

experimental dependences of the active material utilization factor for the
investigated electrode. Analytical equations (24) and (25) were used to
calculate polarization as a function of the oxidation state, and to calculate the
limiting value of the oxidation state as the function of the discharge current
(see Figures 7 and 8).
469
The dimensionless current i was calculated by using equation (17) in the

following way
I 3nFc0 D0 3 G i 0
i ,I mi (28)
I0 0
U i ri 0 2
We have taken c10 , D1 parameters for chloranile as the normalizing values

( c0 c10 , D0 D1 ). As the initial value, we used the following parameters of
0
chloranile (Ox; i=1; m1 , U1 , r10 , G1 ). The initial chloranile mass mi 0 was
0 Q1 M i 0
calculated as mi Q1 = 5.5·10-3 g, where M i is the molecular weight of
nF
chloranile, Q 1 =1, n=2. At small amount of the active mass the appreciable
deviations in the values of the ratio of an active reagent and carbon black are
probable.
We used the following values of other parameters: the density of chloranile
0
U i =1.79 g·cm-3 [16], the solubility of C6Cl4O2 c1 =1.58 10-4 mol /l [17], the
0 -4
solubility of C6Cl4(OH)2 c 2 = 2.58 10 mol/l [17] and the diffusion
-9 2
coefficient of chloranile D1 =.7 10 m /s [18]. We considered that D2 D1 .
As the size of carbon black particles is significantly less than that of
chloranile crystals, we believed that the parameters G1 and G 2 are close to 1.
In calculations the values of G1 G 2 = 0.98 were used. The average size of
0
the crystals ( r § 6.2Pm) have been defined with the help of a
1
photomicrography of the active mass. As concentration of electrolyte did not
change during the reaction, then C k =1, C l =1. The change of solid reagents
solubility was not taken into account either. Hence C1 =1, C 2 =1.
The parameters, which are used in equations (24), (25) and (28), had
0 0
the following values: I 0 = 1.2mA, E10 =1, E 20 | c10 c 20 = 0.61. As one can
see from Figures 7 and 8, the theoretical and experimental curves show very
good agreements.
4. CONCLUSION
1. The proposed model generally describes the electrochemical
process, when the solid reagent and the product have precise phase borders
and the electrochemical reaction taking place on the surface, which does not
vary essentially with the time.
470
2. The analytical expression is derived for a local polarization

characteristic as a function of polarization, the oxidation state of reagents,
and the structural and physicochemical parameters of the system.
3. A simple expression has been derived for the polarization as a
function of the current and the oxidation state in the galvanostatic mode.
4. The expression for the active material utilization factor shows that
in the process considered, it is impossible to achieve full utilization of the
active reagents in the galvanostatic mode.
5. By reducing the crystal size, it is possible to achieve such a state
when even poorly soluble substances become electrochemically active.
ACKNOWLEDGMENTS
This work has been supported by the State Foundation for Basic
Research of the Ministry for Education and Science of Ukraine (grant No.
03.07/00168). The author expresses his gratitude to the U.S. Civilian
Research and Development Foundation for financial support of his
participation in the NATO-CARWC Wokrshop and Conference; to the
Argonne National Laboratory, Superior Graphite Co., and personally to Dr.
C. Johnson and Dr. I. Barsukov. for the organizational support of the visit.
REFERENCES
1. Bogotzky V.S., Skundin A.M. Chemical Power Sources. London-New York: Academic
Press, 1980.
2. Grygar T., Markenand F., Schroder U. and Scholz F. Collect. Czech. Chem. Commun,
2002; 67: 163-205.
3. Matveev V.V. Porous Electrode with Weakly Soluble Nonelectroconducting Reagents:
Polarization Characteristic of an Elementary Cell. Russ. J. Electrochem, 1997; 33: 839-
46.
4. Ksenzhek 0.S. Macrokinetics of Processes on Porous Electrodes. Electrochem. Acta.
1964; 9: 629-37.
5. Micka Ʉ., Rousar I. Theory of Porous Electrodes. XII. The Negative Plate of the Lead-
Acid Battery. Electrochim. Acta. 1974; 19: 499-502.
6. Micka Ʉ., Rousar I. Theory of Porous Electrodes. XVI. The Nickel Hydroxide Electrode.
Electrochim. Acta. 1980; 25: 1085-90.
7. Vest X., Jacobsen Ɍ., Atiung S. Modeling of Porous Insertion Electrodes with Liquid
Electrolite. J. Electrochem. Soc. 1982; 129: 1480-85.
8. Dunning J.S., Bennion D.N., Newman J. Analysis of Porous Electrodes with Sparingly
Soluble Reactant. J. Electrochem. Soc. 1971; 118: 1251-56.
9. Newman J., Teedemann W. Porous-Electrode Theory with Battary Application. A.I. Ch.
Journal. 1975; 21: 25-41.
9. Srinivasan F., Weidner J. W., and Newman J. Hysteresis during Cycling of the Nickel
Hydroxide Electrode. J. Electrochem Soc. 2001; 148: A969-80.
10. Zheng G., Popov B.N. and White R. E. Application of Porous Electrode Theory on
Metal-Hydride Electrodes in Alkaline Solution. J. Electrochem Soc. 1996; 143: 435-41.
471
11. Subramanian V. R. and White R. E. New Separation of Variables Method for Composite
Electrodes with Galvanostatic Boundary Conditions. J. Power Sources. 2001; 96:
385-95.
12. Wu B. and White R. E., Modeling of a Nickel-Hydrogen Cell. Phase Reactions in Nickel
Active Material. J. Electrochem Soc. 2001; 148: A595-609.
13. Matveev V.V. The Influence of the Other Phase Crystals Present on the Crystal
Dissolution. (Rus.). Problems of Chemistry and Chemical Technology /Voprosy Khimii i
Khimicheskoy Tekhnologii/ (Ukrainian Journal); in press.
14. Matveev V.V. Calculation of the Electrochemical Characteristics of Porous Electrodes
with a Polydisperse Crystal Structure (Rus.), Problems of Chemistry and Chemical
Technology /Voprosy Khimii i Khimicheskoy Tekhnologii/ (Ukrainian Journal). 2001; 3:
96-111.
15. Matveev V.V. Calculation of the Electrochemical Characteristics Porous Electrodes with
Polydisperse Crystal Structure. 53rd Meeting ISE, Abstracts, Dusseldorf, Germany 15-20
September. 2002.
16. Chu S.S.C., Jeffrey G.A., Sakurai T. The Crystall Strycture of Tetracloro- p -
Benzoquinono (Chloranil). Acta crystallogr. 1962; 15: 661-71.
17. Dinkevich F.E., Juravel T.A., Ksenzhek 0.S. and Lobach G.A. The Solubility of Some
Quinon Compounds in the Sulphuric Acid Solvents. (Rus.), Problems of Chemistry and
Chemical Technology /Voprosy Khimii i Khimicheskoy Tekhnologii/ (Ukrainian Journal).
1979; 55: 10-12.
18. Lobach G.A. The Disk Microelectrode: Theory and Application. Ph.D. Thesis.
Dniepropetrovsk: DChTI, 1982.
ON THE OPTIMAL DESIGN OF
AMORPHOUS MANGANESE OXIDE
FOR APPLICATIONS IN POWER SOURCES
S.A. Kirillov1*, T.V. Lesnichaya1, N.M. Visloguzova1, S.A. Khainakov1,

O.I. Pendelyuk1, D.I. Dzanashvili2, T.A. Marsagishvili2, V.Z. Barsukov3,
V.G. Khomenko3, A.V. Tkachenko3, and S.I. Chernukhin4
1
Institute for Sorption and Problems of Endoecology,
13, Gen. Naumov St., 03164 Kyiv, Ukraine;
2
Institute of Inorganic Chemistry and Electrochemistry,
11, Mindeli St., 380086 Tbilisi, Georgia;
3
Kiev National University of Technology and Design,
2, Nemirovich-Danchenko St., 01005 Kyiv, Ukraine;
4
Joint Department of Electrochemical Power Engineering,
83A, Vernadsky Ave., 03680 Kyiv, Ukraine
Abstract
Possible advantages of amorphous manganese dioxide as a cathode

material for application in the electrochemical power sources are briefly
outlined. A synthetic routine enabling one to produce amorphous manganese
oxide from aqueous solutions is described. Preliminary galvanostatic tests
reveal better performance of this material in coin-type cells with lithium
anodes, if compared with conventional electrolytic manganese dioxide.
Cycling studies demonstrate an ability of amorphous manganese oxide to
reversibly form compounds with lithium.
1. INTRODUCTION
Crystalline chemical and electrochemical manganese dioxides are
employed in billions of cells of Leclanché type which remain today probably
the most popular and inexpensive power sources for wrist watches, cameras,
flashlights, portable electronic products, etc. Worldwide annual output of
power sources is estimated at 40 bln. pieces; more than 70% of them are
those operating with MnO2 chemistry as a cathode active material or a
*
473
474
catalyst (zinc-manganese dioxide, lithium-manganese dioxide, zinc-air, etc.).

Numerous commercial cells are produced on wide scale by many companies
all over the world.
It is well known that manganese dioxides, being employed in the
power source technologies are crystalline in nature. In the course of
discharge of a power source with lithium anode, manganese oxide “absorbs”
lithium cations and transforms into lithium manganese spinel oxide; when
charging, spinel oxide “desorbs” lithium ions forming initial manganese
oxide. Electrochemical activity of cathode materials made from crystalline
manganese oxides is limited by diffusion of lithium cations inside (outside)
the bulk of a manganese oxide (spinel) crystal.
The main drawback of lithium-ion batteries employing manganese
oxides as cathode materials is their rapid capacity fade during cycling. Such
a disadvantage could be avoided if cathode materials are obtained in their
amorphous state (see, for instance, [1, 2]). This is easy to understand: in
amorphous materials, the crystallites are of tens to hundreds angstrom size,
hence diffusion complications do not exist or, at least, are significantly
eliminated. This enables one to reach higher capacity values in power
sources, to exclude capacity fade, and to improve cycling characteristics. On
the other hand, high values of specific surface area of amorphous manganese
oxides is a prerequisite of increasing catalytic activity of cathode materials in
zinc-air cells, enabling one to reach higher capacity values. Another, and
even better proven, electrochemical application, which explicitly exploits
great specific area of amorphous manganese oxide, is its use in
supercapacitors with non-acidic electrolytes [3].
As presented by this paper, our aim was to describe some
preliminary results demonstrating that amorphous manganese-oxide-based
materials synthesised by means of precipitation technique from aqueous
solutions could serve as effective cathodes in lithium batteries.
2. TEMPLATE SYNTHESIS OF
AMORPHOUS MANGANESE OXIDE
Manganese (IV) oxide enjoys numerous applications in modern
technologies. The most widely known areas of its usage are sorption
processes: one could recall that co-precipitation of contaminating cations
with manganese oxide is still employed as part of the in-tank precipitation in
processes of treatment of supernatant wastes at high concentration.
Furthermore, co-precipitation data are usually used as benchmark results in
studies of novel sorbents for strontium [4].
The legacy of Manganese oxides used in this work is as follows:
initial chemistries were synthesized as part of the project directed towards
optimization of amorphous manganese-oxide-based materials for waste
475
water purification; it has started in 2003, and the works are now under way.
In our preliminary studies we have managed to synthesize some mixed
materials of MnO2-TiO2 system, prove their amorphous nature, estimate their
properties and probe them as sorbents and catalysts [5]. Amorphous xMnO2-
yTiO2 samples have surface areas of up to 400 m2/g. It has been found that
these samples are excellent sorbents for strontium and can be used for
purification of drinkable and waste water contaminated by this hazardous
pollutant. The distribution constants Kd of the samples towards strontium are
ca. 1105 mL/g; this value is comparable with characteristics of best sorbents
for strontium [6]. Moreover, it is well known that manganese oxides are
excellent sorbents for lithium and may be prospective for extraction of
lithium cations from seawater [7]. In view of these facts, it became tempting
to test if materials obtained in our studies could exhibit electrochemical
activity in lithium cells.
Manganese (IV) oxide is known to be obtained by means of red-ox
reactions. Several methods of precipitation of MnO2 are mentioned in the
literature sources: either reduction of manganese (VII) (permanganate) or
manganese (VI) compounds, for example, by means of hydrochloric acid or
hydrogen peroxide,
2MnO−4 + 6HCl → 2MnO 2 + 3Cl2 ↑ +2H 2O + 2OH− , (1)
2MnO−4 + 3H 2O 2 + 2H 2O → 2MnO 2 + 3O 2 ↑ +4OH− , (2)
or oxidation of salts of manganese (II) or manganese (III), for example,

using hydrogen peroxide
Mn 2+ + H 2O 2 → MnO 2 + 2H+
Sometimes, reactions (1) and (3) are combined: permanganate serves

as oxidant and a manganese-II salt as a reducing agent,
3Mn 2+ + 2MnO−4 + 2H 2O → 5MnO 2 + 4H+ . (4)
Main difference between the approaches used in [1-3] and in our

work is as follows: in [1, 2], the authors operate with nonaqueous solvents:
in order to obtain amorphous manganese oxide for battery applications, they
reduce permanganate with lithium iodide and various alcohols in acetonitrile
or propylene carbonate. In [3], the synthesis of amorphous manganese oxide
for supercapacitor applications is made by mixing of KMnO4 and
Mn(CH3COO)2. Since in our studies we are largely interested in mixed
MnO2-TiO2 oxides, the best synthetic approach employs reaction (3):
oxidation of divalent manganese salts in aqueous solutions with hydrogen
peroxide.
476
The discharge cycle of a lithium-ion battery with a manganese oxide

cathode implies diffusion of lithium ions into the bulk of the cathode
material forming lithium manganese spinel, and just structural limitations to
the diffusion process can lead to the mentioned capacity fade during cycling.
In order to facilitate the formation of spinel, it is desirable to create cavities
of prescribed size in the structure of manganese oxide, so as to accommodate
incoming lithium ions. To reach this goal, principles of the so-called
template synthesis can be used. Tunnels or layers are arranged by using
template ions (Li+ in our case) to direct the cavity dimensions in embryos
formed by means of co-precipitation. Depending on future applications,
templates may be removed from the tunnel or inter-layer space by means of
water (or acid) leaching.
As follows from the results obtained, numerous parameters influence
the quality of precipitates, vs raw materials used (such as, manganese
sulphate, nitrate, chloride); pH of initial and final solutions; the excess of the
oxidant; the type of a neutralizing agent (lithium, sodium, potassium
hydroxide); the washing method, and pH of leachates. More detailed
description of the synthetic routine will be published elsewhere.
3. CHARACTERISATION OF MATERIALS OBTAINED

Air-dry samples have been studied using the methods of scanning
electron microscopy, X-ray microanalysis, thermal analysis and X-ray
diffraction. The latter two methods are crucial for conclusions regarding
thermal pre-treatment of samples to be exploited in power sources.
It should be stressed that precipitates obtained do not follow the
stoichiometric formula of MnO2. Hydrous manganese oxides, depending on
the conditions of synthesis, contain manganese in lower oxidation states, and
their formula can be written as MnO2-G. Actually, if the presence of water in
the precipitate is taken into account, its formula should be written as MnOn-
z(OH)2mH2O (z ≤ n). One can easily understand that, depending on z and n,
substances of the variable degree of hydration and the number of hydroxyl
groups could be obtained.
It is considered that hydrous manganese dioxides and their
derivatives used as sorbents belong to the crystallographic group of D-MnO2,
where the atomic ratio O:Mn<2, and the presence of side ions is a
prerequisite of the very existence of this type of the crystalline lattice. For
example, in various synthetic samples, the O:Mn ratio can vary from 1.95 to
1.70. This means that the oxidation state of manganese in synthetic
manganese oxides is seldom equal to 4.
As follows from the data of scanning electron microscopy, the mean
size of particles is at around 2-5Pm. The specific surface area of the
amorphous material is of two orders of magnitude greater than that of
477
crystalline samples. X-ray measurements and thermal analysis reveal that

crystallization of amorphous manganese oxide does not occur up to 350qC.
Even at this and the higher temperatures (up to 550qC), the intensity of
reflexes from Mn2O3, if present, is much higher than from MnO2. The
existence of Mn2O3 can be easily detected by the increase of the mass of the
sample due to an oxidation process taking place at T>350qC; however, we
estimate an initial content of Mn2O3 as less than 3 %. Up to T|600qC, MnO2
remains amorphous or, at least, semi-amorphous, and does not crystallize.
Upon further heating it loses oxygen, and transforms into Mn2O3. Finally,
after 1000qC, all material undergoes transformation into Mn3O4. On the one
hand, these data are significant as a proof of the amorphous nature of
samples. On the other hand, they enable one to follow standard thermal
treatment procedures established in manufacturing of power sources: for
amorphous samples, thermal treatment at ca. 400qC removes all traces of
Mn2O3.
4. ELECTROCHEMICAL TESTS
Preliminary electrochemical tests of materials obtained have been
performed in two types of cells. Primary discharge measurements have been
executed in standard 2325 coin-type cells (23 mm diameter and 2.5 mm
height) with an electrolyte based on propylene carbonate – dimethoxyethane
solution of LiClO4. Cathode materials have been prepared from thermally
treated amorphous manganese oxide in question (0.70r0.02g, 85wt%.)
mixed with a conductive additive (10 % wt.) and a binder (5wt%). Lithium
anodes of 0.45 mm thickness have been of slightly excess mass if compared
to the stoichiometric amount, so as to ensure maximal possible capacity of a
cell and full consumption of the cathode material.
For the sake of comparison, similar cells have been made containing
electrochemical manganese dioxide produced by Chiaturi plant, Georgia.
Just this material is employed in practically all commercial lithium power
sources manufactured in the countries of the former Soviet Union.
Some data obtained at I=0.5 mA (fast discharge) are shown in
Figure 1. It is well seen from this figure and Table 1 that cells with
amorphous manganese oxide demonstrate higher capacity than those with
electrochemical manganese oxide.
Promising results have also been obtained in cycling ability studies
performed in model coin-type cells with this same amorphous cathode
material. Moreover, it has been found that either thermally treated or even
initial amorphous manganese oxide exhibit satisfactory results. Cyclic
voltammetric studies have been carried out with the LP-30 electrolyte
(Merck). Cathode materials have been prepared from amorphous manganese
oxide in question (80wt%) mixed with a conductive additive (10wt%) and a
478
binder (20wt%). Some results of these measurements are summarized in

Table 2. They reveal that amorphous manganese oxide can be reversibly
intercalated and deintercalated with lithium. The best values of the specific
capacity obtained (ca. 40 mAh/g) are comparable with the standard capacity
values of a commercial cathode material, lithiated manganese oxide
(LiMn2O4), equal to 120-140 mAh/g.
3.5
crystalline
amorphous
3.0
U/V
2.5
2.0
0 50 100 150 200 250

Q / mAh/g
Figure 1. Discharge curves for 2325 coin cells with lithium metal anodes and electrolytic
(crystalline) vs amorphous manganese oxide-based cathodes.
For the latter sample, the best results are presented.
Table 1. Performance parameters of 2325 coin cells,

as obtained from galvanostatic cycling regime (0.5 mA).
Short circuit Capacity / Specific

Cathode material Umean/V current / mA mAh capacity /
mAh/g
Crystalline MnO2 2.7 150 134 189

Amorphous MnO2 2.6 120 175 247
Table 2. Cycling ability of amorphous manganese dioxide at various sweep scan rates.
Specific capacity / mAh/g

Cathode 0.5 mV/s 0.2 mV/s 0.1 mV/s
material Anodic Cathodic Anodic Cathodic Anodic Cathodic
Before heat 14.62 12.81 14.89 15.00 - -
treatment
After heat - - 42.92 40.43 42.30 39.56
treatment
479
5. CONCLUSION
Preliminary results presented in this paper seem to prove that
amorphous manganese oxide can serve as a good cathode material in lithium
battery applications. Preliminary galvanostatic tests reveal better
performance of this material in model coin-type cells with lithium anodes, if
compared with conventional electrochemical manganese oxide. Cycling
studies demonstrate an ability of amorphous manganese oxide to reversibly
intercalate and deintercalate lithium. Of course, much work is to be done in
order to optimize the synthetic routine, so as to improve performance of the
amorphous material, especially on cycling. On the other hand, it is clearly
seen that simple aqueous chemistry routes are quite prospective for obtaining
manganese oxide based cathode materials and can successfully compete with
refined synthetic methods based on the use of non-aqueous solvents.
ACKNOWLEDGEMENTS
This research work has been funded by EC within the framework of
the STCU Grant #GR85. S.A.K. is grateful to CRDF for financial support for
visiting the NATO-CARWC. O.I. Pendelyuk acknowledges partial support
of her stay at Argonne obtained through the U.S. Department of Energy
Office of FreedomCAR and Vehicle Technologies, via the Lawrence
Berkeley National Laboratory.
REFERENCES
1. Manthiram, A., Kim, J., Im, D., Amorphous Manganese Oxides for Lithium Ion
Batteries, ITE Letters, 2, 318-325 (2001).
2. Im, D., Manthiram, A., Amorphous Manganese Oxide Cathodes for Rechargeable
Lithium Batteries, Ceramic Trans., 127, 205 (2002).
3. Lee, H. Y., Goodenough, J. B., Supercapacitor Behavior with KCl Electrolyte, J. Solid
State Chem., 104, 220-223 (1999).
4. Hobbs, D. T., Blume, M. S., Tracker, H. L., Screening Evaluation of Alternate Sorbents
and Methods for Strontium and Actinide Removal from Alkaline Salt Solution, WSRC-
TR-2001-00072, 2001, 20 p.p. Available at http://sti.srs.gov/ fulltext/tr2001072/
tr2001072.html.
5. Lesnichaya, T. V., Visloguzova, N. M., Pendelyuk, O. I., Dzanashvili, D. I., Romanova,
I. V., Farbun, I. A., Zazhigalov, V. A., Marsagishvili, T. A., Kirillov, S. A., Optimal
design of nanodispersed amorphous manganese oxide as a sorbent towards strontium,
Trace Element Admixtures in Water, Trans. of Int. Seminar, Kyiv, 2003, p.p. 35-36.
6. Behrens, E. A., Sylvester, P., Clearfield, A., Assessment of a Sodium Monotitanate and
Pharmacosiderite-Type Ion Exchangers for Strontium and Cesium Removal from DOE
Waste Simulants, Environ. Sci. Technol. 32, 101-107 (1998).
7. Ooi, K., Miyai, Y., Katoh, S., Maeda, H., Abe, M., Topotactic Li+ Insertion to O-MnO2 in
the Aqueous Phase, Langmuir, 5, 150-157 (1989).
INVESTIGATION OF CATHODIC MATERIALS
BASED ON DIFFERENT TYPES
OF MnO2/CARBON
I. S. Makyeyeva 1*, N. D. Ivanova 2, G. V. Sokolsky 3

1
Kiev National University of Technologies and Design, Ukraine
2
Institute of General and Inorganic Chemistry of NAS, Kiev, Ukraine
3
National Aviation University, Kiev, Ukraine
Abstract
It is shown that electrolytic manganese dioxide, which has been

obtained from fluorine-containing electrolytes differs from traditional types
of MnO2 obtained by state-of-the-art synthesis methods. This material is
characterized by the increased amount of structural defects. It is shown that
crystalline structure with a large number of defects has a higher catalytic and
electrochemical activity.
The electrochemical testing in coin cells for systems of MnO2 – Zn
and MnO2 – Li has been carried out. The comparison of characteristics of
lithium elements of CR2325 configuration with known analogues
demonstrated an increase by 50% in discharge capacity. The independent
comparative testing of the experimental samples of our power sources of
1142 size and Toshiba’s L 43 was carried out. It was shown that power
sources with the manganese dioxide cathode obtained from fluorine-
containing electrolytes have demonstrated 30% higher capacity in MnO2 –
Zn system. Such outstanding results are explained by apparent higher
catalytic and electrochemical activity of subject MnO2.
1. INTRODUCTION
Manganese dioxide has found wide use in different electrochemical
power sources as cathode material. It is known that the structure and
properties of manganese dioxide depend on the method of its synthesis.
*
481
482
Among five crystalline forms of manganese dioxide electrochemical Ȗ-MnO2

is known to be the most electrochemically active.
Manganese dioxide is the most extensively used electrode for
different types of power sources (alkaline primary and rechargeable, lithium
– and magnesium – based batteries). The production of power sources with
manganese dioxide accounts for about 70 – 80 % of the world total output.
In these system electrolytic Ȗ – manganese dioxide is mainly used [1].
In [2] the authors suggested that naturally occurring and synthetic Ȗ
– MnO2 phases used in batteries might contain lattice defects such as
dislocation, chain defects and other irregular voids and that water, as well as
impurity ions, may be accommodated in such lattice defects. They stated that
such defects might account for the high water content and high
electrochemical activity of Ȗ – MnO2.
In view of this situation we studied a number of samples of MnO2
with different composition and various OH- contents in order to estimate the
correlation between the activity of MnO2 and concentration of OH- ions.
This compound can be electrochemically deposited on the anode from
various aqueous solutions, but electrolytes with sulfate and ammonia sulfate
have found widest application [3]. It has been determined that the
composition and structural parameters of the end-product are governed by
the presence of fluoride ion in electrolyte.
2. EXPERIMENTAL
We investigated the following samples: ʋ 1 – has been synthesized
by electrochemical method from a manganese sulfate-containing electrolyte
with additions of fluoride ion; ʋ 2 – Battery grade Electrolytic Manganese
Dioxide (EMD–2, Georgia); ʋ 3 – Battery grade Electrolytic Manganese
Dioxide Ȗ–MnO2 (Kerr – McGee Corp., USA); ʋ 4 – chemical MnO2
(Pridneprovsky Chemical plant, Ukraine); ʋ 5 – commercial chemical ȕ –
MnO2 (Ukraine).
Chemical analysis of the investigated samples has been carried out
using standard methods [4]. The combined water content was determined by
DTA and Infrared spectral studies in the range of 200 – 4000 cm-1. X – ray
diffraction analysis was carried out with Dron – 3 using MoKĮ radiation.
The discharge test of manganese oxide compounds was performed at
a constant external load at room temperature. Electrochemical behaviour was
studied too. Manganese dioxide compounds were mixed thoroughly with
battery-grade expanded graphite FormulaBT ABG1025 (Superior Graphite
Co., USA) at the ratio 9:1. The purpose of graphite additions is conductivity
enhancement. Conductive graphite has a mean particle size of 25Pm. In
order to stabilize the shape of the electrodes, 0,5wt% of a binder (Teflon
483
suspension) had to be added (only for sample ʋ3). The resulting active
material was pressed at 10MPa into the test cells.

It has been observed that MnO2 samples were dramatically different
from each other. They have different composition, structure and properties,
some of which can be seen in Table 1.
Table 1. Characteristics of Manganese Dioxide Samples.
Sample Mn4+, Mn3+, Mn2+, X– H2O, OH-, k.105,

ʋ % % % MnOx % % s-1
1 50,8 5,4 0,28 1,95 9,9 2,2 21,7

2 58,0 1,8 0,05 1,97 4,9 2,1 10
3 89 - - - 5,3 2,0 9,4
4 61,0 1,7 0,75 1,98 3 1,3 18,1
5 62,1 1,74 - 1,99 0,4 0.2 3
The first sample has the greatest deviation from stoichiometry.

Samples ʋ1 - ʋ4 have been found to be nonstoichiometric oxide –
hydroxide type. They have a mixed conductivity – ionic (ıi) and electronic
(ıe). The ionic one is due to the presence of OH- - groups. Namely, they
stabilize the defects of chemical nature in such compounds. These defects
are determined by the presence of Mn4+ and Mn3+ in the same
crystallographic position.
There is the correlation between water and OH- – groups content,
ionic conductivity and catalytic activity of compound, Figure 1, Table 1. The
data shows that the concentration of OH- - groups tends to increase in the
order 1>2>3>4>5. In this order increases the catalytic activity too (value K,
Table 1). This difference in the activity between samples seems to be related
to the difference in the OH- - group content and Mn3+/Mn4+ concentration.
This means that increasing of structure defects may lead to increasing of
activity of compound. Additional structure distortion has been obtained in
modified sample by insertion of small amount of ions of alkaline metals
(sample ʋ 4).
The observed higher rate of mass transfer in nonstoichiometric
oxides makes it possible to operate at discharge current densities on the order
of 2 – 5 mA/cm2, which is several times greater than that used with the
conventional manganese dioxide. Thus, the purposeful disordering of
structure of a cathode material of manganese dioxide allows to utilize such
power sources in extreme conditions – for element discharge at a low
484
resistance (from 0,3 kOm), Figure 1, a, b. This result indicates that electric
parameters of power sources are in agreement with data on the degree of
deviation from stoichiometric composition.
The comparative testing of the experimental samples of power
sources of LR2325 size (system MnO2-Li) and 1142 size (system MnO2-Zn),
Tables 2, 3, support the above conclusion.
Table 2. Specific capacity, mAh/g, of LR2325 batteries

based on various MnO2 samples at different external loads (MnO2 – Zn).
Load, Sample / Performance

K:
1 2 3 5
0,3 238 136 110 -

1,0 258 177 242 -
5,6 289 245 262 107
Table 3. Technical Characteristics of LR2325-size MnO2-Li and 1142-size MnO2-Zn

cells based on MnO2 synthesized electrochemically from
manganese sulfate-containing electrolytes with additions of fluoride ion.
Electrochemical A.C. R, Q, SE, PD,

System Voltage, : mA h/g Wh/kg W/kg
V
MnO2 – Li 3,5 5600 200 539 1,5

MnO2 – Zn 1,5 5600 280 151 0,25
Table 4. Main performance characteristics of IGIC’s 1142-size MnO2-Zn cells based on

MnO2 synthesized electrochemically from manganese sulfate-containing electrolytes with
additions of fluoride ion vs baseline LR43 Toshib’a cells (cell sizes are similar).
Cell size Diameter, Height, Unom, Uend, Inom, Cnom,

designation mm mm V V mA mAh
LR43 (Toshiba) 11,6 4,2 1,5 1,2 0,12 70
1142 (IGIC) 11,6 4,2 1,5 0,9 5 100

485
a)
b)
Figure 1. Discharge curves of MnO2 – Zn power sources (IGIC 1142 cell size)
at the low load (0,3 k: per cell); (a) - sample ʋ 1; (b) - sample ʋ 3.
The independent comparative testing of the experimental samples of

the internally-made power sources of IGIC 1142 size and the Toshiba`s L43
(both size designations stand for similar cell sizes) were carried out. It was
shown that power sources containing manganese dioxide cathode obtained
486
from fluorine containing electrolyte (sample 1) have 30 % higher capacity in

the MnO2 – Zn system, Table 4.
4. CONCLUSIONS
The effect of structural defects in various types of MnO2 on the
electrochemical behavior has been investigated. It was shown that the basic
types of defects in Mn3+ are in the same crystallographic position as in
Mn4+. Defects were seen to be are stabilized by the OH- - groups. The
catalytic and electrochemical activity of MnO2 increases with increase in
defect content.
A form of electrolytic manganese dioxide obtained by us from
fluorine-containing electrolytes was seen to produce the best cathode active
material for Zn/MnO2 and Li/MnO2 types of power sources.
According to the results presented, it appears that the higher the
number of lattice defects in MnO2 structure, the higher the discharge
capacity will be.
ACKNOWLEDGEMENTS
The authors are grateful to Professor V. Barsukov (Kiev National
University of Technologies and Design, Ukraine) and Dr. I. Barsukov
(Superior Graphite Co., USA) for support of this work.
REFERENCES
1. G. Piao, M. Yoshio, H. Noguchi and A. Kozava. // J. Power Sources, 51 (1994) 391-395.
2. R. G. Burns, A. Kozava and M. Nagayama // Battery Materials Sump., Brussels,
Belgium, 1983, 7-10.
3. K. Takahashi. // Electrochimica Acta, 26, 10 (1981) 1467-1476.
4. A. I. Vogel. // Quantitative Inorganic Analysis (1981).
INVESTIGATION OF THIN-FILM
ELECTRODE MATERIALS AS
CATHODIC ACTIVES FOR POWER SOURCES
Nataliya E. Vlasenko*, Nataliya D. Ivanova,

Evgeniy I. Boldyrev, Olga A. Stadnik
V.I. Vernadskyi Institute of General and Inorganic Chemistry
National Academy of Sciences of Ukraine
32/34 prospect Palladina, 03680 Kyiv-142, Ukraine
Abstract
The electrochemical behavior of thin-film oxide-hydroxide

electrodes containing chromium, nickel and cobalt compounds was
investigated. Experimental results have shown that such compounds can be
successfully used as active cathodic materials in a number of emerging
primary and secondary battery applications.
Keywords
oxides of chromium, nickel and cobalt; cathodic materials; thin

films.
1. INTRODUCTION
Oxide compounds are widely used as cathodic materials in the
power sources and electrochemical generators. Some literature data indicates
that cathodic materials based on nonstoichiometric oxide compounds make it
possible to increase the solid-phase reduction process. The kinetics of
electrochemical reactions and consequently the current density are the
higher, the greater the degree of deviation from stoichiometry, and the lager
the number of the defects in the compounds’ structure [1, 2].
In the rechargeable lithium-ion batteries, in which current-generating
reactions are based on a ''rocking chair'' mechanism, the ions of lithium shift
back and forth between the intercalation hosts of the cathode and anode. The
*
Corresponding author. E-mail: [email protected], [email protected]
487
488
lithium-ion secondary battery mainly consists of a carbonaceous or lithium-

alloy anode and a complex transition metal oxide as the cathode material.
(Figure 1).
Figure 1. Recharging mechanism of batteries with the ''rocking chair'' concept (a)
and structure of a rolled type battery (b), [1].
The use of thin-film cathodes for battery application usually results

in a better performance due to shorter diffusion path of intercalated cation
through solid matrix. The thin film electrodes are used in manufacturing of
rolled type batteries and thin film cells.
The rolled type batteries are general-purpose batteries with high
capacity and good characteristics. Thin film power sources have found a
number of specific applications in the fields of portable accumulators with
high energy density for application in electronic and integral circuits,
sensors, monitors and for medicine [3].
Thin-film electrodes have promise in the development of flexible
power sources (primary and secondary). It must be taken into account that
change in cathodic material crystal lattice must not be over 20% as a result
of intercalation of Li+, Na+ or H+ [4]. The electrodes must be chemically
active and have both electron and ion conductivity. In connection with these
489
demands we have chosen oxide-hydroxide compounds of Cr, Ni, Co as the

object of investigations.
This paper proposes utilization of such compounds as cathodic
materials and identifying dependencies of charge-discharge processes on the
composition of these compounds. We have studied the process of
electrodeposition of thin films of nickel, chromium and cobalt oxides from
aqueous solutions of salts of 3d-metals. The influence of electrolysis
conditions and potentials on surface phase composition has been
investigated. The original method of electrochemical synthesis [5] developed
by us allows nonstechiometric oxide-hydroxide compounds of transition
metals to be obtained. For film production we use cathodic and anodic
electrodeposition on metal surface. For deposition of dense complex oxide
films we have developed a number of original fluorine-containing electrolyte
systems. The use of such complex electrolyte effects in deposition of 3d
metals oxide thin films gives good electrochemical properties.
The mentioned method for synthesis of oxide-hydroxide compounds
(Ni, Cr, Co) is more controllable and enables with production of electrode
films definite amounts of components. Therefore it guarantees the
reproducibility of their compositions and properties. Using the above method
we were able to produce the following oxide compounds: zero valence metal
and lowest oxidation state oxide-hydroxide compounds; in cathode process
and oxide-hydroxide compounds; (in anode process the oxide compounds
consist of highest oxidation state oxide-hydroxide compounds). Both type
compounds possesses electronic and ionic conductivity.
2. EXPERIMENTAL
The film electrodeposition process was studied by means of linear
sweep voltammetry. The rate of electrochemical reaction was determined
from current density (current-potential curves). The film deposits were
characterized by chemical analysis, IR - spectroscopy, XRD, TG, TGA and
SEM methods.
The electrochemical activity of the products (Li intercalation) was
studied in experimental coin cells of 2016 size. In the case of precipitated
powder of Ni the electrode active masses were prepared by mixing the active
material with 10 wt.% of graphite (conductive additive) and 5 wt.% of
polyvinylidene fluoride (binder) stirred in 1-methyl-pyrrolidone solvent.
Thin oxide films on steel lead were used as cathode in the case of dense and
solid film covers. 1M LiPF6 solution in mixture of ethylene carbonate (50
vol.%) and methyl carbonate was used as an electrolyte (Merck product
LP30). The assembled cells were then tested using the method of linear
sweep voltammetry (LSV) (potentiostat PI50) and/or galvanostatic charge-
discharge (PC cotiolled automatic bench) methods. In the case of thin film of
490
oxide-hydroxide of chromium and cobalt, the electrode active masses where

used directly without conductive additives and binders. The electrolyte
contained 1M LiClO4 in propylene carbonate (PC).

On the basis of obtained data of cyclic voltammograms for 3d metals
oxides electrodeposition the optimal conditions (current density, potential,
process time, electrolyte composition, temperature) for dense oxide films
(Ni, Cr and Co) deposition on steel foil have been elaborated. Data relating
to several best films are summarized in Table 1.
0
0
3
3
Mass loss, %
6
Mass loss, %
9
9
12
12
15
15 0 100 200 300 400 500
0 100 200 300 400 500 Temperature, C
Temperature, C
Figure 2. TG (1) and DTA (2) curves for samples 4 (a) and 6 (b).
According to TG and TGA measurements for all obtained samples

(Figure 2) two steps of mass loss were observed. The first step in 80-190qC
temperature range is an effect of bound water loss:
MeOx(OH)y·zH2O o MeOx(OH)y + zH2On;
the second one is hydroxide water formation:
MeOx(OH)y o MeOx + yH2On.

491
As it can be seen from reaction mechanisms, all water can be

removed at up to 300-350qC. The small exothermal area on DTA curve at
400-450qC can be attributed to recrystallization processes after H2O and OH–
groups removal.
Table 1. Thin-film oxide chemical composition.

Deposited
OH- - Bound O2
Main components of metal oxide- Metal
group water cont.
electrolyte, g/l hydroxide content,
content content %
compounds %
% %
1 CCrO3= 300, CHF= 1.2. Cr 70.0 6.0 5.0 19.0

2 CCrO3= 300, CHF= 0.8. Cr 73.0 5.0 4.5 17.5
3 CCrO3= 300, CHF= 0.4. Cr 82.0 4.5 4.0 9.5
4 CNiSO4·7H2O =120, Ni 69.6 6.1 5.1 19.2
CHF=10.
5 CNiSO4·7H2O=120, Ni 72.3 5.2 4.6 17.9
CHF= 8.
6 CNiSO4·7H2O =120, Ni 78.4 4.6 4.2 12.8
CHF= 6.
7 CCoSO4·7H2O=250, Co 71.0 5.8 5.2 18.0
CHF= 6.
8 CCoSO4·7H2O=250, Co 73.5 5.5 5.0 16.0
CHF= 4.
9 CCoSO4·7H2O=250, Co 80.4 4.4 3.8 11.7
CHF= 2.
10 CCoSO4·7H2O=10, Co 46 8.5 10.5 36
CNH4F=20.
11 CCoSO4·7H2O=10, Co 45 9.0 11 35
CNH4F=25
12 CCoSO4·7H2O=10, Co 44 9.5 12.0 34.5
CNH4F=30.
Figure 3. Microphotograph of a thin-film Figure 4. Microphotograph of a thin-film

nickel oxide representing sample # 4. nickel oxide representing sample # 6.
492
250
4.5
4 200
Specific capacity, mAh/g

Potential vs. Li, V
3.5
150
3 Discharge 1 Discharge 4
Charge 1 Discharge 5
100
Discharge 40 Charge 4
2.5 Charge 40 Charge 5
50
2
0 10 20 30 40 50
0 20 40 60
Cycle number
Time, min
Figure 5. Specific capacities vs. Figure 6. Charge discharge profile (1-st and
cycle number for samples #4 and #5 40-th cycles) for sample #1. Charge and
in the 2,5-4,2V range. discharge currents are 100 mA and 80 mA.
The mass losses for all samples are in a good agreement with values
of water and hydroxide group content obtained by chemical analysis
(Table 1). Thus, for electrochemical testing of Li- intercalation activity each
sample was heated at 350qC to remove all types of bound water.
As it can be seen (Figures 3 and 4), film morphology is strongly
dependent on deposition conditions. The film with lower oxygen content has
larger crystallites and defects. The electrochemical activity for lithium
intercalation for samples obtained from fluorine-containing electrolyte with
higher F- content is much higher. The electrochemical data are summarised
in Figures 5 and 6, Table 2.
Table 2. Initial electrochemical capacity and cycle efficiency for different samples.
Sample Specific capacity mAh/g, Cd/Cch,

# first discharge second cycle
1 205 0.971
2 191 0.964
4 234 0.985
5 210 0.979
7 208 0.966
8 188 0.962
It is shown that the first step in the electrochemical reduction of

cathodic material CoxOy(OH)z in power sources is controlled by kinetics.
This process at the interface depends on the catalytic activity of compound,
Figure 7 (a) the reduction of the bulk of cathodic materials is controlled by
diffusion Figure 7 (b).
493
K,(mol·c)-1 9 I, mA
0.00026
8.5
8
0.00022 7.5
7
6.5
0.00018
6
5.5
0.00014
5
Q,mKl 4.5
vV
0.0001 4
0 40 80 120 160 1 1.5 2 2.5 3 3.5 4 4.5 5
(a) (b)
Figure 7. Dependence of charge of the catalytic activity of cobalt oxide-hydroxide (a), I-¥V
dependence of cobalt oxide-hydroxide electroreduction (b).
4. CONCLUSIONS
We managed to obtain dense and solid thin films of 3d-metal oxides
using the techniques of electrochemical deposition from aqueous fluorine-
containing electrolytes. The films have been studied as a possible cathode
material for secondary cells. The best samples show good cycle retention and
acceptable specific capacity in the range of 180 mAh/g. They also feature a
plateau of electrochemical potential at approximately 3,5 V, which is
acceptable for present industrially produced electrochemical devices.
REFERENCES
1. J. O. Besenhard, M. Schwake, 4th Int. Meet. Li batteries. Progr. And Extended Abstr.
May 24-27 (1988).
2. J. V. Goodenough 11th Int. Meet. Li batteries. Munchen (1992).
3. E. A. Nignikovskii Electrochemiya, 34, 7, 722 (1998).
4. I.A. Kedrinsky, V.E.Dmitrenko, I.I.Grydyanov. Li batteries. M.:Energoizdat 1992. 240p.
5. N.D.Ivanova, S.V.Ivanov. Russian Chemical Reviews 62(10) 907-918 (1993).
SYNTHESIS OF MIXED OXIDES
USING POLYBASIC CARBOXYLIC HYDROXY-
AND AMINO-ACID ROUTES:
PROBLEMS AND PROSPECTS
S. A. Kirillov*, I. V. Romanova, and I. A. Farbun

Institute for Sorption and Problems of Endoecology,
13, Gen. Naumov St., 03164 Kyiv, Ukraine
Abstract
Polybasic carboxylic hydroxy and amino acid aided synthetic routes

directed towards obtaining mixed inorganic materials, especially for battery
and fuel cell applications, are overviewed. It has been shown that, in spite of
enormous number of papers on the subject, significant efforts should be
undertaken in order to understand the basic principles of these routes.
Possible influence of the structure of reactants employed in the process
(acids, poly hydroxy alcohols, metal salts) is put forward, and some
directions of future work in the field are outlined.
1. INTRODUCTION
Transition-metal oxides and their mixtures are widely employed in
numerous industrial applications, especially as cathode materials for
batteries and fuel cells [1,2]. Practice poses certain well-known requirements
to oxide materials, first of all, to uniformity of the size distribution of
particles, to homogeneity of mixed oxides, etc. To meet these demands, two
broad categories of methods are now in use, vs (i) mechanical methods and
(ii) chemical methods.
In the first group of methods, small particles are produced from
larger ones by mechanical treatment. Naturally, it is an uneasy task to obtain
homogeneous powders with uniform size distribution by means of
comminution, therefore, in modern applications, chemical methods prevail.
*
495
496
The chemical methods can be divided into three types, in accord with three
aggregate states of the matter: (i) gas phase reactions, (ii) liquid state
reactions, and (iii) solid state reactions.
Gas phase reactions are not so common in producing solid oxides
and will not be considered here. As far as liquid state reactions are
concerned, these can be performed in aqueous and non-aqueous solvents and
can be subdivided into (i) precipitation and co-precipitation, including the
so-called template synthesis, (ii) evaporation, and (iii) sol-gel methods.
Precipitation is the most straightforward approach to the preparation
of oxides, using hydrous oxides, carbonates, oxalates and other insoluble
salts as precursors, with further thermal treatment (decomposition) of
precursors. Co-precipitation is often utilised in order to obtain mixed oxides
or to create tunnels or layers in the mother structure by inserting guest
(template) ions. Depending on future applications, templates may be
removed from the tunnel or inter-layer space by means of water (or acid)
treatment. Precipitation (co-precipitation) can also be performed by means of
complete evaporation of a solvent. Quite close to this method is the so-called
freeze-drying, where the solution in the form of aerosol is sprayed into a
coolant agent (e.g., liquid nitrogen). Another approach based on evaporation
is spray drying where a solution is sprayed into a hot chamber.
In precipitation reactions, powder characteristics depend on the
speed of the nucleation of particles and their growth due to the mass flow to
the surfaces. In freeze drying and spray drying, powder characteristics
primarily depend on the size of droplets, which in turn is determined by the
parameters of a nozzle and characteristics of the flow of a carrier gas. Both
these methods enable one to obtain powders with very high surface area.
Solid state reactions are also very common in producing oxide
materials and are based on thermal treatment of solid oxides, hydroxides and
metal salts (carbonates, oxalates, nitrates, sulphates, acetates, etc.) which
decompose and react forming target products and evolving gaseous products.
Solid-state chemistry states that, like in the case of precipitation, powder
characteristics depend on the speed of the nucleation of particles and their
growth; however, these processes in solids are much slower than in liquids.
There are numerous advantages and shortcomings of liquid state and
solid-state reactions. Briefly speaking, solid-state routes are often
unacceptable since it is notoriously difficult to achieve proper homogeneity
of materials. Liquid state routes are much more flexible and enable one to
find optimal precipitating agent and/or explicitly employ principles of
chemical equilibria (hydrolysis, complex formation, etc.). On the other hand,
some methods of liquid state chemistry, like freeze drying and spray drying,
require the use of quite expensive equipment and therefore cannot be utilised
on industrial scale.
497
2. SOL-GEL (PECHINI) METHODS

During past decades, a relatively new method attracted much of
attention of experimentalists, namely, the sol-gel method widely
implemented in numerous scientific and industrial applications. Actually, it
caused a breakthrough in material science: it is worth mentioning that sol-gel
routes are crucial for synthesis of nanomaterials, and one of methods of the
21st century; namely, commercial nanoporous glasses are readily available
on the market just thanks to sol-gel technologies.
In the sol-gel process, an oxide network is obtained via inorganic
polymerization reactions in solution. It offers many advantages if compared
with conventional methods, first of all, a better control over morphology and
texture of materials. Initially, it was considered that sol-gel routes are
applicable to systems like silica, titania and zirconia, where hydrosols and
hydrogels are formed due to the ability of hydrous oxides to set up a network
where positively charged ions are linked by oxygen bridges. Later on,
similar methods have been extended to almost all metal oxides: it has been
found that a network could be created by hydrogen bonds formed by anions
of polybasic carboxylic hydroxy and amino acids, and cations are
presumably held in this network due to complex formation.
An idea to use polybasic hydroxy carboxylic acids in syntheses of
oxides goes back to Pechini [3]. Evaporating solutions of metal salts in citric
acid at presence of ethylene glycol he obtained a polymeric resin as a
precursor of target oxides. Then this process was extensively used to
manufacture various ceramic oxide powders in several publications [4-8].
One of the modern descriptions of the routine referred to as Pechini
method reads as follows [5]. Citric acid forms chelates with the metal
cations. These chelates undergo esterification, when heated with a poly
hydroxy alcohol, such as ethylene glycol, at a temperature of 110-150°C to
form a polymeric precursor resin. The cations are expected to be dispersed
uniformly throughout the polymeric resin. Additional heating of the resin in
air (at 400°C) results in the removal of organics and the formation of an
amorphous product (sometimes called a char) with a controlled cation
stoichiometry, little cation segregation, and high porosity. Sintering of the
amorphous product at higher temperatures leads to the formation of oxide
ceramics.
Other descriptions of Pechini-type syntheses presented by various
authors are quite similar, and the only difference touches upon the definition
of precursors. Some authors prefer to call them “polymeric resins” (see, for
instance, [6, 8, 9]), although the terms “sol” and “gel” can be found with
even higher probability [4,10-18].
Summarising, the following distinctive features of Pechini-type
syntheses should be mentioned:
498
(1) Dissolution of metal salts in the aqueous solution of polybasic

hydroxy (or amino) carboxylic acids (and poly hydroxy alcohols);
(2) Heating (evaporation) of the solution obtained and formation of a
polymeric resin or a sol (gel) as a precursor;
(3) Decomposition of the precursor to obtain an oxide powder.
Single and mixed oxides synthesised using these procedures are

widely probed as cathode materials, superionics, magnetic materials,
catalysts, and other multipurpose materials [4-20].
In what follows, we try to shed the light on some mysterious features
of Pechini-type methods, to overview major drawbacks in synthetic routines
described in the past decade and put forward some problems to be solved in
order to design advanced oxide materials in an optimal way.
3. UNSOLVED PROBLEMS OF PECHINI-TYPE ROUTES
3.1. Role of Polybasic Carboxylic Hydroxy (amino) Acids and

Poly Hydroxy Alcohols
It should be emphasized that citric acid is not the only possible acid
employed in Pechini-type syntheses. Other polybasic carboxylic hydroxy
acids (malic, tartaric, hydroxyglutaric, etc.) and polybasic carboxylic acids
(e.g., succinic) have been probed in Refs. [4, 13-16]. As far as amino acids
are concerned, glycine seems to remain the only representative [13, 14].
However, the choice of each particular organic acid has never been justified,
and no comparative studies are performed in order to find possible
dependencies of the process (ability to form a sol, a gel, or a resin, easiness
of thermal decomposition of precursors, etc.) on the steric factors,
specifically, on the number of hydroxy and carboxylic groups in the
molecule of an acid, as well as on the length of its carbon skeleton.
It is for long known that all aforementioned acids form stable
complex (chelate) compounds with numerous, especially transition metal
cations (see, e.g., Refs. [21, 22]). This fact is pointed out in practically all
papers on the subject. However, the composition of complex compounds,
their stability, the type of bonding between metal cations and carboxylic
anions on various stages of the synthesis, and possible role of steric factors
in these phenomena are discussed in a very small number of publications.
In [20], the composition of the citrate precursor of CoFe2O4 is
proposed as Co3Fe6O4(C6H6O7)86H2O, i.e., two protons are detached from
each molecule of citric acid, and the complex compound could be classified
as an acidic salt. Distinct signatures of complex formation are obtained by
means of infrared (IR) spectroscopy and nuclear magnetic resonance (NMR)
for citrate complexes of iron and yttrium, potential precursors of YFeO4 and
499
Y3Fe5O12 [7]. These complexes have been obtained at 110-120qC, isolated,

and studied in the crystalline form and in an aqueous solution. IR spectra
suggest that these are monodentate coordinated to the metal ions. NMR data
are in accord with an assumption that hydroxy and carboxylate groups
attached to the central carbon atom of the citric acid anion are coordinated to
the metal cations forming a stable five-membered ring. However, these
important data are obtained at conditions that do not correspond to the
conditions of the synthesis of precursors.
From the point of view of complex formation, homogeneity of
mixed oxides accessible by means of Pechini-type routines could be a result
of the emergence of mixed complex compounds where two different cations
are bonded to one and the same anion. Such a conjecture formulated in Ref.
[7] yet has no definite proofs, and no attempts have been made in order to
quantify the very role of complex formation in Pechini-type syntheses.
Furthermore, the role of a poly hydroxy alcohol, like ethylene
glycol, seems ambiguous. As mentioned above, it is believed that the
presence of ethylene glycol favours esterification of chelates. IR and NMR
studies performed in [7] do not present solid proofs of such a belief.
Synthetic routines with and without alcohols look the same, and the absence
of alcohol seem not to influence the properties of precursors and final
products. Some evidence exists enabling one to consider the esterification
idea liable to more than one interpretation: it has been reported in [4] that the
presence of ethylene glycol does not influence the morphology of oxides.
Accounting for these facts one should not be surprised coming
across entirely different recommendations regarding the metal-acid-ethylene
glycol ratio, which varies from 1-1.88-1.25 [5] or 1-4-1 [8] to 1-4.25-17.5
[7]. It is even sometimes noticed that a carboxylic acid and a poly hydroxy
alcohol simply play the role of a “fuel” to burn a precursor and to form target
oxides [16]. To shed some light on the aforementioned problems, the
following steps could be of some help.
The influence of the structure of the acid could be clarified if
extensive comparative studies of homologous acids would be performed in
Pechini-type processes involving the same cations. From this point of view,
the comparison of related acids [like malonic, HOOC–CH2–COOH, succinic,
HOOC–CH2–CH2–COOH, and glutaric, HOOC–CH2–CH2–CH2–COOH), as
well as acids and hydroxy acids [like succinic, HOOC–CH2–CH2–COOH,
and hydroxysuccinic (malic), HOOC–CH(OH)–CH2–COOH] is required.
As far as the role of poly hydroxy alcohols is concerned, com-
parative studies of products obtained with and without their addition should
be made. The use of homologous alcohols (like low cost substituted glycols
employed as components of brake liquids, and glycerol) could be helpful if
such role would be found significant.
500
In order to co clarify the role of complex formation, the new data on

stability constants should be accumulated, being collected at strictly similar
conditions. It should be also mentioned that any analysis of equilibrium in
solutions involving anions of polybasic hydroxy carboxylic acids requires
the data on the deprotonation constants of the acid in question. This
information would be crucial for conclusions regarding the presence and
stability of mixed complexes in the system. Valuable knowledge about the
structure of complex compounds present in solutions (and in precursors as
well, see later) may be gained by means of vibrational spectroscopy (IR and
Raman spectra) and nuclear magnetic resonance.
3.2. Physical Nature of the Precursors

As it was mentioned above, there is no unanimous opinion
concerning the nature of precursors obtained in Pechini-type syntheses, and
these are called polymeric resins (glasses) or gels. Clearly, this contradiction
has little to do with technical aspects of the synthesis. On the other hand, its
clarification will be of great value for the physics of amorphous state.
In spite of the absence of clear physical evidence, there are little
doubts that the early stage of formation of precursors involves creation of a
network formed by hydrogen bonds between anions of polybasic carboxylic
hydroxy and amino acids. The further direction of this process upon heating
could be twofold: either formation of a sol and then a gel, or creation of a
glass. To distinguish between these two possible directions, optical
measurements, like light scattering, could be helpful. Being colloidal
systems, sols and gels are optically inhomogeneous and exhibit significant
light scattering (a Tyndall cone), whereas glasses are homogeneous and do
not scatter the visible light.
Hydrogen bonded glasses are not so numerous [23], and glycerol is
one of the best-known examples. Should the glassy nature of precursors be
proven, the list of hydrogen bonded glasses would be considerably enlarged,
and the studies of cooperative behaviour of glass forming systems (see, e.g.,
Refs. [24-28]) would get a new impetus.
3.3. Chemical Nature of Precursors, Their Behavior upon

Heating, and Possible Role of Anions of the Metal Salt
One of the most remarkable attributes of Pechini-type routines is the
behaviour of precursor resins (gels) upon heating. On the first stage of
thermal decomposition of samples, being dried at approx. 120qC, they
decompose forming very porous and voluminous sponge-cake-like materials
[6]. Probably, just at this temperature, a highly porous intermediate is
501
created, being a prerequisite of the exceptionally porous nature of target

oxides. At higher temperatures (up to 400qC) this intermediate experiences
further decomposition giving the amorphous oxide. Subsequent heating up to
1000qC is required to crystallise oxides in question and to reduce the content
of carbon to ppm levels. On the other hand, the higher the temperature of
calcination, the lower is the value of the surface area of the material obtained
[5]. No definite influence on the cation of a metal salt taken as a raw material
(nitrate [4-8,14], acetate [13-15], citrate [9,20]) on the properties of target
oxides have been noticed.
It seems well established that decomposition of precursors takes
place in two stages. Unfortunately, little efforts have been undertaken in
order to study either the chemical nature of precursors and the
decomposition process in greater detail. Say, it is found in Ref. [4] by means
of thermal analysis that decomposition of citrate and tartrate precursors
obtained from nitrates first involves the removal of water, and then the
polymer itself decays. Gas analysis during decomposition shows the
presence of H2O, CO2, and the products of the thermal degradation of the
nitrate ion present in the raw material (N2O, NO, NO2). In Ref. [20], the
citrate precursor of CoFe2O4, Co3Fe6O4(C6H6O7)86H2O, first decomposes at
120qC forming an instable complex of acetonedicarboxylate acid and
evolving H2O and CO. On further heating to 220-230qC the citrate network
breaks down giving the target oxides, CO2, and acetone. Interestingly, at
high heating rate (10°C min–1 and above), both processes can appear
simultaneously.
All these data, though being quite incomplete, call for other valuable
questions. Different carboxylic acids decompose in different ways, not
obviously evolving simplest products, like H2O, CO, and CO2, hence it is
clear that a carboxylic acid anion does not simply serve a fuel for burning. It
is therefore tempting to find if an “optimal” carboxylic acid anion exists
from the point of view of decomposition processes, and what is the most
desirable metal-anion ratio. Further, participation of the anion introduced
with the starting material is evident in the burning process. The question
arises of what kind of anions favours performing the thermal decomposition
in the most optimal way. These studies could be carried out by means of the
mentioned methods (thermal analysis, gas chromatography).
4. CONCLUSIONS
In this paper we have tried to present a brief overview of state-of-
the-art of the synthetic chemistry of oxide materials based on polybasic
carboxylic hydroxy (amino) acid routes. It has been shown that, in spite of
enormous number of papers on the subject (we have mentioned just a few
latest references, significant efforts should be undertaken in order to
502
understand the basic principles of this important method. Some gaps in our
knowledge are mentioned, possible influence of the structure of reactants
employed in the process (acids, poly hydroxy alcohols, metal salts) is put
forward, and some directions of future work in the field are outlined.
Specifically,
(1) Extensive comparative studies of salts of homologous acids
involving the same cations should be performed, aimed at
clarification of the influence of the structure of acids on the
properties of precursors and target oxides;
(2) Comparative studies of products obtained with and without addition
of poly hydroxy alcohols should be made, so as to determine their
role in the process;
(3) New data on stability constants accumulated at strictly similar
conditions are required to explain the role complex formation;
(4) Valuable knowledge about the structure of complex compounds
present in solutions and in precursors could be gained by means of
vibrational spectroscopy (IR and Raman spectra) and NMR;
(5) Precursors obtained with various anions should be studied at
comparable conditions, and the quality of target products should be
compared in order to illuminate the role of anion in thermal
decomposition processes.
It should be also mentioned that careful optical determination of the
nature of precursors and solving the gel-glass dichotomy would be of great
value to the physics of amorphous state.
Unravelling these problems one would be able to make well-
grounded predictions and to justify optimal synthetic routes for the directed
synthesis of advanced oxide materials.
ACKNOWLEDGEMENTS
This research has been funded, in part, by the EC, within the
framework of the STCU Grant #GR85. S. A. Kirillov is grateful to CRDF
Travel Grants Program for financial support of his visit to the NATO
CARWC.
REFERENCES
1. Materials for Lithium Ion Batteries, Ed. by C. Julien and Z. Stoynov, NATO Science
Series 3. High Technology, Vol. 85, Kluwer, Dordreht, 2000.
2. Kirillov, S. A., Tsiakaras, P. E., Romanova, I. V., Adsorption and oxidation of methanol
and ethanol on the surface of metallic and ceramic catalysts, J. Mol. Struct., 651-653,
365-370 (2003)
503
3. Pechini, M., Method of preparing lead and alkaline earth titanates and niobates and
coating method using the same form a capacitor, U.S.Pat. # 3330697, July 11, 1967.
4. Weidenkaff, A., Ebbinghaus, S. G., Lippert, T., Ln1-xAxCoO3 (Ln=Er, La; A=Ca,
Sr)/carbon nanotube composite materials applied for rechargeable Zn/Air batteries,
Chem. Mater. 14, 1797-1805 (2002).
5. Tas, A. C., Majewski, P. J., Aldinger, F., Chemical preparation of pure and strontium-
and/or magnesium-doped lanthanum gallate powders, J. Am. Ceram. Soc. 83, 2954-2960
(2000).
6. Alonso, J. A., Martinez-Lopez, M. J., Casais, M. T., Martinez, J. L., Demaseau, G.,
Largeteau, A., Garcia-Munoz, J. L., Munoz, A., Fernandez-Diaz, M. T., High-pressure
preparation, crystal structure, magnetic properties, and phase transitions in GdNiO3 and
DyNiO3 perovskites, Chem. Mater. 11, 2463-2469 (1999).
7. Todorovsky, D. S., Dumanova, D. G., Todorovska, R. V., Getsova, M. M., Preparation
and characterization of yttrium-iron citric acid complexes, Croat. Chem. Acta 75, 155-
164 (2002).
8. Kwon, Y.J., Kim, K.H., Lim, C.S., Shim, K.B., Characterization of ZnO nanopowders
synthesized by the polymerized complex method via an organochemical route, Journ.
Ceram. Proc. Res. 3, 146-149 (2002).
9. Castro, R.H.R., Hidalgo, P., Mucillo, R., Gouvea, D., Microstructure and structure of
NiO-SnO2 and Fe2O3-SnO2 systems, Appl. Surf. Sci. 214, 172-177 (2003).
10. Pereira-Ramos, J.-P., Electrochemical properties of cathodic materials synthesized by
low-temperature techniques, J. Power Sources 54, 120-126 (1995).
11. Franger, S., Bach, S., Pereira-Ramos, J.-P, Baffier, N., Highly rechargeable LixMnO2+G
oxides synthesized via low temperature techniques, J. Power Sources 97-98, 344-348
(2001).
12. Franger, S., Bach, S., Farcy, J., Pereira-Ramos, J.-P, Baffier, N., Synthesis, structural and
electrochemical characterization of the sol-gel birnessite Mn1,840,6H2O, J. Power
Sources 109, 344-348 (2001).
13. Julien, C., Letranchant, C., Rangan, S., Lemal, M., Ziolkiewicz, S., Castro-Garcia, S., El-
Farh, L., Benkaddour, M., Layered LiNi0,5Co0,5O2 cathode materials grown by soft-
chemistry via various solution methods, Mater. Sci. Eng. B76, 145-165 (2000).
14. Julien, C., Letranchant, C., Lemal, M., Ziolkiewicz, S., Castro-Garcia, S., Layered
LiNi0,5Co0,5O2 compounds synthesized by a glycine-assisted combustion method for
lithium batteries, J. Mater. Sci. 37, 2367-2375 (2002).
15. Castro-Garcia, S., Julien, C., Senaris-Rodriguez, M. A., Structural and electrochemical
properties of Li-Ni-Co oxides synthesized by wet chemistry via a succinic-acid-assisted
technique, Int. J. Inorg. Mat. 3, 323-329 (2001).
16. Prabahan, S. R. S., Saparil, N. B., Michael, S. S., Massot, M., Julien, C., Soft-chemistry
syntethis of electrochemically-active spinel LiMn2O4 for Li-ion batteries, Solid State
Ionics 112, 25-34 (1998).
17. Chitra, S., Kalyani, P., Vebka, B., Mohan, T., Haro-Poniatowski, E., Gangadharan, R.,
Julien, C., Synthesis, characterization and electrochemical studies of LiNiVO4 cathode
material in rechargeable lithium batteries, Mater. Chem. Phys. 65, 32-37 (2000).
18. Julien, C., Camacho-Lopez, M. A., Mohan, T., Chitra, S., Kalyani, P., Gopukumar, S.,
Combustion synthesis and characterization of substituted lithium cobalt oxides in lithium
batteries, Solid State Ionics 135, 241-248 (2000).
19. Prasad, S., Vijayalakshmi, A., Gajbhiye, N.S., Synthesis of ultrafine cobalt ferrite by
thermal decomposition of citrate precursor, J. Therm. Anal. Calorim. 52, 5-10 (1998).
20. Kostromina, N. A., Complex compounds of cerium, neodim, and samarium with
glyconic acid, J. Inorg. Chem. (Russ). 5, 95-101 (1960).
21. Kostromina, N. A., Comparative stability of glyconate complexes of rare earth metals,
Ukr. Khim. Zhurn. 26, 299-304 (1960).
22. Frontiers in Material Science: Glasses and Amorphous Materials, Special Issue, Science
267, 1924–1953 (1995).
504
23. Kirillov, S. A., Spatial disorder and low-frequency Raman pattern of amorphous solid,
with special reference to quasi-elastic scattering and its relation to Boson peak, J. Mol.
Struct. 479, 279-284 (1999).
24. Kirillov, S. A., Yannopoulos, S. N., Charge-current contribution to low-frequency
Raman scattering from glass-forming ionic liquids, Phys. Rev. B, 61, 11391-11399
(2000).
25. Kirillov, S. A., Yannopoulos, S. N., Vibrational dynamics as an indicator of short-time
interactions in glass-forming liquids and their possible relation to cooperativity, J. Chem.
Phys., 117, 1220-1230 (2002).
26. Kalampounias, A. G., Yannopoulos, S. N., Steffen, W., Kirillova, L. I., Kirillov, S. A.,
Short-time dynamics of glass-forming liquids: Phenyl salicylate (salol) in bulk liquid,
dilute solution, and confining geometries, J. Chem. Phys., 118, 8340-8349 (2003).
27. Kirillov, S. A., Novel approaches in spectroscopy of interparticle interactions. Raman
line profiles and dynamics in liquids and glasses, J. Mol. Liq. 110, 99-103 (2004).
IMPROVED ELETROCHEMICAL PROPERTIES
OF SURFACE-COATED Li(Ni,Co,Mn)O2
CATHODE MATERIAL
FOR Li SECONDARY BATTERIES
Sun-Ho Kang* and Khalil Amine

Chemical Engineering Division, Argonne National Laboratory, Argonne, IL 60439, USA
Abstract
The surface of Li(Ni0.4Co0.2Mn0.4)O2 was coated with amorphous

aluminum oxide using Al-isopropoxide. The effect of the surface coating on
the cycling performance and high-temperature storage characteristics was
investigated using coin-type cells. The surface-coated materials showed
superior capacity retention at 55°C as well as at room temperature and
limited impedance rise during high-temperature storage compared with the
uncoated material. From the a.c. impedance measurement, it was found that
the uncoated material exhibited very large increase in the cathode/electrolyte
interfacial impedance during cycling whereas the surface-coated materials
showed a limited increase in the interfacial impedance.
Keywords
Li-ion secondary batteries; Li(Ni,Co,Mn)O2; surface coating;

interfacial impedance.
1. INTRODUCTION
High-power lithium-ion batteries are promising alternatives to the
nickel metal hydride batteries which are currently used for energy storage in
hybrid electric vehicles (HEVs). Currently, Li(Ni,Co)O2-based materials are
the most widely studied cathode materials for the high-power lithium-ion
batteries [1-4]. Although Li(Ni,Co)O2-based materials meet the initial power
requirement for the HEV application, however, it has been reported that they
*
505
506
have major concerns related with calendar life, safety, and cost [1,4,5,6]. To
address these concerns, we have been investigating the applicability of
Li(NixCo1-2xMnx)O2 (1/3 x <1/2) materials to the high-power applications
due to their cost advantage and improved thermal safety compared with the
Li(Ni,Co)O2 materials [7-9]. In this article, we report the effect of surface
modification on the electrochemical properties of the Li(NixCo1-2xMnx)O2
cathode materials. Recently, a lot of research results have been reported on
the improvement of electrochemical properties of cathode materials by
modifying their surface with various metal oxides such as Al2O3, MgO, SiO2,
ZnO, and ZrO2 [10-18]. In this work, we coated the surface of
Li(Ni0.4Co0.2Mn0.4)O2 particles with amorphous aluminum oxide using Al-
isopropoxide and investigated its impact on the cycling performance,
impedance, and high-temperature storage characteristics.
2. EXPERIMENTAL
2.1. Materials
Li(Ni0.4Co0.2Mn0.2)O2 powder was prepared by the solid state
reaction method using Li2CO3, Ni(OH)2, Co(OH)2, and MnCO3 as starting
materials. An appropriate amount of each starting powder was mixed using
zirconia balls in acetone. The mixed powder was calcined at 700qC for 30 h
and then at 1100qC for 5 h in air. The synthesized Li(Ni0.4Co0.2Mn0.2)O2
powder was well dispersed in an ethanol solution of Al-isopropoxide (AIP,
[(CH3)2CHO]3Al), dried at 100qC, and then heat-treated at 300°C for 3hr in
air. The amounts of AIP used were 0.5 and 1.0 wt%. Phase purity of the
prepared materials was examined with a powder X-ray diffractometer using
Cu KD radiation.
The electrochemical properties of the prepared materials were

evaluated using coin-type cells. The positive electrode consisted of 80 wt%
oxide powder, 10 wt% carbon, and 10 wt% polyvinylidene difluoride
(PVDF) binder on aluminum foil. The negative electrode was either metallic
lithium or carbon on copper foil. The electrolyte was 1 M LiPF6 in a 1:1
mixture of ethylene carbonate (EC)/diethyl carbonate (DEC). The coin cells
were galvanostatically cycled at 2.8a4.3 V. The a.c. impedance of the coin-
type cells was measured in the frequency range of 10mHz-100KHz using an
impedance analyzer (BAS-ZAHNER IM6).
507
The high-temperature storage characteristic was evaluated as

follows. After initial formation cycles, the C/Li(Ni0.4Co0.2Mn0.2)O2 cells were
charged to 3.8V and stored at 50qC; every couple of days, the cells were
cooled to room temperature and the impedance of the cells was measured.
By repeating the above steps, the change of the cell impedance with the
50qC-storage time was obtained.
3.1. Crystal Structure

Figure 1 shows the powder X-ray diffraction (XRD) pattern of the
as-prepared Li(Ni0.4Co0.2Mn0.4)O2 material. All of the peaks could be indexed
based on the D-NaFeO2 structure (R 3 m). The lattice parameters in
hexagonal setting obtained by the least square method were a=2.868Å and
c=14.25Å. Since no second-phase diffraction peaks were observed from the
surface-coated materials and it is unlikely that the Al ions were incorporated
into the lattice at the low heat-treatment temperature (300qC), it is
considered that the particle surface was coated with amorphous aluminum
oxide.
(003)
L i N i0 .4 C o 0 .2 M n 0 .4 O 2
a = 2 .8 6 8 A
Intensity (arbitrary unit)
c = 1 4 .2 5 A
(104)
(101)
(012)
(110)
(018)
(006)
(107)
(015)
(113)
(021)
(116)
10 20 30 40 50 60 70 80
2T
Figure 1. The power XRD pattern of the Li(Ni0.4Co0.2Mn0.4)O2 material

prepared at 1100°C for 5hr in air.
508
3.2. Cycling Performance

Cycling performances of the uncoated and coated cathode materials
cycled versus Li at room temperature and 55qC are given in Figs. 2(a) and
2(b), respectively. It can be seen that the coated materials exhibit cycling
performance superior to the uncoated one. At room temperature, the
uncoated material delivered only 27% of the initial capacity at the 100th
cycle whereas the materials coated with 0.5 and 1.0 wt% AIP retained 71
and 75% of their initial capacities, respectively, at the 100th cycle. We
investigated the XRD pattern of each cathode after 100 cycles; neither
impurity peaks nor changes in peak breadth and I(003)/I(104) ratio were
observed in all of the cycled cathodes. This observation indicates that the
significant capacity fading of the uncoated material was not due to the
degradation of the sample bulk such as phase segregation, amorphization and
cation mixing during cycling.
3.3. High-temperature storage characteristic

In Figure 3 are given the variations of the area specific impedance
(ASI) of the C/Li(Ni0.4Co0.2Mn0.2)O2 cells as a function of 50qC-storage time.
The ASI was determined by A·'V/I, where A is the cross-sectional area, 'V
the voltage variation during current interruption for 30 s at each state of
charge (SOC), and I the current applied during the galvanostatic cycling. The
data shown in Figure 3 are the ASI values at 60% SOC. Initially, all of the
materials showed similar ASI values (35-40 :-cm2). However, the uncoated
material showed very large increase in the ASI; the ASI after storage at 50qC
for 90 days increased to 160:-cm2. On the contrary, the coated materials
exhibited limited increase in the ASI with storage time compared with the
uncoated one.
3.4. Impedance
As can be seen in Figure 3, the surface coating improved the high-
temperature storage characteristic so that the surface-coated materials
possessed enhanced calendar life. The variation of the ASI during room
temperature cycling was also found to be significantly suppressed by the
surface coating. The ASI is the overall impedance from various impedance
sources in a cell that are related with cathode(anode) /electrolyte interfacial
overpotential, electrolyte resistance, and Li-diffusion kinetics in the oxides.
To distinguish the major source responsible for the very large impedance rise
of the cell with uncoated material, we measured the a.c. impedances of the
509
coin-type cells. Figure 4 shows the variations of the a.c. impedance spectra
during cycling. The impedances were measured at the open cell voltage of
3.7V at each cycling number, which corresponds to ~60% SOC [6].
200
RT a
150
-1
Capacity / mAh g
100
Li(Ni0.4Co0.2Mn0.4)O2
50
uncoated
coated with 0.5 wt% AIP
coated with 1 wt% AIP
0
0 20 40 60 80 100
Cycle Number
200
o
55 C b
150
-1
Capacity / mAh g
100
50 Li(Ni0.4Co0.2Mn0.4)O2
uncoated
coated with 0.5 wt% AIP
0
0 20 40 60 80
Cycle Number
Figure 2. Change of discharge capacity with cycle number of

Li/Li(Ni0.4Co0.2Mn0.4)O2 cells in the voltage range of 2.8-4.3V at 0.2mA/cm2. (a) at room
temperature; (b) at 55°C.
510
Li(Ni0.4Co0.2Mn0.4)O2
150
2
ASI / :-cm
100
uncoated
50 coated with 0.5wt% AIP
coated with 1.0wt% AIP
0 20 40 60 80 100
Storage time / day

Figure 3. Changes of the ASI at 60% SOC with storage time at 50°C.
It can be seen in Figure 4(a) that the uncoated material showed

significant increase in diameter of the semicircle at medium frequency,
which is generally attributed to Li-ion transport and charge-transfer kinetics
at the cathode/electrolyte interface Consequently, the very large increase of
the ASI during the 50qC-storage and galvanostatic cycling of the uncoated
material was caused mainly by the impedance rise at the oxide particle
surface/electrolyte interface.
The experimental results in Figures 2-4 clearly show that the surface
phase suppresses the degradation of the cathode/electrolyte interface,
resulting in improved capacity retention and limited impedance rise. At this
stage, the detailed mechanism for the role of the surface phase is not clear.
One possible scenario is the role of the surface aluminum oxide as an acid
scavenger. When water molecules are present as an impurity in the
electrolyte, they react with LiPF6 salt to produce HF. The generated HF
attacks the cathode surface, which results in destruction of the SEI layer and
dissolution of transition metal ions into the electrolyte. We speculate that the
aluminum oxide protects the surface from the acid attack by reacting with
the acid to form another compound such as AlF. A similar mechanism was
also proposed by Zheng et al. [14] and Sun et al. [15] for SiO2- and ZnO-
coated spinel, respectively.
511
4. SUMMARY
The surface of Li(Ni0.4Co0.2Mn0.4)O2 was coated with amorphous
aluminum oxide using Al-isopropoxide. The as-prepared and coated
materials exhibited almost no differences in initial capacity and impedance.
However, during the extended cycling, the uncoated material exhibited
significant capacity fading.
50
(a) uncoated
40
30
-Z" / :
20
10 60
40
20 30
0
0 20 40 60 80 100
50
(b) coated with 0.5 wt% AIP
40
-Z" / :
30
20
60
30 40
10
0
0 10 20 30 40 50 60 70 80 90 100
50
(c) coated with 1.0 wt% AIP
40
-Z" / :
30
20
30 40
10 60
0
0 10 20 30 40 50 60 70 80 90 100
Z' / :
Figure 4. Variations of the a.c. impedance with cycling measured with

C/Li(Ni0.4Co0.2Mn0.4)O2 cells at the open cell voltage of 3.7V.
512
On the other hand, the coated cathodes showed superior cycling

performance. The surface-coated materials also exhibited better high-
temperature storage characteristics than the uncoated material. From the
variations of the a.c. impedance with cycling, we observed that the cell with
the uncoated cathode showed a significant increase in cathode/electrolyte
interfacial impedance during cycling. In other words, the poor performance
of the uncoated material was caused by the significant degradation of the
cathode surface in contact with the electrolyte, and the surface-coated phase
stabilized the cathode/electrolyte interface. We speculate that the surface
aluminum oxide reacts with HF present in the electrolyte so that the cathode
surface is protected from the acid attack. Further studies are needed to
explore the specific role of the surface phase in improving the
electrochemical properties of the cathode material. In summary, the surface
coating of cathode materials is a simple and effective way to improve cycle
life as well as calendar life of the Li-ion cells.
ACKNOWLEDGMENTS
This paper has been created by the University of Chicago as
Operator of Argonne National Laboratory under contract No. W-31-109-
Eng-38 with the U. S. Department of Energy. This work was supported by
the U.S. Department of Energy, FreedomCAR and Vehicle Technologies
Program, under contract No. W-31-109-ENG-38. The authors also wish to
thank Dr. S. –Y. Yoon (Quallion LLC, Saugus, CA) and Dr. H. –J. Kwon
(Samsung SDI, South Korea) for helpful discussions.
REFERENCES
1. K. Amine et al., J. Power Sources 2001; 97: 684.
2. M. Balasubramanian, X. Sun, X. Q. Yang, J. McBreen, J. Power Sources 2001; 92: 1.
3. X. Zhang et al., J. Electrochem. Soc. 2001; 148: A463.
4. D. P. Abraham et al., Electrochem. Commun. 2002; 4: 620.
5. A. M. Andersson et al., J. Electrochem. Soc. 2002; 149: A1358.
6. D. P. Abraham et al., J. Power Sources 2003; 119: 511.
7. S. –H. Kang et al., J. Power Sources 2002; 112: 41.
8. S. –H. Kang, K. Amine, J. Power Sources 2003; 119: 150.
9. I. Belharouak, Y. –K. Sun, J. Liu, K. Amine, J. Power Sources 2003; 123: 247.
10. J. Cho, Y. J. Kim, B. Park, Chem. Mater. 2000; 12: 3788.
11. L. Liu, Z. Wang, H. Li, L. Chen, X. Huang, Solid State Ionics 2002; 152: 341.
12. H. –J. Kwon, D. G. Park, Electrochem. Solid-State Lett. 2000; 3: 128.
13. Z. Wang et al., J. Electrochem. Soc. 2002; 149: A466.
14. Z. Zheng et al., Solid State Ionics 2002; 148: 317.
15. Y. –K. Sun, K. –J. Hong, J. Prakash, K. Amine, Electrochem. Commun. 2002; 4: 344.
16. Z. Chen, J. R. Dahn, Electrochem. Solid-State Lett. 2002; 5: A213.
17. Y. J. Kim, J. Cho, T. –J. Kim, B. Park, J. Electrochem. Soc. 2003; 150: A1723.
18. M. M. Thackeray et al., Electrochem. Commun. 2003; 5: 752.
INDEX
A amorphous manganese dioxide,
33, 473
a.c. impedance, 506 ampere-hour capacity, 133
AC sputtering chamber, 447 anatase, 433
acceptor properties, 397 angle tolerance, 413
acceptor type GIC, 385 anion intercalation overpotential,
acceptor-type graphite 279
intercalation compounds (A- anion intercalation, 277
GIC), 389 anions, 76
acetylene black, 164, 451 anisotropic rim, 416
activated carbon, 24, 65, 73, 435 anisotropic, 66
activated reserve power supply, anode material, 189, 379
132 anode, 389, 411, 488
activation energy, 281 anodic intercalation, 281
active carbon, 132, 139 anodic potential threshold, 279
active mass, 469 anolyte, 428
active material, 213, 473 area specific impedance (ASI),
active sites at the surface of 289
carbon, 241 aerogels, 411
active surface area, 232 aromatic layers, 413
adhesion of solution reduction artificial graphites, 161
products, 197 asymmetric capacitors, 33
adsorption, 335, 425 ATD Program, 447
aerogel, 73, 379 atomic force microscopy (AFM),
air electrodes, 137 197, 201, 445, 447
Air gas-diffusion electrode; 105, Auger spectroscopy, 349
131 average effective pore radius, 61
air-zinc batteries, 141
Al-air cell, 135
alkaline cells, 158 B
alkaline Zn-MnO2 cells, 158
all-electric vehicles (EV), 445 basal plane of the carbon/graphite
alloys, 297 particles, 161
alternating current method, 22 basal plane surfaces in graphite,
aluminum-air battery, 134 180
ammonium chloride, 425 basal plane, 160, 161, 241, 408
amorphization of graphite basal planes of graphite, 206
particles, 198 Basic Structural Units (BSUs),
amorphization, 198 416
amorphous carbon, 379, 407, 433 Butler-Volmer equation, 454
amorphous dust, 199 batteries, 137, 495
amorphous graphitic carbon, 400
513
514
batteries, high rate performance, carbon coating, 361, 400

164 carbon composites, 190, 433
BET specific surface area, 241, carbon fibers, 22, 190, 379
392 carbon layers, 242
bituminous coal, 65 carbon material RIECP, 268
boron-doped carbon, 189, 190 carbon materials, 283
boron-doped carbonaceous carbon microtexture, 5
material, 190 carbon nanotubes, 5, 33, 168
boron-doped graphite, 189 carbon paste electroactive
Boron-doped natural graphite, 24 electrodes, 454
boron-doped purified flake carbon retreat, 445
graphite, 25 carbon retreat/redistribution, 450
Bragg angle, 413 carbon skeleton, 411
Bragg’s equation, 405 carbon structure, 271
carbon, 137, 337, 358, 379, 411,
433
C carbon/carbon composite, 5
carbonaceous anode, 488
calendar life, 446, 506 carbonaceous material, 190,
capacitance, 17, 36, 44, 51 213, 267, 345, 358
capacity fade, 197, 474 carbon-based catalysts, 117, 135
capacity loss, 395 carbon-carbon composite, 380,
capacity, 446 413, 433
carbamide melts, 426 carbon-carbon type
carbamide, 425 electrochemical capacitors, 23
carbamide-chloride melt, 380 carbon-carbon ultracapacitor, 168
carbamides, 380 carbon-inoculated complex
carbide-derived carbon (CDC), compounds, 335
400 carbonization, 190, 434
carbides, 399 carbonized polyacrylonitrile, 5
carbon / ionic liquid capacitor, 81 Carbon-Ni Oxide electrochemical
carbon additive retreat, 451 capacitor, 26
carbon aerogel, 379, 411, 421 carbon-zinc cells, 158
carbon atoms, 180 carrier gas, 426
carbon bands, 448 catalysts, 137
carbon black, 22, 38, 164, 390, catalytic activity,89
453 catalytic layers, 127
carbon cloth, 415 cathode impedance rise, 450
carbon coated natural graphite, cathode impedance, 446
292 cathode materials, 495, 506,
carbon coated Si, 296 cathode, 445, 488, 506
carbon coated Si-based anode cathodic materials, 487
materials, 329 catholyte, 428
carbon coated silicon, 317 cations, 76, 497
515
cell impedance, 451, 507 crevices in edge planes of

cellulose, 165 graphite, 206
charge-discharge cycling, 203 crystal planes, 450
charging process, 133 crystal structure, 199
chemical structure, 63 crystallite size, 407
chemical vapor decomposition, crystallization, 453
415 crystallographic position, 483
chemical vapor deposition, 317 Cu K radiation, 506
chloranile, 453 Cu-metallized graphite, 361
chromatograph, 426 current density range, 123
chromatographic methods, 426 current density, 487
chronopotetiometry, 201 current-sensing AFM, 445
coal semi-coke, 65 current-sensing atomic force
coal, 380 microscopy (CSAFM), 447
coal-tar pitch, 190 current-voltage characteristics,
coated Al- modified graphite, 296 115
coated carbon, 379 cycle life, 167, 198, 389
coated Si- modified graphite, 296 cycle retention of battery, 342
coating machine, 165 cycleability, 46
coating, 283 cyclic voltammetry, 48, 379, 389,
cohesion of solution reduction 425, 426
products, 197 cyclic voltammograms, 277, 393
cointercalation, 199, 381 cyclic voltamperic curves, 426
coke, 358
complex formation, 496
composite electrode, 33 D
compound, 425
concentric microtexture, 411 dangling bonds, 241
conductance image, 447, 450 degree of carbon ordering, 402
conducting network, 33 de-intercalation, 164
conducting polymers, 89, 297 dendrite, 160
conductive additive, 31, 477 depassivator, 426
conductive graphite, 451 depth of discharge (DOD), 289
conductivity enhancement, 482 desorption electrode, 389
conductivity, 22, 445 diagnostic evaluations, 446
conductivity: electronic, 483 diamond, 400
conductivity: ionic, 483 diethyl carbonate (DEC), 173
convection, 454 diffusion, 454
copper metallization, 358 dimethyl carbonate (DMC), 173
co-precipitation, 496 discharge capacity, 486
co-products, 201 disordered carbon, 242, 271, 406,
co-pyrolysis, 190 411
corridor model, 420 dissolution, 453
corrosion of graphite, 395 doped carbons, 190
516
double electric layer, 379 electrode polarization, 426, 465

double layer capacitance, 51 electrode potentials, 252
double layer capacity, 21 electrode resistance, 90
double-layer capacitors, 73 electrode swelling, 28
dropping mercury electrode, 425 electrodeposition, 489
DTA spectral studies, 482 electroless plating, 350
electrolysis, 380, 425, 429
electrolyte additive, 171
E electrolytic manganese dioxide,
473
edge planes of graphite, 197 electron acceptor, 190
edge planes, 161, 373 electron density, 95
electric cars, 133 electron energy loss spectroscopy
electric scooters, 133 (EELS), 401
electric vehicles (EV), 445 electron microscopy, 349
electrically rechargeable batteries, electronic barrier, 446
138 electronic conductivity, 445, 447
electroactive, 430 electronic contact, 451
electroactivity, 90 electronically conducting
electrocatalyst, 103, 335 polymers, 41, 168
electrocatalytic activity, 336 electrosorption electrode, 389
electrochemical capacitor, 33, 63, electrosorption, 398
297 elementary cell, 455
electrochemical characteristics, elevated-temperature calendar
455 life, 446
electrochemical corrosion, 379 end-of-charge voltage, 166
Electrochemical energy sources; energy density, 198
105 energy dispersive X-ray (EDX)
electrochemical intercalation of analysis, 435
Li, 337 energy storage capacity, 21
electrochemical oxidation, 28, environmental compatibility, 137
379 epoxy resin, 165
electrochemical performance, 379 equivalent circuit, 54
electrochemical process, 469 ethylene carbonate (EC), 173,
electrochemical properties, 489, 197, 489
506 eutectic point, 409
electrochemical reaction, 453 exfoliated graphite, 433, 435
electrochemical reduction, 131 exfoliation of graphite particles,
electrochemical stability, 75 199
electrochemical synthesis, 398 exfoliation of the graphene
electrochemical techniques, 197 planes, 198
electrochemically active material, exfoliation, 198, 381
453 expandable graphite, 379
electrode coating operations, 164
517
expanded graphite, 25, 89, 158, gas diffusion electrode, 87, 98,
165, 168, 265, 297, 379, 406, 105, 117
482 gas-transport in porous media,
ex-situ x-ray diffractometry, 245 117
extraction of metals, 399 Gatan image, 401
Gibbs energy, 401
glassy carbon, 425
F glove box, 191
grafted carbons, 336
face-centered cubic (fcc) graphene layers, 414, 419
structure, 363 graphene planes, 198
failure of graphite electrodes, 197 graphite - Sn composite materials,
faradaic redox reactions, 33 350
faradaic yield, 281 graphite anodes, 445
Faraday’s constant, 457 graphite basal plane, 408
fibers, 411 graphite crystals, 405
fibers, boron-doped, 164 graphite electrode, 197, 245
fibers, surface coating, 164 graphite embryo, 165
fibrous microtexture, 411 graphite exfoliation, 181
first cycle loss, 161 graphite grains, 252
flat discharge curve, 137 graphite intercalation compound,
flotation technologies, 214 214, 380
fluid bed reactor, 361 graphite lamellae, 406
fluidized bed, 358 graphite macroplanes, 348
Fourier transform (FT) level, 413 graphite nitrate, 379, 381, 437
freeze drying, 496 graphite nucleate, 407
fringe superimposition, 413 graphite particles, 198
FTIR spectroscopy, 197 graphite reservoirs of lithium, 214
fuel cells, 103, 134, 495 graphite, 22, 96, 160, 166, 171,
fully lithiated graphite, 203 197, 214, 259, 270, 340, 345,
346, 358, 379, 381, 389, 390,
400, 405, 415, 433, 451, 482
G graphite/metal composites, 317
graphite’s ash, 215
galvanostatic charge/discharge, graphite-based composite
216 materials, 330
galvanostatic cycling, 201, 319, graphite–silicon nanocomposite
395 materials, 355
galvanostatic discharge, 43 graphite-tin nanocomposite
galvanostatic energy storage, 423 materials, 355
galvanostatic polarization, 208 graphitic carbon, 407
galvanostatic, 473, 489 graphitic crystallites, 412
gas chromatography, 380 graphitic materials, 269
graphitic spheres, 407
518
graphitization, 271, 400, 408 hydrogen storage capability, 380

graphitized carbon fibers, 270 hydrogen-containing materials,
graphitized mesophase carbon 269
microbeads (MCMB), 214 hydrolysis, 496
graphitized mesophase, 216 hydrophobic gas layer, 119
graphitized milled carbon fiber, hydrophobic pores, 131
214 hydrophobicity, 131
hysteresis, 415
H
I
half-cells, 216
hard carbon, 161, 166, 190, 214, ICP inductively coupled plasma
217, 269, 345, 411 spectrometer, 215, 216
heterogeneous decomposition, ideal gases, 401
131 impedance spectra, 329
high capacity of portable impedance spectroscopy, 15
electronic devices, 167 impedance, 451
high energy density batteries, 227 in situ 7Li NMR, 231, 419
high energy density, 317 in situ prepared activated carbons,
high power battery, 227, 283 435
high power portable electronic inert electrode, 133
devices, 167 infrared spectral studies, 482
high resolution transmission in-situ AFM imaging, 197
electron microscopy in-situ x-ray diffraction (XRD),
(HRTEM), 239 249
highly conductive graphite, 451 insulating surface, 451
highly ordered carbons, 5 intercalant, 381
Highly Oriented Pyrolytic intercalation compounds, 407
Graphite (HOPG), 161 intercalation, 160, 270, 379, 407
high-power Li-ion cells, 445 interfacial impedance, 505
high-power lithium-ion batteries, inter-granular spaces, 450
505 interlayer distance, 347, 381
high-resolution transmission interlayer insertion, 271
electron microscope (TEM) interlayer spacings, 415, 419
images, 359 internal surface of the electrode,
HRTEM image, 420 454
hybrid electric vehicles (HEV), interplanar distance, 408
283, 445, 505 ion thinning, 413
hybrid power supply, 51 ionic barrier, 446
hybrid vehicles, 167 ionic liquid, 81
hydrated manganese dioxide, 33 IR spectroscopy, 380, 425, 428,
hydrogen bonding, 381, 385 489
hydrogen intercalation, 423 iron carbides, 379
519
irreversible capacity loss, 267, lithiation/delithiation voltage, 246

379 lithium alloy, 171
irreversible capacity, 160, 166, lithium batteries, 358, 408
199, 214, 241, 259, 415 lithium insertion potentials, 198
isotropic carbon fiber, 413 lithium insertion/extraction, 271
Isotropic, 66 lithium intercalation compounds,
213
lithium ion clusters, 346
K lithium metal deposition, 160
lithium potential, 161
Karl Fisher moisture titration, 216 lithium storage alloys, 180, 186
kinetics of particle cracking, 205 lithium storage metals, 180
kish graphite, 379 lithium-alloy anode, 488
Knight shift, 419 lithium-ion batteries, 171, 189,
KOH-activated carbons, 63 213, 297, 309, 317, 355, 379,
411, 415, 474, 487
lithium-ion cell, 259, 283
L lithium-ion secondary battery, 488
lithium-ion shuttling batteries,
lacey carbon film, 401 245
lamellar microtexture, 411, 416 local electrochemical
lamellar structure, 214 characteristics, 455
lattice defects, 486 local polarization, 455
lattice organization, 217
lattice structure of graphite, 198
lattice, 161
lead acid automotive batteries, M
158
Li intercalation, 164 macroanions, 281
Li metal, 171 magnesium anode, 105, 108
Li+ deintercalation, 345 magnesium-air cells, 106, 134
Li+ intercalation, 345 manganese dioxide, 473, 481
lifecycle, 446 manganese oxide, 475
Li-graphite electrodes, 209 market segments, 167
Li-Ion batteries, 159, 198, 345, mass interchange, 463
445 mass transfer, 483
Li-ion cell, 189 MCMB carbons, 161
Li-ion secondary batteries, 505 Mechanically rechargeable zinc-
Li-Ion system, 160 air cell, 105, 135
linear sweep voltammetry (LSV), melt, 429
489 melt electrolysis, 426, 429
liquid- state reactions, 496 mesocarbon microbeads
lithiated graphite electrodes, 209 (MCMB), 197
lithiated graphite, 166, 198
520
Mesophase low temperature multi-walled carbon nanotubes,

fibers, 161 17, 38
mesophase microbeads, 411
mesophase pitch fibers, 168
mesophase pitch, 19 N
mesopore, 63, 421
metal coatings, 358 nano-compounds, 380
metal powders, 22 nanocrystalline diamond, 400
metal-air batteries, 88, 89, 117, nanocrystalline graphite flakes,
297 409
metal-free rechargeable batteries nanocrystalline graphite ribbons,
(MFRB), 390 409
metallic lithium, 160 nanocrystalline graphitic carbon,
metallized graphite, 361 400
methyl carbonate, 489 Nanomaterials, 497
Mg-air cells, 135 nanoporous glasses, 497
microcrystalline graphite, 409 nanoscrolls, 380, 433
micropore, 380, 421 nanostructures, 380
microporous carbon fibers, 412 nanotube-like structures, 400
microporous carbons, 5, 411 nanotubes, 190, 262, 345, 400,
microporous polypropylene film, 411
261 natural crystalline flake graphite,
microporous, 63 346
microporous/mesoporous solids, natural flake graphite, 391
411 natural flakes, 197
MicroRaman surface-average natural graphite anodes, 168
spectra, 448 natural graphite, 161, 190, 214,
microstructure, 406 267, 283, 359
microtexture, 411 rounded edge graphite, 359
migration, 454 negative active material, 309
misoriented graphitic-like neutral electrolyte, 105
crystallites, 412 nickel metal hydride batteries,
mixed solvent electrolytes, 173 505
modification of carbon, 343 Ni-metallized graphite, 361
modification of graphite, 343 NiOx-Carbon Ultra Capacitors, 23
modified carbon paste electrodes, non-graphitised carbons, 190
454 non-stacked layers, 417
Modified graphite, 297 nonstoichiometric oxide, 483,
modifier, 379 487, 483
molten carbamide, 380, 426 nucleation, 496
molten salts, 425 Nyquist plots, 247
morphological aspects, 199
morphology, 197
mother pitch, 165
521
O polymeric resin, 497

Polypyrrole, 41
optical microscopy, 416 polyvinylidene difluoride, 360
order degree, 418 porosity characteristics, 63,
oxidation of graphite, 389 porous electrode, 140, 453
oxidation resistance, 31 porous microtexture, 411
oxidation, 389 potassium chloride, 425
oxides of chromium, 487 powder coating, 361
oxides of cobalt, 487 power fade, 446
oxides of nickel, 487 power requirement, 505
power sources, 159, 481
power tools, 167
P primary alkaline Zn-MnO2 cell,
167
PANI/expanded graphite primary batteries, 138
composites, 89 primary carbonization, 271
partial pressure, 124 primary cells, 89
partially graphitized cokes, 270 primary zinc-air cells, 132
particle orientation, 165 prismatic (edge) surfaces in
passivation of graphite particles, graphite, 180
203 probe solution, 197
passivation, 197 propylene carbonate (PC), 180,
peak output power, 51 197
Pechini method, 497 proteins, 165
phase composition, 199 proton affinity, 381
phase purity, 506 pseudo-capacitance, 33
phase transitions, 198 pulse-type profile, 446
phenol resin, 165 purified flake graphite, 25
photomicrography, 469 purified graphite, 217
physicochemical properties, 429, purified natural graphite, 317
446 pyridine-borane complex, 190
pitch mesophase, 65 pyro-carbon, 263
pitch semi-coke, 65 pyrolysis, 131, 190
planar graphite nucleate, 407 pyrolytic carbon coating, 415
polarization characteristic, 453 pyrolytic carbon, 345, 415
polarization curves, 123
polarization of a graphite, 245
polarization rates, 205 Q
polarization, 90, 108
polarized graphite electrodes, 198 quasi-metallic lithium clusters,
polarographic investigations, 425 242, 419
polyaniline, 41, 89 quenched melts, 426
polyaromatic layers, 411
polycarbon, 379
522
R Sn-graphite composite, 297

soft carbon, 190, 270, 411
Raman analysis, 405 sol-gel method, 497
Raman microscopy, 401, 445, 446 solid electrolyte interface, SEI,
Raman spectra, 408 161, 171, 232, 341, 252
Raman spectroscopy, 406 solid reagent, 453, 469
randomly oriented microstructure solid state reaction, 496, 506
of graphite, 166 solid state synthesis of graphite,
rate capabilities, 198 400
rechargeable batteries, 379, 389, solid-phase reduction, 487
390, 398 solution, 199
rechargeable cells, 89 solvated lithium ions, 239
reciprocal node elongation, 413 solvation ligands, 385
reference electrode, 426 solvent co-intercalation, 181
reserve power supply, 132 solvent, 173
reversible alloys, 213 specific capacity, 259
reversible capacity, 214, 241, 270, specific conductivity, 77
342, 415 specific energy, 102
rigid carbon coating of disordered specific impedance (ASI), 289
nature, 330 specific power, 51
''rocking chair'' mechanism, 487 specific resistance, 21
rocking-chair battery, 189 spectrally pure graphite, 266
spheroidized purified natural
graphite, 327
S spinel, 474
spinel-structured 3d-metal oxides,
saline electrolytes, 135 334
scanning electron microscope spray drying, 496
(SEM), 197, 201, 360, 435 stabilizing carbon coating, 327
secondary ion mass spectrometry, stable morphology, 208
349 staging process, 217
segmentation, 167 state-of-charge (SOC), 446
SEI layer, 451, 510 stationary electrode potentials,
semi-coke, 190 139
sequential cyclic voltammetry, stationary platinum electrode, 426
389 Stefan-Maxwell equations, 454
shelf life, 137 structure, 411
Sherrer formula, 406 sugars, 165
shuttle-cock battery, 189 supercapacitor, 5, 21, 33, 41, 51,
Si-graphite composite, 297 295, 379, 389, 408, 475,
single-layer hard carbons, 269 supercapacity, 271
Single-walled carbon nanotubes supporting electrolyte, 425
(SWNT), 277 surface area, 199
slit-shaped pores, 419 surface coating, 505
523
surface conductance, 450 transport hindrances, 124, 135

surface electronic conductivity, Tuinstra & Koenig’s equation,
445 406
surface films, 197, 203
surface functional groups, 241
surface modification, 283 U
surface reaction, 205
surface-electrolyte interface (SEI) ultramicrotomy, 413
layer, 214 universal gas constant, 457
swing battery, 189
symmetric capacitors, 33
synthetic flakes, 197 V
synthetic graphite, 190, 359
systematic Raman mapping, 445 Van der Waals-type interaction in
carbons, 277, 414
vinylene additives, 180
T voltammetry curve, 422
volume changes, 203
Tafel plots, 123
template synthesis, 474
template technique, 5 W
ternary solvated graphite-
intercalation compounds, 180 Warren-Bodenstein formula, 406
texture, 411 weakly organized carbons, 413
theoretical model, 453 well-graphitized synthetic
thermal purification of graphite, graphite, 379
215 well-ordered graphite, 399
thermal safety, 506 well-staked layers, 422
thermal vacuum deposition, 349 Wurster coating, 361
thermally exfoliated graphite, 42,
379, 391,
thermally expanded graphite X
(TEG), 265, 302, 389
thermoexfoliated graphite (TEG), X-ray diffraction (XRD)
347 spectroscopy, 197, 201, 347,
thickness control, 165 359, 399, 446, 482, 506
thin film electrodes, 487
three-electrode cell, 44
tin-based alloys, 309 Z
transfer coefficients, 457
transition metal oxide, 488, 495 zinc-air batteries, 87
Transmission Electron zinc-air fuel cell, 134
Microscopy (TEM), 379, 401, Zn-MnO2 cells, 158
435

New Carbon Based Materials For Electrochemical and Fuel Cells Energy Storage Systems: Batteries, Supercapacitors

Uploaded by

Copyright:

Available Formats

New Carbon Based Materials For Electrochemical and Fuel Cells Energy Storage Systems: Batteries, Supercapacitors

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

New Carbon Based Materials For Electrochemical and Fuel Cells Energy Storage Systems: Batteries, Supercapacitors

Uploaded by

Copyright:

Available Formats

New Carbon Based Materials for Electrochemical

Energy Storage Systems: Batteries, Supercapacitors

I. Life and Behavioural Sciences IOS Press

Series II: Mathematics, Physics and Chemistry – Vol. 229

Published in cooperation with NATO Public Diplomacy Division

ISBN-10 1-4020-4811-4 (PB)

Printed on acid-free paper

All Rights Reserved

Printed in the Netherlands.

Novel Carbonaceous Materials for Application in the

Effect of Carbonaceous Materials on Performance of

Hybrid Supercapacitors Based on a-MnO2/Carbon

Development of Supercapacitors Based on

Supercapacitors: Old Problems and New Trends...................... 51

Modeling Porosity Development During KOH

General Properties of Ionic Liquids as Electrolytes

New Concept for the Metal-Air Batteries Using Composites:

Mechanically Rechargeable Magnesium-Air Cells with

Application of Carbon-Based Materials in Metal-Air

Metal – Air Batteries with Carbonaceous Air Electrodes and

Carbonaceous Materials for Batteries............................................ 157

Anode-Electrolyte Reactions in Li Batteries:

Performance of Novel Types of Carbonaceous Materials in

Why Graphite Electrodes Fail in PC Solutions: An

New Developments in the Advanced Graphite for

Mechanisms of Reversible and Irreversible Insertion in

Some Thermodynamics and Kinetics Aspects of the Graphite-

Characterization of Anodes Based on Various Carbonaceous

A Carbon Composite for the Negative Electrode of Li-Ion

Electrochemical Intercalation of PF 6 and BF 4

Surface Treated Natural Graphite as Anode Material

On The Theoretical Prerequisites for Application of Novel

Capabilities of Thin Tin Films as Negative Electrode Active

Composite Anode Materials for High Energy Density

Electrochemical Activity of Carbons Modified by d-Metal

Metal-Graphite Composites as Materials for Electrodes of

Electrochemical Performance of Ni/Cu-Metallized & Carbon-

Stabilization of Graphite Nitrate via Co-intercalation of Organic

Electrochemical Stability of Natural, Thermally Exfoliated and

Low Temperature Synthesis of Graphite from Iron

High Resolution Transmission Electron Microscopy Image

Electrolytes of Carbamide-Chloride Melts at Inert

Graphite Intercalation as a Way to Carbon-Carbon

Diagnostic Evaluation of Power Fade Phenomena and

Modeling of Electrochemical Processes in the Electrodes

On the Optimal Design of Amorphous Mangaense Oxide

Investigation of Cathodic Materials Based on Different

Investigation of Thin-Film Electrode Materials as

Synthesis of Mixed Oxides using Polybasic Carboxylic

Improved Electrochemical Properties of Surface-Coated

Index ....................................................................................................... 513

Carbonaceous materials play a fundamental role in electrochemical

Technology (Poland) for their participation and assistance in identifying

October 20, 2003

On behalf of Argonne National Laboratory, I would like to welcome you to Argonne

I am confident that this meeting will be a stimulating and productive one. As

OPERATED by THE UNIVERSITY of CHICAGO for THE UNITED

Group Photograph, Argonne National Laboratory, Argonne, IL USA (Courtesy C. Johnson)

Technical discussions during the poster ses-