Theory and Modeling of Electronic Excitations in Nanoscienceook

NANOEXC2004 CONFERENCE
Programme, Abstracts, and Participants List of the conference
Theory and Modeling of Electronic

Excitations in Nanoscience
19 23 September 2004
Conference Organizers
Pablo Garcia-Gonzalez (Madrid, Spain)
Maurizia Palummo (Rome, Italy)
Olivia Pulci (Rome, Italy)
John J. Rehr (Seattle, US)
Arno Schindlmayr (Julich,

Germany)
Nathalie Vast (Palaiseau, France)
Local Organizers Committee
Rodolfo Del Sole (Rome, Italy)
Maurizia Palummo (Rome, Italy)
Olivia Pulci (Rome, Italy)
Sponsored by
European Network of Excellence Nanoquanta
European Science Foundation
PSI-K: Electronic structure calculations for elucidating the complex atomistic behaviour
of solids and surfaces
INFM UdR Roma Tor Vergata
Universita degli Studi di Roma Tor Vergata
Centre National de la Recherche Scientifique, Commissariat lEnergie Atomique, Ecole
Polytechnique
Computational Materials Science Network
supported by University of Washington, Office of the Basic Energy Sciences U.S.
Department of Energy
Contents
Conference Programme
List of Posters
Abstracts of talks and posters
List of participants
46
Conference Programme
Monday, September 20 2004
8:50
Presentation and general announcements
9:00-9:40
9:40-10:00
R. W. Godby
Self-consistency in GW calculations of spectral and ground-state properties
T. Kotani
Quasi-particle self-consistent GW approximation
10:00-10:20
P. Rinke
Exact-Exchange-Based Quasi-particle Calculations of II-VI Compounds and Group III Nitride
10:20-10:40
M. Marsili
Electronic band structure of the C(111)2x1 surface within an iterative
GW scheme
10:40-11:10
* Coffee break *
11:10-11:50
J. Rehr
Local-Field Effects in Calculations of Optical Constants from UV-Xrays
11:50-12:10
G.P. Brivio
Ab initio treatment of adatom Auger decays
12:10-12:50
E. Shirley
Bethe Salpeter treatment of core excitations: multiplet effects
13:00-14:30
* Lunch break *
16:00-16:40
E.K.U. Gross
Ab-initio theory of superconductivity
16:40-17:00
E. Chulkov
Electron-phonon interaction and its role in the lifetime broadening
mechanism of electron and hole excitations in metallic materials
17:00-17.20
* Coffee break *
17:20-18:00
J. Palacios
Electronic transport in nanoscale systems from first principles
18:00-18:20
P. Bokes
Conductance of metal-vacuum-metal interface from the polarisation
function
18:20-18:40
H. Mera
Exchange in non-equilibrium systems
20:00-21:00 * Dinner *
21:00 EVENING POSTER SESSION
Tuesday, September 21 2004

9:00-9:40
F. Bechstedt
Optical spectra of three-, two-, and zero-dimensional systems treated
using many-body perturbation theory
9:40-10:00
R. Del Sole
Calculation of surface optical properties within TDDFT: the bound
exciton of Si(111)2x1
10:00-10:20
P. Eggert
Long-ranged interactions in a supercell approach
10:20-10:40
L. Chiodo
The energy loss of (110)Ag and (001)Al surfaces
10:40-11.10
* Coffee break *
11:10:11:50
G. Rignanese
First-principles study of crystalline and amorphous transition metal
oxides and silicates
11:50-12:10
L. Dash
Ab initio spectra of the low-pressure phases of zirconia
12:10-12:30
A. Incze
Ab initio study of reflectance anisotropy spectra of a sub-monolayer
oxidized Si(100) surface
13:00-14:30
* Lunch break *
15:00-15:40
S. Baroni
Time-dependent density-functional perturbation theory
15:40-16:00
Y. Pouillon
Understanding the dissociation of diatomic molecules in the DFT-ACFD
formalism : ground-state and excited-state studies
16:00-16:20
S. Kurth
Orbital Functionals in Current-Density Functional Theory
16:20-16:40
* Coffee break *
16:40-17:20
L. Reining
TDDFT and Many-Body Perturbation Theory: comparisons and combinations
17:20-17:40
A. Marini
A threepoint ManyBody vertex function from TimeDependent
DensityFunctional Theory
17:40-18:20
G. Stefanucci
Excitons from TDDFT as generated by the variational approach to
many-body perturbation theory
18:30-20:30
POSTER SESSION (continued)
20:30 * Dinner *
Wednesday, September 22 2004

8:30-9:10
C. Delerue
Quasi-particle gap, dielectric properties and electronic correlations in
semiconductor nanostructures
9:10-9:50
L. Benedict
Optical absorption in hydrogenated Si clusters: BSE vs. TDLDA
9:50:10:10
L. Ramos
Excitation energies and optical properties of Si nanocrystallites
10:10-10:30
S. Ossicini
Ab-initio calculation of the structural, electronic and optical properties
of Silicon nanoclusters in their ground and excited states
10:3011:00
* Coffee break *
11:00-11:40
A. Franceschetti
Electronic Excitations in Semiconductor Quantum Dots
11:40-12:00
C. Verdozzi
Tuning the optical and/or magnetic properties of correlated nanoclusters
12:00-12:20
M. Gatti
Effects of oxygen on the structural, electronic and optical properties of
silicon nanocrystals
12.20-12:40
E. Degoli
Impurities states in hydrogenated silicon nanoclusters
13:00
* Lunch break *
16:00-16:40
S. Ismail-Beigi
First-Principle Greens function methods: electronic and optical properties of carbon nanotubes and excited-state forces
16:40-17:00
E. Chang
Quasiparticle energies and optical properties in carbon nanotubes
17:00-17:40
L. Wirtz
Excitons in BN-nanotubes: dimensionality effects
17:4018:00
* Coffee break *
18:00-18:20
M. Marques
Absorption spectra of biological systems from TDDFT
18:20-18:40
A. Seitsonen
Recent applications of TD-DFT in molecular, liquid and solid phases
18:40- 18:50
Conclusive Remarks
18:50
20:00
Closing of nanoexc2004 and presentation of the 2005 workshop
**Pizza Party**
Thursday, September 23 2004: DEPARTURES

4
List of Posters
1 Andrey Bogdanov Coherent exchange between the forward and backward waves in Bistable
optical response of a thin glassy film
2 Silvana Botti Optical properties of complex ystems from TDDFT
3 Fabien Bruneval Electronic Excitations of Cu O within GW approximation
4 Mauro Bruno Ab-initio electronic and optical properties of Germanium Quantum wires
5 Gabriele Butti Image potential states and surface electronic structure calculation on the
Na/Cu(111) system
6 Sebastiano Caravati Selectivity of auger decays to the local surface environment
7 Nuno Carneiro A novel generalized Kohn-Sham approach for total energy calculations within
Density Functional theory
8 Louise Dash Choosing ultrasoft pseudopotentials for spectroscopy calculations
9 Kris Delaney Self-consistent Solutions of Dysons Equation and Total Energies from the
Approximation
10 Christoph Freysoldt GW calculations for surfaces

11 Martin Friak Ab-initio calculations of the half-metal to metal transition in magnetite
12 Christophe Friedrich Photocatalitic oxidation of CO at the V O (010) surface: cluster model
studies
13 Maja Garcia-Vergniory Self-energy and lifetimes of surface states beyond the GW approximation
14 Conor Hogan Theory of electron energy Loss at surfaces
15 Aleksey Kuznetsov Ab initio study of Fe O (111) surface
16 Roman Leitsmann Nonlinear Optical coefficients from first-principles calculations: Different
numerical approaches
17 Sergey Lisenkov Ab-initio study of new polimeric crystals based on fullerenes C and carbynes
18 Manfred Niesert Ab-initio calculations of the optical properties of phase-change materials
19 Enrico Perfetto Microscopic model for a strongly correlated superconductor single-electron
transistor
20 Jose M. Pitarke Large crystal local-field effects in the dynamic structure factor of TiO
21 Stefano Pittalis Optimized effective potentials in current-density functional Theory
22 Victor A. Popa Band gaps in GW self-consistent calculations
23 Sampsa Riikonen Ab-initio studies of the Si(557)/Au and Si(553)/Au Reconstructions
5
24 Arno Schindlmayr Vertex Corrections in a systematic self-energy perturbation theory

25 Paolo Emilio Trevisanutto A bi-exciton mechanism of photo-induced desorption from MgO
26 Daniele Varsano A study of excitonic properties of low dimensional systems with TDDFT
27 Nathalie Vast Phase stability in titanium and its dioxides with ab initio methods
28 Wojciech Welnic Ab-initio calculations of amorphous phase change materials
Abstracts of talks and posters
Time-dependent density-functional perturbation theory

Marco Saitta (1,2), Brent Walker(2,3), Ralph Gebauer(2,3), and
Stefano Baroni(1,2,4)
(1) Physique des Milieux Condenss, Universit Pierre et Marie Curie, Paris, France (2) DEMOCRITOS National
Simulation Center, Trieste, Italy (3) The Abdus Salam International Center for Theoretical Physics, Trieste, Italy
(4) Scuola Internazionale Superiore di Studi Avanzati, Trieste, Italy
In this talk I will discuss some ideas that would allow an extremely efficient calculation of
optical spectra within TDDFT. Once some preliminaries have been set up with a numerical
effort typical of a static DFPT calculation, the dynamical polarizability can be obtained at any
(not too large) frequency, essentially for free. As it is the case for static DFPT, our method does
not make any explicit reference to unoccupied molecular orbitals, so that it is well suited for
large-scale plane-wave calculations. At the time when this abstract is being written, we are still
working on the algorithm and testing it on a simple toy model. It is hoped that, by the time of
the meeting, some meaningful results on a realistic system will be available.
Optical spectra of three-, two-, and zero-dimensional systems treated

using many-body perturbation theory
F. Bechstedt, P. Hahn, and W.G. Schmidt
Friedrich-Schiller-Universitat Jena, Institut fur Festkorpertheorie und -optik, Max-Wien-Platz 1, 07743 Jena,
Germany
The aim of the present paper is to demonstrate the power of the many-body perturbation theory
to describe accurately optical spectra of systems with arbitrary confinement. The method
is based on solutions of the Dyson and Bethe-Salpeter equations in the framework of the
GW approximation. The dynamical screening is treated by a model dielectric function or
by the full frequency-dependent dielectric matrix calculated within the independent-particle
approximation. An initial-state formulation of time-dependent polarizability allows spectra
computations for systems with many atoms in the unit cell. The treatment of the quasiparticle
self-energy effects beyond the first-order perturbation theory is discussed for molecular systems.
Applications of the method are demonstrated for bulk compound semiconductors without
confinement, hydrogenated Si surfaces with one-dimensional confinement, and molecular
systems such as silan, H O molecules as well as molecular crystals (ice) with more or less
three-dimensional confinement. Problems of the approximations concerning e.g. the frequency
dependence of the self-energy and the mixing of wave functions are discussed.
Optical absorption in hydrogenated Si clusters: BSE vs. TDLDA

L. X. Benedict
Lawrence Livermore Lab, L-045, 7000 East Ave, USA
We have performed calculations of the optical absorption spectra of nm-sized hydrogenated Si

clusters using two related approaches, an approximate Bethe-Salpeter equation treatment, and
TDLDA. We find that the gross features of the results are similar for the larger clusters, but are
8
highly discrepant for . The similarities and differences will be discussed, as they illustrate
the relation between the two theoretical approaches. In the end, I offer some suggestions as to
why the TDLDA spectra for and other small systems seems to be in such good agreement
with experiments.
Coherent exchange between the forward and backward waves in Bistable

optical response of a thin glassy film
A. A. Bogdanov and A. I. Zaitsev
A.I.Herzen Russian State Pedagogical University, naberezhnaya reki Moiki 48, St. Petersburg, 191186 Russia
[email protected]
The generation of reflected and transmitted waves in the dense two-level absorber for an intense
plane wave under normal-incidence conditions is investigated. A region of parameters is
intentionally selected for which one needs to take into account strong atom-field coupling,
which results in the appearance of polaritons. The near-zone dipole-dipole interaction is
taken into account within the local-field approach. We also discover the possibility of optical
hysteresis of transparence and the dependence of the instabilities of several states on length of
a sample. If the length of the sample increases, time of transition to the mode of transparency
increases too. Thus the impulse regime of the reflected field is formed. Possibility of change
of conditiones of bistables behaviour under influence of the inhomogeneous broadening is
researched too.
The nonlinear resonant optical properties of dense dielectric two-level media have been a subject
of interest during the past two decades. We mainly focus our study on the analysis of the
nonlinear reflection and transmission from a vacuum to two-level medium, varying the driving
parameters over a region in which the system shows bistable behavior. This problem was partly
studied by us in a recent publication [2]. The coherent cooperative phenomenon has been
mainly studied in the unidirectional approximation, where the reflection from the boundaries
of an inverted medium is ignored [1]. The amplitudes are usually assumed to be slowly
varying in both time and space. The inclusion of reflection implies that we should abandon
the approximation of slowly varying amplitudes in space [3]. In the presence of reflection, an
atomic system may generate waves propagating in both directions, although the external field
initiates generation of only one wave.
The medium is described by the modified Bloch equations in which the field acting on each
atom of the system consists of two parts: the first is the average (Maxwell) field, and the second
is the local-field correction (4 /3)P (where P is the electric polarization of the medium), when
one takes into consideration the resonant dipole-dipole interaction.

! #"$!"% &'(*)
9
(1)
(2)

)

)

!" $# $
%&'# !
(3)
where the dots denote time derivatives; are the elements of the atomic density matrix at a
point with coordinate x at the moment of time t; is the detuning of an atom and light. We
(
is
include effects related to homogeneous and inhomogeneous broadening in our analysis.
the characteristic time. The amplitude of the quasi-resonant external field increases and then
decreases, as we want to research the phenomenon of a hysteresis.
Since the fields of transmitted and reflected waves are determined by the values of the field
on the boundaries of the sample, the kinetics of these fields can be understood only through
the analysis of field and atomic characteristics in the medium. From our point of view, the
investigation of spatial and kinetic characteristics of our system should be supplemented by the
analysis of the kinetics of the mode structure of the field, i.e., the spatial spectrum of the field
inside the sample: Many features of the studied phenomena can be better understood in terms
of the normalized spatial spectrum of the field inside the sample. For the analysis of integral
dependences of the field kinetics, we will also employ the fraction of the forward wave.
In our opinion, it is also useful for the analysis of the process under study to separate the field
inside the sample into the forward and backward waves, which can be reconstructed by the
inverse Fourier transform over the spatial spectrum in the positive and negative regions of wave
numbers. From our point of view, the use of these quantities allows the process under study to
be understood in a more detailed way.
The shift of the resonance frequency, which is due to the dipole Vdipole interaction, as well
as the dependence of the resonance frequency on the population, which reflects the nonbosonic
character of the collective excitations (excitons), are described in an adequate manner. We
use the concept of polaritons to analyze qualitatively the spectral regions into which one
can separate the effects generated by the local-field non-linearity from those caused by selfreflection, as well as when such separation is impossible. The possibility of a stable reflectivity
hysteresis loop is also demonstrated. The mutual action of the local-field and the saturation
nonlinearities is analyzed. We pursued only one aim, namely, to check how the results obtained
in our previous study will change when the local-field correction is considered.
[1] M. G. Benedict, A. M. Ermolaev, V. A. Malyshev, I. V. Sokolov, and E. D. Trifonov, Super-radiance:
Multiatomic coherent emission, Institut of Physics Publishing, Bristol - Philadelphia, 1996.
[2] A.A.Bogdanov,I.V. Ryzhov, A.I.Zaitsev,V. A. Malyshev,J. Knoester, Optical response and bistability
of an thin film consisting of molecular aggregates, Proc. SPIE, 4605, pp. 325-332, 2001.
[3] A.A.Bogdanov, A.I.Zaitsev, Generation of Transmitted and Reflected Waves in Induced Superradiance, Laser Physics, Vol. 11, No. 3, pp. 1 V15, 2001.
10
Conductance of metal-vacuum-metal interface from the polarisation

function
Peter Bokes and Rex Godby

(1) Department of Physics, Slovak University of Technology (FEI STU), Ilkovicova 3, 812 18 Bratislava, Slovakia.
(2) Department of Physics, University of York, Heslington, York YO10 5DD, U.K.
Recently we have suggested a novel approach for calculation of the conductance of quantum
junctions from the polarisation function [1]. The method, unlike all other available approaches,
does not rely on a single-particle approximation and can effectively take advantage of well
developed many-body techniques, particularly the GW method. Here we present results of
calculation for a metal-vacuum-metal interface at the SCF-DFT level and discuss problems met
in actual numerical implementation of the approach.
[1] P. Bokes and R. W. Godby, arXiv:cond-matt/0403204, to appear in Phys. Rev. B (2004).
Optical properties of complex systems from TDDFT

Silvana Botti, Fabien Bruneval and Lucia Reining
Laboratoire des Solides Irradies, UMR 7642 CNRS/CEA, Ecole

Polytechnique, F-91128 Palaiseau, France
Time-dependent density functional theory (TDDFT) allows to study the electronic excitations
involved in spectroscopic experiments, possibly conserving a computational effort comparable
to that of ground-state density functional theory (DFT). In practice, its use is still restricted to
some classes of systems, due to the difficulty in finding general approximations for the exchange
and correlation (xc) terms.
Here we discuss applications of TDDFT to the calculation of the optical properties of a variety
of systems: from clusters and nanostructured materials to bulk crystalline and amorphous
systems. We will consider different approximation schemes, showing that results in agreement
with experiments do not necessarily come from a single scheme. Finally, we present a simple
model to account for the energy dependence of the long range xc kernel of TDDFT.
Ab initio treatment of adatom Auger decays

G.P. Brivio
1-Dipartimento di Scienza dei Materiali - Universit di Milano Bicocca, Italy
We calculate the core-core-valence (CCV) Auger profiles of an adatom on a periodic surface

within the DFT framework by using an approximate method based on the local density of states
(LDOS), with a core hole on the atom, in the neighbourhood of the impurity. To verify the
validity of this simpler approach we compare the Auger rates worked out with the Fermi golden
rule with those obtained via the LDOS for a substrate jellium model. In this way we show that
the radius of the sphere in which we compute the LDOS is the relevant parameter of the latter
simpler approach. We investigate the CCV lineshapes of Na on Al(111), a system in which two
11
ordered phases are observed at room temperature. From a detailed comparison of the calculated
spectra with the experimental ones measured at the Elettra synchrotron for the KL23V lineshape
of Na, we prove that a joint experimental and theoretical effort (based on DFT) can allow for
determining different surface geometries for a given adsorption system, and also for unravelling
different adsorbate components coexisting in the signal.
Electronic Excitations of Cu O

within GW approximation
F. Bruneval, N. Vast, L. Reining
1-Laboratoire des Solides Irradies, CNRS, CEA, Ecole Polytechnique, Palaiseau, France
Cuprous oxide has been extensively studied during the last decades, mainly because of its
exciton series in the optical range. Cu O is a good starting point to address the fundamental
issue of 3d electrons of metals in oxides. This semiconductor material has indeed a cubic
structure, a closed d shell, and is non-magnetic. This is now a very important topic: it is known
that density functional theory fails to predict a gap in various insulating oxides like CoO, CuO.
We performed band structure calculations on Cu O within Density Functional Theory and GW
approximation. We carefully studied the role of semicore states (3s 3p ). Though deep in
energy, these states have a large overlap with valence bands. Their influence is slight on the
Kohn-Sham band structure. However, we state that the semicore states have to be included in
the GW calculationto get meaningful results.
Even a GW calculation including semicore states largely underestimates the quasiparticle gap.
Further approximations are usually used to perform a standard GW calculation: the use of a
plasmon pole model to describe the dynamical screening, a first-order perturbation scheme, an
assumption that LDA and GW wavefunctions are equal... We extensively discuss many of them
and conclude that they are reliable.
Then the failure of GW may lie in the neglect of the vertex part of the self-energy.We therefore
propose a scheme to include some local vertex corrections, thanks to Time-Dependent DFT.
Ab-initio Electronic and Optical Properties of Germanium Quantum

Wires
Mauro Bruno , Maurizia Palummo , Rodolfo Del Sole , Stefano Ossicini
1-Dipartimento di Fisica Universita Tor Vergata Roma
2-INFM-S3 nanoStructures and bioSystems at Surfaces, Dipartimento di Scienze e Metodi dell Ingegneria,
Universita di Modena e Reggio Emilia
We have performed first principles calculations, in order to understand the influence of spatial
orientation on the electronic band structure and optical properties of hydrogenated germanium
quantum wires. Such calculations have been performed by means of a plane wave basis set code
[1]. In particular we have analyzed how the geometrical relaxation affects the electronic and
optical properties. Further we have compared our results with existing similar calculations (see
12
Ref. [2]) where ideal bulk truncated atomic positions have been used. What we have found is
that atomic relaxation determines a small changing (modification) in the wires band structure.
Moreover some optical spectra beyond the one-particle approach will be shown. Important
GW correction and strong excitonic effects have been found. Nevertheless an almost complete
compensation of these two contributions seems to happen.
[1] www.abinit.org
[2] A. N. Kholod, V. L. Shaposhnikov,V. E. Borisenko, F. Arnaud D Avitaya, S. Ossicini, accepted on
Phys. Rev. B
Image potential states and surface electronic structure calculation on the

Na/Cu(111) system
G. Butti , S. Caravati , M.I. Trioni
1-Dipartimento di Fisica - Universit di Milano Bicocca
Image potential states are a common feature of all metal surfaces; these states are originated
by the long range character of the image potential. This long range behavior is usually not
accounted for within standard DFT calculations because of the local character in the common
approximation of the exchange and correlation potential. In our Greens function embedding
approach we overcome this limitation imposing the correct long range behavior by gradually
matching the DFT effective potential with a model one. Within this framework we manage
to reproduce the Rydberg series of image states and obtain an excellent agreement with
experimental results, in a fully ab-initio way. Unfortunately given the static character of our
calculation, we are not able to estimate any intrinsic linewidth, a feature which seem to have
drag very recently a lot of theoretical and experimental attention. The system we chose for
testing our approach is Na/Cu(111) for which a consistent volume of data is available; as an
aside we also give a detailed picture of its surface electronic structure.
Selectivity of Auger decays to the local surface environment.

S. Caravati , G.P. Brivio , M.I. Trioni
1-Dipartimento di Fisica - Universit di Milano Bicocca
In recent years, alkali metals adsorption on metal surfaces has been a matter of both theoretical
and experimental investigation. In particular Na on Al(111) is a nice prototype to study the
interaction between simple metals and alkali atoms. This system show unusual adsorption
geometries inducing different kinds of surface reconstructions depending on temperature and
coverage conditions.
We focus on two ordered structures, both observed at room temperature: a ( R with

a Na atom in a substitutional site and a structure presenting an intermixing between Al and
Na atoms, resulting in a surface alloy. We investigated the electronic properties of the Al(111)
surface and of the two systems described above, applying the embedding method, proposed by
Inglesfield and extended by Ishida. In such a way we are able to consider a real, semi-infinite
13
substrate within the DFT framework, using a Greens function formalism and adopting a fullpotential LAPW basis set.

We then calculated the Auger lineshape for the Na

transition for each of two
reconstructions considered. Although in principle, in this work one have to deal with
excited state due to the presence of a core hole, and with the consequently screening effects,
nevertheless, because of the simple electronic structure of the particular system studied, we
show that a good approximation of the Auger lineshape can be obtained within standard KohnSham approach from the local DOS around the excited sodium adatom.
Calculated Auger spectra present marked differences suggesting the possibility to use such
experimental technique to investigate structural properties. We also show here the results
of the experimental investigation performed on the same kind of systems at the TASC Lab,
using ALOISA beamline. The measured Auger lineshapes perfectly agree with theoretical
predictions, confirming the suitability of combining Auger lineshape analysis and theoretical
calculations to determine local surface morphology.
A novel generalized Kohn-Sham approach for total energy calculations

within Density Functional theory
N. Carneiro and R. Godby
Department of Physics, University of York, Heslington, York YO10 5DD, UK
Generalised Kohn-Sham (GKS) theory [1],[2], provides a variety of possible generalisations of

the well-known Kohn Sham implementation of density functional theory for the calculation of
total energies of systems of interacting electrons such as molecules and solids. We present an
assessment of various GKS schemes for a one-dimensional model semiconductor. Comparison
is made with previous quantum Monte Carlo calculations [3] and with previous GW calculations
within many-body perturbation theory for this system. Further GW calculations suggests a new
GKS approach based on a GW formulation of the total energy, which will lead to a new version
of density functional theory for efficient practical calculations which circumvents the problem
of approximating the usual Kohn Sham exchange correlation energy functional.
[1] A. Seidl, A. Gorling, P. Vogl, and J. A. Majewski and M. Levy; Physical Rev. B 53, 3764 (1996)
[2] P. Sanchez-Friera and R. W. Godby, Phys. Rev. Letters 85, 5611 (2000)
[3] W. Knorr and R.W. Godby. Phys. Rev. Lett. 68, 639-641 (1992)
Quasiparticle energies and optical properties in carbon nanotubes: a

first-principle study
Eric K. Chang , Giovanni Bussi, Alice Ruini, and Elisa Molinari
INFM National Center on nanoStructures and bioSystems at Surfaces (S3)
and Physics Department, University of Modena and Reggio Emilia
41110 Modena, Italy
We present the results of our calculations of excitonic effects and optical properties of carbon
14
nanotubes. We adopt a first-principles approach and use many-body methods such as the BetheSalpeter Equation (BSE), which)has been used with considerable success to calculate excitonic
and most recently, in one-dimensional systems, such as
effects in bulk semiconductors

polymers. The success of the latter reinforces the applicability of the BSE to one-dimensional
systems such as nanotubes. The novelty of our approach consists in the symmetry analysis of
the nanotube in which we take advantage of the invariance of the tube with respect to screwsymmetry operations. Such an analysis not only reduces the computation effort required, but
also offers a better understanding of the physics involved, including the nature of the excitons
and of the optical transitions. We show results for the GW quasiparticle corrections, optical
absorption and Raman spectra.
[1] M. Rohlfing and S.G. Louie, Phys. Rev. Lett. 80 (11): 2312 (1998).
[2] L.X. Benedict, E.L. Shirley, and R.B. Bohn, Phys. Rev. Lett. 80 (20): 4514 (1998).
[3] S. Albrecht, L. Reining, R. Del Sole, et al., Phys. Rev. Lett. 80 (20): 4510 (1998).
[4] A. Ruini, M. J. Caldas, G. Bussi and E. Molinari, Phys. Rev. Lett. 88, 206403 (2002).
The energy loss of (110)Ag and (001)Al surfaces

by ab initio, all electron calculations.
L. Chiodo , P. Monachesi , C.Hogan , R. Del Sole
1-Dipartimento di Fisica - Universit di Roma, Tor Vergata and Istituto Nazionale per la Fisica della Materia
I-00133 Roma (Italy)
2-Dipartimento di Fisica - Universit dellAquila and Istituto Nazionale per la Fisica della Materia ,I-67100
LAquila (Italy)
Since the milestone paper by A. Liebsch [1], aimed to overcome the limitation of the jellium
model to evaluate the loss properties of transition and noble metals, further improvements to
the model [2] yield only qualitative agreement with experimental results. For the purposes
of material science, however, model representation of metals are highly unsatisfactory and
unsuitable for comparison with to-date experiments. We present energy loss functions and
surface plasmon dispersions of (110)Ag and (001)Al surfaces, based for the first time on ab
initio, all electron calculations, performed within DFT-LDA. We use a Full-Potential Linear
Muffin Tin Orbital code, particularly efficient for metals. In this way, the simple metal

aluminum, considered a prototype of the jellium model, and silver, with -electrons, are treated
on the same footing. The linear upward dispersion of the surface plasmon of Ag(110) is
recovered, in quantitative agreement with measurements [3]. It originates from the specific
dielectric response of the surface layer.
[1] A. Liebsch, Phys. Rev. Lett. 71, 145 (1993).
[2] C. Lopez Bastidas et al, Phys. Rev. B 63, 165407 (2001)
[3] M. Rocca et al, Phys. Rev. Lett. 69, 2122 (1992)
15
Electron-phonon interaction and its role in the lifetime broadening

mechanism of electron and hole excitations in metallic materials
E.V.Chulkov, I. Yu.Sklyadneva, A. Leonardo, and P.M. Echenique
Departamento de Fsica de Materiales and CentroMixto CSIC-UPV/EHU, Faculatad de Ciencias Qumicas,
UPV/EHU, Apdo.1072, 20080 San Sebastin/Donostia, Basque Country, Spain; Donostia International Physics
Center (DIPC), Paseo de Manuel Lardizabal 4, 20018 San Sebastin/Donostia, Basque Country, Spain
Departamento de Fsica de Materiales and CentroMixto CSIC-UPV/EHU, Faculatad de Ciencias

Qumicas, UPV/EHU, Apdo.1072, 20080 San Sebastin/Donostia, Basque Country, Spain;
Donostia International Physics Center (DIPC), Paseo de Manuel Lardizabal 4, 20018 San
Sebastin/Donostia, Basque Country, Spain Ab-initio study of electron-phonon interaction and
its contribution to the lifetime broadening of excited electrons and holes in metallic materials
is reported. The energy and momentum resolved Eliashberg function and electron- phonon
coupling parameter are calculated for simple and transition metals. The electron-phonon
contribution calculated at different temperatures is analyzed for sp and d-bands as a function of
momentum and energy. It is shown that the electron-phonon coupling matrix elements strongly
modify phonon density of states supressing some phonon modes contribution. The relative
role of electron-electron and electron-phonon interaction in the lifetime broadening of excited
quasiparticles is analyzed.
Ab initio spectra of the low-pressure phases of zirconia

Louise Dash, Tiphaine Tesson, Virginie Quequet
and Nathalie Vast
Laboratoire des Solides Irradies, CNRS, CEA, Ecole Polytechnique, Palaiseau, France
We have developed ultrasoft pseudopotentials intended to be used in calculations for ab

initio spectra beyond DFT. Ultrasoft pseudopotentials can be of several forms, but the most
widely used are those using methods developed by David Vanderbilt [1] and by Rappe, Rabe,
Kaxiras and Joannopoulos (RRKJ) [2], both of which have produced excellent results for many
problems. For spectroscopy calculations using plane-wave techniques, however, we demand a
high level of convergence in the calculations in order to obtain reliable results. We have found
that many Vanderbilt-type ultrasoft pseudopotentials suffer from poor absolute convergence,
even at extremely high plane-wave cutoffs, and are therefore unsuitable for many applications
in spectroscopy.
Using examples of both types of ultrasoft pseudopotentials, as well as norm-conserving
Bachelet, Hamann and Schluter (BHS) pseudopotentials, we demonstrate this convergence
problem and the implications for calculations of electronic spectra.
References:
[1] David Vanderbilt, Phys. Rev. B 41 7892 (1990)
[2] Andrew M. Rappe, Karin M. Rabe, Efthimios Kaxiras and J. D. Joannopoulos, Phys. Rev.
B 41 1227 (1990); G. Kresse and J. Hafner, J. Phys. Cond. Mat. 6 8245 (1994)
16
Choosing ultrasoft pseudopotentials for spectroscopy calculations

Louise Dash, Tiphaine Tesson, Virginie Quequet
and Nathalie Vast
Laboratoire des Solides Irradies, CNRS, CEA, Ecole Polytechnique, Palaiseau, France
We have developed ultrasoft pseudopotentials intended to be used in calculations for ab

initio spectra beyond DFT. Ultrasoft pseudopotentials can be of several forms, but the most
widely used are those using methods developed by David Vanderbilt [1] and by Rappe, Rabe,
Kaxiras and Joannopoulos (RRKJ) [2], both of which have produced excellent results for many
problems. For spectroscopy calculations using plane-wave techniques, however, we demand a
high level of convergence in the calculations in order to obtain reliable results. We have found
that many Vanderbilt-type ultrasoft pseudopotentials suffer from poor absolute convergence,
even at extremely high plane-wave cutoffs, and are therefore unsuitable for many applications
in spectroscopy.
Using examples of both types of ultrasoft pseudopotentials, as well as norm-conserving
Bachelet, Hamann and Schluter (BHS) pseudopotentials, we demonstrate this convergence
problem and the implications for calculations of electronic spectra.
References:
[1] David Vanderbilt, Phys. Rev. B 41 7892 (1990)
[2] Andrew M. Rappe, Karin M. Rabe, Efthimios Kaxiras and J. D. Joannopoulos, Phys. Rev.
B 41 1227 (1990); G. Kresse and J. Hafner, J. Phys. Cond. Mat. 6 8245 (1994)
Impurities states in hydrogenated silicon nanoclusters

E. Degoli , E. Luppi , R. Magri , O. Bisi , G. Cantele , D. Ninno , G.
Iadonisi , and S. Ossicini
1-INFM-S3 and Universita di Modena e Reggio Emilia, Dipartimento di Fisica, Modena, ITALY,
2-INFM-S3 and Universita di Modena e Reggio Emilia, Dipartimento di Scienze e Metodi dellIngegneria,
Reggio Emilia, ITALY
3-INFM and Universita di Napoli Federico II - Dipartimento di Scienze Fisiche, Complesso Universitario
Monte S. Angelo, Via Cintia, I-80126 Napoli, ITALY
In this work we show an ab-initio study of impurity states in silicon nanocrystals. It is

known that doping semiconductor nanocrystals results in a much less efficient increasing of
the carriers density than for the bulk crystal. In fact, impurities states can be far apart from
band edges, due to confinement effects, so leading to ionisation thresholds much higher than
for the bulk system [1]. Studies on doped semiconductor nanocrystals are still lacking [2].
Here, we study the main structural and electronic properties of doped silicon nanoclusters from
ab initio calculations. The PWscf package [3] has been used, based on density functional
theory, pseudopotentials (the GGA approximation for the exchange-correlation functional is
considered) and plane waves basis set. Spherical silicon nanocrystals are built adding bulk
silicon shells around a central Si atom, up to a given radius. The surface is then passivated
with hydrogen atoms, in order to prevent the cluster from surface reconstruction phenomena.
17
Next, doped nanocrystals are obtained replacing one atom by a trivalent or pentavalent impurity
(substitutional B and N respectively). Non equivalent positions of the impurity within each
cluster are considered, with the aim of investigating the dependence of the calculated properties
on the distance of the impurity itself from the surface. Moreover, the impurity states are
investigated as a function of the nanocluster size. The starting hydrogenated nanoclusters

which have been considered are Si H and Si87H76, whose radii are 0.45 and 0.71 nm
respectively. For each cluster, full relaxation with respect to the atomic positions is carried out.
The structural properties (reconstruction around the impurity, bond lengths and angles, etc.)
have been determined. It is obviously expected that the starting Td symmetry is broken, giving
rise to atoms arrangement. Neutral impurities show lower symmetry due to a missing or an extra
electron. Charged impurities (negatively charged B and positively charged N), instead, retrieve
the original symmetry (tetravalent electronic configuration). Formation energies for both neutral
and charged impurities have been determined. Care has been taken in determining total energy
of charged systems within a supercell calculation, applying the Makov-Payne correction [4].
Finally, an attempt has been done to calculate ionisation thresholds of each impurity species as
a function of size.
[1] M. Lannoo, C. Delerue and G. Allan, Phys. Rev. Lett. 74, 3415 (1995) [2] D. V. Melnikov
and J.R. Chelikowsky, Phys. Rev. Lett. 92, 046802 (2004). [3] S. Baroni, A. Dal Corso, S. de
Gironcoli, and P. Giannozzi, http://www.pwscf.org. [4] G. Makov and M.C. Payne, Phys. Rev.
B 51, 4014 (1995).
Self-consistent Solutions of

Dysons
Equation and Total Energies from the

Approximation
Kris Delaney , Patrick Rinke , P. Garc

a-Gonzalez

Godby
and R. W.
(1) Department of Physics, University of Illinois at Urbana-Champaign, Urbana, IL 61801, U.S.A.

(2) Fritz-Haber-Institut der Max-Planck-Gesellschaft, Faradayweg 4-6, D-14195 Berlin-Dahlem, Germany
(3) Departamento de Fsica Fundamental, U.N.E.D., Apartado de Correos 60141, Madrid, Spain
(4) Department of Physics, University of York, Heslington, York YO10 5DD, U.K.
We present details of the use of many-body perturbation theory, within the

approximation,
in computing ground-state total energies for finite-sized systems. We study the total energy
using the Galitskii-Migdal formula with a Green function derived from the non-, partially- and
fully-self-consistent
approximation. total energy calculations have been successfully
applied for systems with high symmetry [1,2,3], motivating further investigation.
We present the results of these calculations for a family prototype finite-sized systems known as
jellium spheres, and an analysis of the role of the approximate exchange-correlation functional
of the preceeding DFT calculation in the accuracy of the
total energies.
[1] B. Holm and U. von Barth, PRB 57, 2108 (1998)
[2] P. Garca-Gonzalez and R. W. Godby, PRB 63, 075112 (2001)
[3] P. Garca-Gonzalez and R. W. Godby, PRL 88, 056406 (2002)
18
Quasi-particle gap, dielectric properties and electronic correlations in

semiconductor nanostructures.
C. Delerue(*), G. Allan(*) and M. Lannoo(+)
(*) EMN, Dept. ISEN, Lille, (+) L2MP, Marseille
Electronic excitation spectra of semiconductor nanostructures are particularly interesting

quantities to calculate because they can be related to many experimental results in optical or
electrical spectroscopy. On a fundamental point of view, it is also instructive to understand
how the spatial and quantum confinement affect the spectra, in particular the gap. In this talk,
we present recent theoretical works [1-4] concerning the quasi-particle gap of nanostructures,
which required to consider the dielectric properties of the nanostructures.
Thus, in a first part, we consider the following question: how the dielectric properties of
semiconductor nanostructures differ from those of the bulk material? We present calculations
explaining why the average dielectric constant changes with size of the systems. We
demonstrate that this evolution is not due to quantum effects (opening of the gap) and that
it may be explained starting from the bulk dielectric function [1].
In a second part of the presentation, we discuss the influence of these dielectric properties
on the exchange-correlation potential Vxc which is a central quantity in density functional
theory since there is no explicit definition of this quantity. It is well-known that Vxc has a
discontinuity within the gap of semiconductors and insulators. In this context, we have studied
the dependence of this discontinuity in confined systems. Using a tight binding version of
the GW theory [2,3], we show that the discontinuity of Vxc has a non monotonic behavior with
respect to the dimensionality of the systems, from 0D (= zero dimension = dots), 1D (wires), 2D
(wells) to 3D (bulk). We demonstrate that this behavior is due to the presence of polarization
charges at the surface of the nanostructures arising from the exchange-correlation hole. We
explain why the local density approximation gives comparatively better results for 0D systems
than 2D or 1D ones. Finally, we hope to present preliminary results concerning metallic dots.
1.C. Delerue, M. Lannoo, and G. Allan, Phys. Rev. B 68, 115411 (2003). 2.C. Delerue, M.
Lannoo, and G. Allan, Phys. Rev. Lett. 84, 2457 (2000); Erratum: Phys. Rev. Lett. 89, 249901
(2002). 3.C. Delerue, G. Allan, and M. Lannoo, Phys. Rev. Lett. 90, 076803 (2003). 4.C.
Delerue and M. Lannoo, Nanostructures - Theory and Modeling (Springer-Verlag 2004).
Calculation of surface optical properties within TDDFT: the bound

exciton of Si(111)2x1
O. Pulci, M. Palummo, A. Marini, and R. Del Sole
Dipartimento di Fisica Universita di Roma Tor Vergata, Via della Ricerca Scientifica 1, 00133 Italy
In recent years it has been possible to develop methods [1] for calculating optical spectra with
inclusion of the electron hole interaction by solving the Bethe Salpeter equation (BSE) within
the framework of Greens function theory. The calculated spectra are in good agreement with
experiments, although the computational effort is very heavy, since one has to solve a twoparticle equation, for the electron and the hole. A promising alternative method is that based on
19
Time Dependent Density Functional Theory (TDDFT)[2], which treats an effective one-body
problem, where all many-body effects are included in the so called exchange-correlation kernel,

. The kernel is the crucial -yet unknown- quantity needed to calculate accurate
spectra.
In the last three years a breakthrough in the knowledge of has been obtained by comparing
TDDFT with the well tested -yet more complex-BSE method. Two different approaches [3,4]
have yielded the same (approximate) formula for fxc; this has been shown to work very well for
bulk systems, for both semiconductors and insulators, for bound excitons and unbound electronhole pairs[3,4,5]. It also works, within the simplest TDLDA approximation, for small systems
as clusters. However, the TDDFT approach has never been applied so far to surfaces.
In this work we consider the test case of the Si(111)2x1 surface, where an exciton with binding
energy of about 0.27 eV results from BSE calculations and we show that the same result is
obtained within TDDFT. To our knowledge, this is the first time that TDDFT is applied to
calculate surface optical properties.
[1] G. Onida, L. Reining, A. Rubio, Rev. Mod. Phys. 74, 601 (2002)
[2]E. Runge, E.K.U. Gross, Phys. Rev. Lett. 52, 997 (1984)
[3] G. Adragna, R. Del Sole, A. Marini, Phys. Rev. B 68, 165108 (2003)
[4] F. Sottile, V. Olevano, L. Reining, Phys. Rev. Lett. 91, 056402 (2003)
[5] A. Marini, R. Del Sole, A. Rubio, Phys. Rev. Lett. 91, 256402 (2003)
Long-ranged interactions in a supercell approach

P. Eggert , Christoph Freysoldt , Patrick Rinke , Arno Schindlmayr and
Matthias Scheffler

Fritz-Haber-Institut der Max-Planck-Gesellschaft, Faradayweg 46, 14195 Berlin-Dahlem, Germany, Institut

fur Festkorperforschung, Forschungszentrum Ju lich, 52425 Julich, Germany
In todays ab initio electronic structure calculations plane wave basis sets are widely used due
to certain numerical advantages. For systems without translational invariance such as surfaces,
clusters and quantum wires, however, the plane waves introduce an artificial periodicity. A
surface for example is thereby represented by a repeated slab geometry, with vacuum separating
the individual slabs. Vacuum and slab form the repeated unit, the so called supercell for the
calculation. On the level of DFT the issue of spurious long-ranged interactions between these
repeated units has been studied in detail. Through the non-local screened Coulomb potential,
however, for non-charged systems GW becomes more long-ranged than the local DFT-LDA
formalism. In a cluster GW calculation for the sodium tetramer long-ranged interactions up to
40 bohr of vacuum have been observed [1]. Our GW calculations using the space-time approach
[2] indicate that long-ranged interactions also play a dominant role for surfaces. We explain this
20
behaviour by an electrostatic effect und show that only a cutoff of the long-ranged interactions
leads to converged quasiparticle energies. The need to cutoff the long-ranged interactions has
been also discussed very recently for semiconducting nanotubes [4].
As an application, we compare our quasiparticle band structure for the Si(001) c(4 2) surface
with recent two-photon photoemission data [3].
[1] G. Onida, L. Reining, R.W. Godby, R. Del Sole and W. Andreoni, Phys. Rev. Lett. 75, 818 (1995) .
[2] M. M. Rieger, L. Steinbeck, I. D. White, H. N. Rojas and R. W. Godby,
Comput. Phys. Commun. 117, 211 (1999) .
[3] M. Weinelt, M. Kutschera, T. Fauster and M. Rohlfing, Phys. Rev. Lett. 92, 126801 (2004) .
[4] C. D. Spataru, S. Ismail-Beigi, L. X. Benedict and S. G. Louie, Appl. Phys. A 78, 1129 (2004) .
Electronic Excitations in Semiconductor Quantum Dots

A. Franceschetti
Oak Ridge National Laboratory, Oak Ridge TN 37831 , USA
Semiconductor quantum dots can now be synthesized with a high degree of control over
their size, shape, and composition. Recent advances in measurement and characterization
techniques have lead to a wealth of experimental information on the physics of electronic
excitations in these nanostructures. The challenge for theory is to accurately predict the
behavior of complex electronic excited states (such as multi-excitons, charged excitons, etc.)

in semiconductor quantum dots containing typically 10 to 10 atoms. These nanostructures are
too large for first-principles calculations to be practical, and too small for continuum models,
such as effective mass or k.p, to be accurate. I will discuss here a semi-empirical method,
based on pseudopotential wave functions, to calculate electronic excitations in million-atom
semiconductor quantum dots with accuracy comparable with computationally more expensive
first-principles methods. This approach requires the solution of an atomistic pseudopotential
Schroedinger equation to determine the single-particle energies and wave functions of the
quantum dot, followed by a configuration-interaction-like expansion of the excited-state wave
functions to obtain the many-particle excitations. I will present results on: (i) Multi-exciton
radiative and non-radiative recombination in colloidal and self-assembled quantum dots, (ii)
Auger electron-hole scattering and electron cooling in CdSe free-standing nanocrystals, (iii)
Dark and bright exciton lifetimes as a function of temperature and size, and (iv) Mechanisms
for direct carrier multiplication in quantum dots.
21
GW calculations for surfaces
Christoph Freysoldt , Philipp Eggert ,

Patrick Rinke , Arno Schindlmayr , Matthias Scheffler
The combination of density functional theory (DFT) and many-body perturbation theory in
the GW approximation has become an important tool for ab-initio band structure calculations
and is typically in very good agreement with experiment [1]. In order to treat surfaces one
often employs a repeated slab geometry (for computational convenience). However, electric
multipole moments in the slabs lead to a slowly decaying electrostatic interaction between the
slabs. This can be corrected for in DFT total energy calculations in order to obtain a fast
convergence to the limit of an isolated slab [2]. Within GW, multipoles may occur dynamically
even if the time-averaged ground state exhibits no dipole.
Although we avoid direct interactions between periodic images in our approach [3] by
integrating the self-energy in real space, our calculations showed no convergence with respect
to the vacuum spacing in the numerically accessible range. A step-by-step analysis reveals that
the divergent behaviour is introduced by the head and wing elements of the dielectric matrix
perturbation theory. In anisotropic systems such as surfaces,
which are calculated using
head and wings become directionally dependent, which reflects the non-analytic behaviour of
. We show how the full angular dependence of head and
the dielectric function for
wings can be incorporated into the inverse dielectric matrix in an efficient manner. However,
the convergence problem persists in the angle resolved implementation.

An alternative approach to suppress the unwanted interaction that leads to a sufficiently fast
convergence with respect to the slab separation is presented for silicon surfaces.
[1] W.G. Aulbur, L. Jonsson, J.W. Wilkins. Quasiparticle Calculations in Solids, volume 54,
edited by H. Ehrenreich (Academic, Orlando), 1999
[2] J. Neugebauer, M. Scheffler, Phys. Rev. B 46, 16067 (1992)
[3] M.M. Rieger, L. Steinbeck, I.D. White, H.N. Rojas, R.W. Godby, Comp. Phys. Com. 117,
211 (1999)
Ab initio study of the electronic structure of magnetite

Martin Friak , Arno Schindlmayr , and Matthias Scheffler

We analyzed the influence of internal and external conditions on the electronic structure of
magnetite, specifically the half-metal to metal transition. The latter leads to a loss of the
materials desirable half-metallic properties. Total energies, electronic structure, and magnetic
22
moments are calculated by density functional theory (DFT) using the FP-LAPW method.
Results obtained within the generalized gradient approximation are in excellent agreement
with experimental findings. In response to uniaxial, biaxial or triaxial pressure, a half-metal
to metal transition occurs, which shifts the Fermi energy from the gap of the majority-spin
electrons under the top of the valence band so that both spin channels become metallic. We also
performed an analysis of different structural and spin configurations simulating the substitution
Fe X O (X = Mn, Co, Ni). The Mn-doped crystal is metallic whereas Co- and Ni-doped
ones are half-metallic.
Photocatalytic Oxidation of CO at the V O (010) Surface: Cluster Model

Studies
C. Friedrich and K. Hermann
1-Institut f r Festk rperforschung, Forschungszentrum J lich, 52425 Jlich, Germany
2- Abt. Theorie, Fritz-Haber-Institut der Max-Planck-Gesellschaft, Faradayweg 4-6, 14195 Berlin, Germany
While the catalytic behavior of vanadium oxide has been known and utilized for a long
time many microscopic details of oxidation reactions at vanadium oxide surfaces are still
not fully understood. Experiments indicate that CO binds weakly at the clean surface
but can adsorb strongly at oxygen vacancies of the reduced surface. In addition, there is
experimental evidence that local electronic excitations at the clean surface can facilitate
CO oxidation involving surface oxygen. In this work we use embedded cluster models as large
) ) )
together with DFT methods to examine the effect of local electronic excitations
as
and the influence of O vacancies on the interaction of CO with the (010)surface. The
electronic structure of the surface in its ground state is modelled by the singlet cluster while local
excitations are introduced by singlet to triplet excitations. Our calculations show, in qualitative
agreement with experiment, that local excitations of the surface can promote oxidation
of CO while oxygen vacancies do not facilitate this reaction. Molecular CO adsorption can,
however, take place at oxygen vacancy sites.
Self-energy and lifetimes of surface states beyond the GW approximation

M. Garcia-Vergniory , J. M. Pitarke , and P. M. Echenique
1 Materia Kondentsatuaren Fisika Saila Zientzi Fakultatea Euskal Herriko Unibertsitatea
We report many-body calculations of the self-energy and the lifetime of Shockley and image
states on the (100) and (111) surfaces of Cu. The self-energy is computed either in the absence
of short-range correlation effects (GW approximation) or by including short-range correlation
effects both in the screened interaction W (beyond the random-phase approximation) and in
the expansion of the self-energy in terms of the W (beyond the GW approximation). Different
approximations have been chosen to describe the exchange-correlation kernel, and the impact
of these approximations will be discussed.
23
Effects of oxygen on the structural, electronic and optical properties of

silicon nanocrystals
M. Gatti, G. Onida
Istituto Nazionale per la Fisica della Materia, Dipartimento di Fisica dellUniversita di Milano, via Celoria 16,
I-20133 Milano, Italy
Silicon clusters are ideal models to study the effects related to the reduction in size of semiconductors. In these systems, both quantum confinement effects (opening of quasiparticle gap,
enhancement of excitonic effects), and surface effects (passivation, oxidation, reconstruction,
defects) play a fundamental role.
Starting from the equilibrium geometry of the ground state of Si H computed with the CarParrinello method in DFT-LDA, we have studied how the structure of this cluster changes when
an O atom replaces a couple of H atoms, in the case of formation of either a double Si=O bond
or a bridge Si-O-Si bond.
Then we have calculated the absorption spectra of these clusters, at different degrees of
approximation. In particular, we have studied within RPA the effects of the non locality of
the pseudopotentials, those due to local fields and, finally, the effects of geometrical relaxation
in the excited state.
Self-consistency in GWcalculations of spectral and ground-state properties

Kris Delaney , Patrick Rinke , Pablo Garca-Gonzalez , Angel

Rubio , Tim Gould and Rex Godby

Department of Physics, University of York, Heslington, York YO10 5DD, United Kingdom
Physics Department, 1110 West Green Street, University of Illinois at Urbana-Champaign, Urbana, 61801
Illinois, USA
Fritz-Haber-Institut der Max-Planck-Gesellschaft, D-14195 Berlin, Germany

Departamento de Fsica de la Materia Condensada, Universidad Auto noma de Madrid, 28049 Madrid, Spain
Departamento de Fsica Fundamental, Universidad Nacional de Educacio n a Distancia, Apartado de Correos

60141, 28080 Madrid, Spain
Dpto. Fsica Materiales, UPV, Centro Mixto CSIC-UPV, and DIPC, San Sebasti a n, Spain
The
approximation is, of course, the first-order term in an expansion of the self-energy
operator in powers of the dynamically screened electron-electron interaction, . From ,
the one-electron Greens function may be calculated, from which various spectral and groundstate properties are available. Different ways of deriving the approximation from exact theory
yield different choices of whether and/or are made to be consistent with the Greens
function that arises from : , and fully self-consistent , where generally
indicates the LDA Greens function. We present results for finite and infinite systems at all three
levels of self-consistency, including both spectral and total-energy calculations [1-3]. These
include converged total energies resulting from the incorporation of self-consistency and

total-energy techniques into our general-purpose GWST space-time supercell code suite [4],

24
which interfaces with pseudopotential plane-wave DFT calculations as its input. This opens
the possibility of applications to systems whose energetics is not described sufficiently reliably
within the usual DFT-based approaches.
We also discuss the role of pseudopotentials in
consistency[5].
calculations and the interplay with self-
[1] Many-body GW calculations of ground-state properties: quasi-2D electron systems and van
der Waals forces, P. Garca-Gonzalez and R. W. Godby, Phys. Rev. Lett. 88 056406 (2002)
[2] Accurate total energies from the self-consistent method: application to silicon, Tim
Gould, Kris Delaney, P. Garca-Gonzalez, and R.W. Godby, to be submitted
[3] Total Energy Calculations for Inhomogeneous Systems, Kris Delaney, Patrick Rinke,
P. Garca-Gonzalez, Tim Gould and R. W. Godby, to be submitted [4] Space-time method for
ab initio calculations of self-energies and dielectric response functions of solids, H. N. Rojas,
R. W. Godby and R. J. Needs, Phys. Rev. Lett. 74 1827 (1995)
[5] Comment on Band-Gap Problem in Semiconductors Revisited: Effects of Core States
and Many-Body Self-Consistency , Kris Delaney, P. Garca-Gonzalez, Angel Rubio, Patrick
Rinke and R.W. Godby, accepted for publication in Phys. Rev. Lett.
Further details at http://www-users.york.ac.uk/ rwg3/.
Ab-initio theory of superconductivity

E.K.U. Gross
Institute for Theoretical Physics, Free University Berlin, Arnimallee 14, 14195 Berlin, Germany
One of the great challenges of modern condensed-matter theory is the prediction of materialspecific properties of superconductors. The traditional model of Bardeen, Cooper and Schrieffer
(BCS) properly describes the universal features of (conventional) superconductors but is
not able to make accurate predictions of material-specific properties such as the critical
temperature. To tackle this problem, a density functional formalism has been developed which
describes superconductors in thermal equilibrium in terms of three quantities: the ordinary
electron density, the superconducting order parameter, and the nuclear N-body density matrix.
These three densities are determined by a set of Bogoliubov-type Kohn-Sham equations
representing the electronic degrees of freedom, and a Schrodinger equation with an N-body
interaction describing the nuclear motion. These equations are coupled to each other via
exchange-correlation (xc) potentials which are universal functionals of the three densities.
Approximations of these xc functionals will be derived on the basis of many-body perturbation
theory. In this way, a true ab-initio description is achieved which does not contain any empirical
parameters. Numerical results will be presented for the critical temperatures and the gaps

of simple metals, of
, and of Li and Al under pressure. Superconductors with strong
and with weak electron-phonon coupling are equally well described. For other than phonondriven superconductors, we will discuss how an effective particle-particle interaction and hence
the superconducting mechanism can be identified, once an approximation for the universal
exchange-correlation functional is given.
25
Theory of electron energy loss at surfaces

C. Hogan, R. Del Sole
Department of Physics and INFM, University of Rome Tor Vergata, Italy
Electron energy loss spectroscopy is a frequently exploited technique for characterizing

materials. When carried out at low energies, a substantial proportion (about 1%) of electrons
are backscattered inelastically as they excite a variety of transitions within the surface region.
Depending on the incident and loss energy regimes, these excitations may be related to
very different processes, such as atomic vibrations, bulk or surface plasmons, interband and
intraband transitions (both core and valence), and so on.
Correspondingly, different theoretical approaches have been developed to interpret the observed
spectra. In the 1-20 eV range, where spectral features are predominantly related to interband
transitions and surface plasmons, bandstructure effects should be most important. Hence, a
proper description of the loss function should incorporate the full microscopic inverse dielectric
function of the vacuum-surface interface, calculated ab-initio, and including the effects of local
fields and q-dependence.
We present a new method that accounts for these features within the limit of dipole scattering
theory, improving on previous approaches based on three-layer models. The flexibility of the
model is illustrated via studies of the loss anisotropy of some typical semiconductor surfaces
and the surface plasmon dispersion in a simple metal.
Ab initio study of reflectance anisotropy spectra of a sub-monolayer

oxidized Si(100) surface
A. Incze

, R. Del Sole , G. Onida

1-Istituto Nazionale per la Fisica della Materia, Dipartimento di Fisica dellUniversita di Milano, via Celoria
16, I-20133 Milano, Italy
2-Istituto Nazionale per la Fisica della Materia e Dipartimento di Fisica, dellUniversita di Roma Tor Vergata
Via della Ricerca Scientifica 1, I-00133 Roma, Italy
The effects of oxygen adsorption on the reflectance anisotropy spectrum (RAS) of reconstructed
Si(100):O surfaces at sub-monolayer coverage (first stages of oxidation) have been studied by an
ab initio DFT-LDA scheme within a plane-wave, norm-conserving pseudopotential approach.
Dangling bonds and the main features of the characteristic RAS of the clean Si(100) surface
are mostly preserved after oxidation of 50% of the surface dimers, with some visible changes:
a small red shift of the first peak, and the appearance of a distinct spectral structure at about
1.5 eV. The electronic transitions involved in these changes have been analyzed through stateby-state and layer-by-layer decompositions of the RAS. We suggest that new interplay between
present theoretical results and reflectance anisotropy spectroscopy experiments could lead to
further clarification of structural and kinetic details of the Si(100) oxidation process in the submonolayer range.
26
First-Principle Greens function methods : electronic and optical

properties of carbon nanotubes and excited-state forces
S. Ismail-Beigi
Departement of Applied Physics, Yale University P.O. Box 208284, New Haven, USA
Ab initio predictions of electronic and optical properties of materials within the GW-Bethe
Salpeter Equation (GW-BSE) formalism have proved accurate and reliable. The formalism
allows for the study of electronically excited states and outputs experimentally measurable
quantities such as band gaps and optical spectra. We present two recent research results in
this area: the optical spectra of carbon nanotubes and the calculation of excited-state forces.
The work on small-diameter single-walled carbon nanotubes underlines the physics of Coulomb
interactions of electron-hole pairs in one-dimensional systems and its crucial effect on optical
properties. In addition, inclusion of all the relevant excited-state many-body effects produces
results in quantitative agreement with experiment. This accuracy allows for theoretical
assignment of various observed absorption lines to particular nanotube chiralities.
The ability to compute excited-state forces allows us to move beyond fixed atomic configurations. This advance opens the door to the calculation of the atomic rearrangements of electronically excited-systems as well as concomitant modifications of optical properties such as Stokes
shifts and changes in luminescent intensity. We outline the theoretical background and tests on
molecular systems. We illustrate the method with an application to an interesting solid-state
problem: exciton self-trapping in quartz.
Quasi-particle self-consistent GW approximation

T. Kotani , Mark van Schilfgaarde ,Sergey V. Faleev , Takashi Miyake
1 Osaka University, Toyonaka 560, Japan, 2 Arizona State University, Tempe, AZ 85284, US, 3 Sandia National
Laboratories, Livermore, CA 94551, USA 4 Tokyo Institute of Technology, Tokyo 152-8551, Japan
We present some progress in the development of a semi self-consistent GW method based on

the all-electron, full-potential LMTO method. By iterating the GW Hamiltonian, whose energy
dependence is ignored in a few reasonable ways (we tested some possibilities), self-consistency
in both the eigenfunctions and eigenvalues is achieved. Our intent is to obtain a reasonable
approximation for the best quasi-particle (QP) eigenfunctions and eigenvalues. QPscGW gives
directly the QP density of states (QP-DOS) in the sense of independent-particle picture, which
can be justified by the Landaus Fermi liquid theory. Thus we call our semi self-consistent GW
method as the QP self-consistent GW (QPscGW) method. The QP bands can be used directly
within a one-particle picture to evaluate quantities, e.g. response functions, optical properties,
magnetic moments, transport, and in the construction of the Boltzmann equation.
In addition, we think that the independent-particle picture should be a rigid starting point even
when we try to go beyond the one-particle picture. The method may be viewed as a way
construct the most suitable quadratic part of the full Hamiltonian for many-body perturbation
theory. As for the spectrum DOS (SP-DOS), the imaginary part of the full Green function
27
G, it can be calculated by a one-shot GW calculation (or GW+extensions) from the QPscGW

result. This situation is in contrast with the full self-consistent GW method (full scGW), which
makes SP-DOS self-consistent. Compared the full scGW, QPscGW is advantageous in two
ways: (1) It is numerically easier and more stable. We have successfully applied it to wide
range of materials. (The ability to do so is also due to a mixed-basis expansion of the Coulomb
interaction, as will be described.) (2)There is self-consistency in both the QP bands and the
dynamically screened Coulomb interaction W (they are obtained at the same time). By contrast,
the full scGW may give problematic W, or more precisely it should be identified not as the
dynamical screened interaction but as an intermediate construction during calculation. There
are theoretical predictions that the full scGW would give poor G.
We applied our QPscGW to metals, semiconductors, oxides, magnetic materials, f-electron
materials (in progress) and so on. Our results show very systematic improvement over LDA
compared with experiments. We present two key findings: (1) large improvements are found for
materials where one-shot GW fails. (2) The discrepancies with experiment are very systematic,
and can be explained in terms of what GW theory omits from the exact hamiltonian. As for
semiconductors, not only band gaps but many other properties are improved, such as effective
mass. For NiO and MnO, d-d splitting and relative positions of oxygen 2p bands move closer
to experimental results. Transition metals show a small, systematic d-band width narrowing
relative to LDA. We also present recent results for MnGaAs and half-metallic compounds.
Finally, we report some progress on attempts to solve the Luttinger-Ward functional to compute
the total energy.
Orbital Functionals in Current-Density Functional Theory

Stefan Kurth, Stefano Pittalis, and Nicole Helbig
A proper description of many-electron systems in an (external or internal) magnetic field within

a density functional framework requires the use of the current-density functional theory (CDFT)
of Vignale and Rasolt [1]. Unlike in ordinary spin-density functional theory, where only the
coupling of the magnetic field to the spin degrees of freedom is taken into account, in CDFT
one also allows for coupling to the orbital degrees of freedom. Such a CDFT formalism most
likely is required for a proper description of open-shell systems where the ground state may
carry a finite current. The usefulness of CDFT in practice depends on the availability of accurate
approximations to the exchange-correlation (xc) energy. A local-density approximation based
on the xc energy per electron of the uniform electron gas exposed to a uniform magnetic field
has been suggested. However, since this xc energy has derivative discontinuities as function
of the field whenever a new Landau level is occupied, the corresponding xc potential becomes
singular and difficult to handle in practice.
Here we investigate the use of explicitly orbital-dependent functionals as alternative route. In
fact, one might expect orbital functionals to be well suited for this situation since the appearance
of Landau levels intrinsically is an orbital effect. We show the equivalence of two different
formulations of CDFT if certain conditions on the xc energy functional are satisfied. We also
28
elucidate the connection to ordinary spin-DFT for ground states with vanishing currents. We
then use the optimized effective potential (OEP) formalism for CDFT recently developed in
our group [2] to derive integral equations for the xc scalar and vector potentials. Simplifying
approximations to these integral equations lead to a numerically tractable scheme to calculate
the xc potentials. First results on atomic systems will be presented.
Ab initio study of the Fe O (111) surface

A.E. Kuznetsov, M. Friak, M. Scheffler

Fritz-Haber-Institut der Max-Planck-Gesellschaft, Faradayweg 4-6, D-14195 Berlin, Germany
Metal oxides represent an important class of materials found in environmental science,

electrochemistry, biology, chemical sensor technology, catalysis, magnetism and other fields.
However, our understanding of basic metal oxide surface chemistry and mechanisms of catalytic
reactions on metal oxides greatly lags behind that of semiconductors and metals. Magnetite
(Fe3O4), the oldest known magnet, is one of the most abundant transition metal oxides. The
Fe3O4 (111) surface can be considered as a stack of layers containing either metal cations or
oxygen anions. The stacking of layers with opposite charges gives rise to a large dipole moment
perpendicular to the surface, which the system may reduce by large interlayer relaxations.
Moreover, for such oxide surfaces different surface compositions are possible and should be
taken into account. Despite a large amount of research, the Fe3O4 surface structures and
properties have not been understood fully yet. Our calculations are performed employing
the full-potential linearized augmented plane-wave method (WIEN2k). We will describe our
project and show first results for the relative stability of different Fe3O4 (111) terminations and
their electronic and magnetic properties.
Nonlinear optical coefficients from first-principles calculations:

Different numerical approaches
R. Leitsmann, W.G.Schmidt, and F. Bechstedt
Friedrich-Schiller-Universitat, Institut fur Festkorpertheorie und -optik, Max Wien Platz 1, 07743 Jena, Germany
Different numerical and methodological approaches for the ab initio calculations of the
nonlinear optical response are used in the literature[1,2]. On the one hand, the formulas
used are not easily seen to be equivalent. On the otherhand, the numerical results even for
well characterized bulk systems like GaAs do neither agree well with each other nor with
experiment[3]. We revisit the derivation of the relevant expressions for the second harmonic
generation and show analytically the equivalence of the approaches[1,2] within the independentparticle approximation. In addition, numerical results for bulk GaAs are presented which show
the large influence of technical parameters on the final results.
[1] J.L.P. Hughes and J.E. Sipe, Phys. Rev. B 53, 10751 (1996).
[2] B. Adolph and F. Bechstedt, Phys. Rev. B 57, 6519 (1998).
[3] S. Bergfeld and W. Daum, Phys. Rev. Lett. 90, 036801 (2003).
29
An ab initio study of new polymeric crystals based on fullerene

carbynes
and
S. Lisenkov, L.A. Chernozatonskii

Institute of Biochemical Physics of RAS, 119991, Moscow, Russian Federation
The most basic fullerene structure is the molecule. Since its discovery in 1985, has
been used as the fundamental component of a variety of new carbon nanostructures. Examples
include endohedral fullerenes, peapod nanotubes, dimers, and single-molecule transistors.
Using density-functional calculations, we study a new polymeric form based on fullerene

and linear chain of carbon atoms (cumelens). The structures considered in the work contain
from two to six carbon atoms in cumelens chain between fullerenes . Two different stable
states of these structures have been found. For odd numbers of carbon atoms, adjacent fullerenes
in the chain are rotated in with respect to each other. For even numbers of carbon atoms
between fullerenes the effect did not occur.
We found, that all structures are semiconductors with band gap of the range 1.19 - 1.36 eV. It
also shown that structure containing two carbon atoms between fullerenes is energetically most
favorable. We have compared the energy per atom of all polymeric structures with different
dimensional with two carbon atoms between fullerenes. The most energetically favorable
structure is 3D structure.
A threepoint ManyBody vertex function from TimeDependent

DensityFunctional Theory
A. Marini, A. Rubio
Departamento de Fsica de Materiales and Donostia International Physics Center. E20018 San Sebasti a n,
Basque Country, Spain
Thanks to the recently developed ManyBody expressions for the exchangecorrelation kernel
TimeDependent DensityFunctional Theory (TDDFT) has become an extremely powerful
technique to calculate correlation effects in the microscopical polarization function. In this
talk we will investigate the relation between the ManyBody and the selfenergy vertex
function. We will show that it is possible to obtain a practical and general expression for the
threepoint vertex function based on the twopoint exchangecorrelation kernel. Taking LiF as
a test case we will prove that, in contrast to previous results for the electron gas, simple metals
and semiconductors, vertex corrections in the selfenergy and in the screening function do not
compensate each other.
30
Absorption spectra of biological systems from TDDFT

M. Marques, X. Lopez, D. Varsano, A. Castro, and A. Rubio
Departamento de Fsica de Materiales and Donostia International Physics Center. E20018 San Sebasti a n,
Basque Country, Spain
Not surprisingly, the theoretical understanding of biophysical processes is a very active field
of research. In particular, there have been spectacular advances in the characterisation of
structural and dynamical properties of complex biomolecules by a combination of quantummechanical and classical-molecular mechanics methods (QM-MM). However, and in spite of
the large amount of experimental work in photo-active molecules, the theoretical description
of the interaction of these molecules with external time-dependent fields is very much in its
infancy. Photo-active molecules relevant for biology include the green fluorescent protein
(GFP), retinal, chlorophyll, etc. On the other hand, time dependent density functional (TDDFT)
theory has proved to be an invaluable tool for the calculation of excitation spectra of molecules.
We will present a way to combined QM-MM methods (for the ground state) with TDDFT (for
the description of the excited states) to calculate optical absorption spectra. Our first test case,
the GFP, yielded remarkably good results. Very recent work on the blue fluorescent protein and
on the DNA basis will also be reviewed.
Electronic band structure of the C(111)2x1 surface within an iterative

GW scheme
M. Marsili , O.Pulci , F. Bechstedt , R. Del Sole
1-Dipartimento di Fisica - Universit di Roma, Tor Vergata and Istituto Nazionale per la Fisica della Materia
I-00133 Roma
2-Friedrich-Schiller-Universitat Jena, Institut fur Festkorpertheorie und Theoretische Optik, Max-Wien-Platz 1,
D-07743 Jena Germany
Theoretical calculations show that the most stable reconstruction for the (111) surface of
diamond is consistent with a Pandey Chain model, being the atoms on the chain neither buckled
nor dimerized. Experimental data, while confirming the absence of dimerization, cannot
exclude the presence of buckling. The electronic band structure related to the fully relaxed
(111) surface of diamond appears semimetallic within DFT, in contrast with experiments which
show a gap of at least 0.5 eV. The underestimation of the electronic gap could be due to a
wrong theoretical equilibrium geometry (absence of buckling) and/or to the usual DFT band
gap problem. Introducing quasiparticle corrections within a self-consistent GW scheme, using
an iterative procedure for updating the quasiparticle energies, we found for this surface an
electronic gap of 0.5 eV. Preliminary results on the optical properties are also presented and
discussed.
31
Exchange in non-equilibrium systems

H. Mera and R. Godby
Department of Physics, University of York, Heslington, York YO10 5DD, UK
State-of-the-art simulation tools for non-equilibrium quantum transport systems typically take
the current-carrier occupations to be described in terms of the equilibrium distribution functions
of two particle reservoirs characterised by two different electrochemical potentials, while for the
description of electronic exchange and correlation, the Local Density Approximation (LDA) to
Density Functional Theory is generally used[1],[2]. However,this involves an inconsistency
because the LDA is based on the homogeneous electron gas inequilibrium, while the system is
not in equilibrium (and may be far from it).
In this work we remove this inconsistency by deriving local density approximations for the
non-equilibrium exchange potential, based on non-equilibrium homogeneous electron systems
modelled in the spirit of Landauer-Buttiker theory. I-V characteristics calculated with both,
equilibrium and non-equilibrium, local exchange potentials are then compared for a simple
model non-equilibrium metal-vacuum-metal junction.
[1] M. Brandbyge et al, Phys. Rev. B. 65, 165401 (2002)
[2] N. D. Lang, Phys. Rev. B. 52, 5335 (1995)
Ab-initio calculations of the optical properties of phase-change materials
M. Niesert, G. Bihlmayer, S. Blugel

Institut fur Festkorperforschung, Forschungszentrum Ju lich, 52425 Julich, Germany
The results of ab-initio calculations of the dielectric function and the reflectivity of bulk GeTe
and AgTe systems will be presented. These materials are promising candidates for memory
devices exploiting the change of the optical properties depending on the structural phasechange and are thus subject of intensive investigations. Different atomic configurations have
been computed and the influence on the dielectric function has been investigated. We focus
on the degree of symmetry, comparing cubic with lower symmetry. The dielectric function
considering the long-wavelength limit and neglecting local-field effects, is implemented within
the FLEUR code, an implementation of the full-potential linearized augmented plane-wave
(FLAPW) method.
Ab-initio calculation of the structural, electronic and optical properties of

Silicon nanoclusters in their ground and excited states
S. Ossicini
INFM-S3 and Universita di Modena e Reggio Emilia-DISMI, Reggio Emilia, ITALY
Silicon-based light emitting materials, such as porous silicon and Si nanocrystals, have attracted
a lot of interest in the last years because of the promising applications in advanced electronic
32
devices based on nanoscale technology. Understanding and improving the properties of these
novel semiconductor materials requires to learn more about the atomistic aspects of their
structure and fabrication. Passivated-surface silicon clusters are the ideal theoretical model
to achieve this aim. The most part of first principle calculations and empirical simulations
have used hydrogen as passivating agent for the Si surface, focusing the efforts mainly on
the role of quantum confinement effects. Yet, it has been recently pointed out that also
the surface chemistry produces substantial impact on the Si clusters properties. Here we
investigate the structural, electronic and optical properties of small Si clusters using different
approaches (pseudopotential methods, TDDFT, GW-BSE) within the Density Functional
Theory. To explain the observed Stokes shifts between the absorption and emission energies
we have performed calculations for each cluster both in ground and excited state configuration
considering as the excited state the electronic configuration in which the highest occupied
single-particle state (HOMO) contains a hole (h), while the lowest unoccupied single-particle
state (LUMO) contains the corresponding electron (e). Structural relaxations have been fully
taken into account in all cases through total energy minimization calculations. We have first
considered clusters covered by hydrogen and then we have substituted some H with O both
double bonded or in a bridge configuration with respect to Si. The substitution of H with O
as passivating agent results in a different cluster geometry and in a reduction of the energy
band gap depending on the type of O-Si bond. Moreover also the optical properties strongly
depend on the different O-Si bond type. The results provide a consistent interpretation of
the photoluminescence redshift observed in oxidized samples and of recent outcomes on Si
single quantum dot photoluminescence bandwidth. We have also performed calculations for
the optoelectronic properties of small Si clusters embedded in a SiO2 matrix; the equilibrium
structure shows a deformation of the SiO2 cage only in the neighborhood of the cluster. This
strained interface region, about 1 nm thick, participates to the light emission processes as
supported by experimental results, too.
Work done in collaboration with R. Magri (1), E. Degoli(2), E. Luppi(1), M. Luppi(1), O. Bisi(2), G. Cantele(3),
D. Ninno(3), G. Onida(4), M. Gatti(4), A. Incze(4), O. Pulci(5), R. Del Sole(5)
(1)INFM-S3 and Universita di Modena e Reggio Emilia, Dipartimento di Fisica, Modena, ITALY
(2)INFM-S3 and Universita di Modena e Reggio Emilia-DISMI, Reggio Emilia, ITALY
(3)INFM and Universita di Napoli Federico II - Dipartimento di Scienze Fisiche, Complesso Universitario
Monte S. Angelo, Via Cintia, I-80126 Napoli, ITALY
(4)INFM and Universita di Milano, Dipartimento di Fisica, Milano, ITALY
(5)INFM and Universita di Roma Tor Vergata, Roma, ITALY
Electronic transport in nanoscale systems from first principles.

J. J. Palacios
Departamento de Fisica Aplicada, Universidad de Alicante, San Vicente del Raspeig, Alicante 03690, Spain
I will review some of the key issues concerning ab initio electronic transport in nanosystems,
focusing on the achievements and limitations of Landauer formalism. This widely accepted
formalism, when properly combined with first principles calculations, can give very satisfactory
33
results both at zero and finite bias voltages in a range of situations. Its application in the latter
case, however, cannot be fully justified and remains an open issue. As illustration, various
results for metallic nanocontacts will be presented.
Microscopic model for a strongly correlated superconductor

single-electron transistor
E. Perfetto, M. Cini
Dipartimento di Fisica - Universit di Roma, Tor Vergata and Istituto Nazionale per la Fisica della Materia
I-00133 Roma (Italy)
We show how we can model a Superconducting Single-Electron Transistor (S-SET) operating

by purely repulsive electron-electron interactions. The model consists of properly linked
Hubbard clusters placed between electrods and capacitively coupled to a gate potential. In
the linear regime the conductance exhibits well separated peaks due to Andreev reflection,
reproducing the typical response of a S-SET. This illustrates the W=0 pairing mechanism and
its potential to describe the superconducting behavior of properly symmetric Hubbard sytems.
Large crystal local-field effects in the dynamic structure factor of TiO

I. G. Gurtubay, J. M. Pitarke , W. Ku, A. G. Eguiluz, B.C. Larsom, J. Z.

Tischler, and P. Zschack
Department of Physics, University of the Basque Country, Bilbao Spain
We present ab initio time-dependent-density-functional calculations and non-resonant inelastic

x-ray scattering measurements of the dynamical structure factor of rutile TiO . Our calculations
are in very good agreement with experiment and prove the presence of large crystal local-field
effects below the Ti M-edge, which yield a sharp loss peak at 14 eV featuring a remarkable nonmonotonic dependence on the wave vector. These effects, which impact the excitation spectra
in the oxide more dramatically than in transition metals, provide a signature of the underlying
electronic structure.
Optimized Effective Potentials in Current-Density Functional Theory

S. Pittalis, S. Kurth, N. Helbig
In a density functional framework, the description of many-electron systems moving in (external

or internal) magnetic field is achieved by current-density functional theory (CDFT) [1]. Using
the Optimized Effective Potential (OEP) formalism for CDFT [2] with the exact exchange
functional,we investigate an old problem in DFT: in an open-shell atom with, e.g., one occupied
p-orbital, the density will be different depending on whether the p-orbital with magnetic
or
is occupied. Therefore, ordinary DFT will incorrectly give
quantum number
different total energies. The problem can be alleviated [3], by using current density functionals.
34
Here we present a practical scheme which allows for the calculation of the exchange-correlation
scalar and vector potentials of the OEP-CDFT approach as well as first numerical results for
atoms.
[1] G. Vignale, M. Rasolt, Phys. Rev. Lett. 59, 2360 (1987); Phys. Rev. B 48, 8547 (1988)
[2] N. Helbig, E.K.U. Gross, to be published
[3] A.D. Becke, J. Chem. Phys. 117, 6935(2002)
Band gaps in GW self-consistent calculations

V.A. Popa, G. Brocks, P.J. Kelly
University of Twente, Enschede, The Netherlands
The GW approximation of the self-energy has been very successful in describing single
particle excitations. For example it leads to correct band gaps in inorganic semiconductors
and insulators. However, for systems wrongly predicted by LDA to be metallic (LDA
wave functions and eigenvalues are used as starting point for GW calculations), the standard
perturbative approach of GW is no longer effective. We go beyond this approach, by iterating
the GW procedure in two alternative schemes: (i) only on quasiparticle eigenvalues and (ii)
on eigenvalues and eigenfunctions (full charge self-consistency). In this contribution the two
iteration schemes are compared. In addition, the effect of semicore states is investigated.
The calculations are performed on a range of systems and the band gaps are compared to the
experimental values.
Understanding the dissociation of diatomic molecules in the DFT-ACFD

formalism : ground-state and excited-state studies
Y. Pouillon, F. Tournus , Y.-M. Niquet, X. Gonze
Unite de Physico-Chimie et Physique des Materiaux (PCPM) Universite Catholique de Louvain 1348
Louvain-la-Neuve (Belgium)
There is a growing need for exchange-correlation (XC) functionals going beyond the LocalDensity and Generalized-Gradient Approximations (LDA and GGA), as these do not account for long-range correlations occuring for instance in weakly-bonded compounds and
biomolecules. In such systems van der Waals (vdW) interactions play a crucial role and can
only be taken into account by means of fully non-local XC functionals. Contrary to the LDA
and GGA cases, most of the latter depend on the Kohn-Sham orbitals and their energies. One
way to build such a functional is to resort to the Adiabatic-Connection Fluctuation-Dissipation
(ACFD) theorem [1] [2] in conjunction with the Random-Phase Approximation (RPA). The
RPA captures the physics of the vdW interactions and has been tested non self-consistently on
some atoms, diatomic molecules, and solids.
However, a spurious maximum in the dissociation curve was evidenced for the N [3] and H [4]
molecules. It may result from either (a) the RPA being insufficient to describe the dissociation
35
regime, or (b) the lack of self-consistency, as well as (c) H and N being pathological
cases, without forgetting (d) a possible bug in the implementation. n parallel with studies
concerning points (a), (b) and (d), we are tackling point (c) by examining the dissociation
of several diatomic molecules within the ACFD-RPA framework as well as their excitation
energies through Time-Dependent DFT (TDDFT) and SCF calculations. We compare some
of their ground-state properties with the results obtained from LDA/GGA approximations too.
In addition of being heteroatomic, and composed of low Z atoms, the LiH molecule presents
very peculiar properties and is often used as a test case for new computational methods. That is
why we have decided to put a particular emphasis on it. To further enhance our understanding
of weakly-bonded systems we are also considering the van der Waals molecules Be and Mg .
In particular, it is known that the vdW interaction is quite larger in the latter than in all those
previously tried.
[1]W. Kohn, Y. Meir, and D.E. Makarov, Phys. Rev. Lett. 80, 4153 (1998).
[2] M. Fuchs and X. Gonze, Phys. Rev. B 65, 235109 (2002).
[3] F. Furche, Phys. Rev. B 64, 195120 (2001).
[4] M. Fuchs, K. Burke, Y.-M. Niquet, and X. Gonze, Phys. Rev. Lett. 90, 189701 (2003).
Excitation energies and optical properties of Si nanocrystallites
F. Bechstedt
L.E. Ramos, J. Furthmuller,
Friedrich-Schiller-Universitat Jena, Institut fur Festkorpertheorie und Theoretische Optik, Max-Wien-Platz 1,
D-07743 Jena Germany
Nanostructures and porous materials based on Si are an appealing research topic due to their
promising applications in optoelectronics, data storage devices, and solar cells. By exploiting
the three-dimensional quantum confinement of nanocrystallites, their luminescence properties
can be enhanced and optical gain can in principle be achieved. Recent theoretical investigations
on nanocrystallites have been focused in the electronic excitations and screening effects. Several
attempts have been made in order to model a dielectric function of Si nanocrystallites, but
so far the applicability of these methods is not clear. We present an investigation of Si
nanocrystallites by means of an ab initio pseudopotential planewave method. By applying the
-selfconsistent field method in the framework of the density-functional theory and the localdensity approximation, we calculate the electron-hole pair-excitation energies. We also discuss
different approaches to compute the quasiparticle gaps by combining a GW approximation and
an effective medium theory applied to the crystallites. An estimate of the excitonic effects is
made by evaluating the Coulomb integral with the highest occupied molecular orbital and lowest
unoccupied molecular orbital. The quasiparticle gap calculated by means of the difference
between ionization potentials and electron affinities, and the influence of Coulomb effects are
taken in consideration with respect to the supercell size. Our results intend to find reliable
procedures to calculate excitation energies in Si nanocrystallites and to model screening effects
in these structures.
36
Local-Field Effects in Calculations of Optical Constants from UV X-rays

J. Rehr, and A. L. Ankudinov
Dept. of Physics, Univ. of Washington, Seattle, WA 98195-1560, USA
There has been dramatic progress in recent years both in calculations and in the interpretation
of x-ray absorption spectra (XAS). For deep core x-ray absorption an independent-electron
approximation with final state potentials including a core-hole is often adequate [1]. However,
corrections to the independent particle approximation due to local fields are often important,
even far above an absorption edge. These local fields arise from the screening of the external
x-ray field and the coupling to the core-hole. Their treatment requires a theory that goes beyond
the independent-electron approximation. We have developed an efficient approach for treating
these effects in solids and molecules based on a combination of the Bethe-Salpeter equation
(BSE) and time-dependent density functional theory (TDDFT). First the TDDFT response
function is calculated with a local exchange-correlation functional [2]. This first step alone
is adequate over a wide spectral range above the edge. In the near-edge region where nonlocal effects can be important, the full BSE response is then calculated using a local projection
operator approach, with a screened core-hole potential W calculated within the RPA [3]. This
combined approach can be appended straightforwardly to our self-consistent real-space Greens
function code FEFF8, and makes possible efficient calculations of optical constants from the UV
to x-ray energies. The approach is illustrated with a number of examples.
[1] J. J. Rehr and R. C. Albers, Rev. Mod. Phys. 72, 621 (2000). [2] A. L. Ankudinov et
al., Phys. Rev. B. 67, 115120 (2003). [3] A. L. Ankudinov, J. J. Rehr and Y. Takimoto, [UW
Preprint 2004].
*Supported by DOE Grant DE-FG03-97ER45623 and facilitated by the DOE CMSN.
TDDFT and Many-Boby Perturbation Theory: comparisons and

combinations
Lucia Reining, Silvana Botti, Fabien Bruneval, Louise Dash, Apostolos
Marinopoulos, Valerio Olevano, Francesco Sottile, Nathalie Vast,
Wojciech Welnic
Laboratoire des Solides Irradies, UMR 7642 CNRS/CEA, Ecole

Polytechnique, F-91128 Palaiseau, France
Today, one of the big challenges of theoretical condensed matter physics is to find ways for
describing accurately and efficiently the response of electrons to an external perturbation.
In fact, the knowledge of response functions allows one to directly derive spectra (such as
absorption or electron energy loss); moreover, response functions enter the description of
correlation effects, for example through the screened Coulomb interaction W in Hedins GW
approximation to the electron self-energy. Two main developments for the ab initio calculation
of response functions of both finite and infinite systems are, on one side, the solution of the
Bethe-Salpeter equation (BSE), and, on the other hand, Time-Dependent Density Functional
Theory (TDDFT). Both approaches are promising, but suffer from different shortcomings: the
37
solution of the Bethe-Salpeter equation is numerically very demanding, whereas for TDDFT,
despite recent progress a generally reliable but at the same time very efficient description of
exchange-correlation effects has still to be developed.
We will review the two approaches focussing on their comparison. The meaning and importance
of different contributions to the induced potentials will be analyzed for various systems, ranging
from atoms to bulk semiconductors and insulators, from graphite to nanotubes. We will then
show various ways to combine TDDFT and the BSE approach, that are all leading to a similar
class of exchange-correlation kernels. These kernels yield excellent results for absorption and
energy loss spectra; moreover, they can be used to determine vertex corrections beyond the GW
approximation. We will discuss the links between these methods as well as the relation with
other work (e.g., Goerlings EXX), show results and give perspectives.
First-principles study of crystalline and amorphous transition metal

oxides and silicate
G.M. Rignanese, F. Detraux, X. Rocquefelte, J. Bouchet, X. Gonze, A.
Bongiorno, F. Giustino, and A. Pasquarello
Unite de Physico-Chimie et de Physique des Materiaux, Universite Catholique de Louvain, 1 Place Croix du Sud,
B-1348 Louvain-la-Neuve, Belgium and Research Center on Microscopic and Nanoscopic Materials and
Electronic Devices (CERMIN), Universite Catholique de Louvain, B-1348 Louvain-la-Neuve, Belgium
Until recently, silicon dioxide has been the dielectric of choice in integrated MOS devices.
By 2004, the thickness of this insulating layer is expected to decrease to 1.0 nm to 1.5 nm
(about five silicon atoms deep) precluding its effective use due to severe leakage problems. An
alternative high-permittivity (high- ) gate dielectric is required. In this framework, transition
metal oxides and silicates have attracted considerable attention as possible substitutes. Their
dielectric properties constitute an issue of great practical relevance. In particular, it is highly
desirable to develop a better understanding of how the permittivity is related to the underlying
microstructure. This is the aim of the present work.
Using first-principles calculations, we investigate the electronic and dielectric properties of
group IVB transition metal (Ti, Zr, and Hf) oxides and silicates. We consider various crystalline
phases for the oxides (cubic, tetragonal, and rutile) and for the silicates (zircon), as well as
amorphous silicates. The electronic properties are first calculated in the framework of the
density functional theory, and then corrected using the approximation. For the dielectric
properties, we introduce a scheme which relates the dielectric constants to the local bonding
of silicon and metal atoms. This scheme is based on the definition of parameters characteristic
of the basic structural units formed by silicon and metal atoms and their nearest neighbors,
and allows us to avoid heavy large-scale calculations, which are beyond current computational
capabilities. Applied to amorphous Zr silicates, our scheme provides a good description of the
measured dielectric constants, and highlights the role of sixfold coordinated metal atoms.
38
Ab-initio studies of the Si(557)/Au and Si(553)/Au Reconstructions

Sampsa Riikonen , Daniel Sanchz-Portal
1. Donostia International Physics Center (DIPC), Paseo Manuel de Lardizabal 4, 20018 Donostia-San Sebastin,
Spain
2. Centro Mixto CSIC-UPV/EHU and Donostia International Physics Center (DIPC), Donostia International
Physics Center (DIPC), Paseo Manuel de Lardizabal 4, 20018 Donostia-San Sebastin, Spain
Gold deposited on silicon surfaces at the monolayer regime results in a variety of reconstructions. The most interesting ones are the (quasi-) one dimensional phases, featuring monatomic
wires of gold. They promise exotic physical phenomena, such as the Luttinger liquid [1]. These
wires can be grown on the flat Si(111) surface which yields three possible directions for the
wires to grow [2]. Stepped silicon surfaces avoid this problem; the wires grow parallel to the
step edges. These systems have now been manufactured for years [3].
We present ab-initio studies of the Si(557)/Au and Si(553)/Au reconstructions using the Siesta
ab-initio code [4].
The Si(557)/Au surface, resulting from the deposition of 0.2 ML of gold, has been intensively
studied, due to the rich physical phenomena observed in it. Our ab initio calculations show
that spin-orbit coupling is crucial to understand the electronic structure of the Si(557)-Au
surface [5]. The spin-orbit splitting originates the two one-dimensional bands observed in
photoemission, which were previously attributed to spin-charge separation in a Luttinger liquid
[6]. This spin splitting might have relevance for future device applications. We also show that
the apparent Peierls-like transition observed in this surface by scanning tunneling microscopy
is a result of the dynamical fluctuations of the step-edge structure, which are quenched as the
temperature is decreased [7].
In the case of the Si(553)/Au reconstruction, we study a recently proposed model [3] and
give some alternative models that yield bandstructures that are in good agreement with the
photoemission data [8]. The role of the spin-orbit coupling is also studied.
Exact-Exchange-Based Quasiparticle Calculations of II-VI Compounds

and Group III Nitride
P. Rinke , A. Qteish , J. Neugebauer
, M. Scheffler
1- Fritz-Haber-Institut der Max-Planck-Gesellschaft, Berlin

2- Department of Physics, Yarmouk University, Irbid - Jordan
3- Department of Theoretical Physics, University of Paderborn
We present a systematic ab initio study of the electronic structure for a wide range of IIVI compounds and group III nitrides in the zinc-blende structure. The challenge from a
computational point of view is to capture the exchange-correlation (XC) effects arising from

the shallow semicore -electrons of the cation, which we explicitly treat as valence states in our
pseudopotential approach.
In order to correctly describe the dominant exchange interaction we apply density-functional
theory (DFT) in the exact-exchange (EXX) approximation. Although the EXX Kohn-Sham
39
bandstructure compares well with experiment for standard semiconductors [1] the bandstructure
as measured by photoemission is an excited state property. We therefore apply many-body
perturbation theory in the approximation to the EXX groundstate.
The hirarchy of our approach allows us to systematically investigate the role of XC in these
materials from first principles. Our results show that it is crucial to treat XC on the same
footing at every stage. To achieve this in our EXX calculations we employ the newly developed
EXX-pseudopotentials [2]. Furthermore we obtain bandstructures in very good agreement
with existing all-electron calculations.
Our results indicate that, in contrast to common believe, a pseudopotential approach including

only the -electrons of the semicore shell in question, is sufficient to accurately describe the
electronic structure if treated in the EXX+ formalism.
[1] W. G. Aulbur et al, Phys. Rev. B 62 7121 (2000)
[2] M. Moukara et al, J. Phys.: Condens. Matter 12 6783 (2000)
Vertex Corrections in a systematic self-energy perturbation theory

A. Schindlmayr
Institut fr Festkrperforschung, Forschungszentrum Jlich, 52425 Jlich, Germany
The perturbative implementation of the GW approximation, in which the self-energy is

evaluated non-self-consistently with Kohn-Sham orbitals, has proved to be very successful
in practical applications and yields accurate quasiparticle band structures for a wide range
of materials. While this approach originally suggested itself naturally due to its numerical
simplicity, expanding computational resources have now made it possible to perform selfconsistent calculations and to explore vertex corrections beyond the simple functional form
of the GW approximation. However, it has proved hard to achieve systematic improvements,
which depend on a careful balance between counteracting self-energy contributions. Starting
from density-functional theory, I here develop a perturbation expansion of the self-energy that
converges systematically. The first-order correction to the Kohn-Sham eigenvalues resembles
the GW approximation but includes some modifications that make the approach internally
consistent. The implications are illustrated by numerical results for the homogeneous electron
gas.
Recent applications of TD-DFT in molecular, liquid and solid phases

Hutter
Ari P. Seitsonen, Thomas Chassaing, Jurg
Physikalisch Chemisches Institut (PCI), Universitt Zrich (UniZh)
Time-dependent density functional theory (TD-DFT) can be used to efficiently calculate

excitation energies, with excitonic effects included. In molecular systems the energies of
valence excitations are usually quite accurate, however Rydberg and especially charge-transfer
excitations are badly described with the adiabatic LDA or GGA functionals. Here we compare
the GGA-PBE functional [John P Perdew et al, Phys. Rev. Lett. 77 (1996) 3865], the TPSS
40
meta-functional [Jianmin Tao et al, Phys. Rev. Lett. 91 (2003) 146401], the asymptotically
corrected SAOP potential [P R T Schipper et al, J. Chem. Phys. 112 (2000) 1344] and the
PBE0 hybrid functional [C. Adamo and V. Barone, J. Chem. Phys. 110 (1999) 6158]. Except
for the charge-transfer excitations the excitation energies are generally improved by the SAOP
potential and PBE0 functional.
Bethe Salpeter treatment of core excitations: multiplet effects

E.L. Shirley
Optical Technology Division NIST 100 Bureau Dr., MS 8441 Gaithersburg, MD 20899-8441 USA
We present a Bethe-Salpeter treatment of core excitations in solids that can be used to include
multiplet effects. This accounts for electron spin degrees of freedom, as well as core-hole
quantum numbers, L, M and S. The advantage of this approach over other approaches (atomic,
Anderson impurity model) is that it allows for the natural inclusion of many chemical and solid
state effects. The approach should be usable in molecular and cluster calculations as well, so
that it can help study multiplet effects in core-hole spectra for nanosystems. Issues regarding
charge-transfer excitations and multiple electron excitations will also be discussed.
Excitons from TDDFT as generated by the variational approach to

many-body perturbation theory
G. Stefanucci, Ulf Von Barth
Lund Sweden
It is rather well known that the optical or density response function of electronic systems can
be obtained from time-dependent density-functional theory (TDDFT) in terms of an exchangecorrelation kernel (fxc). This approach is computationally far simpler as compared to solving
the Bethe-Salpeter equation of many-body perturbation theory (MBPT). In actual fact, within
TDDFT, no systematic route toward successively better approximations has, so far, been
available. In the present work we construct such a scheme based on our variational action
functionals. In terms of the one-electron Green function of MBPT these functionals give
stationary expressions for the total action of the system at hand - or the total energy in the case
of time independent problems. From a stationary action it is rather straight-forward to construct
the time dependent density response function. Building the functionals from the PHI derivable
theory of Baym and Kadanoff, always results in response functions which obey a large number
of physical constraints like particle or energy conservation. The simple idea of the present work
is to restrict the variational freedom of the functionals to the domain of Green functions which
are non-interacting and given by a local one-electron potential. According to the Runge-Gross
theory this restriction immediately results in a density-functional theory the quality of which is
determined by the sophistication which is build in to the choice of PHI derivable approximation
for the action functional.
41
A bi-exciton mechanisms of photo-induced desorption from MgO

P. E. Trevisanutto, P. V. Sushko, A. L. Shluger
Department of Physics & Astronomy, University College London, Gower Street, London, WC1E 6BT, UK
Recent experimental data clearly demonstrate that using both femto- and nano-second laser
pulses with photon energies tuned to excite particular surface features at the MgO surface, one
can achieve a very effective desorption of atomic O and Mg species. In this presentation we
demonstrate a new bi-exciton mechanism responsible for desorption of oxygen atoms and draw
analogies with the exciton mechanism of desorption of alkali halides.
Recently we demonstrated that photo-induced desorption can be used for probing surface
spectroscopic features [1-3]. The technique is sensitive enough to detect energy shifts between
the bulk and surface excitons as small as 0.2 eV. In addition, a significant degree of control
over the desorbed species can be achieved if the knowledge of the surface electronic structure
is fully exploited [1-3]. We extended this approach to more complex oxide surfaces. The
experimental and theoretical studies demonstrate that excitation energies of corner and step
sites at the MgO (001) surface are strongly shifted with respect to the bulk and terrace exciton
energies [4]. Recent experiments demonstrated that selective excitation of the kink and corner
surface features using 4.7 eV laser pulses leads to desorption of O atoms with average kinetic
energies of about 0.3 eV. Since the valence of anions in MgO is twice that of anions in alkalihalides, the desorption mechanism is expected to be more complex and involve several steps.
To model the photon-induced processes in MgO we employ an embedded cluster approach,
which combines a quantum-mechanical treatment of a site of interest with a self-consistent
classical description of the remaining part of the system. The method is implemented in
a computer code GUESS; it allows one to account consistently for the ionic and electronic
contributions to the defect-induced lattice polarisation and for their effect on the defect itself,
and to treat excited states using different ab initio techniques. We modelled desorption from
a 3-coordinated oxygen corner site at an MgO surface and developed a two-step mechanism
for this process. The results of our calculations suggest that the first photon of ca. 4.5 eV
can induce an optical transition at the oxygen corner site. This first excited state relaxes with
a significant displacement of the corner oxygen away from its original site. The relaxed state
corresponds to a charge transfer exciton configuration O V E Mg+. This system can be further
excited with another photon at about 4.8 eV creating a bi-exciton, where the oxygen species
becomes effectively neutral. The adiabatic potential energy surface for the second excited state
is such that oxygen atom prefers to leave the surface with the maximum kinetic energy of several
tenths on an eV. The surface site relaxes with the formation of a neutral F centre at the corner
site previously occupied by the desorbed oxygen atom. Our results are in good agreement with
the experimental photon energies and kinetic energies of desorbed species.
[1] V. E. Puchin, A. L. Shluger, N. Itoh, Phys. Rev. B, 47, 10760 (1993)
[2]K. M. Beck, A. G. Joly, N. F. Dupuis, P. Perozzo, W. P. Hess, P. V. Sushko, A. L. Shluger, J. Chem.
Phys., 120, 2456 (2004)
[3]W. P. Hess, A. G. Joly, K. M. Beck, P. V. Sushko, A. L. Shluger., Surf. Sci. (in press)
[4] A. L. Shluger, P. V. Sushko, L. N. Kantorovich, Phys. Rev. B. 59, 2417 (1999)
42
A study of excitonic properties of low dimensional systems with TDDFT

D. Varsano, A. Marini and A. Rubio
Departamento de Fsica de Materiales, Facultad de Qumicas, Universidad del Pas Vasco and Donostia
International Physics Center (DIPC), 20080 San Sebastian (Spain)
We investigate the optical, electron energy-loss spectra and static polarizability, of polyacetylene, a conjugated polymer, via Time Dependent Density Functional Theory (TDDFT). The
results obtained using our recently developed frequency dependent and non-local exchange
correlation kernel
, are compared with the stateoftheart Many Body Perturbation Theory (MBPT) calculations. The TDDFT optical and energy-loss spectra of the isolated
infinite chain turns out to be practically indistinguishable from the MBPT results and from the
experiment. As the proposed xc-kernel has been tested only on bulk solids, the present results confirm its applicability and stabulity for the case of quasi one dimensional systems. We
present also results for the static polarizability of the finite chain as a function of the numbers of the monomers. Compared to the Adiabatic Local Density Approximation (ALDA) that
systematically overestimate for such systems, the present approximation for yields very
good results when compared with quantum chemical calculations.
Phase stability in titanium and its dioxides with ab initio methods

Virginie Quequet, Nathalie Vast and Marc Hayoun
Laboratoire des Solides Irradies, CEA-DSM, CNRS UMR 7642, Ecole Polytechnique, F-91128 Palaiseau, France
We have studied the crystal structures of titanium metal and dioxides by means of ab initio
total-energy calculations. The phases of interest for titanium metal are the phase (hcp), the

high-temperature phase (bcc), and the high-pressure phase (hex), and for the rutile
phase (stable) and the anatase phase (metastable). For rutile the electron energy loss spectra
have already been studied in our laboratory [1].
We have used density functional theory (DFT) with a plane-wave basis set and the pseudopotential method [2]. We have calculated the relaxed lattice parameter for each structure and
the energy differences between the phases. Different pseudopotentials were used, where 3
semicore states were explicitly treated, or implicitely via a non linear core correction. The relative stability of the different phases is found to change a lot depending on the treatment of the
semicore states.
In order to provide a better treatment of the semicore states, we have taken into account an
approximate self-interaction correction (SIC) for the semicore states in our pseudopotentials.
The results for titanium metal will be presented.
References:
[1] N. Vast, L. Reining, V. Olevano, P. Schattschneider et B. Jouffrey Phys. Rev. Lett. 88,
037601 (2002).
[2] Using PWSCF (www.pwscf.org) and FHIPP (www.fhi-berlin.mpg.de/th/ fhi98md/fhi98PP)
softwares.
43
Tuning the optical and/or magnetic properties of correlated nanoclusters

C. Verdozzi and C.O.Almbladh
Solid State Theory, Department of Physics, Lund University, Lund Sweden
Due to its technological relevance, the microscopic tailoring of nanoclusters electronic properties is currently one of the most active areas of condensed matter research.
We report here of recent and in progress activity in the characterization of the optical and magnetic response of nanoclusters. The theoretical approach is based on exact diagonalization (ED),
supplemented by many-body perturbation theory, depending on the systems and phenomena
considered.
In one part of the presentation we will show results for magnetic clusters* with dense Kondotype impurities. For such systems, controlling upon varying the energy spacing is acquiring
increasing importance since it may allow to tune the cluster magnetic behavior and to encode
quantum information. Our calculations reveal that tunes the interplay between the local Kondo
and non-local RKKY interactions, giving rise to a Doniach phase diagram for a nanocluster.
We will also show that for specific couplings and parameters, electron- phonon coupling may
enhance the local moment and affect the RKKY vs. Kondo interplay, thus changing the phase
boundaries domains of the Doniach phase diagram.
In the other part of the presentation, we will illustrate our progress about a characterization of
the optical and Auger response of supported semiconductor quantum dots. We are addressing
single-dot and clustered dots geometry, without specific assumptions about dot shapes, in the
effective mass-approximation. This is conveniently done using a Gaussian basis approach and
with the inclusion of e-e, h-h, e-h correlation (Coulomb) effects. We will characterize the
methodology with some numerical examples and overview the experimental phenomena and
subsequent technical developments we plan to address.
* Work done by CV with Y.Luo and N.Kioussis
Ab initio Calculations of amorphous phase change materials

Wojciech
Wenic , Silvana Botti , Matthias Wuttig and Lucia Reining

I. Physikalisches Institut, RWTH Aachen, 52056 Aachen, Germany Laboratoire des Solides Irradies, Ecole
Polytechnique, 91128 Palaiseau cedex - France
Chalcogenide alloys, which show a significant change of optical reflectivity and electric
conductivity under phase transformation from the amorphous to the crystalline structure are
widely used for optical data storage today (e.g. DVD-RW). As they can also be used as non
volatile memories (Phase-Change-RAM), they are furthermore promising candidates for future
electronic data storage techniques. In this work ground state and excited state calculations are
presented for GeTe, the basic phase-change-material. It is reported to crystallize in a cubic
rocksalt phase and in a trigonal phase, which can be described as a distorted cubic structure.
From literature relatively little is known about the amorphous structure, but the available data
hints towards a tetragonal, diamond-like local coordination. The focus of the project is on the
44
relation between optical properties and geometrical structure in the amorphous phase compared
to the crystalline phases. The origin for the difference of the optical and electronic properties
in the crystalline and the amorphous state -which is crucial for the application- is examined.
The problem of obtaining a realistic amorphous structure is approached in two different ways:
on one side we employ ab initio molecular dynamics within a 64-atom supercell and on the
other side we use a simple model structure which reproduces the local configuration reported
in earlier experimental work. The electronic structure is presented in the GW-correction and
the theoretical spectra are calculated within TDDFT and GW-RPA. They are compared with
experimental data of thin film GeTe-samples. Differences between theory and experiment are
discussed as well as the changes in the optical and electronic properties upon phase transition
from the crystalline to the amorphous state.
Excitons in BN-nanotubes: dimensionality effects

L. Wirtz, A. Marini, A. Rubio
Department of Material Physics, University of the Basque Country and Donostia International Physics Center
(DIPC), Po. Manuel de Lardizabal 4, 20018 San Sebastian, Spain
Optical absorption spectroscopy is an important tool for the characterization of nanotubes.

The first order explanation in terms of single-particle excitations from occupied to unoccupied
states is straight-forward. For quantitative predictions, however, a simple calculation based on
the assumption of non-interacting particles and on the electronic band-structure on the DFT
level is not accurate enough. In particular, electron-hole interaction has a large impact on
the absorption spectra of BN structures and leads to excitonic peaks with a binding energy
of up to 2 eV. We have calculated these excitonic effects through the solution of the BetheSalpeter equation. We discuss the resulting spectra of bound excitonic peaks in the quasi-one
dimensional tubes and compare with the spectra of the quasi-two dimensional single sheet of
hexagonal BN and of bulk hexagonal BN. The sheet and tube exhibit strong excitonic effects
that overcome the quasiparticle correction to the DFT-bandstructure. This give rises to bound
states in the quasiparticle gap that are not present in the case of the layered structure.
This work was supported by the European Community Research Training Network
NANOPHASE (HPRN-CT-2000-00167) and Network of Excellence NANOQUANTA (NOE
500198-2). Computer time was granted by DIPC and CEPBA (Barcelona).
45
List of participants
Carl-Olof ALMBLADH, University of Lund, SWEDEN;
email: [email protected]
Stefano BARONI, INFM DEMOCRISTOS, Trieste ITALY;
Friedhelm BECHSTEDT, Friedrich Schiller Universitat, Jena, GERMANY;
Lorin BENEDICT, Lawrence Livermore National Laboratory, Livermore, CA, USA;
Andrey BOGDANOV, Russian State Pedagogical Univ., Saint-Petersburg, RUSSIAN FEDERATION; email: [email protected]
Peter BOKES, Slovak University of Technology, Bratislava, SLOWAKIA;
Silvana BOTTI, LSI Ecole Polytechnique, Palaiseau, FRANCE;
Johann BOUCHET, Universite Catholique de Louvain, Louvain-la-neuve, BELGIUM;
Gian Paolo BRIVIO, University of Milano Bicocca, Milan, ITALY;
Fabien BRUNEVAL, LSI Ecole Polytechnique, Palaiseau, FRANCE;
Mauro BRUNO, University of Rome Tor Vergata, Rome, ITALY;
Gabriele BUTTI, University of Milano Bicocca, Milan, ITALY;
Sebastiano CARAVATI, University of Milano Bicocca, Milan, ITALY;
Nuno CARNEIRO, University of York, UNITED KINGDOM;
Michele CINI, University of Rome Tor Vergata, Rome, ITALY;
Eric CHANG, INFM-S3, University of Modena and Reggio-Emilia, Modena, ITALY;
Letizia CHIODO, University of Rome Tor Vergata, Rome ITALY;
46
Eugueni CHULKOV, University of the Basque Country, San Sebastian, SPAIN;

Louise DASH, LSI Ecole Polytechnique, Palaiseau, FRANCE;
Elena DEGOLI, University of Modena-Reggio Emilia, Modena, ITALY;
Kris DELANEY, University of Illinois, Urbana, USA;
Christophe DELERUE, IEMN Dept. ISEN, Lille, France;
Rodolfo DEL SOLE, University of Rome Tor Vergata, Rome, ITALY;
Philipp EGGERT, Fritz Haber Institut, Berlin, GERMANY;
Alberto FRANCESCHETTI, Oak Ridge National Laboratory, Oak Ridge, USA;
Christoph FREYSOLDT, Fritz-Haber-Institut, Berlin, GERMANY;
Martin FRIAK, Fritz-Haber-Institut, Berlin, GERMANY;
Christoph FRIEDRICH, Institut fr Festkrperforschung, Julich, GERMANY;
Pablo GARCIA-GONZALEZ, UNED, Madrid SPAIN;
Maja GARCIA-VERGNIORY, University of the Basque Contry, San Sebastian, SPAIN;
Matteo GATTI, University of Milan, iMilan, ITALY;
Christine GIORGETTI, CNRS-LSI-LURE, Palaiseau, FRANCE;
Rex GODBY, University of York, York, UNITED KINGDOM;
Xavier GONZE, Universite Catholique de Louvain, Louvain-la-neuve BELGIUM;
Eberhard K.U. GROSS, Freie Universitat, Berlin, GERMANY;
47
Conor HOGAN, University of Rome Tor Vergata, Rome, ITALY;

Andrei INCZE, INFM-University of Milan, Milan, ITALY;
Sohrab ISMAIL-BEIGI, Dept. of Applied Physics, Yale University, USA;
Krister KARLSSON, Natural Science, Skovde, SWEDEN;
Takao KOTANI, Osaka University, Osaka, JAPAN;
Stefan KURTH, Freie Universitaet, Berlin, GERMANY;
Aleksey KUZNETSOV, Fritz-Haber-Institute, Berlin, GERMANY;
Roman LEITSMANN, Friedrich Schiller Universitat, Jena, GERMANY;
Sergey LISENKOV, Inst. of Bioc. Phys. of RAS, Moscow, RUSSIA;
Andrea MARINI, University of Rome Tor Vergata, Rome, Italy;
Miguel MARQUES, DIPC, San Sebastian, SPAIN;
Margherita MARSILI, University of Rome Tor Vergata, Rome, ITALY;
Hector MERA, University of York, York, UNITED KINGDOM;
Manfred NIESERT, Institut fur Festkorperforschung, Julich, GERMANY;
Ola NYQUIST, Natural Science, Skovde, SWEDEN;
Valerio OLEVANO, Ecole Polytechnique, Palaiseau, FRANCE;
Giovanni ONIDA, University of Milan, Milan, ITALY;
Stefano OSSICINI, INFM- S3, University of Modena and Reggio Emilia, ITALY;
48
Juan Jose PALACIOS, Universidad de Alicante, Alicante, SPAIN;

Maurizia PALUMMO, University of Rome Tor Vergata, ITALY;
email : [email protected]
Enrico PERFETTO, University of Rome Tor Vergata, ITALY;
Jose PITARKE, University of the Basque Country, Bilbao, SPAIN;
Stefano PITTALIS, Freie Universitaet, Berlin, GERMANY;
Victor A. POPA, University of Twente, Enschede, THE NETHERLANDS;
Yann POUILLON, Universite Catholique de Louvain, Louvain-la-neuve, BELGIUM;
Olivia PULCI, University of Rome Tor Vergata, Rome, ITALY;
Louis RAMOS, FSU Jena IFTO, Jena, GERMANY;
John REHR, University of Washington, Seattle, USA;
Lucia REINING, Ecole Polytechnique, Palaiseau, FRANCE;
Gian Marco RIGNAGNESE, Universite Catholique de Louvain, BELGIUM;
Sampsa RIIKONEN, DIPC, San Sebastian, SPAIN;
Patrick RINKE, Fritz Haber Institut, Berlin, GERMANY;
Angel RUBIO, Donostia International Physics Center, SPAIN;
Arno SCHINDLMAYR, Institut fur Festkorperforschung, Julich GERMAY;
Ari SEITSONEN (University of Zurich, Zurich, SWITZERLAND);
Eric SHIRLEY (NIST, Gaithersburg, USA);
49
Francesco SOTTILE, Ecole Polytechnique, Palaiseau, FRANCE;

Martin STANKOVSKI, University of York, York, UNITED KINGDOM;
Gian Luca STEFANUCCI, University of Lund, Lund, SWEDEN;
Paolo Emilio TREVISANUTTO, University College of London, London, UNITED KINGDOM;
Daniele VARSANO, University of Basque Country, San Sebastian, SPAIN;
Nathalie VAST, Ecole Polytechnique, Palaiseau, FRANCE;
Claudio VERDOZZI, University of Lund, Lund, SWEDEN;
Ulf von BARTH, University of Lund, SWEDEN;
Wojciech WELNIC, Ecole Polytechnique, Palaiseau, FRANCE;
Ludger WIRTZ, DIPC, San Sebastian, SPAIN;
Hongqi Xu, University of Lund, Lund, SWEDEN;
50

Theory and Modeling of Electronic Excitations in Nanoscienceook

Uploaded by

Copyright:

Available Formats

Theory and Modeling of Electronic Excitations in Nanoscienceook

Uploaded by

Document Information

Original Description:

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

Theory and Modeling of Electronic Excitations in Nanoscienceook

Uploaded by

Copyright:

Available Formats

NANOEXC2004 CONFERENCE

Programme, Abstracts, and Participants List of the conference

Theory and Modeling of Electronic

Arno Schindlmayr (Julich,

Abstracts of talks and posters

Presentation and general announcements

Tuesday, September 21 2004

POSTER SESSION (continued)

Wednesday, September 22 2004

Closing of nanoexc2004 and presentation of the 2005 workshop

Thursday, September 23 2004: DEPARTURES

10 Christoph Freysoldt GW calculations for surfaces

24 Arno Schindlmayr Vertex Corrections in a systematic self-energy perturbation theory

Abstracts of talks and posters

Time-dependent density-functional perturbation theory

Optical spectra of three-, two-, and zero-dimensional systems treated

Optical absorption in hydrogenated Si clusters: BSE vs. TDLDA

We have performed calculations of the optical absorption spectra of nm-sized hydrogenated Si

Coherent exchange between the forward and backward waves in Bistable

Conductance of metal-vacuum-metal interface from the polarisation

Peter Bokes and Rex Godby 

Optical properties of complex systems from TDDFT

Laboratoire des Solides Irradies, UMR 7642 CNRS/CEA, Ecole

Ab initio treatment of adatom Auger decays

1-Dipartimento di Scienza dei Materiali - Universit   di Milano Bicocca, Italy

We calculate the core-core-valence (CCV) Auger profiles of an adatom on a periodic surface

Ab-initio Electronic and Optical Properties of Germanium Quantum

Image potential states and surface electronic structure calculation on the

Selectivity of Auger decays to the local surface environment.

We then calculated the Auger lineshape for the Na

A novel generalized Kohn-Sham approach for total energy calculations

Generalised Kohn-Sham (GKS) theory [1],[2], provides a variety of possible generalisations of

Quasiparticle energies and optical properties in carbon nanotubes: a

The energy loss of (110)Ag and (001)Al surfaces

Electron-phonon interaction and its role in the lifetime broadening

Departamento de Fsica de Materiales and CentroMixto CSIC-UPV/EHU, Faculatad de Ciencias

Ab initio spectra of the low-pressure phases of zirconia

We have developed ultrasoft pseudopotentials intended to be used in calculations for ab

Choosing ultrasoft pseudopotentials for spectroscopy calculations

We have developed ultrasoft pseudopotentials intended to be used in calculations for ab

Impurities states in hydrogenated silicon nanoclusters

In this work we show an ab-initio study of impurity states in silicon nanocrystals. It is

Self-consistent Solutions of

Kris Delaney , Patrick Rinke  , P. Garc

(1) Department of Physics, University of Illinois at Urbana-Champaign, Urbana, IL 61801, U.S.A.

We present details of the use of many-body perturbation theory, within the 

Quasi-particle gap, dielectric properties and electronic correlations in

Electronic excitation spectra of semiconductor nanostructures are particularly interesting

Calculation of surface optical properties within TDDFT: the bound

Long-ranged interactions in a supercell approach

Fritz-Haber-Institut der Max-Planck-Gesellschaft, Faradayweg 46, 14195 Berlin-Dahlem, Germany,  Institut

Electronic Excitations in Semiconductor Quantum Dots

GW calculations for surfaces

Christoph Freysoldt , Philipp Eggert ,

Ab initio study of the electronic structure of magnetite

Martin Friak , Arno Schindlmayr  , and Matthias Scheffler

Fritz-Haber-Institut der Max-Planck-Gesellschaft, Faradayweg 46, 14195 Berlin-Dahlem, Germany,  Institut

Photocatalytic Oxidation of CO at the V O (010) Surface: Cluster Model

Self-energy and lifetimes of surface states beyond the GW approximation

Peter Bokes and Rex Godby

1-Dipartimento di Scienza dei Materiali - Universit di Milano Bicocca, Italy

Self-consistent Solutions of

Kris Delaney , Patrick Rinke , P. Garc

We present details of the use of many-body perturbation theory, within the

Fritz-Haber-Institut der Max-Planck-Gesellschaft, Faradayweg 46, 14195 Berlin-Dahlem, Germany, Institut

Martin Friak , Arno Schindlmayr , and Matthias Scheffler

Fritz-Haber-Institut der Max-Planck-Gesellschaft, Faradayweg 46, 14195 Berlin-Dahlem, Germany, Institut

Kris Delaney , Patrick Rinke , Pablo Garca-Gonzalez , Angel

Using density-functional calculations, we study a new polymeric form based on fullerene

Local-Field Effects in Calculations of Optical Constants from UV X-rays