Arsenic Water Technology Partnership

Aerogel & Iron-Oxide Impregnated

Granular Activated Carbon Media For
Arsenic Removal
Aerogel & Iron-Oxide Impregnated
Granular Activated Carbon Media For
Arsenic Removal
Prepared by:

Paul Westerhoff
Department of Civil and Environmental Engineering, Arizona State University
Tempe, AZ 85287-5306

Tanju Karanfil
Department of Environmental Engineering and Science
Clemson University, Anderson, SC 29625

And

John Crittenden
Department of Civil and Environmental Engineering, Arizona State University
Tempe, AZ 85287-5306

Jointly Sponsored by:

Awwa Research Foundation
6666 West Quincy Avenue, Denver CO 80235-3098

and

U.S. Department of Energy

Washington, D.C. 20585-1290

Published by:

WERC, a Consortium for Awwa Research Foundation

Environmental Education and
Technology Development at
New Mexico State University
(add logo)

ii
About the Awwa Research Foundation

The Awwa Research Foundation (AwwaRF) is a member-supported, international, nonprofit

organization that sponsors research to enable water utilities, public health agencies, and other
professionals to provide safe and affordable drinking water to consumers.

The Foundation’s mission is to advance the science of water to improve the quality of life. To
achieve this mission, the Foundation sponsors studies on all aspects of drinking water, including
supply and resources, treatment, monitoring and analysis, distribution, management, and health
effects. Funding for research is provided primarily by subscription payments from approximately
1,000 utilities, consulting firms, and manufacturers in North America and abroad. Additional
funding comes from collaborative partnerships with other national and international
organizations, allowing for resources to be leveraged, expertise to be shared, and broad-based
knowledge to be developed and disseminated. Government funding serves as a third source of
research dollars.

From its headquarters in Denver, Colorado, the Foundation’s staff directs and supports the
efforts of more than 800 volunteers who serve on the board of trustees and various committees.
These volunteers represent many facets of the water industry, and contribute their expertise to
select and monitor research studies that benefit the entire drinking water community.

The results of research are disseminated through a number of channels, including reports, the
Web site, conferences, and periodicals.

For subscribers, the Foundation serves as a cooperative program in which water suppliers unite
to pool their resources. By applying Foundation research findings, these water suppliers can save
substantial costs and stay on the leading edge of drinking water science and technology. Since its
inception, AwwaRF has supplied the water community with more than $300 million in applied
research.

More information about the Foundation and how to become a subscriber is available on the Web
at www.awwarf.org.

iii
 DISCLAIMER

This study was jointly funded by the Awwa Research Foundation (AwwaRF) and the U.S.
Department of Energy (DOE) under Grant No. DE-FG02-03ER63619 through the Arsenic Water
Technology Partnership. The comments and views detailed herein may not necessarily reflect
the views of the Awwa Research Foundation, its officers, directors, affiliates or agents, or the
views of the U.S. Federal Government and the Arsenic Water Technology Partnership. The
mention of trade names for commercial products does not represent or imply the approval or
endorsement of AwwaRF or DOE. This report is presented solely for informational purposes.

Copyright 2006
By Awwa Research Foundation
and Arsenic Water Technology Partnership

All Rights Reserved

Printed in the U.S.A.

iv
 FOREWORD

The Awwa Research Foundation is a nonprofit corporation that is dedicated to the

implementation of a research effort to help utilities respond to regulatory requirements and
traditional high-priority concerns of the drinking water community.

The Arsenic Water Technology Partnership (AWTP) program is a partnership between

AwwaRF, Sandia National Laboratories (SNL) and WERC, a Consortium for Environmental
Education and Technology Development at New Mexico State University that is funded by DOE
and the Awa Research Foundation. The goal of the program is to provide drinking water
utilities, particularly those serving small and rural communities, with cost-effective solutions for
complying with the new 10 ppb arsenic MCL. This goal is being met by accomplishing three
tasks: 1) bench-scale research to minimize operating, energy and waste disposal costs; 2)
demonstration of technologies in a range of water chemistries, geographic locales, and system
sizes; and 3) cost effectiveness evaluations of these technologies and education, training, and
technology transfer.

The AWTP program is designed to bring new and innovative technologies developed at the
laboratory and bench-scale to full-scale implementation and to provide performance and
economic information under actual operating conditions. Technology transfer of research and
demonstration results will provide stakeholders with the information necessary to make sound
decisions on cost-effective arsenic treatment.

AwwaRF participates in the overall management of the program, helps to facilitate the program’s
oversight committees, and administer the laboratory/bench-scale studies. SNL conducts the
pilot-scale demonstrations and WERC oversees the education, training, economic analysis, and
outreach activities associated with this program.

Walter J. Bishop Robert Renner

Chair, Board of Trustees Executive Director
Awwa Research Foundation Awwa Research Foundation

v
 ACKNOWLEDGEMENTS

This report is the product of a collaborative effort between the members of the Arsenic Water
Technology Partnership and was made possible by funds from congress and the drinking water
community. A special thanks to U.S. Senator Pete Domenici for his support and assistance in
helping to bring low-cost, energy efficient solutions for the removal of arsenic from drinking
water.

The authors wish to acknowledge the involvement of Sabre Coleman from Lawrence Livermore
National Laboratory for assistance on the initial proposal and providing samples of aerogel-GAC
materials. The City of Scottsdale provided access to its groundwater wellsites and historic data
which was valuable throughout this project. Paul Sylvester and Owen Boyd (Solmetex Inc.)
generously provided their technical assistance and time to develop a marketable Fe-GAC
product.

Much of the laboratory work was conducted by Jorge Christian Navarro Aragon at Arizona State
University, who received his Masters of Science in Engineering degree in December 2005.
Further assistance from Andrew Baumgardner and Mohammad Badruzzaman at ASU is
appreciated, as is research by a visiting student (Renaud Bordier) from the University of Poitiers
(France). Additional gratitude is extended to the graduate students (Wei Cheng, Abhishek
Yadav, and Zaheer Sheikh) and post-doctoral researchers (Dr. Seyed Dastgheib and Dr. Yanping
Guo) at Clemson University.

Technical advice and guidance of the AwwaRF project officer, Albert Ilges, is appreciated.

vi
 CONTENTS
LIST OF TABLES......................................................................................................................... ix
LIST OF FIGURES ....................................................................................................................... xi
EXECUTIVE SUMMARY ......................................................................................................... xiii
Research Objectives......................................................................................................... xiii
Approach.......................................................................................................................... xiii
Conclusions...................................................................................................................... xiii
Recommendations for future Research............................................................................. xv
CHAPTER 1 INTRODUCTION AND OBJECTIVES...................................................................1
Statement of Problem.......................................................................................................... 1
Project Objectives ............................................................................................................... 3
CHAPTER 2 EXPERIMENTAL AND ANALYTICAL METHODS............................................5
Iron Modified Media........................................................................................................... 5
Spectroscopic Characterization........................................................................................... 7
XPS Analysis ................................................................................................................ 7
XRD Analysis ............................................................................................................... 8
Stock Solution Preparation for Experiments....................................................................... 8
Batch Arsenic Adsorption Experiments.............................................................................. 8
Batch Kinetic Test Procedure ....................................................................................... 8
Batch Isotherm Test Procedure..................................................................................... 9
Batch Isotherm Test using Modified NSFI-53 challenge water ................................... 9
Sample Filtration Procedure ....................................................................................... 10
Rapid Small Scale Column Tests...................................................................................... 10
Small Scale Column Packing Procedure..................................................................... 10
Sampling Procedure and Flow Rate Setting for Rapid Small Scale Columns............ 11
Analytical Methods........................................................................................................... 12
Sample Preservation.................................................................................................... 12
Glassware Cleaning Procedure ................................................................................... 12
Arsenic Analysis by Graphite Furnace Atomic Adsorption (GFAA)......................... 12
Gas Chromatography (GC) Analysis .......................................................................... 13
Other Measurements ................................................................................................... 14
CHAPTER 3 BATCH EXPERIMENT RESULTS .......................................................................23
Kinetic Tests ..................................................................................................................... 23
Screening Tests ................................................................................................................. 23
Clemson University Media ......................................................................................... 23
Arsenite ....................................................................................................................... 23
Solmetex Fe-GAC Media ........................................................................................... 24
LLNL Fe-GAC Media ................................................................................................ 24
Isotherms in Model Solutions and Natural Waters .......................................................... 25
Model Solutions .......................................................................................................... 25
Groundwaters.............................................................................................................. 26
pH Effects ................................................................................................................... 27
Silicate (Si) and Vanadate (V) Effects........................................................................ 27
Co-Adsorption of Organic Chemicals and Arsenic .......................................................... 27
Summary ........................................................................................................................... 28

vii
CHAPTER 4 ARSENIC REMOVAL IN CONTINUOUS FLOW COLUMN TESTS................45
Fe-GAC Media Performance with Clemson University Fe-GAC Composites ................ 45
Effect of EBCT ................................................................................................................. 45
Clemson University Fe-GAC Composites.................................................................. 46
Solmetex Fe-GAC Composites................................................................................... 46
Comparison of Fe-GAC Composites ................................................................................ 46
Concurrent Arsenic and Organic Removal ....................................................................... 47
Summary ........................................................................................................................... 47
CHAPTER 5 DISCUSSION OF RESULTS .................................................................................55
Batch and RSSCT Adsorption Capacities......................................................................... 55
Model Solutions .......................................................................................................... 55
Groundwaters.............................................................................................................. 56
Effects of pH and Ion Competition................................................................................... 59
Material Properties............................................................................................................ 59
CHAPTER 6 RECOMMENDATIONS FOR UTILITIES AND FUTURE RESEARCH............69
Conclusions....................................................................................................................... 69
Recommendations for future Research............................................................................. 71
REFERENCES ............................................................................................................................. 73
APPENDIX A –Isotherms for Comparison of Fe-GAC Completed After Composite Selection for
RSSCTs......................................................................................................................................... 77
ABBREVIATIONS ...................................................................................................................... 81

viii
 List of Tables

2.2 Fe-GAC media properties produced by Solmetex and ASU .............................................16

2.3 Characteristics of aerogel-GAC composites......................................................................16
2.4 Elemental surface composition of samples........................................................................17
2.5 NSFI-53 challenge water formulation ...............................................................................17
2.6 RSSCT scaling equations...................................................................................................18
2.7 RSSCT design parameters for simulating a 4 minute full-scale pilot system....................19
2.8 MDL for GFAA analysis ...................................................................................................19
4.1 RSSCT design parameters for EBCT of 2.5, 4 and 10 minutes.........................................49
5.1 Isotherm parameters for different Fe-GAC composites in model waters, blended
groundwater and surface water. ...................................................................................61
5.2 Calculated adsorption capacities for Fe-GAC composites evaluated in RSSCTs .............62
5.3 Summary of bed volumes treated during RSSCT..............................................................63

ix
x
 List of Figures
2.1 Schematic of a typical set up for a column study ..............................................................20
2.2 Variability of slopes of standard calibration curves for arsenic analysis...........................21
2.3 Summary of QC results for GFAA analysis ......................................................................21
2.4 Standards calibration curve for MTBE ..............................................................................22
3.1 Kinetic tests using different Fe-GAC composites. ............................................................30
3.3 Batch Screening of Clemson University Fe-GAC samples ...............................................31
3.4 Batch screening of Fe-GAC samples from Clemson University .......................................32
3.5 Screening test for As(III) removal using Fe-GAC samples from Clemson University.....32
3.6 Batch adsorption test with virgin HD3000, LLNL aerogel 294 and two Solmetex Fe-
GAC composites in Nanopure water system.. .............................................................33
3.7 Adsorption test with different Fe-GAC materials from Clemson University....................34
3.8 Log form of Freundlich adsorption isotherm with different Fe-GAC materials from
Clemson University .....................................................................................................34
3.9 Isotherm test in model solution comparing different Fe-GAC media from Solmetex
and Clemson University...............................................................................................35
3.10 Adsorption test in model solution for arsenic removal using Fe-GAC media from
Solmetex and Clemson University...............................................................................36
3.11 Isotherm test comparing best performance Fe-GAC composites for arsenate removal
in nanopure water from Clemson University and Solmetex........................................36
3.12 Log version of isotherm in a model solution using ASU produced composites and S-
4-11 from Solmetex.. ...................................................................................................37
3.13 Adsorption test in Scottsdale groundwater with C-II4B-3 from Clemson University.......37
3.14 Effect of natural water matrix composition on competition and removal of As(V)
using 33% or 67% Scottsdale Site 4E groundwater blended with DI water................38
3.15 Effect of pH on arsenate adsorption by Fe-GAC media from Clemson University ..........38
3.16 Comparison of arsenic removal as a function of pH with C-II4B-3 in nanopure water....39
3.17 Effect of pH and silica on As(V) removal using a base NSFI challenge water matrix .....40
3.18 Effect of pH and vanadium on As(V) removal using a base NSFI challenge water
matrix ...........................................................................................................................41
3.19 Isotherm test in Rooselvelt Lake water..............................................................................41
3.20 Comparison of media performance for arsenate removal in different water systems.
All three cases are from different batch experiments at pH of 7 .................................42
3.21 Comparison of C-II4B-3 and virgin GAC for DOC removal in surface water .................42
3.22 Adsorption of SOCs (MTBE and benzene) from groundwater (blended at 33% with
nanopure) containing arsenate by several Fe-GAC materials......................................43
3.23 As(V) adsorption in groundwater (33% natural, 67% nanopure) spiked with SOCs by
Fe-GAC materials from Clemson University and ASU (Solmetex process)...............44
3.24 Summary of Fe-GAC performances removing arsenate in model solutions based on
iron content. . ...............................................................................................................44
4.1 RSSCTs with Fe-GACs from Clemson University in Scottsdale groundwater.................50
4.2 Effect of EBCTLC on As(V) breakthrough with C-II4B-3 ................................................50
4.3 Arsenic breakthrough curves for RSSCTs using S-1-7 (EBCTLC= 4 min) and a blend
(S-1&2) of S-1-7 and S-2-4 (EBCTLC = 10 minute)....................................................51

xi
4.4 Arsenic breakthroughs by four Fe-GAC composites. RSSCTs run in parallel from
common water source. 33% of Scottsdale groundwater plus 67% nanopure water
was used. All media were 100x140 mesh size.............................................................52
4.5 RSSCTs for As(V) removal in modified Scottsdale groundwater ( 33% groundwater +
67% nanopure water) spiked with SOC using Fe-GACs from Clemson University
and ASU (EBCTLC=4 min) ..........................................................................................53
4.6 Concurrent As(V) and SOC removal by two Fe-GAC composites. Initial MTBE and
benzene concentrations were 2.8 mg/L and 1.5 mg/L, respectively............................54
5.2 RSSCT vs. batch adsorption capacities for arsenic removal based on iron content ..........64
5.4 SEM/EDX surface analysis for A-3-7 (Norit HD3000 GAC) ...........................................66
5.5 SEM/EDX surface analysis for A-J1-6 (Jacobi GAC) ......................................................66
5.6 Elemental profile of cross-section through a Fe-GAC material (Jacobi) ..........................67

xii
 EXECUTIVE SUMMARY
RESEARCH OBJECTIVES

The goal of this project is to validate proof-of-concept testing for iron enriched granular
activated carbon (GAC) composites (aerogel-GAC or iron-oxide impregnated) as a viable
adsorbent for removing arsenic from groundwater and conduct technical and economic
feasibility assessments for these innovative processes. Specific project objectives include:
• Conduct batch experiments for aerogel-GAC and Fe-oxide impregnated GAC
composites to evaluate their performance removing arsenic.
• Evaluate Fe-GAC media performance in rapid small scale column tests (RSSCTs) to
assess arsenic removal in a more dynamic treatment system.
• Evaluate Fe-GAC potential for removal of other contaminants (e.g., methyl tertiary
butyl ether, dissolved organic carbon).
• Characterize Fe-GAC media.
• Correlate performance and media characterization for possible selection of two media
for a future second phase of this project.

APPROACH

Three different techniques to impregnate iron onto GAC was evaluated: 1) iron containing
aerogel impregnation of GAC (Lawerence Livermore National Laboratory); 2) four methods of
impregnating iron oxides or zero valent iron onto GAC (Clemson University); 3) impregnating
iron hydroxide onto GAC (Solmetex Inc. & ASU). Iron content of each sample was determined
before conducting batch adsorption experiments and/or column tests. Effects of water chemistry
(pH, competing ions) and operating conditions (empty bed contact times) were evaluated. Co-
removal of arsenic and synthetic organic chemicals (MTBE and benzene) were evaluated.

CONCLUSIONS

• Fe-GAC composite media contained <1% to 15% iron by dry weight. There were no
friability issues during use of the media, nor did iron appear to “flake” off of the
composite media. Thus Fe-GAC media appear suitable for batch or packed-bed use
during water treatment.
• Arsenic removal by Fe-GAC composites depended highly upon the process used to
impregnate iron. Iron-aerogel materials had the lowest iron content and exhibited the
lowest arsenic adsorption capacities. Roughly six other techniques to impregnate iron
into GAC were successful and were the focus of the study. For these materials iron
impregnation did not alter the surface area significantly (<15% change), as measured by
nitrogen deposition. However, arsenic adsorption varied significantly based upon the
technique to impregnate iron.
• Arsenic adsorption was not a function of the iron content (i.e., mgFe/mgGAC). This
implies that the technique to impregnate Fe-GAC composites and the resulting
mineralogy and structure is probably more important than the iron content.
• The best performing Fe-GAC composite for As(V) from Clemson University (C-II4 and
C-II4B) was produced via an iron hydroxide precipitation process. This was superior to

xiii
 iron impregnated via ion exchange, organic-ligand based, or zero valent iron based
techniques. The zero valent iron based technique result achieved comparable or slightly
better As(III) removal than other Clemson University Fe-GAC composites.
• The best overall performing Fe-GAC composite was synthesized by Arizona State
University and Solmetex Inc. using a patented process. Generally, this process involves
in-situ precipitation of iron (hydr)oxide using an oxidant and reduced iron species.
• Extensive media characterization was undertaken (surface area, pore size distributions,
zeta potential, and various spectroscopy measurements). These proved valuable in
assessing differences between iron deposition techniques.
• Batch kinetic experiments indicated that 100x140 mesh Fe-GAC composites reached
pseudo-equilibrium for As(V) adsorption within 7 days of contact time. Adsorption
isotherms were fit well by the Freundlich Isotherm model (q = KCe1/N).
• In batch experiments with Fe-GAC composites, As(V) adsorption improved as solution
pH decreased. For example, the adsorptive capacity constant (K) decreased by a factor of
ten for each full pH unit increase, between pH 5 and 9.
• In batch experiments with Fe-GAC composites and silicate (20 ppm) at pH 6.5 and 8.5,
addition of silicate decreased As(V) adsorption at both pH levels. All Fe-GAC
composites evaluated exhibited similar pH, silicate, vanadate and other competing ion
effects.
• In batch and column experiments with Fe-GAC composites, As(V) adsorption was lower
in groundwaters or NSFI-53 challenge water compared model solutions prepared from
nanopure water at an equivalent pH as the groundwater. Thus, ions present in these more
complex water matrices compete for arsenic adsorption sites on the Fe-GAC composite.
Addition of 100 ppb vanadate to the NSFI-53 challenge water had almost no effect on
As(V) removal. Addition of 20 ppm silicate to the NSFI-53 challenge water decreased
As(V) adsorption at both pH 6.5 and 8.5, compared against the NSFI-53 challenge water
prepared without silicate added. Thus silicate appears to be a major foulant/competitor
for As(V) adsorption sites on Fe-GAC.
• In batch and column tests, decreasing the ionic strength of a groundwater through
blending with nanopure water resulted in improved As(V) adsorption. These experiments
maintained the same ratio of ions, but demonstrate that the concentration of competing
ions affect As(V) adsorption.
• Rapid small scale column tests (RSSCTs) were conducted in continuous flow operation
with empty bed contact times (EBCTs) of 2.5 to 10 minutes. Fe-GAC composites that
performed best in batch testing, also had the best performance in RSSCTs. The ASU and
Solmetex Inc. Fe-GAC composites achieved the longest run length (i.e., bed volumes
treated). Lengthening the EBCT resulted in a larger number of bed volumes treated. The
minimum EBCT recommended EBCT is 4 minutes; at 2.5 minute EBCT the mass
transfer zone was not captured within the column.
• RSSCTs operated for 50,000 to over 100,000 bed volumes before complete As(V)
breakthrough.
• RSSCTs were also conducted with a surface water containing dissolved organic carbon
(DOC) and model groundwaters spiked with MTBE and benzene. Both waters also
contained As(V). As(V) was removed effectively, and did not seem to be impacted by
the presence of DOC, MTBE or benzene. Fe-GAC removed not only the As(V),
presumably due to adsorption onto iron (hydr)oxides, but also removed the DOC and

xiv
 synthetic organic chemicals (MTBE and benzene). Organics were presumably removed
by adsorption onto the GAC. Therefore, iron impregnation does not appear to hinder
organic removal by GAC, but adds the capability to remove arsenic.
• GAC is a relatively inexpensive substrate to impregnate with iron. For water treatment
situations were co-removal of arsenic and other contaminants suitable for removal by
GAC is desired (e.g., SOCs, radionuclides, DOC) Fe-GAC composites are a feasible
option.

RECOMMENDATIONS FOR FUTURE RESEARCH

• Optimization of the selection of base GAC substrate and iron deposition process needs to
be conducted. This should follow the basic technique which resulted in the best
performing Fe-GAC composites (produced by ASU and Solmetex Inc.).
• Batch equilibrium tests proved to be excellent indicators of more labor-intensive RSSCT
performance. Future work should use batch tests to screen new composites. However,
optimization of Fe-GAC composites must also consider that deeper iron penetration into
the GAC base material may impact arsenate mass transfer. Therefore, RSSCTs should
continue to be conducted to assess mass transport limitations.
• Because Fe-GAC composites are new materials, there is a need to compare arsenic
breakthrough achieved from pilot tests against those achieved by RSSCTs to validate the
scaling approach.
• Improved spectroscopic analysis of Fe-GAC composites is warranted to understand the
nano-structure of the iron deposits (coated surface versus crystals crossing through the
pore space). This could help optimize the iron deposition process (e.g., duration of iron
treatment). Ultimately the goal should be to maximize iron content for the best iron
mineralogy to result in the most arsenic removal.
• Forensic analysis of spent media would provide additional insight into the role of foulants
and changes in mineralogy over time.
• A detailed economic analysis of iron-oxide impregnated GAC composites should be
conducted. Without full-scale production of these products the economic analysis could
be conducted based upon costs of raw materials, costs for similar processes applied for
other products, and lifecycle cost of the media.

xv
xvi
 CHAPTER 1 INTRODUCTION AND OBJECTIVES
STATEMENT OF PROBLEM

Arsenic is a ubiquitous metalloid that naturally occurs in surface and groundwaters. The
speciation of arsenic depends on the pH and redox potential (e.g., Eh) of the water body.
Arsenic usually occurs as arsenate [As(V)] or arsenite [As(III)] in potable water supplies.
Arsenate (H3AsO4, H2AsO4-, HAsO42-, or AsO43-) is typically present in the mono- and
divalent anionic forms in oxygenated waters while arsenite (H3AsO3, H2AsO3-, and
HAsO32-) occurs primarily in the neutral and in lower Eh waters. Arsenic is classified as a
human carcinogen (prone to cancer of the bladder, lungs, skin, kidney, liver, and prostate)
by the International Agency for Research on Cancer and the National Research Council
(NRC, 1999, 2002; Pontius et al., 1994). In 2002, the USEPA lowered the maximum
contaminant level (MCL) for arsenic in drinking water from 50 μg/L to 10 μg/L, and the
new MCL became effective in January 2006 The WHO, the European Union, and several
countries recently lowered the recommended or required arsenic limit to 10 μg/L.
Because of the lower MCL, potable water suppliers have an increased need for arsenic
removal processes suitable for treating water sources with low ambient arsenic
concentrations (10 to 50 μg/L).

Arsenic can be removed from surface and groundwater supplies by iron or alum
coagulation with filtration, reduced Fe/Mn oxidation systems, and lime softening (Cheng,
Wang et al. 1994; McNeill and Edwards 1995; Scott, Green et al. 1995; Hering, Chen et
al. 1997; McNeill and Edwards 1997; Chen, Fields et al. 2002). Anion-exchange resin
and activated alumina packed-bed systems are traditionally used for wellhead or point-of-
entry groundwater treatment systems (Lin and Wu 2001; Wang, Chen et al. 2002). In
recent years, several iron-based adsorbents have been developed that typically require
less chemical pretreatment and/or can treat a greater number of bed volumes than anion
exchange resins or activated alumina. Porous iron oxide or hydroxide (GFH, Bayoxide
E33) packed-bed adsorbent systems are used in Europe for arsenic treatment (Jekel and
Seith 2000). Zero valent iron (ZVI) systems are used for in-situ and above ground
arsenic mitigation (Lackovic, Nikolaidis et al. 2000; Farrell, Wang et al. 2001; Su and
Puls 2001; Bang, Meng et al. 2002; Manning, Hunt et al. 2002; Meng, Bang et al. 2002).
The porous nature of the iron oxides/hydroxides presumably results in arsenic adsorption
at internal iron complexation sites, whereas ZVI relies on iron corrosion, reprecipitation,
and arsenic adsorption on the external surface of the adsorbent. Bench-scale comparisons
of porous and non-porous iron, sulfur-modified iron, and activated alumina adsorbents
for arsenic removal suggest a large number of newly developed adsorbents may be
economically feasible in achieving the new arsenic MCL (Sinha, Lee et al. 2002).

Iron has a high affinity for arsenic and forms strong inner-sphere complexes. Based on
the literature, iron (hydr)oxides have pKa1 and pKa2 values of ~7.3 and 8.9, respectively,
resulting a pHZPC on the order of 8.1. Coulombic forces favor association of anionic
arsenate with positive surface sites (e.g., MeOH2+). Arsenic adsorption onto TiO2 has not
yet been modeled, but probably mirrors other inner-sphere metal (hydr)oxide surface site
(SOH) complexes with arsenic:

≡FeOH2+ ' ≡FeOH + H+ pKa1 (1.1)

- +
≡FeOH ' ≡FeO + H pKa2 (1.2)
SOH + H3AsO4 ' SAsO42- + H2O + 2H+ (1.3)
SOH + H3AsO3 ' SAsO3- + H2O + H+ (1.4)

In addition to this electrostatic bonding between arsenic species and mineral surfaces,
arsenic will also form covalent bonds with some surfaces. These include monomolecular
monodentate and monomolecular bidentate bonds. Whereas electrostatic bonds form
rapidly (seconds) and depend on the charge difference between the arsenic and the
surface, covalent bonds depend on their respective molecular structure and form less
rapidly. Covalent bonds are stronger (i.e., irreversible) than electrostatic attractions. As
covalent bonds form, surface sites can become available for electrostatic bonding again.
The kinetics of bond formation may affect the optimal contact time required for a specific
media in a column operation. Overall, media taking advantage of the strong affinity
between arsenic and iron will lead to good treatment strategies.

This project involves a new array of granular activated carbon (GAC) based adsorbents
that contain iron for arsenic removal. GAC has low capacity for arsenic, but provides a
porous substrate to which iron or iron compounds can be added. Iron enriched GAC
would remove both arsenic and organics (VOCs, SOCs, DOC). By combining the
benefits of GAC and iron (hydr)oxides a new class of arsenic adsorption media can be
developed with the following benefits:

• GAC has much higher surface area (700-1200 m2/g) than commercially available
metal oxide media (GFH (226 m2/g), E33 (130 m2/g), FS-50 (232 m2/g), iron-
modified AA (161 m2/g) and activated alumina (265 m2/g)). The higher GAC surface
area has the potential for a larger number of arsenic adsorption sites.
• Iron enriched GAC (Aerogel or iron-oxide impregnated) will simultaneously remove
VOC’s, SOC’s, uranium, and potentially radionuclides in addition to arsenic
• Iron enriched GAC (Aerogel or iron-oxide impregnated) have apparent densities of
~0.6 g/cc and are lower than metal oxides (~1.5 g/cc), so less absolute mass of GAC
is required to fill a packed-bed vessel of a given size.
• Commercial iron-oxide adsorbents cost between $4/lb to $8/lb, whereas GAC
substrate is < $1/lb and the cost of iron enrichment will probably mean a less
expensive product.
• Iron enriched GAC can be manufactured to optimize iron content
• Aerogel-GAC composites are a new and novel treatment process that maintains a
high porosity and high surface area; iron-oxide impregnation also maintains most of
the surface area.
• Iron enriched GAC can be distributed via a well developed GAC process platform or
replace media in existing metal-oxide packed bed treatment systems.

2
• Aerogel-GAC composites can be optimized for high arsenic selectivity (low
competition by competing ions) by changing the ratio of Fe/Mn content
• Aerogel-GAC composites can serve as a platform product and aerogel could be
reformulated to remove other contaminants (hexavalent chromium, perchlorate)
• The infrastructure for GAC supplies, markets, treatment infrastructure (vessels,
pumps, handling, etc), and disposal is very well established and would provide rapid
deployment (market penetration) of iron-modified GAC composites into the water
industry.

PROJECT OBJECTIVES

The goal of this project is to validate proof-of-concept testing for iron enriched granular
activated carbon (GAC) composites (aerogel-GAC or iron-oxide impregnated) as a viable
adsorbent for removing arsenic from groundwater and conduct technical and economic
feasibility assessments for these innovative processes. Specific project objectives
include:
• Conduct batch experiments for aerogel-GAC and Fe-oxide impregnated GAC
composites to evaluate their performance removing arsenic.
• Evaluate the best GAC media performances in rapid small scale column tests
(RSSCTs) to assess arsenic removal in a more dynamic treatment system.
• Evaluate Fe-GAC potential for removal of other contaminants (e.g., methyl
tertiary butyl ether, dissolved organic carbon).
• Characterize spent Fe-GAC media.
• Correlate performance and media characterization for possible selection of
two media for a future second phase of this project.

3
4
 CHAPTER 2 EXPERIMENTAL AND ANALYTICAL METHODS
This chapter describes the types of media used for batch and column experiments.
Details of analytical methods are also presented. All laboratory waters were prepared
using nanopure water (NANOpure Infinity Ultra-pure Water System).

IRON MODIFIED MEDIA

Clemson University provided 17 different iron-impregnated GAC samples to ASU. Norit

HD3000 carbon was acid washed, heat or steam treated, and sieved prior to iron
treatment. These 17 samples were prepared using four different preparatory approaches,
plus control samples without iron impregnation (Table 2.1). All materials provided to
ASU were 100x140 mesh. Table 2.1 shows the Fe-GAC composites prepared by
Clemson University with the different iron impregnation processes. Coal-based
mesoporous GAC HD3000 (Norit Inc.) was used as precursor material in this study. The
different Fe-GAC samples were assigned acronyms based upon the institution preparing
the sample (“A” represents Arizona State University; “C” represents Clemson University;
“L” represents Lawerence Livermore National Laboratory; “S” represents Solmetex Inc.)
and an internal sample identification code. The majority of samples were prepared using
a single, common base GAC material (HD = Norit HD3000); for one set of samples a
second GAC material was used (Jacobi Carbons Ltd.).

Clemson University prepared all samples using acid washed Norit HD3000 material
(sample C-1). Some samples were also pretreated using heat (HDHT; sample C-2) or
stream (HDST; sample C-7). Then four different types of iron impregnation were
undertaken using these materials:
I. Ion exchange technique (HDHTFeS): acid-washed and heat treated Norit
HD3000 was mixed with FeCl3. The solution was evaporated at 800 oC under
a nitrogen gas flow stream.
II. Iron hydroxide precipitate technique (HDHTFeOH): acid-washed and heat- or
steam- treated Norit HD3000 was mixed with an aqueous FeCl3 solution. The
pH of the solution was increased to either pH 9 or 11 using sodium hydroxide
and then filtered. Another sample (HDHTFeOHI) was prepared in a same
fashion, but the GAC but carbon was separated from iron solution before
immersing in a 0.1N NaOH solution followed by filtration, washing and
drying.
III. Organic-iron complexed technique (HDHTFeOX): acid-washed and heat
treated Norit HD3000 was mixed with and organic-iron complex solution in
the presence of hydrochloric acid. After exposure to the solution, moisture
was evaporated at 800 oC.
IV. Zero valent iron technique (HDHTFe0): acid-washed and heat treated Norit
HD3000 was mixed with FeCl3. Then the solution underwent wet chemical
reduction prior to filtration. Another sample (HD-16NOFe2A) was prepared

5
 by mixing FeSO4 solution under nitrogen atmosphere, followed by air
oxidation under controlled pH (HD,16NOFe2A).

After preparation, all of the samples were extensively washed with distilled and
deionized water until no iron could be detected in the filtrate, then dried at 90°C for
samples treated by I and II and 40°C for III and IV. After sieving, a 100x140 mesh size
was used for the experiments. Later in the project Clemson University produced three
new Fe-GACs (C-16A-3, C-17A-3 and C-IV13B-6). These new media were designed
based upon the results of characteristics of C-II4-11 and C-IV13.

Solmetex Inc. and Arizona State University prepared a second set of materials used a
different process for depositing iron and is based on its patent license to impregnate iron
on to ion exchange resin (i.e., ArsenXnp media). This Solmetex iron-deposition process
produced four Fe-GAC samples, all with 12x60 mesh sizes. The 12x60 mesh size
material was crushed and sieved to 100x140 mesh. Table 2.2 presents Fe-GAC
composites manufactured by Solmetex and ASU following the patented process of
Solmetex. HD3000 GAC was the base for all composites with the exception of A-J1-6
which was prepared using Jacobia GAC as the precursor.

Lawerence Livermore National Laboratories (LLNL) provided 4 aerogel-GAC composite

samples containing iron and/or manganese. Like GAC, the aerogel has very large surface
area (approximately 1000 m2/g), but it has the added advantage of being hydrophobic
with a low density. Unlike GAC, the composite can remove both volatile organic
compounds (VOCs) and metals. The composite consists of approximately 10% of a
functionalized hydrophobic silica aerogel and 90% virgin coconut shell GAC. Aerogels
are open foam like structures with high surface areas, low densities and high porosity.
The composite materials are formed by mixing hydrophobic sol-gel precursors with
chemical agents including those with iron and manganese functional groups, gelation
catalysts, and GAC. After gelation of the sol-gel, the mixtures are supercritically dried to
form aerogel-GAC composites.

Table 2.3 summarizes some of the characteristics of the aerogel-GAC composites. The
pretreated GAC material with the higher iron content (294 and 295) removed the most
arsenate. Without pretreatment (virgin GAC) the presence of manganese seemed
important in aiding in arsenate removal (285 versus 291). The 285 GAC also had
significantly more surface area than 291.

Fe-GAC samples from Solmetex Inc. and LLNL were crushed in a ceramic mortar and
pestle and then sieved to 100x140 mesh size. This separation procedure of the media was
conducted wet using nanopure water to wash the Fe-GAC material to prevent entrainment
of fines. Most media provided to ASU from Clemson University and Solmetex had a
moisture content within 30-40%; however some samples had moisture content out of this
range. For this reason the experiment results were interpreted on a dry weight basis.
Before performing each experiment the specific media to be used was tested for moisture
content. Each time one gram of the sample was weighed in an analytical balance, and

6
then placed in an aluminum weighing dish covered with aluminum foil. These samples
then were placed in a Thermolyne 6000 oven at temperature of 100OC for approximately
12 hours. After this step the samples were left to cool for one hour and then weighed
again. This difference in weight was used to calculate the moisture content of the initial
sample. This dry weight determination procedure was done in duplicate to assure its
accuracy and then an average of the samples taken as the final value.

SPECTROSCOPIC CHARACTERIZATION

X-ray diffraction (XRD) analysis and X-ray photoelectron spectroscopy (XPS) analysis
were performed on Fe-impregnated GAC samples for characterization of the materials.
These iron-oxide impregnated GAC composites from Clemson University were modified
thorough the duration of the experimental work as a means of changing structural
properties to improve arsenic removal. The XPS and XRD tests were helped to measure
the critical parameters of each different composite and were preformed after each
modification. The instruments used were a scanning auger microprobe surface analysis
system (XRA TOS XSAM 800) and a high resolution X-ray diffract meter (Micro
photonics, Allenton, PA) equipped with Cu Kα radiation (40KV, 25mA). Eight of the
fourteen Fe-impregnated GAC samples provided by Clemson University were included:
one virgin sample (C-1), one ion exchange treated sample (C-I3-4), two iron hydroxide
precipitated samples (C-II4-11 and C-II14-2), two iron-deposited samples (C-III5A and
C-III10A) and two zero-valent impregnated samples (C-IV6-8 and C-IV13-6).

XPS Analysis

XPS was employed to determine elemental surface composition of Fe-impregnated

samples. Characterization results are summarized in Table 2.4. All samples contained
oxygen (O1s), carbon (C1s), silicon (Si 2p), and iron (Fe2p3/2). Information for the
surface characterization is limited for up to 5 nm depths from the external surface of the
particle.

Iron content of the samples varied from 6 to 11%. Sample C-5A contained the highest
amount of Fe, which is about 11%, followed by C-I3-4 (9.5%) and C-III10A (8.9%). All
other samples contained similar amounts of iron (~6-7%). The silicon/silica content
varied from 2-16%. Sample C-II4 contained the highest amount of silica (i.e., 16.5%),
followed by C-III10A (i.e., 5%), C-II14 (i.e., 5%) and C-IV13 (i.e., 4.5%). All other
samples contained similar amounts of silica (~2-4%).

Peaks for iron (i.e., Fe2p3/2) are detected at different binding energy levels depending
upon the chemical form on the surface (e.g., elemental iron at 707eV, Fe(II) as FeO at
709.3eV, Fe(III) as Fe2O3 at 710.8eV, FeOOH at 711.5eV). As the iron peaks are weaker
relative to carbon and oxygen peaks and become influenced by the background noise, any
information obtained from XPS regarding the bonding structure of Fe has limited
applicability.

7
XRD Analysis

The same eight samples mentioned in the XPS tests were also analyzed for X-ray
diffraction pattern. Peaks of graphite, quartz, magnetite and elemental iron were observed
in different samples. The width of the peaks suggests that samples are not purely
crystalline but slightly amorphous. All samples have a peak at diffraction angle (2θ) of
26.6, which corresponds to the graphite crystalline structure of carbon. Silicon (i.e., 2θ -
20.85) as quartz was detected at all samples. Iron as a form of magnetite structure (i.e., 2θ
- 34.4, 62.62) was noted in samples C-III10A, C-III5A, C-IV13, C-IV6 and C-II4,
whereas iron in a range of different forms (i.e., 2θ - 44.35) was noted in C-I3, C-III10A,
and C-III5A. Iron peaks were not noted in the virgin GAC sample or C-II14. As all
samples were attached to a glass plate with gel before analysis, area coverage on the glass
plates for C-II4 and C-II14 was not adequate due to non-sticky nature of the particle. All
samples were 100x140 mesh, a smaller particle size is expected to provide better area
coverage.

STOCK SOLUTION PREPARATION FOR EXPERIMENTS

All stock solutions used for batch experiments and column studies were prepared using
reagent grade chemicals. Arsenate and arsenite stock solutions were prepared using
Fisher Scientific reagent-grade arsenic acid (Na2HAsO4•7H2O) and sodium arsenite
(NaAsO2), respectively. 5mM of NaHCO3 was used for pH buffering and nitric acid (JT
Baker ultra pure reagent grade Ultrex II Nitric) to limit the pH within 7.0-7.5 units.
MTBE and benzene were from Fisher Scientific. These stock solutions (small volumes
~3-4L) were mixed using a VWR Scientific standard 360 stirrer and magnetic stir bars.
For column studies in which large volumes of water were needed, a Cole-Parmer Stir-
Park mixer was used for mixing by placing this tool directly in the tank containers.

All glassware needed for the preparation of these solutions was cleaned prior to use in
accordance with glassware cleaning procedures as described in the Analytical Methods
section.

BATCH ARSENIC ADSORPTION EXPERIMENTS

Batch Kinetic Test Procedure

60-mL HDPE Nalgene bottles were used for batch experiments. 10-mg (dry weight) of
Fe-GAC was added to each bottle. 40-mL of a pH buffered solutions containing arsenate
were placed in the bottles. The samples were placed in a box and agitated to provide
mixing. An orbital shaker (Barnstead/Thermolyne, model M49235) was used. Samples
were sacrificed after 0, 1-hr, 6-hr, 1 day, 3 days, 7 days,14 days and 21 days. Fe-GAC
media was separated by filtration using a Whatman 0.45 µm filter. Arsenic content in the
filtrate was measured.

8
Batch Isotherm Test Procedure

Fixed masses of Fe-GAC were added to HDPE Nalgene bottles (250-mL, 1-L) containing
the test solution with arsenic. The sample bottles were placed in a mechanical shaker at
low speed for approximately seven days. Then the samples were filtered (0.45µm
Whatman) to remove the Fe-GAC media. For experiments with arsenic, benzene and
MTBE, 100-mL clear glass serum bottles (Fisher Scientific) with gas-tight lids were
used.

Batch Isotherm Test using Modified NSFI-53 challenge water

Several batch tests were conducted to determine the water quality factors found in natural
waters that affect arsenic removal by Fe-GAC media. A modified NSFI-53 challenge
water was used to determine the significance of sulfate, fluoride, phosphate, bicarbonate
and other ions (Table 2.5). The effects of silica concentrations (0 and 20 mg/L) and
vanadium concentrations (0 and 100 µg/L) were also studied, separately, by preparing the
NSFI-53 water with or without these metals. pH effect in arsenic removal was also tested
(6.5, 8.5) at the same time with the parameters above mentioned. The procedure to make
and use the NSFI-53 water is as follows:
1. Chemicals were added to nanopure water to meet the specific characteristics
shown in Table 2.5.
2. Add sodium silicate (Na2SiO3•9H2O) to achieve a test tank concentration of 93
mg/L Na2SiO3•9H2O. Stir and transfer the solution to the test container.
3. Add sodium bicarbonate (NaHCO3) to achieve a test concentration of 250 mg/L
NaHCO3. Stir and transfer to the test container.
4. Separately add magnesium sulfate (MgSO4•7H2O), sodium nitrate (NaNO3) and
sodium fluoride (NaF) to achieve test concentrations of 128mg/L, 12mg/L and 2.2
mg/L, respectively.
5. Add sodium phosphate (NaH2PO4•H2O) to achieve a test concentration of 0.18
mg/L NaH2PO4•H2O. Stir and transfer to the test container.
6. Add calcium chloride (CaCl2) to achieve a test concentration of 111mg/L CaCl2.
Stir and transfer to the test container.
7. Adjust the pH of the solution using HCl or NaOH to 6.5±0.25 for the low pH test
and to pH 8.5±0.25 for the high pH test.
8. Add arsenic solution in the test container, depending on the desired concentration
of arsenate in the water. For these experiments enough arsenic solution was added
to reach a concentration of 100 µg/L of As(V).
9. Mix and measure the final pH and adjust as needed.
10. For vanadium tests, instead of adding sodium silicate in step 2, sodium vanadium
(Na3VO4) is added to achieve 0.36 mg/L Na3VO4.

9
Sample Filtration Procedure

All samples from the batch experiments were filtered with a Gelman 25 mm filter holder
and a 0.45 µm Nuclepore polycarbonate membrane filter (0.45 µm Whatman). A 60-mL
BD syringe with plastic tubing was used to take the sample out of the bottle, and then the
60-mL water sample was passed through the filter to remove the adsorbent media and
poured directly into a cleaned 60-mL HDPE Nalgene bottle or in cleaned glass 100-mL
serum bottles in the case of organic samples. For every water sample the syringe was
rinsed with nanopure water and a new 0.45µm Whatman filter loaded in the filter holder.

RAPID SMALL SCALE COLUMN TESTS

Rapid Small Scale Column Tests (RSSCTs) were developed in the 1980’s to evaluate
removal of organics by GAC in hydrodynamically-scaled laboratory columns based upon
pilot- and full-scale operational parameters (GAC media size, loading rates, EBCT, etc.).
This concept was extended for arsenic testing using porous metal oxide (Badruzzaman
2002). Scaling can be based either upon proportional (PD) or constant (CD) diffusivity
assumptions (Table 2.6). For arsenic removal by porous metal oxides it was concluded
that PD scaling was most appropriate (Badruzzaman 2002). Proportional diffusivity (PD)
scaling equations were used for perchlorate removal on iron-preloaded GAC. It is
believed that PD scaling is appropriate for arsenic (Westerhoff and Badruzzaman 2003;
Westerhoff, De Haan et al. 2006). All RSSCTs employed herein will use PD scaling
designs.

RSSCTs using Fe-GAC materials were sized with PD assumptions. Columns studies
were performed with EBCTLC’s of 2.5 minutes (samples C-II4, C-II9, C-IV13 and C-
II14), 4 minutes (samples C-II4, C-III5, C-IV6, C-IV11, S-1, and S-3) or 10 minutes
(samples S-1, S-2). Flow rate was set to approximately 11-mL/-min with a Reynolds-
Schmidt number of 1000 for all of these experiments. A spreadsheet was designed (Table
2.7) based on the scaling equations presented above in which design parameters are
calculated for small scale columns tests.

Small Scale Column Packing Procedure

Figure 2.1 shows the main components required for a column experiment and a schematic
of a typical set up. The list of parts and procedure to pack a glass column is described as
follows:
1. Material and Components:
• Small glass column with diameter of 1.1 cm and a length of 30 cm (Ace
Glass)
• Inlet endcaps with filter ring and outlet endcaps with o-ring (Spectrum
Chromatography)
• Glass wool
• 5 mm diameter-borosilicate glass beads (VWR)
• Teflon tubing of 3.2 mm diameter (Spectrum Chromatography)

10
 • QG150 piston pumps (Fluid Metering Incorporated, FMI)

2. Sequence Procedure:
• A clean glass column with the outlet endcap is filled with nanopure water and
glass wool added.
• 5 mm diameter glass beads are added to disperse the influent flow.
• Approximately 1 cm of glass wool is added at the top of glass beads to
support the media.
• The Fe-GAC media (100x140 mesh) is added using nanopure water to flush
the material down until reaching the required length or mass.
• After packing the media, it was backwashed to remove the fines at a low flow
rate (~10-mL/min) in up flow mode with nanopure water until the effluent
looked clear.
• Once the media is backwashed, 1 cm of glass wool is added followed by glass
beads.
• The endcaps are connected to the Teflon tubing (inlets to the pumps and raw
outlets to sample waste/collection points). It is critical to maintain the sample
waste/collection tube elevation at an elevation above the inlet end-cap of the
RSSCT column to prevent air bubbles from entering the packed bed.

The tank containers for the groundwater used for these column studies were HDPE
reservoirs from Nalgene. The volumes of these containers were 50 and 150 gallons. All
the tanks were cleaned with nanopure water several times prior to use.

Sampling Procedure and Flow Rate Setting for Rapid Small Scale Columns

All column studies were conducted in the water quality laboratory at ASU. The inlet and
outlet tubing were 1/8th inch inside diameter Teflon tubing. Influent groundwater was
stored and pumped to the RSSCTs from either a 50 or 150 gallon HDPE tanks. The
pumps used in these experiments were QG150 piston pumps with Q2CSC pump heads
purchased from Fluid Metering Incorporated (FMI). Water samples were collected
directly from the outlet Teflon tubing line into 60-mL HDPE Nalgene bottles as required
and then preserved with 0.15% nitric acid according to the procedures described in the
Analytical Methods section. Samples collected for MTBE-benzene adsorption
experiments were stored in clear glass 100-mL serum bottles for analysis and stored at
4oC.

The flow rate of the water thorough the columns was measured in a very simple
procedure. Using a stopwatch and a 100-mL graduated cylinder from VWR the water
from the outlet line was collected in the graduated cylinder for 3 minutes. Flow rate was
measured and adjusted throughout the course of the RSSCT. However, the FMI piston
pumps were found to be extremely stable and rarely needed flow adjustment during the
course of an RSSCT experiment.

11
ANALYTICAL METHODS

Sample Preservation

Samples that were collected for trace metals analysis or for arsenic measurements were
preserved with nitric acid (JT Baker ultra pure reagent grade Ultrex II Nitric). 0.15% total
volume of nitric acid was added to the 250-mL and 60-mL sample bottles for a result of
approximately 350 µL and 90 µL of acid, respectively. All samples acidified achieved a
pH of around 2.0 units. Samples were stored in the dark at 4oC until the time of analysis.

Glassware Cleaning Procedure

All glassware, bottles and glass RSSCT were rinsed with nanopure water before being
submerged in a 0.1 % nitric acid bath for 24 hours prior to their use. Then equipment was
immediately rinsed with nanopure water at least three times and allowed to air dry. Clean
bottles were stored and placed in ziplock plastic bags before their use to prevent
contamination.

Arsenic Analysis by Graphite Furnace Atomic Adsorption (GFAA)

The instrument used for arsenic analysis is a Varian SpectrAA 400 Zeeman graphite
furnace atomic adsorption spectrometer equipped with correction capabilities (zeeman
background). This particular instrument has an Arsenic super lamp with a photron
boosted device. All water samples were dispensed into the furnace with the auto sampler
device included in this instrument. All samples were acidified with nitric acid before
analysis.

The instrument was calibrated using freshly prepared arsenic standards using an atomic
spectral standard from VHG labs. The solution contains an arsenic concentration of 1000
µg/L and a matrix of 5% HNO3 that was blended to reach the appropriate operating range
for the instrument (2, 5, 10, 20, and 30 µg/L). 0.1% nitric acid was also added to the
standards to reach the same characteristics of the water samples. The 1000 µg/L arsenic
solution from VHG labs was blended to achieve a stock solution of 500 µg/L with 0.1%
nitric acid. A five-point calibration curve was developed. All volumetric glassware used
for the arsenic stock solution and the standards preparation was properly cleaned
following the same procedure described in the Glassware Cleaning Procedure prior to
use.

The Varian SpectrAA GFAA uses the analytical method 200.9 for trace elements; the
method detection limit described and calculated in the Detection Limit Procedure was
determined to be approximately 1 µg/L. However the actual detection limit and linear
ranges are dependent on the sample matrix and instrumentation parameters selected for
each operation. GFAA sensitivity and limited linear range generally implies the need to
dilute the samples before every analysis. For this project all samples with concentrations
within the range 50-250 µg/L were blended from 10 to 1 times, but most of the samples
fell within the range of the standard curve calibration of 2-30 µg/L. For the sample

12
dilution preparation, a stock of nanopure water solution with 0.1% nitric acid was made
every time. An auto pipette was then used to dispense the required nanopure solution in
the 2-mL disposable cups used for the samples analysis, the second step was to add (with
a different clean pipette) the required amount of the water sample depending of the
dilution factor. Then the total volume of the sample in the cup was mixed using the auto
pipette to remove approximately one-third of the volume and then re-dispensed again into
the cup several times to allow a complete and efficient dilution of the water sample. The
procedure was repeated for each sample using clean pipettes every time.

Water samples were analyzed by GFAA with the use of a standard calibration curve with
a range of 2-30µg/L. The calibration curve should show a linear response for accurate
results when measuring trace metals in drinking water samples. Figure 2.2 shows the
variability of the slope of the standards calibration curve from several GFAA analyses
performed during 2005. An external Quality Control (QC) Check Standard was
purchased from VHG Labs and prepared for an arsenic concentration of 10 µg/L
following the proper instructions in order to compare and verify the accuracy of the
calibration standards. This verification with the quality control sample was realized every
nine samples through the complete analysis. In every analysis the standard deviation of
these external quality control samples was calculated and the variation was always less
than 10%. Figure 2.3 presents a summary of QC values obtained from several GFAA
analyses performed during 2005. The variability of this sample fluctuated between 8 to
11 ppb.

The determination of the Method Detection Limit (MDL) was followed in accordance to
the guidelines described in Appendix B of the Federal Register. Seven aliquots of the
estimated MDL value were prepared at five times the desired detection limit, in this case
of 1 µg/L. The MDL was determined to be 0.54 μg/L from the deviation standard of the
seven replicates as shown in Table 2.8.

Gas Chromatography (GC) Analysis

Water samples were analyzed for organics removal by using headspace gas
chromatography (GC) procedures. The GC instrument (SRI Model 8610C) was equipped
with a 0.53mm ID x30m MXT-Vol glass capillary column with a 2.0 mm. 0.5-mL of
headspace gas was injected into the GC instrument, which was set isothermally at 80oC
for 2 -min in each sample.

Standards of MTBE and benzene were prepared in methanol. The density of MTBE
(ρMTBE = 0.743 g/cm3) or benzene (ρbenzene = 1.01 g/cm3) were used to prepare 10,000
mg/L stock solutions in methanol. For example, 0.471-mL of MTBE was placed into 35-
mL of methanol to prepare a 10,000 mg/l stock solution. Figures 2.4 and 2.5 present
calibration curves.

13
Other Measurements

pH Measurement

For water samples required for temperature and pH measurements, a φ240 Beckman
basic meter which includes a pH indicating electrode, an automatic temperature probe
and a different range of pH buffers (1.68, 4.00, 7.00, 10.01, 12.45) was used. This
instrument is equipped with selectable resolution, automatic temperature compensation
(ATC), Auto-Find that automatically recognizes stored buffers, and Auto-Read to lock
TM TM

onto stable readings. The electrode and automatic temperature compensator probe were
rinsed with deionized water and the excess blotted free using KimwipesTM. Then the
electrodes were stirred briefly in the samples and the readings recorded. The same
procedure was followed for each sample.

Conductivity Measurements

Conductivity measurements were done in the central laboratory using a conductivity

meter (VWR model 2052-B) which uses a Pt-plated dip electrode. The instrument was
standardized using a factory provided 718 μΩ standard solution. Measurements were
made in a 20 mL test tube. Between each sample the dip electrode was rinsed with
ultrapure water and dried with KimwipesTM.

Dissolved Organic Carbon (DOC) Analysis

Dissolved organic carbon (DOC) was analyzed with a Shimadzu TOC 5050A carbon
analyzer, and UV absorbance of DOC at 254 nm (UVA254) was measured with a
Shimadzu UV/Vis 1601 Spectrophotometer. For the UV absorbance analysis the
wavelength was set to 254 nm. Optically cells were purchased from Fisher Scientific for
the analysis; these cells were rinsed several times with ultrapure water before use. All
samples expected to be over 1.0 absorbance were blended with ultrapure water prior to
analysis. Specific absorbance (SUVA) was calculated as UVA254 divided by DOC. The
MDL for DOC was 0.2 mg/L.

14
 Table 2.1 Fe-GAC media properties provided by Clemson University

Iron Impregnation Sample Surface Total Micropore Micropore pHZPC Iron

Approach ID Area Pore Volume + Content
(m2/g) Volume (cm3/g) Macropore (mg Fe/g -
(cm3/g) Volume GAC)
(cm3/g)
HD3000 GAC C-1 686 0.712 0.266 0.446 5.3 <3
HD Heat Treated C-2 768 0.726 0.298 0.428 7 <3
HD Steam treated C-7 853 0.790 0.325 0.465 7.4 <3
I. Ion exchange C-I3-4 778 0.724 0.295 0.429 6.8 36.2
and heat treatment C-I8-4 860 0.766 0.334 0432 5.6 40.9
HDHTFeS
II.Iron hydroxide C-II4-11 646 0.649 0.249 0.400 6.8 116.3
precipitation C-II4B-3 679 0.642 0.247 0.385 7.2 34.0
HDHTFeOH C-II9-8 762 0.726 0.296 0.430 6.9 78.8
HDSTFeOHI C-II14-2 832 0.752 0.321 0.431 7.5 20.9

III.Organic comp C-III5-16 619 0.587 0.231 0.356 6.5 163.5

to deposit iron C-III10-9 772 0.690 0.301 0.389 6.1 87.3
HDHTFeOX

IV.Zero valent C-IV6-8 653 0.696 0.241 0.455 6.3 83.3

iron surface C-IV11-7 742 0.704 0.288 0.416 6.6 74.3
HDHTFe0 C-IV13-6 788 0.728 0.302 0.426 8.5 58
HDST,16NO, C-16-3 836 0.783 0.31 0.473 5.4 31.6
Fe2A C-17-3 863 0.781 0.325 0.456 5.5 30.9

15
 Table 2.2 Fe-GAC media properties produced by Solmetex and ASU

Iron content
Supplier Carbon Type Sample ID
(mgFe/g-GAC)
Solmetex Inc. Norit HD3000 S-1 67.6
Norit HD3000 S-2 38.4
Norit HD3000 S-2B 112
Norit HD3000 S-4 105
ASU (Solmetex Fe-Treatment) Norit HD3000 A-3 71.8
Jacobi GAC A-J1 63

Table 2.3 Characteristics of aerogel-GAC composites

Aerogel-GAC Iron content Manganese Granular Activated Surface Area

composite ID (%) Content (%) Carbon (m2/g)
Virgin GAC 0 0 Coconut shell --
L-285 0.8 0.02 Virgin coconut 739±20
shell
L-291 0.8 0 Virgin coconut 450±11
shell
L-294 0.75 0 Pretreated --
L-295 0.96 0 Pretreated --
L-265 0.61 1.2 Pretreated 853±23
L293-2 0.15 0 Virgin coconut --
shell

16
 Table 2.4 Elemental surface composition of samples

Sample ID O (%)* C (%) Si (%) Fe (%)

C-1 33.63 58.96 4.03 0.0
C-I3-4 39.24 50.1 2.62 6.93
C-II4-11 19.68 55.33 16.51 6.78
C-II14 23.78 61.93 5.09 6.5
C-III5A 29.9 53.71 2.77 11.05
C-III10A 27.97 55.54 5.09 8.86
C-IV6 31.11 55.7 2.69 7.56
C-IV13 23.73 53.97 4.57 9.54
* Atomic percent basis

Table 2.5 NSFI-53 challenge water formulation

Parameter Target Value

Mg+2 12 mg/L
SO4-2 50 mg/L
NO3-1-N 2.0 mg/L
F-1 1 mg/L
SiO2 or 0 & 20 mg/L
Na3VO4 0 & 100 µg/L
PO4-3-P 0.04 mg/L
Ca+2 40 mg/L
As(V) 100 µg/L
Temperature 20oC±2.5oC
Turbidity < 1 NTU
pH 6.5±0.25 or 8.5±0.25

17
 Table 2.6 RSSCT scaling equations

Scaling CD or PD Specific Relationships General Relationships

Assumption
Constant 2− X
EBCTSC ⎡ d p,SC ⎤ tSC
Diffusivity =⎢ ⎥ = Eqn 2.1
(CD: X=0) EBCTLC ⎢⎣ d p,LC ⎥⎦ tLC

VSC ⎡ d p,SC ⎤ ⎡ ⎤
Eqn 2.2 VSC = d p ,SC × Re SC × Sc Eqn 2.5
=⎢ ⎥ ⎢ ⎥
VLC ⎢⎣ d p,LC ⎥⎦ VLC ⎣⎢ d p ,LC ⎦⎥ Re LC × Sc
Proportional X V × ρL × d p Eqn 2.6
DSC ⎡ d p ,SC ⎤ Eqn 2.3 Re =
Diffusivity =⎢ ⎥ μ
(PD; X=1) DLC ⎢⎣ d p ,LC ⎥⎦
μ
⎡d ⎤ t Sc = Eqn 2.7
EBCTSC
= ⎢ p ,SC ⎥ = SC
Eqn 2.4 DL × ρ L
EBCTLC ⎣⎢ d p ,LC ⎦⎥ t LC

18
Table 2.7 RSSCT design parameters for simulating a 4 minute full-scale pilot system

Simulated RSSCT
Design Parameters
Pilot Scale (PD)
Input Parameters
Particle radius (cm) 0.0513 0.0064
Column diameter (cm) 30.48 1.10
Column area (cm2) 729.29 0.9499
Wet density (g/mL) 1.79 1.40
EBCT (min) 4.00 0.50
Loading rate (m/h) 10.11 7.03
Loading rate (cm/min) 16.85 11.72
BV to be processed 30000.00 30000.00
Flow rate (gpm) 3.25
Flow rate (gpm/ft2) 4.14
Re*Sc (200,000-200) 1000.00
U (Pa s) 8.94E-04 0.00
P (kg/m3) 1000 1000.00
Dl (Free liquid Diffusivity) (m2/s) 1.00E-09 0.00
d (m) 0.001026 0.00
v (m/s) 0.002807 0.00
e (porosity) 0.25 0.25
Re 1.29E+01 1.12
Sc 893.7 893.70
Re*Sc 11522.10 1000.00
Design Data
Flow rate (mL/min) 12285 11.13
Volume of media (mL) 49140 5.6
Length of the column (cm) 85.34 5.8
Mass of the media (g) 87960.6 7.8
Water required (L) 1474200 166.6
Water required (gals) 390000 44.1
Duration of experiment (days) 83.33 10.4

Table 2.8 MDL for GFAA analysis

Sample Type Concentration
(5.0 ppb)
Replicate 1 5.4
Replicate 2 5.5
Replicate 3 5.1
Replicate 4 5.0
Replicate 5 5.2
Replicate 6 5.3
Replicate 7 5.3
Mean 5.25
Std Deviation 0.1718
MDL 0.54

19
 INLET ENDCAP

1.1 X 60 cm GLASS COLUMN

1.5 cm
O-RING

10.0 cm
GLASS BEAD

3.0 cm GLASS WOOL

23.4 cm GRANULAR ACTIVATED CARBON

(GAC)

0.5 cm WHITE SAND

1.0 cm TEFLON SLEEVE, 0.3 cm LONG
0.1 cm 100 MESH S.S SCREEN

3.0 cm TEFLON SLEEVE, 1.0 cm LONG

19.0 cm
GLASS BEAD

O-RING
1.5 cm

OUTLET ENDCAP

Figure 2.1 Schematic of a typical set up for a column study

 0.006

0.005

0.004
Slope

0.003

0.002 y = mx + b

0.001

0
29-Dec 17-Feb 8-Apr 28-May 17-Jul 5-Sep 25-Oct

Date

Figure 2.2 Variability of slopes of standard calibration curves for arsenic analysis
by GFAA during 2005

14

12
QC Values (~10ppb)

10

6
QC samples during 2005
QC samples from VHG Labs
4
Dilution to 10 ppb
Analyzed every 10 samples
2

0
29-Dec 17-Feb 8-Apr 28-May 17-Jul 5-Sep 25-Oct
Date
Figure 2.3 Summary of QC results for GFAA analysis during 2005

21
 100

Conc(mg/L) 10

1
(mg/l)

y = 0.0058x
Conc.

0.1 2
R = 0.99

0.01

0.001
0 500 1000 1500 2000
Area counts

Figure 2.4 Standards calibration curve for MTBE

100

10
(mg/L)
Conc. (mg/l)

1
Conc.

y = 0.0007x
0.1 2
R = 0.99

0.01

0.001
0 5000 10000 15000 20000
Area counts

Figure 2.5 Standards calibration curve for benzene

22
 CHAPTER 3 BATCH EXPERIMENT RESULTS
KINETIC TESTS

Most of the kinetic experiments were conducted using Clemson University samples with
consistent 100x140 mesh size. Nanopure water, pH = 7.00 controlled with 5 mM of
NaHCO3 and spiked with As(V) concentrations of 100 µg/L was used. Adsorbent dosages
of 0.25 g dry/L were set for each Fe-GAC sample (Figure 3.1). Samples C-2, C-II4-11,
C-I8-4, C-III10-9 and C-IV11-7 were selected for kinetic experiments. Each composite
represents a different iron deposition process for its impregnation (Table 2.1). C-II4-11
and C-IV11-7 had the highest adsorption capacity. Kinetic tests were conducted also to
determine if some samples had lower capacity, but improved kinetics. However in all
cases arsenic adsorption appeared to reach pseudo-equilibrium in approximately 3 days.

For comparison against existing commercial media, kinetic tests were conducted with
granular ferric hydroxide (GFH) which is a commercial non GAC based pure iron (hydr)
oxide media. From previous work, it was known that lower GFH dosages than Fe-GAC
composites would be required to show any observable kinetics. Therefore a GFH dosage
approximately 25 % less than that from Fe-GAC was tested (Figure 3.2). Additional tests
show that GFH does not reach equilibrium until 20 to 30 days (Badruzzaman, Westerhoff
et al. 2004). For the same mesh size GFH has a higher adsorption capacity than Fe-GAC
samples tested, but the Fe-GAC appeared to have faster adsorption kinetics.

SCREENING TESTS

Clemson University Media

Batch screening experiments testing different Fe-GAC media (100x140 mesh size) were
conducted in nanopure systems, pH = 7.00 buffered with 5 mM of NaHCO3 with an
initial arsenic concentrations of 100 µg/L (Figure 3.3). The contact time was set at 7 days
based on previous experience. Subsequent kinetic tests verified that this duration was
sufficient to reach equilibrium. Fe-GAC dosage of 0.25 mg dry/L was used. Samples C-
II4-11, C-IV13-6, C-II9-8, C-II14-2, C-IV6-8 and C-IV11-7 removed ~75% of the initial
As(V). A second batch screening test was conducted partway through the study as
Clemson University produced three new Fe-GACs (C-16A-3, C-17A-3 and C-IV13B-6).
These new media were designed based upon the results of characteristics of C-II4-11 and
C-IV13. During these tests a subset of Fe-GACs from the first screening were used as a
reference. Figure 3.4 shows that these new adsorbents had similar kinetic performance,
although the initial C-II4-11 and C-II9-8 samples had higher adsorption capacity.

Arsenite

Screening tests were conducted for As(III) removal using selected Fe-GAC composites.
Experimental conditions mirrored those for arsenate screening tests. Results are shown in
Figure 3.5. C-IV13B-6 (zero valent iron) was slightly better than C-II4-11 (iron

23
hydroxide precipitation). C-17A, C-16A and C-II4B-3 (all three treated by iron hydroxide
precipitation) had similar performance for arsenite removal.

Solmetex Fe-GAC Media

Two samples of Solmetex Fe-GAC were provided, with and without acid pretreatment,
these samples were manufactured using Norit HD3000. The non-acidified Solmetex Fe-
GAC (S-1-7) contained 6.7% iron by dry weight (67.6 mgFe/g dry composite), and the
acid-treated Solmetex Fe-GAC (S-2-4) contained only 3.8% iron (38.4 mgFe/g dry
composite). This range of iron content is on the lower end of the materials provided by
Clemson University that contained 6% to 13% iron by weight.

Batch tests were conducted with Solmetex Fe-GAC composites and the base GAC
material (virgin HD3000 carbon). Figure 3.6 shows the results. Virgin HD3000 had no
As(V) removal. The two Solmetex Fe-GAC composites had significant As(V) removal.
S-1-7 and S-2-4 had similar performance despite the difference in iron content. The
Solmetex Fe-GAC composites removed more (~90% for 250 mg/L of media) arsenic
from the solution than the aerogel material (~45% removal), that was based upon Carbon
Activated Corporations coconut-shell based GAC.

LLNL Fe-GAC Media

Sample L-294-1 is an iron functionalized hydrophobic silica aerogel composite provided

by Lawrence Livermore National Laboratories (LLNL). The composite consists of
approximately 10% silica aerogel and 90% virgin coconut shell GAC. This composite
was tested in parallel with Fe-GAC media from Solmetex (Figure 3.6). L-294-1 adsorbed
less arsenic(~45% removal) than S-1-7 and S-2-4.

Based upon adsorption experiments conducted in the past, aerogel L-294-1 sample was
selected to perform new isotherm tests in parallel with the C-II4-11 composite from
Clemson University. A nanopure system spiked with 100 ppb of arsenate and pH of 7.00
was used. The aerogel sample from LLNL was used in full mesh size due to a limited
volume of the media for reducing it to 100x140 size; the Fe-GAC sample from Clemson
University was used in 100x140 mesh size, no full size particle media was available.
Doses were on the range of 50-2000 mg dry/L and 50-500 mg dry/L for LLNL and
Clemson University composites respectively. Freundlich adsorption isotherm model (q =
KCe1/n) was used to fit the results obtained. Good (R2 > 0.94) isotherm fits were observed
for both composites. C-II4B-3 appears as the stronger adsorbent for arsenate with a
significant difference to the aerogel L-294-1 media, K values were 0.01 and 0.09 for
LLNL (1/n=0.51) and Clemson University (1/n=0.80) samples, respectively, indicating
the latter as the one with higher adsorption capacity for arsenic. The large difference in
adsorption efficiency between these media may be attributed to the larger particle size of
the aerogel sample due to pore blockage or bad distribution of the silica gel into the GAC
material.

24
Previous work at LLNL has investigated uranium removal by a phosphorous enriched
hydrophobic aerogel/GAC composite. Producing a P-enriched aerogel-GAC composite
improved uranium removal above the GAC without the composite. Varying the type and
amount of phosphorous (1% or 17% H3PO4) affected uranium removal, based upon
analysis of isotherm data. Additional absorbance was confirmed via column tests and a
field demonstration. This uranium-related work will be briefly discussed to demonstrate
the potential for aerogel-GAC composite materials to remove oxo-anions.

More tests were conducted for arsenic and uranium removal using aerogel-GACs.
Nanopure water with pH of 7.0 (5 mM NaHCO3) spiked with 100 ppb of arsenate and 50
ppb of uranium (Aldrich standard solution) was used. The particle size of the media was
10x60 (due to lack of enough mass of the media this sample was not sieved to 100x140
mesh). Doses used were within the range of 50 to 2000 mg dry/L. A phosphate
functionalized hydrophobic silica aerogel-GAC (sample L-05) and an iron functionalized
aerogel (sample L-06) were selected for these experiments; L-05 contained 17% H3PO4.
Both composites adsorb uranium from the solution, when equilibrium is reached 70% of
U was removed. Sample L-05 had better potential for removing the contaminant and
faster kinetics. This sample removed approximately 80% of arsenate in the solution
which is comparable with other Fe-GAC materials from Clemson University. The better
adsorption capacity of L-06 than L-05 might be explained by better impregnation of iron
on the GAC particles or the poor distribution of silica aerogel on the composite, as
visually observed by “patch” coverage of the Fe-GAC with whitish aerogel.

ISOTHERMS IN MODEL SOLUTIONS AND NATURAL WATERS

Model Solutions

Isotherm experiments were conducted for most Fe-GAC composites for arsenate removal
based on screening and kinetic tests. Nanopure water with an arsenic concentration of
100 µg/L and 5 mM NaHCO3 was used; the solution pH was regulated to 7.00±0.2.
Media doses (100x140 mesh sizes) varied between 10 mg/L to 500 mg/L dry weight. 7-
day contact times were used.

Clemson University Fe-GAC Media

Initially C-II4-11, C-II9-8, C-IV13-6 and C-II14-2 were tested and the results are
presented in Figure 3.7. Figure 3.8 is a plot of the same adsorption data from which
Freundlich Isotherm parameters were obtained. C-II4-11 and C-II9-8 performed better
than C-IV13-6 and C-II14-2. This data is consistent with the pattern observed from
screening tests.

Solmetex Fe-GAC media

S-1-7 and S-2-4 performed very similar in batch screening experiments with model
solutions. Sample S-1-7 was selected for an isotherm test along with sample C-II4-11
from Clemson University and ArsenXnp media (Figure 3.9). ArsenXnp is currently a

25
commercial hybrid ion exchange media from Solmetex which has an iron content of
approximately 20%. ArsenXnp had a higher adsorption capacity than the rest of the
samples. S-1-7 (which is 6.7% iron by dry weight) had a slightly higher capacity than C-
II4-11 (11% iron content).

Two more Fe-GAC samples were produced by Solmetex and tested in isotherm
experiments. These composites were prepared using virgin Calgon GAC instead of Norit
HD3000. S-2B-11 and S-4-11 have 11.2% and 10.5% iron, respectively. C-II4-11 sample
was included in this experiment for direct comparison of these media as arsenic
adsorbents. Figure 3.10 shows similar performance between samples S-2B-11 and S-4-
11. C-II4-11 had slightly lower capacity for arsenic adsorption (see also Figure 3.11).

ASU produced Fe-GAC composites following Solmetex iron deposition procedures. This
process is similar to the one used in impregnating nano-iron onto ion exchange media
(ArsenXnp media). ASU used virgin Norit HD3000 and Jacobi GAC as base material to
produce samples A-3-7 and A-J1-6, respectively. A-3-7 and A-J1-6 contain 7% and 6%
Fe, respectively. Figure 3.12 shows an isotherm in a model solution comparing both
composites from ASU with S-4-11. The graph indicates that S-4-11 has slightly higher
arsenic adsorption capacity. A-3-7 and A-J1-6 performed in a similar trend as the sample
from Solmetex but with lower potential for arsenic removal.

Groundwaters

Isotherms were conducted in groundwater collected from Scottsdale, AZ 4E well site to

test the effectiveness of the Fe-GAC composites when competing with other constituents
found in natural waters. Scottsdale groundwater had an arsenic concentration of
approximately 33 µg/L, TDS of 800 mg/L (conductivity ~1.18 mS) and silica
concentrations of around 25 ppm. The pH of the water was lowered from 8.7 to 7.8±0.2.
Media doses varied between 25-500 mg dry/L 100x140 mesh sizes. Contact time for
these experiments was 7 days.

Arsenic adsorption was significantly lower than expected from the isotherm with C-II4B-
3 in nanopure water as shown in Figure 3.13, suggesting that competing ions present in
the water affected the performance of the Fe-GAC composite. The difference in pH (7.0
and 7.8) may also affect the capacity of the media for arsenic removal. Although these
experiments were conducted at different initial As(V) concentrations, had no ion
competition occurred the percentage removals should not have differed by the observed
amount. Likewise, the slight pH difference between the tests should not have affected
As(V) removal by the observed amount.

Figure 3.14 shows isotherms for As(V) on Solmetex samples in two dilutions of a local
groundwater which has a TDS of ~800 mg/L (conductance = 1.18 mS). For these batch
experiments S-1-7 and S-2-4 were combined in order to yield enough material for
subsequent RSSCTs. The local groundwater was blended with DI water to evaluate the
effect of naturally occurring competitive ions on As(V) removal. As(V) was maintained

26
at a similar level in the experiments (~28 ppb) by spiking sodium arsenate. The adsorbent
appears quite sensitive to the presence of naturally occurring ions, with lower adsorption
capacity in the presence of a higher content of groundwater and its competing ions.

pH Effects

pH effects on performance of Fe-GAC composites for arsenate removal were tested with
isotherm experiments. Model solutions (5 mM NaHCO3; 100 µg/L As(V)) at pH of 5 to 9
were used. Media doses varied between 10-250 mg dry/L using 100x140 mesh sizes for a
contact time period of 8 days (7-days would have been used, but it feel on a holiday). Fe-
GAC composites from Clemson University were selected for this particular experiment
due to performance in past adsorption experiments.

Figure 3.15 shows similar trends of adsorption efficiency of Clemson University

composites as observed in past isotherm experiments (Figure 3.8). C-II4B-3 has slight
better adsorption capacity for arsenate at pH of 5. The effect of pH on As(V) adsorption
is observed clearly in Figure 3.16 in which the adsorbent performs better at pH levels of
5.0 and 6.5. Arsenic removal decreased at increasing pH values.

Silicate (Si) and Vanadate (V) Effects

To quantify the effect of pH and competing ions (hydroxide, silica, vanadium) batch
experiments were conducted with C-II4-11. Effects of pH and silica (270 g/L NaOHSiO2
Aldrich solution) were evaluated using a modified NSFI-53 challenge water matrix
containing different ion concentrations (Table 3.4 in Experimental Methods). The results
of adsorption tests using C-II4-11 are presented in Figure 3.17. As(V) adsorption was
greater at pH 6.5 compared to pH 8.5. At pH 6.5 the addition of 20 ppm SiO2 decreased
As(V) adsorption by 25%, and by 80% at pH 8.5.

Effects of pH and vanadium were similarly evaluated using the same NSFI-53 challenge
water matrix as above, but containing Na3VO4; no silica was added during vanadium
experiments. Figure 3.18 demonstrates that 100 µg/L of vanadium decreases As(V)
adsorption by less than 10%. If higher vanadium concentrations were used a greater
competition may have been observed, but it was decided to evaluate a As(V) to V ratio
that is common in western USA groundwaters.

CO-ADSORPTION OF ORGANIC CHEMICALS AND ARSENIC

Surface water from Rooselvelt Lake, AZ was used for the performance testing of Fe-
GAC composites performance on As(V) removal when organics are present in the
solution. C-II4B-3 and virgin Norit HD3000 (100x140 mesh size) were used for this
batch isotherm experiment. Natural water was spiked with 100µg/L As(V) and pH
adjusted to 7.00 with HNO3. Results show that C-II4B-3 efficiently removed As(V) while
no arsenic adsorption capacity was observed for virgin media in the surface water (Figure
3.19). Compared with previous experiments in model solutions the adsorption capacity of
C-II4B-3 was 15 to 20 % lower in the surface water, again due to competing ions.

27
Comparing Rooselvelt Lake water to Scottsdale 4E well site groundwater, C-II4B-3 had
less capacity (~20 %) for arsenic removal when tested in the groundwater (Figure 3.20).
Figure 3.21 shows that both media removed similar amounts of dissolved organic carbon
(DOC).

Several Fe-GAC composites were evaluated with Scottsdale groundwater spiked with
arsenate, MTBE, and benzene. The results are illustrated in Figure 3.22 and Figure 3.23.
Both Fe-GAC composites removed arsenate and the synthetic organic chemicals (SOCs).
Because of its higher octanol-water partition coefficient, benzene adsorption capacity was
significantly greater than MTBE adsorption capacity.

A-3-7 had a higher capacity for As(V) than C-II4B-11 (Figure 3.23). A-3-7 from ASU
contains ~7% iron by weight. Both Clemson University and ASU materials were
modified from the same base carbon material (Norit HD3000) and as a result had similar
adsorption of the SOCs. This work shows that modifying GAC with iron does not impair
the ability of the composite to continue to remove SOCs, while also removing arsenic.

SUMMARY

• Kinetic results showed that samples C-2, C-II4-11, C-I8-4, C-III10-9 and C-IV11-
7 reached pseudo-equilibrium at approximately 7 days of contact time.
• GFH which is a commercial pure iron (hydr) oxide media was tested in parallel
with the Fe-GACs and had higher adsorption capacity, but the Fe-GAC appeared
to have faster adsorption kinetics.
• Screening test results for arsenate show that C-II4-11 and C-II9-8 had higher
adsorption capacity (80% As(V)removal) than the rest of the Fe-GACs.
• In arsenite-based screening tests C-IV13B-6 had slightly better adsorption
capacity than C-II4-11.
• Screening tests using S-1-7 and S-2-4 media had higher adsorption capacity for
arsenic (90% As(V) removal) than L-294-1 and virgin Norit GAC samples.
• Initial batch isotherms on a model solution shows that C-II4-11 and C-II9-8
performed better than C-IV13B-6 and C-II14-2. This data is consistent with the
pattern observed from screening tests.
• S-1-7 and C-II4-11 were tested in parallel with ArsenXnp (a Solmetex Inc. hybrid
ion exchange product with a 20% dry weight iron content). ArsenXnp had much
greater adsorption capacity than the other two. S-1-7 had higher adsorption
capacity than C-II4-11.
• Two Fe-GACs from ASU along with S-4-11 were used for batch isotherms. S-4-
11 (K value of 0.29) has slightly higher arsenic adsorption capacity. A-3-7 (K
value of 0.18) has higher adsorption capacity than A-J1-6 (K= 0.13).
• An adsorption test in Scottsdale groundwater with C-II4-11 shows that this
media’s performance for arsenic removal is significantly reduced when
competing with other ions. pH variations between 7.0 to 7.8 also affected the
adsorption capacity of the media.

28
• Experiments at different pHs show similar trends of adsorption efficiency of
Clemson University composites as observed in past isotherms. C-II4-11 has
slightly better adsorption capacity for arsenate at pH of 5 than the rest of Fe-
GACs. Arsenic removal decreased at increasing pH values.
• For evaluation of vanadate and silicate influence, C-II4-11 was tested. As(V)
adsorption was greater at pH 6.5 compared to pH 8.5 with no silica. At pH 6.5 the
addition of 20 ppm SiO2 decreased As(V) adsorption by 25%. At pH 8.5 As(V)
adsorption was diminished by 80%.
• Vanadate concentrations of 100 µg/L decreased As(V) adsorption by less than
10%.
• When testing in a surface water (Roosevelt Lake, AZ), C-II4B-3 removed As(V),
while no As(V) removal was observed for virgin GAC-media. However, the
arsenic adsorption capacity by C-II4B-3 was 15 to 20 % lower in the surface
water than in model solutions, probably due to competing ions. K values were
0.13, 0.09 and 0.006(µg/mg)/(µg/L)1/N for C-II4B-3 when testing in nanopure,
surface and groundwaters, respectively.
• C-II4B-3 and virgin media removed DOC similarly in surface water.
• Batch adsorption tests with arsenate, MTBE and benzene in Scottsdale, AZ
groundwater (33%) demonstrated that both A-3-7 and C-II4B-3 remove arsenate
and SOCs. Benzene had more affinity to both Fe-GACs than MTBE.
• A-3-7 had a higher capacity for As(V) removal than C-II4B-3. This work shows
that modifying GAC with iron does not impair the ability of the composite to
continue to remove SOCs, while also removing arsenic by the added iron.

29
 140
As (V) solution ~ 100 to 120 ppb C-2
pH= 7.00 (5 mM NaHCO 3) C-II4-11
120
Fe-GAC dose = 0.25 g Dry /L C-I8-4
Arsenic Remaining (µg/L)

100 C-III10-9
C-IV11-7
80

60

40

20

0
0 5 10 15 20 25
Time (Days)

Figure 3.1 Kinetic tests using different Fe-GAC composites. Each composite
represents a different approach for production

30
 120

100 As(V) solution~ 110 ppb

GFH 100x140
pH = 7.00 (5mM NaHCO 3
A rsenic Remaining (μg/L)

Media dose = 0.063 g Dry /L

80

60

40

20

0
0 5 10 15 20 25
Time (Days)

Figure 3.2 Kinetic tests with granular ferric hydroxide (GFH) at lower dosage than
Fe-GAC tests
100
90 Nanopure Sy stem
A rsenic Remaining (μg/L)

80 As (V) ~ 100 μg/L

Fe-GAC Dose = 0.25g-Dry /L
70 pH= 7.0 (5mM NaHCO3)
60
50
40
30
20
10
0
-1

-2

C -7
C '

l
C 6

C 6-8

C 1-7

C -16

C 9

C "
C -11

C -8

-4

-4
C -2

-4

tro
-4
0-
-

C
-I8

-I3
13

9
14

-I8
-I8

on
-I V
-II

1
I5

I1
4

-IV
-IV
-II

-II

C
-II

-II
C

Figure 3.3 Batch Screening of Clemson University Fe-GAC samples (Oct-04).

Control sample contained no Fe-GAC adsorbent

31
 120
Nanopure Sy stem
100 As(V) ~ 100 μg/L
Arsenate Remaining (μ g/L)

Fe-GAC Dose = 0.25g dry /L

pH = 7.00(5mM NaHC O 3)
80 C ontact Time = 7 Day s

60

40

20

0
C-II4-11 C-II9-8 C-17-3 C-IV13-6 C-16-3 Control

Figure 3.4 Batch screening of Fe-GAC samples from Clemson University (Jan-05).
Sample 16A was tested in triplicate along with controls

120
A rsenite Remaining (μg/L)

100 Nanopure Sy stem

As(III) ~ 100 μg/L
Media Dose = 0.25 g dry /L
80 pH = 7.00 (5 mM NaHCO3)
Contact Time = 7 day s
60

40

20

0
C-IV13-6 C-II4-11 C-17-3 C-II4B-3 C-16-3 C-III10-9 Control 1 Control2

Figure 3.5 Screening test for As(III) removal using Fe-GAC samples from Clemson
University (July-05). Control samples were tested in duplicate with no adsorbent

32
 100
A rsenic Remaining (μg/L)
90 Nanopure Sy stem
80 As(V) ! 100ppb
70 pH = 7.01(5 mMNaHCO3)
60 Media Size: 100x140
50 Contact Time = 8 Day s
40
30
20
10
0
)

up
)

nt)
g/L

g/L
g/L

g/L

g/L

g/L

g/L

g/L

-D
rbe
0m

0m

0m

0m

0m

0m
0m

0m

ol
so
10

25
(10

(25

(10

(25

(10

(25

ntr
ad
-1(

-1(

Co
gin

gin

1-7

1-7

2-4

2-4

no
94

94
Vir

Vir

S-

S-

S-

S-

ol(
L-2

L-2

ntr
Co
Figure 3.6 Batch adsorption test with virgin HD3000, LLNL aerogel 294 and two
Solmetex Fe-GAC composites in Nanopure water system. Values in parenthesis are
adsorbent dosages.

33
 100
90 CII9-8
As(V)~ 100 µg/L
80 pH = 7.00 (5 mM NaHCO 3) CIV13-6
A rsenic Remaining (µg/L)

70 Media size: 100x140 mesh CII14-2

Contact Time = 7 day s
60 CII4-11
50
40
30
20
10
0
0 100 200 300 400 500 600
Fe-GA C Dose (mg dry/L)

Figure 3.7 Adsorption test with different Fe-GAC materials from Clemson
University. Jan-05

10
C-II4-11 q = 0.13Ce 0.54 ;R2= 0.99
As(V)~ 100 µg/L
pH = 7.00 (5 mM NaHCO3) C-II9-8 q = 0.09Ce 0.55 ;R2 =0.94
Media size: 100x140 q =0.06Ce 0.59 ;R2 =0.89
Contact Time : 7 days C-I13-6
q = 0.02 Ce0.7 ;R2 = 0.92
q(µgAs/mgdry adsorbent)

C-II14-2
1

0.1

0.01
1 10 100
Ce(µgAs/L)

Figure 3.8 Log form of Freundlich adsorption isotherm with different Fe-GAC
materials from Clemson University. Jan-05

34
 10
q (µgA s/mgdry adsorbent)

0.1
As(V) ~ 100 ppb ArsenXND q = 3.2Ce0.28; R2 =0.91
pH = 7.01 (5 mM NaHCO 3)
S-1-7 q = 0.22Ce 0.56; R2 =0.93
Contact Time = 7 Day s
100x140 mesh C-II4-11 q =0.13Ce 0.54; R2 = 0.99

0.01
1 10 100

Ce (µgAs/L)

Figure 3.9 Isotherm test in model solution comparing different Fe-GAC media from
Solmetex and Clemson University. March 2005

35
 80

70 As (V) ~ 100 ppb S-2B-11

Arsenic Remaining (μg/L)

pH= 7.0 (5mM NaHCO 3)

S-4-11
60 100x140 mesh
C-II4-11
Contact time = 6 day s
50

40

30

20

10

0
0 100 200 300 400 500 600
Fe-GAC Dose (mg dry/L)

Figure 3.10 Adsorption test in model solution for arsenic removal using Fe-GAC
media from Solmetex and Clemson University

10
As (V) ~ 100 ppb
pH= 7.0 (5mM NaHCO 3)
100x140 mesh
q (μg As/mg adosrbent)

Contact time = 6 day s

1

0.1
0.59
S-4-11 q =0.29Ce ; R2 =0.93

C -II4-11 q =0.13Ce
0.54
; R2 =0.99

0.01
1 10 100
Ce (ug/L)

Figure 3.11 Isotherm test comparing best performance Fe-GAC composites for
arsenate removal in nanopure water from Clemson University and Solmetex. July
2005

36
 10
Nanopure System
As(V) Solution = 95ug/L
pH= 7.00 (5 mM NaHCO3)
q (µgAs/ mg adsorbent)

Contact Time = 8 days

1

0.1
A-J1-6 q =0.14 Ce0.49
A-3-7 q =0.18 Ce 0.64
S-4-11 0.59
q =0.29 Ce
0.01
1 10 100
Ce (µg/L)

Figure 3.12 Log version of isotherm in a model solution using ASU produced
composites and S-4-11 from Solmetex. Media was a 100x140 mesh size.

120
Scottsdale Groundwater
As(V) ~ 33 ppb; pH = 7.80
100
Nanopure water
As(V) ~ 100 ppb; pH = 7.00
% A s Removal

80

60

40 Media size = 100x140 mesh

Contact Time = 7 day s
Groundwater (pH = 7.8)
20
Nanopure (pH= 7.00)

0
0 100 200 300 400 500 600
Fe-GAC Doses (mg Dry/L)
Figure 3.13 Adsorption test in Scottsdale groundwater with C-II4B-3 from Clemson
University.

37
 10
Scottsdale Groundwater Water 1(33%gw + 67%D.I.)
As(V)-33% = 28 µg/L
Water 2(67%gw+ 33% D.I.)
As(V)-67% = 25 µg/L
q (µg A s/mg adsorbent)

pH = 7.01
1

0.1
W1 : q = 0.15Ce0.69 ; R2 = 0.99
W2: q = 0.01Ce1.78 ; R2 = 0.93

0.01
0 1 Ce (µg/L) 10 100

Figure 3.14 Effect of natural water matrix composition on competition and removal
of As(V) using 33% or 67% Scottsdale Site 4E groundwater blended with DI water.
As(V) spiked to nearly equivalent concentrations to that of the ambient
groundwater (23-28 ppb). S-1-7 and S-2-4 were combined in these tests. 100x140
mesh material was used.

10
Nanopure Sy stem
As(V) = 100 ug/L
pH = 5.00 (5mM NaHCO3)
q (µg A s/ mg adsorbent)

Media size: 100x140

1 Contact Time = 8 Day s

0.1
C-II4-11 q= 0.12Ce 0.7 ; R2 = 0.94
C-II9-8 q= 0.22Ce 0.52; R2 = 0.97
C-IV13-6 q= 0.002Ce1.4 ; R2 = 0.62
C-II14-2 q= 0.06Ce 0.69; R2 = 0.73
0.01
1 10 100
Ce (µg/L)
Figure 3.15 Effect of pH on arsenate adsorption by Fe-GAC media from Clemson
University

38
 100
Nanopure Water
90
As(V)~ 100 ppb
80
Arsenic Remaining (µg/L)

Media: GAC4(100x140 mesh)

C ontact Time : 8 Day s
70
60
50
40
30 pH = 5.0
pH = 6.5
20
pH = 7.5
10 pH = 9.0
0
0 50 100 150 200 250 300
Fe-GAC Doses (mg dry/L)

1.0
Nanopure Water pH = 5.0
0.8
As(V) = 100 ppb q = K ce 1/N
pH = 6.5
0.6 pH = 5.00 (5mM NaHCO 3)
pH = 7.5
Log q(µgA s/mgGA C)

0.4 Media size: 100x140

pH = 9.0
Contact Time = 8 Day s
0.2
0.7
0.0 (5.0): q= 0.12Ce
0.99
R2 = 0.94 (7.5)q = 0.007Ce
-0.2 R2 = 0.56
-0.4
-0.6 (6.5)q = 0.23Ce
0.43
3.3
(9.0)q= 1E-07Ce
-0.8 R2 = 0.997 R2 = 0.87
-1.0
0 0.5 1 1.5 2 2.5
Log Ce(µgA s/L)

Figure 3.16 Comparison of arsenic removal as a function of pH with C-II4B-3

(similar trends observed for samples #9, 13, and 14) in nanopure water

39
 10
0.36
W1(pH=6.5,SiO2=0) q = 0.26 Ce ; R2 = 0.96
0.49
W2(pH=6.5,SiO2=20) q = 0.08 Ce ; R2 = 0.94
0.56
W3(pH=8.5,SiO2=0) q = 0.07 Ce ; R2 = 0.98
q(µgA s/mg adsorbent)

-0.0013
q = 1.49 Ce ; R2 = 0.72
1 W4(pH=8.5,SiO2=20)

0.1

As(V) Solution = 100 ppb

Adsorbent = GAC4
Contact Time = 7 Day s

0.01
1 10 100
Ce (µgAs/L)

Figure 3.17 Effect of pH and silica on As(V) removal using a base NSFI challenge
water matrix (contact time = 7 days)

40
 90

80
As(V) Solution = 100 µg/L W1( pH=8.5 , V=100 ppb)
70 Media: C-II4-11(100x140 mesh) W2 (pH=6.5, V=100 ppb)
Arsenic Remaining (µg/L)

Contact Time = 7 Days

60 W3(pH=8.5, V=0 ppb)
W4(pH=6.5, V=0 ppb)
50

40

30

20

10

0
0 100 200 300 400 500 600
Fe-GAC Doses (mg dry/L)

Figure 3.18 Effect of pH and vanadium on As(V) removal using a base NSFI
challenge water matrix (contact time = 7 days)

100
90
80
A rsenic Remaining (µg/L)

70 Surface Water from Rooselvelt Lake C-II4B-3

As(V) spiked to ~ 100ppb
60 C-1
pH = 7.00 ( lowered with HNO3)
50
40
30
20
10
0
0 100 200 300 400 500 600
Fe-GAC Doses (mg dry/L)

Figure 3.19 Isotherm test in Rooselvelt Lake water. Media mesh size was 100x140

41
 10
Scottsdale Groudwater Groundwater q = 0.006C e 0.95; R2 = 0.95
As(V) = 110.4 ug/L
Nanopure q: 0.13 C e0.54; R2= 0.99
pH = 6.94 (5 mM NaHC O 3)
Surface Water q = 0.09 C e0.33; R2 = 0.80
Media: C -II4B-3
q (µg A s/mg adsorbent)

0.1

0.01
1 10 100
Ce (µg/L)

Figure 3.20 Comparison of media performance for arsenate removal in different

water systems. All three cases are from different batch experiments at pH of 7

7
Rooselvelt Lake water C-II4B-3
6 As(V) spiked - 100 ppb
Virgin Norit
DOC Remaining (mg/L)

DOC of Raw Water - 6.00 mg/L

5 pH= 7.00 (HNO3)

0
0 100 200 300 400 500 600
Media Doses (mg dry/L)

Figure 3.21 Comparison of C-II4B-3 and virgin GAC for DOC removal in surface
water

42
 4.5
A-3-7-MTBE
4.0
A-3-7-Benzene
Organics Remaining (µg/L)

3.5 C-II4B-3-MTBE
3.0 C-II4B-3-Benzene

2.5

2.0 Scottsdale groundwater ( 33%)

As(V) Solution = 100 µg/L
1.5 MTBE-Benzene = 5 mg/L
1.0 pH= 7.00 (5 mM NaHCO3)

0.5

0.0
0 100 200 300 400 500 600
Fe-GAC Doses (mg Dry/L)

Figure 3.22 Adsorption of SOCs (MTBE and benzene) from groundwater (blended
at 33% with nanopure water) containing arsenate by several Fe-GAC materials.

43
 10 0.61
As(V) Solution = 100 ug/L A-3-7 q = 0.09 Ce ; R2 = 0.99
MTBE = 5 mg/L C-II4B-3 q = 0.02 Ce0.86; R2 = 0.72
Benzene = 5 mg/L
pH= 7.0 (5 mM NaHCO3)
q (µgAs/ mg adsorbent)

100x140 mesh
1

0.1

0.01
1 10 100
Ce (µg/L)

Figure 3.23 As(V) adsorption in groundwater (33% natural, 67% nanopure) spiked
with SOCs by Fe-GAC materials from Clemson University and ASU (Solmetex
process)

14
Model Solution Clemson
12 As~100ppb Solmetex
pH = 7.0 ASU
10
Mesh size: 100x140
q (ugAs/mgFe)

0
0 2 4 6 8 10 12 14
Iron Content (%)

Figure 3.24 Summary of Fe-GAC performances removing arsenate in model

solutions based on iron content. Adsorption densities for Clemson University,
Solmetex and ASU composites were calculated at final arsenic concentrations of 10
ppb.

44
 CHAPTER 4 ARSENIC REMOVAL IN CONTINUOUS FLOW
COLUMN TESTS
Rapid Small Scale Column Tests (RSSCTs) were conducted using Arizona groundwater
collected from the City of Scottsdale (4E site). This water typifies groundwaters in the
southwestern USA impacted by the arsenic regulation and contains ~25 µg/L of arsenic,
low DOC (<1 mg/L), pH of 8.5, alkalinity ~156 mg CaCO3/L, total dissolved solids
(TDS) of ~800 mg/L, vanadium of 80 µg/L and silica of 30 mg/L. RSSCTs were
conducted in parallel, allowing comparison for different media on a common influent
water source.

FE-GAC MEDIA PERFORMANCE WITH CLEMSON UNIVERSITY FE-GAC

COMPOSITES

Based on prior batch adsorption experiments conducted (Chapter 3), several Fe-GAC
composites from Clemson University were selected to compare performance in a more
dynamic system (continuous flow column tests). RSSCTs were scaled (Table 4.1) to
simulate a pilot scale column having an EBCT of 2.5 minutes and a loading rate
comparable with that employed for GFH or E33 (~8 gpm/ft2). A Reynold Schmidt
product of 1000 was used. 12x40 mesh size full-scale media was used in the scaling
analysis. RSSCT results for four Fe-GACs are presented in Figure 4.1. C-II4-11 and C-
II9-8 exhibited a short lag period of near complete arsenic removal, before arsenic
breakthrough. Arsenic breakthrough was reached after approximately 500 to 1000 BV.

C-II4-11 and C-II9-8 performed better than C-IV13B-6 or C-II14-2. This conclusion was
consistent with the pattern observed from batch experiments (Chapter 3, Fig. 3.4 and 3.8).
An isotherm was also conducted on this same water using C-II4-11 (Figure 6.13). The
arsenic adsorption was significantly lower than would have been expected from the
isotherm test for C-II4-11 in nanopure water (Figure 3.7-3.8), suggesting that competing
ions were present in the water that affected the performance of the Fe-GAC in the
RSSCT. An effluent arsenic concentration of 10µg/L was reached immediately for
samples 13B and #14, and after ~500 bed volumes for samples 4 and 9.

For comparison, GFH or E33 would be expected to treat approximately 20,000 bed
volumes before reaching 10 µg/L on a comparable groundwater. However, the results of
Fe-GAC isotherm adsorption densities were also approximately an order of magnitude
lower than GFH or E33, so the duration of arsenic removal in the RSSCTs is not
unexpected.

EFFECT OF EBCT

Scottsdale 4E groundwater blended to 33% with nanopure water (77% of final volume)
was used for these experiments, with an adjusted pH of 7.0 and spiked with sodium
arsenate to reach ~25 ppb. These RSSCTs with 100x140 mesh size materials were

45
conducted simulating a full scale system with the following operating characteristics
(radius 0.513 mm, loading rate of 4.1 gpm/ft2) with a Re-Sc value of 1000 and using
proportional diffusivity scaling equations as stated before, (Table 4.1). EBCTLC’s from
2.5 to 10 minutes were tested.

Clemson University Fe-GAC Composites

C-II4-11 (100x140 mesh size) was one of the best performing media, so Clemson
University prepared a larger mass of a similar material (C-II4B-3). This material was
used in subsequent tests. For EBCT of 4 and 10 minutes the columns were operated in
parallel but were fed from different tank containers.

Figure 4.2 presents the results of both tests using an EBCT of 4 -min and 10 min. Prior
RSSCT at EBCT of 2.5 was plotted as a reference. The 10 minute EBCT experiment had
a slightly different breakthrough curve than the 4 minute EBCT. Both reached 10 ppb
As(V) in the effluent after approximately 5000 bed volumes. Complete break through
was reached after 15,000 bed volumes.

Solmetex Fe-GAC Composites

RSSCTs were conducted with S-1-7 for the 4 minute EBCT and a mixture of S-1-7 and
S-2-4 as used in one batch isotherm test (Figure 6.14). From past batch adsorption
experiments (Figure 3.6) it appeared that S-1-7 and S-2-4 would have very similar
performance. There was insufficient mass of S-1-7 alone to conduct a RSSCT with a
simulated EBCT of 10 minutes. The influent water contained 33% Scottsdale Site 4E
groundwater plus 67% nanopure water, adjusted to pH 7.0 with NaHCO3 and spiked with
sodium arsenate to ~25 ppb.

Arsenic breakthrough curves as a function of bed volumes (BV) treated are shown in
Figure 4.3. At a 4 minute simulated EBCT approximately 15,000 BV were treated before
reaching 10 ppb, whereas 22,000 BV were treated at 10 minutes before reaching 10 ppb
effluent arsenic. This represents a 46% increase in treated bed volumes with the longer
EBCT. The area above the arsenic breakthrough curves for 4 and 10 minute EBCTs
would be equal for a synthetic organic compound on GAC alone.

COMPARISON OF FE-GAC COMPOSITES

To facilitate direct comparison of different Fe-GAC media (Clemson University,

Solmetex LLNL and ASU) performance, four RSSCTs with an EBCTsc designed to
simulate a full-scale system with EBCTLC of 4 minutes (Table 4.1) were conducted in
parallel from a common water source. Scottsdale groundwater was blended with 33%
GW plus 67% nanopure water and As(V) spiked to 25 µg/L with sodium arsenate. pH of
this water was lowered to 7.00 with CO2.

Figure 4.4 presents the results. The aerogel Fe-GAC composite from LLNL (L-294-1)
and the sample from Clemson University (C-II4B-3) had similar breakthrough curves and

46
exhibited complete breakthrough after ~15,000 bed volumes. Two Fe-GAC composites
produced by ASU using the Solmetex process on Jacobi base material (A-J1-6) and Norit
HD3000 (A-3-7) achieved significantly longer operational run times. A-J1-6 operated for
approximately 65,000 bed volumes before approaching the influent As(V) concentration.
Even after 70,000 bed volumes A-3-7 only had reached 60% of the influent As(V)
concentration. Some of the variations from a smooth breakthrough curve occurred as two
of the columns (A-J1-6 and A-3-7 at 22,000 bed volumes) were temporarily turned off
for a day while new water was collected. Overall, the Fe-GAC composites produced
using the Solmetex process, which yield ~8% to 10% iron content by dry weight, appear
to have the greatest capacity for arsenic removal.

CONCURRENT ARSENIC AND ORGANIC REMOVAL

RSSCTs with As(V) and MTBE (polar compound) and benzene (non-polar compounds)
in solution were conducted to test the potential of Fe-GAC composites (A-3-7 and C-
II4B-3) for removing both SOCs and arsenic in a more dynamic system. Once again
Scottsdale groundwater blended at 33 % with DI water was used for these experiments
adjusting pH of the water to 7.00 with HNO3. As(V) at 25 ppb, MTBE at 2.8 ppm and
benzene at 1.5 ppm concentrations were all spiked into the feed-water. 10 mM NaHCO3
was added and the pH adjusted to 7.0. Over the course of the RSSCT run, SOC
concentrations in the influent tank were observed to decrease slowly by ~40% based
upon daily measurements.

Figure 4.5 presents breakthrough curves for As(V). A-3-7 achieved better arsenic
removal than sample C-II4B-3, treating ~25,000 bed volumes (4000 bed volumes for C-
II4B) before exhibiting a tailing breakthrough curve. Both Fe-GAC composites had
similar GAC base materials (Norit HD3000) and achieved comparable MTBE and
benzene removal (Figure 4.6). C-II4B-3 treated almost 5000 bed volumes before
complete breakthrough for MTBE while A-3-7 treated around 2000 bed volumes. For
benzene, C-II4B-3 reached complete breakthrough after 15,000 bed volumes whereas A-
3-7 broke through at 8000 bed volumes. As expected the more polar compound (MTBE)
broke through much sooner than the more hydrophobic compound (benzene) Figure 4.6.

SUMMARY

An initial RSSCT was conducted with Fe-GAC composites from Clemson University
using Scottsdale groundwater at pH 7.8, arsenate concentration of 33 ppb and EBCTLC of
2.5 minutes. The following observations were made:
• Arsenic breakthrough for C-II4-11 and C-II9-8 was reached after approximately
500 to 1000 bed volumes
• The pattern of media performance was consistent with prior screening tests and
batch isotherm experiments in both a model solution and natural waters.

47
Additional RSSCTs using Scottsdale groundwater blended with nanopure water, pH of
7.0 and arsenate concentration of ~25 ppb were conducted. The following observations
were made:
• When using C-II4B-3 as adsorbent, the 10-minute EBCTLC experiment had a
slightly longer runtime breakthrough curve than the 4-minute EBCTLC. Both
reached 10 ppb As(V) in the effluent after approximately 5000 bed volumes.
Complete break-through was reached after 15,000 bed volumes
• Two Fe-GAC composites (S-1-7 and S-1&2) were used to assess the impact of
EBCTLC (4- and 10-minute) on As(V) removal. EBCTLC of 4-minutes
approximately 15,000 BVs were treated before reaching 10 ppb, whereas 22,000
BV were treated with an EBCTLC of 10-minutes. This represents a 46% longer
runtime with the longer EBCTLC.

Four Fe-GAC composites were evaluated in RSSCTs with EBCTLC = 4-minutes and a
common influent water source. The influent water source was Scottsdale groundwater
was blended to achieve 33% GW and 67% nanopure water before spiking in 25 µg/L
As(V). pH of this water was lowered to 7.00 with CO2. The following was observed:
• L-294-1 and C-II4B-3 had similar breakthrough curves and exhibited complete
breakthrough after ~15,000 bed volumes
• A-J1-6 operated for approximately 65,000 bed volumes before approaching the
influent As(V) concentration. Even after 70,000 bed volumes A-3-7 only had
reached 60% of the influent As(V) concentration.

RSSCTs with Scottsdale groundwater blended at 33% with 67% nanopure water were
conducted for concurrent arsenic and SOCs removal. pH of the water was 7.0 and As(V),
MTBE and benzene were all spiked into the feed-water. The following was observed:
• A-3-7 achieved higher arsenic removal than sample C-II4B-3, treating ~25,000
bed volumes compared to 4000 bed volumes with C-II4B-3, before exhibiting a
tailing breakthrough curve
• C-II4B-3 treated almost 5000 bed volumes before complete breakthrough for
MTBE while A-3-7 treated around 2000 bed volumes before complete
breakthrough
• For benzene C-II4B-3 reached complete breakthrough after 15,000 bed volumes
whereas A-3-7 reached complete breakthrough at 8000 bed volumes.

48
 Table 4.1 RSSCT design parameters for EBCT of 2.5, 4 and 10 minutes
Simulated Simulated Simulated
RSSCT RSSCT RSSCT
Design Parameters Pilot Pilot Pilot
(PD) (PD) (PD)
Scale Scale Scale
Input Parameters EBCTLC = 2.5 min EBCTLC = 4 min EBCTLC = 10 min
Particle radius (cm) 0.0513 0.0064 0.0513 0.0064 0.0513 0.0064
Column diameter (cm) 30.48 1.10 30.48 1.10 30.48 1.10
Column area (cm2) 729.29 0.9499 729.29 0.9499 729.29 0.9499
Wet density (g/mL) 1.79 1.40 1.79 1.40 1.79 1.40
EBCT (min) 2.50 0.31 4.00 0.50 10.00 1.25
Loading rate (m/h) 20.21 7.03 10.11 7.03 10.11 7.03
Loading rate (cm/min) 33.69 11.72 16.85 11.72 16.85 11.72
BV to be processed 30000 30000 30000 30000 30000 32000
Flow rate (gpm) 6.50 3.25 3.25
Flow rate (gpm/ft2) 8.28 4.14 4.14
Re*Sc (200,000-200) 1000 1000 1000
U (Pa s) 8.94E-04 0.00 8.94E-04 0.00 8.94E-04 0.00
P (kg/m3) 1000 1000 1000 1000 1000 1000
DL (Liquid diffusivity) (m2/s) 1E-09 0.00 1E-09 0.00 1E-09 0.00
d (m) 0.00103 0.00 0.0010 0.00 0.0010 0.00
v (m/s) 0.00562 0.00 0.0028 0.00 0.0028 0.00
e (porosity) 0.25 0.25 0.25 0.25 0.25 0.25
Re 25.8 1.12 1.30 1.12 1.3 1.12
Sc 893.7 893.7 893.7 893.7 893.7 893.7
Re*Sc 23044.2 1000 11522 1000 11522.1 1000
Design Data
Flow rate (mL/min) 24570 11.13 12285 11.13 12285 11.13
Volume of media (mL) 61425 3.5 49140 5.6 122850 13.9
Length of the column (cm) 85.34 3.7 85.34 5.8 85.34 14.6
Mass of the media (gm) 109950.7 4.9 87960.6 7.8 219901.5 19.4
Water required (L) 1842750 104.2 1474200 166.6 3685500 444.4
Water required (gals) 487500 27.6 390000 44.1 975000 117.6
Duration of test (days) 52.08 6.5 83.3333 10.4 208.333 27.7

49
 40
Scottsdale Groundwater
35
As(V) ~ 33 ppb
30 pH = 7.83
Effluent A s(V) (µg/L)

EBCT LC = 2.5 min

25

20
C-II4-11
15 C-II9-8
10 C-IV13B-6
C-II14-2
5

0
1 10 100 1000 10000 100000
Bed Volume

Figure 4.1 RSSCTs with Fe-GACs from Clemson University in Scottsdale

groundwater

35

30

25
Effluent As(V) (µg/L)

20 Scottsdale Groundwater (33%)

As(V) = 25.5 ppb
15 pH = 7.01 (C O 2)
100x140 mesh
10
EBCT= 4min(June-05)
5 EBCT= 10min(June-05)
EBCT= 2.5 min(Jan-05)

0
0 5000 10000 15000 20000 25000
Bed Volume

Figure 4.2 Effect of EBCTLC on As(V) breakthrough with C-II4B-3 ( 100%

groundwater, pH= 7.8 and 33 µg/L of arsenate).

50
 30
Scottsdale Groundwater (33% GW + 67% nanopure water)
25 As(V) -33% = 25 µg/L
Media: 100x140 mesh
Effluent As(V) (µ g/L)

pH = 7.00
20

15

10

S-1+ 2
5
S-1-7

0
0 5000 10000 15000 20000 25000 30000 35000 40000

Bed Volume

Figure 4.3 Arsenic breakthrough curves for RSSCTs using S-1-7 (EBCTLC= 4 min)
and a blend (S-1&2) of S-1-7 and S-2-4 (EBCTLC = 10 minute) (100x140 mesh)

51
 30

25

20
Effluent As(V) (µg/L)

C-II4B-3
15 A-3-7
A-J1-6

10 L-294-1

Scottsdale groundwater diluted at 33%

As(V) Influent: 25 to 30 ppb
5 pH = 7.00 ( CO 2)
EBC T LC = 4 min

0
0 10000 20000 30000 40000 50000 60000 70000 80000 90000
Bed Volume

Figure 4.4 Arsenic breakthroughs by four Fe-GAC composites. RSSCTs run in

parallel from common water source. 33% of Scottsdale groundwater plus 67%
nanopure water was used. All media were 100x140 mesh size.

52
 30

25
Effluent A s(V) (µg/L)

20

15

10

C-II4B-3
5
A-3-7
0
0 5000 10000 15000 20000 25000 30000 35000
Bed Volume

Figure 4.5 RSSCTs for As(V) removal in modified Scottsdale groundwater ( 33%
groundwater + 67% nanopure water) spiked with SOC using Fe-GACs from
Clemson University and ASU (EBCTLC=4 min)

53
 3.5
C-II4B-3-MTBE
C-II4B-3-Benzene
Effluent SOC Concentration (mg/L) .

3.0
A-3-7-MTBE
2.5 A-3-7-Benzene

2.0

1.5

1.0

0.5

0.0
0 5000 10000 15000 20000 25000 30000 35000
Bed Volume

Figure 4.6 Concurrent As(V) and SOC removal by two Fe-GAC composites. Initial
MTBE and benzene concentrations were 2.8 mg/L and 1.5 mg/L, respectively.

54
 CHAPTER 5 DISCUSSION OF RESULTS

This chapter presents a discussion of results obtained for arsenic adsorption by Fe-GAC
composites in both batch isotherm experiments and RSSCTs. The chapter is subdivided
into experiments with either model solutions or groundwaters. Finally, the observed test
results are related with available characterization data for the media.

BATCH AND RSSCT ADSORPTION CAPACITIES

Adsorption data were interpreted using the Freundlich model (Faust and Hoigne 1987).
The linearization of the log vs. log data indicated reasonable application of the model to
this work. Evaluation of performance of the composites tested was represented by the
following equation.
qo = K Ce 1/N (5.1)

where qo is arsenic adsorption density on the media (µg/mg), Ce is the equilibrium arsenic
concentration in solution (µg/L), 1/N is the adsorptive intensity constant, and K is the
adsorptive capacity constant (µg/mg)/(µg/L)1/N. Batch adsorption tests were conducted to
evaluate Fe-GAC composites for As(V) removal in model solutions and groundwaters.
The iron content and values for isotherm parameters for the different Fe-GACs tested in
model solutions are summarized in Table 5.1.

Adsorption capacities were also calculated from RSSCTs. Adsorption densities (qo in
µgAs/mg dry-mass of GAC) from column studies were calculated based upon the area
above the As(V) breakthrough curve, volume of the packed bed, and mass of Fe-GAC in
the packed bed (Table 5.2). Based upon the iron content qo values were calculated at Ce
of 10 µg/L and at Ce=Co where Co is As(V) concentration at the influent.

Model Solutions

Preliminary kinetic studies and batch screening tests in model solutions demonstrated the
order of efficiency of different Fe-GAC composites from Clemson University and
Solmetex for arsenic removal. C-II4-11>C-II9-8>C-IV13-6>C-II14-2 removed more
arsenic than other Clemson University media. C-II4-11 and C-II9-8 (both ferrihydrate)
have higher iron content (11% and 8%) than C-IV13-6 and C-II14-2 (6% and 2%)
however the latter have larger surface area and higher pore distribution than C-II4-11 and
C-II9-8 (Table 3.2).

Subsequent isotherm tests with S-1-7 and S-4-11 in parallel with C-4II-11 (Norit HD3000
GAC based) demonstrated better arsenic adsorption capacity for the Solmetex
composites. K values for S-1-7 and S-4-11 are 0.22 and 0.29 (µg/mg)/(µg/L)1/N,
respectively. C-II4-11 had a K value of 0.13 (Table 5.1).

55
Isotherm tests with Fe-GACs prepared by ASU and the best Fe-GAC media from
Solmetex Inc. were conducted for comparison. S-4-11 had a higher adsorption capacity
(K = 0.29(µg/mg)/(µg/L)1/N) than A-3-7 and A-J1-6 (K values of 0.18 and
0.14(µg/mg)/(µg/L)1/N respectively). S-4-11 and A-3-7 were Norit HD3000 GAC based.
A-J1-6 was Jacobia GAC based. These Fe-GAC samples from ASU and Solmetex were
prepared with the same treatment approach for iron deposition. It can be argued that the
difference in arsenic adsorption capacities might rely on both the iron content and
treatment for each composite. S-4-11 contains 10.5% Fe, A-3-7 contains 7.2% Fe and A-
J1-6 contains 6.3% Fe.

Overall results from isotherms in model solutions showed that S-4-11, A-3-7, A-J1-6 and
C-II4-11 had the best arsenic adsorption capacities with 1.13, 0.79, 0.43 and 0.45
µgAs/mg-adsorbent, respectively. Based on these results C-II4-11 was selected to
perform extensive RSSCTs and for this goal a larger volume of this composite was
prepared and provided by Clemson University. The new Fe-GAC was identified as C-
II4B-3. Solmetex sample S-4-11 was not used for RSSCTs (among best media
performance) due to lack of sufficient material for these experiments.

Groundwaters

Effect of natural water matrix composition on competition and removal of As(V) was
evaluated in isotherm tests. Waters with 33% and 67% Scottsdale Site 4E groundwater
blended with DI water were used. A combined mixture of S-1-7 and S-2-4 was used due
to the limited availability of these media. Adsorptive capacities for As(V) were 0.15 and
0.01 (µg/mg)/(µg/L)1/N for 33% and 67% groundwaters respectively. These K values
suppose greater competition of other ion constituents in the less blended groundwater and
less blockage of pores on the GAC material by these ions in the more blended
groundwater.

Initially, RSSCTs were conducted using samples from Clemson University in 100%
groundwater from Scottsdale at EBCTLC of 2.5 minutes. C-II4-11 and C-II9-8 treated 500
bed volumes before reaching 10 ppb effluent arsenic. C-IV13-6 and C-II14-2 broke
through almost immediately. The qo value for C-II4 and C-II9 was 0.009 µgAs/mg dry-
adsorbent at 10 ppb of As(V) in the column effluent confirming better adsorptive
capacity than C-IV13 and C-II14 (qo ~0 in both cases; Table 5.2). In batch experiments
with model solutions the adsorption capacities for these composites were 0.45, 0.32, 0.23
and 0.07 µgAs/mg dry-adsorbent at As(V) Ce of 10 ppb, respectively. These results
confirmed the trend observed for these media in screening and isotherm tests.

Table 5.3 summarizes the number of bed volumes processed in RSSCTs before reaching
effluent As(V) concentrations of 10 or 20 μg/L. The number of bed volumes processed
ranges from < 5 to nearly 75,000. Water quality and technique of iron impregnation
affects this performance of the Fe-GAC materials.

56
The effect of water chemistry on As(V) removal by Fe-GACs was indirectly evaluated by
RSSCTs. One set of RSSCTs were conducted with 100% Scottsdale groundwater at
pH=7.8 using four Fe-GAC samples from Clemson University. One of the Clemson
University media formulation (C-II4-11 & C-II4B-3)) was also evaluated in an RSSCT
using blended Scottsdale groundwater (33% groundwater plus 67% nanopure water; pH
of 7.0 and EBCTLC of 4 min). The higher value of qo on the second study (~0.07 vs.
0.009 µgAs/mg-adsorbent) suggests that reducing the ion competition (silica, vanadium,
phosphate, manganese, etc.) in the influent water plus reducing the pH and increasing the
bed depth of the adsorbent (4min-EBCTLC) increased the potential of the solute to be
adsorbed to the Fe-GAC material. Other competing ions such as bicarbonate and
hydroxide ions may also compete with arsenate for surface adsorption sites reducing
surface diffusion mass transport.

Effect of EBCT on As(V) removal by Fe-GACs was also evaluated. C-II4B-3 from
Clemson University and S-1-7 and S-(1+2) from Solmetex were used for column studies
at different EBCTLC (4 and 10 minutes) in 33% of Scottsdale groundwater. Significant
design parameters of these experiments are indicated in Chapter 4, Table 4.4.

C-II4B-3 had qo values of 0.065 and 0.089 µgAs/mg dry GAC for EBCTLC of 4 and 10
minutes. S-1-7 had qo values of 0.23 and 0.32 µgAs/mg dry GAC for EBCTLC’s of 4 and
10 minutes. For both samples the adsorptive capacity for As(V) increases approximately
50% with increasing bed depth (Table 5.2). Since there is a benefit in the longer EBCT, a
surface reaction may be rate limiting. Compared with iron (hydr)oxide commercial
products (E33 or GFH) that have 100 to 250 m2/g of surface area and 50% iron by dry
weight, the Fe-GAC materials with >700 m2/g appear to require longer EBCTs
approaching times conventionally used for groundwater treatment (10 to 20 minutes).

Isotherm tests and RSSCTs were conducted to analyze co-adsorption of organics and
arsenic. Initially C-II4B-3 and virgin Norit HD3000 GAC were tested in isotherm tests in
surface water (Roosevelt Lake). As expected, C-II4B has higher As(V) adsorption
capacity than the virgin GAC (Faust, S. D. and O. M. Aly 1987), but its efficiency was
reduced by the presence of other constituents present in the water (natural organic
matter). However, this test demonstrated that C-II4B and the virgin GAC have
comparable adsorption capacities for DOC. These results indicate that Fe-GACs retain
As(V) adsorption capacities when removing organics from water as well. Comparing C-
II4B-3 capacity for As(V) removal in different water systems, results show that this
media had better performance in Lake Roosevelt than in Scottsdale groundwater with qo
values of 0.19 and 0.06 µgAs/mg-adsorbent, respectively. These results suggest that ionic
composition affects As(V) adsorption onto Fe-GAC; the groundwater has higher TDS
than Lake Roosevelt.

Secondly, SOCs (MTBE and benzene) and arsenic removal was evaluated in isotherm
tests conducted in parallel to RSSCTs. Isotherms show that A-3-7 and C-II4B-3 can
remove As(V), MTBE and benzene from groundwater at the same time. Benzene though
had more affinity than MTBE to be adsorbed to the Fe-GACs as observed in other

57
investigations. Yet SOCs might reduce Fe-GACs adsorption capacities for As(V).
Results showed that A-3-7 and C-II4B-3 had K values of 0.09 and 0.02(µg/mg)/(µg/L)1/N,
respectively for As(V). In model solutions these composites had K values of 0.18 and
0.13(µg/mg)/(µg/L)1/N without spiked SOCs. RSSCTs results confirmed those of
isotherm tests. A-3-7 and C-II4B-3 demonstrated capacity for As(V) removal from water
containing SOCs keeping the same order of efficiency as in isotherm tests. However
presence of MTBE-benzene chemicals reduced arsenic adsorption of both samples
approximately 31% and 43% (at 10 ppb of arsenate in the effluent) than when testing in
the absence of these organics. This can be observed in the volume of water treated
(Figures 4.4 and 4.5). In batch tests qo values for A-3-7 and C-II4B-3 were 0.37 and 0.14
while in RSSCT these values were 0.35 and 0.05 µg/g at As(V) Ce of 10 ppb. These
differences in adsorption capacities might indicate that SOCs and other chemicals
contained in Scottsdale groundwater occupy considerable sites in the Fe-GAC material
decreasing the rate of adsorption for As(V).

One last set of RSSCTs were conducted in parallel with several Fe-GACs for direct
comparison of performances (Figure 4.4). Scottsdale groundwater at 33% was the water
source and based on the effect of EBCT on mass transfer zone, an EBCTLC of 4 minutes
was selected. A-3-7 and A-J1-6 had better As(V) removal with qo values of 0.625 and
0.459 than C-II4B-3 and L-294-1 with 0.08 and 0.12 µgAs/mg dry-adsorbent at As(V) Ce
of 10 ppb. Batch tests in model solutions indicated qo values of 0.79, 0.43, 0.25 and 0.06,
for these composites, respectively. Iron content for these samples is 7%, 6%, 4% and 1%.
A possible reason for much lower adsorption capacity for C-II4B-3 (besides the
percentage of iron impregnated ~3%) might be problems with iron dispersion or
penetration into the GAC material. RSSCTs and isotherms showed the same trend of
effectiveness for As(V) removal by C-II4B-3 and A-3-7.

Figure 5.1 presents the relationship between iron content and adsorption capacity in
isotherms (model solutions) and RSSCTs (groundwater) for Clemson University,
Solmetex and ASU composites. Iron content of the media is plotted on the x-axis.
Adsorption capacities (y-axis) are expressed as µgAs/mgFe) for an RSSCT effluent
concentration (Ce) of 10 µg/L. Because arsenic adsorption to iron (hydr)oxides is the
presumed mechanism for arsenic removal, it could be hypothesized that arsenic removal
then be independent of iron content (x-axis). This does not appear to be the case.
Therefore, it must be the iron mineral surface area, iron mineralogy, and/or blockage of
pores by deposited iron as a function of depth into the Fe-GAC composite that influence
arsenate adsorption capacity, and not only the iron content.

Figure 5.2 compares q values from batch and RSSCT experiments. The adsorption
capacities (µgAs/mgFe) indicate that the Clemson University media had low values for
both type of experiments. Solmetex media had better performance in batch experiments
than in RSSCTs. Both ASU media had similar adsorption values in both studies with S-7-
3 having the highest removal of arsenic.

58
EFFECTS OF PH AND ION COMPETITION

Evaluation of pH and ion competition influence was tested in isotherms. C-II4B-3 along
with C-II9-8, C-IV13-6 and C-II14-2 were used for evaluation of pH effects in model
solutions. C-II4B-3 was the best adsorbent for arsenate. As(V) adsorption capacities (qo)
were 0.12, 0.23, 0.007 and ~0.0002 (µg/mg)/(µg/L)1/N at pH levels of 5, 6.5, 7.5 and 9.0,
respectively for C-II4B-3 (Chapter 3, Table 3.16).

The pHzpc is the pH at which the adsorbent has a net zero surface charge (at pH < pHzpc,
the surface has a net positive charge; at pH > pHzpc, the surface has a net negative
charge. Aqueous arsenate speciation is controlled by the pH dependent dissociation of
two ionic species (H2AsO4- ↔ HAsO4-2 + H+; pKa = 7). With increasing pH, the net
surface charge of the C-II4B-3 becomes less positive (i.e., decreasing attractive
electrostatic force) until the pHzpc (7.0) is surpassed. At pH > pHzpc, the net surface
charge is negative, producing a repulsive electrostatic force between HAsO4-2 and the
adsorbent. This was demonstrated by the performance of GAC4 in this experiment, where
there was better arsenate removal at lower pHs (5.0 and 6.5). Figure 5.3 presents a plot of
adsorption capacities of the composites at their different pHZPC. The qo values were
obtained at Ce of 10 µg/L of As(V). This graph shows C-II4B-3 with similar capacities
(~18 µgAs/mg-Fe) at low pHs (5.0 and 6.5), and decreasing capacities as the pH
increases to 7.5 and 9.0.

Additional isotherms were used to test the influence of silicate and vanadate on arsenic
removal. Modified NSFI-53 challenge water and C-4IIB-3 was used for this purpose.
Arsenic adsorption by C-II4B-3 was diminished by 25% with silica present at 20 mg/L
and pH=6.5. At pH=8.5 adsorption capacity of C-II4B-3 diminished by 80%. Silica
competes with arsenate for surface adsorption sites in Fe-GAC materials and also
decreases the surface charge, which results in greater electrostatic repulsion of arsenate.
At the low vanadate concentrations tested, vanadate had only a small detrimental impact
on arsenate adsorption in the NSFI-53 challenge water. As(V) adsorption capacities for
the media were 0.22, 0.48, 0.25 and 0.52 µgAs/mg-adsorbent for waters at pH=8.5 and
100ppb of V, pH=6.5 and 100 ppb of V, pH= 8.5 without V , and pH= 6.5 without V,
respectively.

MATERIAL PROPERTIES

Surface analyses by scanning electron microscope (SEM) coupled to an energy-

dispersive X-ray spectrometer (EDX) were conducted on ASU samples. Figures 5.4 and
5.5 show graphs of these analyses. Both A-3-7 and A-J1-6 composites had a homogenous
distribution of iron on the surface area of the GAC material. The X-axis represents a
section of the material (line indicated on the photo) whereas the Y-axis displays energy
counts.

In addition a cross-section analysis by SEM/EDX was conducted on A-J1-6. Figure 5.6

shows that most of the iron is deposited near the outer portions of the composite. Thus, it

59
may be possible to improve arsenic removal through a slower deposition process which
allows the iron to penetrate deeper into the GAC. Samples A-3-7 and S-4-11 had similar
iron deposition treatment. Not enough mass of S-4-11 was available to perform cross
section analysis this composite for comparison of iron distribution to the ASU samples.

60
 Table 5.1 Isotherm parameters for different Fe-GAC composites in model waters,
blended groundwater and surface water.

qo at Ce 10ppb qo at Ce
Iron
(µgAs/mgGAC) 10ppb
Manufacturer Sample ID Content K 1/N R2
Dry basis (µgAs/mgFe)
%
Dry basis

Clemson University C-II4-11 11.63 0.13 0.54 0.99 0.45 3.9

C-II4B-3 3.4 0.07 0.55 0.99 0.25 7.3
C-II9-8 7.9 0.09 0.55 0.94 0.32 4.1
C-IV13-6 5.8 0.06 0.59 0.89 0.23 4.0
C-II14-2 2.1 0.01 0.87 0.76 0.07 3.5
C-II4B (GW) 3.4 0.0064 0.95 0.95 0.057 1.7
C-II4B (Surface) 3.4 0.09 0.33 0.8 0.19 5.7
Solmetex S-1-7 6.8 0.22 0.56 0.93 0.80 11.8
S-2-4 3.8 0.18 - - - -
S-2B-11 11.2 0.032 1.19 0.92 0.50 4.4
S-4-11 10.5 0.29 0.59 0.92 1.13 10.7
ASU A-3-7 7.2 0.18 0.64 0.94 0.79 10.9
A-J1-6 6.3 0.14 0.49 0.9 0.43 6.9
LLNL L-294-1 1 0.01 0.8 0.96 0.063 6.3

As-SOC (Groundwater)
ASU A-3-7 7.2 0.09 0.61 0.99 0.37 5.13
Clemson University C-II4B-3 3.4 0.02 0.86 0.72 0.14 4.11

61
 Table 5.2 Calculated adsorption capacities for Fe-GAC composites evaluated in
RSSCTs (* indicates experiments with spiked MTBE and benzene)

Initial Complete As(V)

Media ID EBCT pH At Ce = 10 µgAs(V)/L
As(V) Breakthrough
q q q q
(μg/L) (min)
(µgAs/mgGAC) (µgAs/mgFe) (µgAs/mgGAC) (µgAs/mgFe)

C-II4-11 33 2.5 7.83 0.009 0.08 0.039 0.34

C-II4B-3 26 4.0 7.01 0.065 1.91 0.097 2.85
C-II4B-3 26 10.0 7.01 0.089 2.62 0.110 3.24
C-III5-13 25 4.0 7.00 0.024 0.15 0.037 0.23
C-IV6-8 25 4.0 7.00 0.055 0.66 0.057 0.68
C-II9-8 33 2.5 7.83 0.009 0.11 0.021 0.27
C-IV11-7 25 4.0 7.00 0.053 0.71 0.062 0.83
C-1V13-6 33 2.5 7.83 -- -- 0.021 0.36
C-II14-2 33 2.5 7.83 -- -- 0.014 0.67
C-II4B-3 27 4.0 7.01 0.075 2.21 0.106 3.12
A-3 27 4.0 7.01 0.625 8.70 0.955 13.30
A-J1 27 4.0 7.01 0.459 7.29 0.707 11.23
L-294 27 4.0 7.01 0.060 8.00 0.125 16.67
S-1 25 4.0 7.00 0.230 3.40 0.321 4.75
S-1+2 25 10.0 7.01 0.318 4.43 0.375 5.22
C-II4B-3* 26 4.0 6.8-7.1 0.052 1.53 0.112 3.29
A-3* 26 4.0 6.8-7.05 0.352 4.90 0.409 5.70

62
 Table 5.3 Summary of bed volumes treated during RSSCT (aScottsdale
groundwater blended with nanopure water; b spiked with MTBE and benzene)

Water Bed Volumes Treated To

Initial
Media ID EBCT Matrix a pH Reach Effluent As(V)
As(V)
Concentration
(percentage Effluent Effluent
(μg/L) (min) groundwater) As(V) = 10 As(V) = 20
μg/L μg/L
100% 1 20
C-II4-11 33 2.5 7.83
33% 500 750
C-II4B-3 26 2.5 7.01
33% 4,000 7,600
C-II4B-3 26 4.0 7.01
33% 5,000 8,000
C-II4B-3 26 10.0 7.01
33% 750 1,250
C-III5-13 25 4.0 7.00
33% 1,500 2,500
C-IV6-8 25 4.0 7.00
100% 700 900
C-II9-8 33 2.5 7.83
100% 1,500 3,000
C-IV11-7 25 4.0 7.00
100% 10 100
C-1V13-6 33 2.5 7.83
100% 1 20
C-II14-2 33 2.5 7.83
33% 4,000 7,500
C-II4B-3 27 4.0 7.01
33% 39,000 75,000
A-3 27 4.0 7.01
33% 25,000 49,000
A-J1 27 4.0 7.01
33% 4,000 8,500
L-294 27 4.0 7.01
33% 20,000 >35,000
S-1 25 4.0 7.00
33% 15,000 30,000
S-1+2 25 10.0 7.01
33% 3,500 13,000
C-II4B-3 b 26 4.0 6.8-7.1
33% 23,000 >30,000
A-3 b 26 4.0 6.8-7.05

63
 14
Model Solution-33% Groundwater
Clemson-Batch
12 As~ 100ppb
pH = 7.0 Clemson-GW
10 Mesh size: 100x140
Clemson-Surface
q0 (µgA s/mgFe)

8 Clemson-RSSCT
q = K Ce 1/N
Solmetex-Batch
6
Solmetex-RSSCT
4
ASU-Batch
2 ASU-RSSCT

0
0 2 4 6 8 10 12 14
Iron Content (%)

Figure 5.1 Media performance based on iron content for both RSSCT and batch
tests. Scottsdale groundwater (33%) and model solutions were used for both studies
respectively; test conditions were As(V) = 100 µg/L and pH = 7

14
Batch vs. RSSCT
12 S-1-7
A-3-7
10
q Batch, Ce=10 ppb

8
A-J1-6
C-II4B-
6
Model solutions-33%groundwater
4 As (V) = 100 µg/L
Ce = 10 µg/L
pH = 7.0
2 C-II4B-3~GW 100x140 mesh size

0
0 2 4 6 8 10
q RSSCT, Ce=10ppb

Figure 5.2 RSSCT vs. batch adsorption capacities for arsenic removal based on iron
content

64
 20
18 pH = 5.0
C-II4B-3
pH = 6.5
16
pH = 7.5
14
q0 (µgA s/mg Fe)

pH = 9.0
12
C-II14-2
10 5 mM NaHCO 3
As = 100 µg/L
8
Media: 100x140 C-II9-8
6
4
2 C-IV13-6
C-II4B-3
0
0 2 4 6 8 10
pH ZPC

Figure 5.3 Batch adsorption capacities of media at different pH levels plotted vs. the
pH of zero point charge of the composites.

65
 Fe

Figure 5.4 SEM/EDX surface analysis for A-3-7 (Norit HD3000 GAC)

Figure 5.5 SEM/EDX surface analysis for A-J1-6 (Jacobi GAC)

66
 Fe-Profile S-Profile Mn-Profile Cl-Profile

120

100

80
Counts

60

40

20

0
0 200 400 600 800 1000 1200 1400 1600
Distance (um)

Figure 5.6 Elemental profile of cross-section through a Fe-GAC material (Jacobi)

67
68
 CHAPTER 6 RECOMMENDATIONS FOR UTILITIES AND
FUTURE RESEARCH

This project evaluated the “proof-of-concept” for using Fe-GAC composites as

adsorbents for arsenic removal and organic contaminants. Batch adsorption experiments
and column studies were conducted in different water sources to evaluate and identify the
performance of Fe-GACs produced by Arizona State University, Clemson University,
Solmetex Inc. and LLNL.

CONCLUSIONS

• Fe-GAC composite media contained <1% to 15% iron by dry weight. There were
no friability issues during use of the media, nor did iron appear to “flake” off of
the composite media. Thus Fe-GAC media appear suitable for batch or packed-
bed use during water treatment.
• Arsenic removal by Fe-GAC composites depended highly upon the process used
to impregnate iron. Iron-aerogel materials had the lowest iron content and
exhibited the lowest arsenic adsorption capacities. Roughly six other techniques
to impregnate iron into GAC were successful and were the focus of the study. For
these materials iron impregnation did not alter the surface area significantly
(<15% change), as measured by nitrogen deposition. However, arsenic
adsorption varied significantly based upon the technique to impregnate iron.
• Arsenic adsorption was not a function of the iron content (i.e., mgFe/mgGAC).
This implies that the technique to impregnate Fe-GAC composites and the
resulting mineralogy and structure is probably more important than the iron
content.
• The best performing Fe-GAC composite for As(V) from Clemson University (C-
II4 and C-II4B) was produced via an iron hydroxide precipitation process. This
was superior to iron impregnated via ion exchange, organic-ligand based, or zero
valent iron based techniques. The zero valent iron based technique result
achieved comparable or slightly better As(III) removal than other Clemson
University Fe-GAC composites.
• The best overall performing Fe-GAC composite was synthesized by Arizona State
University and Solmetex Inc. using a patented process. Generally, this process
involves in-situ precipitation of iron (hydr)oxide using an oxidant and reduced
iron species.
• Extensive media characterization was undertaken (surface area, pore size
distributions, zeta potential, and various spectroscopy measurements). These
proved valuable in assessing differences between iron deposition techniques.
• Batch kinetic experiments indicated that 100x140 mesh Fe-GAC composites
reached pseudo-equilibrium for As(V) adsorption within 7 days of contact time.
Adsorption isotherms were fit well by the Freundlich Isotherm model (q =
KCe1/N).

69
• In batch experiments with Fe-GAC composites, As(V) adsorption improved as
solution pH decreased. For example, the adsorptive capacity constant (K)
decreased by a factor of ten for each full pH unit increase, between pH 5 and 9.
• In batch experiments with Fe-GAC composites and silicate (20 ppm) at pH 6.5
and 8.5, addition of silicate decreased As(V) adsorption at both pH levels. All Fe-
GAC composites evaluated exhibited similar pH, silicate, vanadate and other
competing ion effects.
• In batch and column experiments with Fe-GAC composites, As(V) adsorption
was lower in groundwaters or NSFI-53 challenge water compared model solutions
prepared from nanopure water at an equivalent pH as the groundwater. Thus, ions
present in these more complex water matrices compete for arsenic adsorption sites
on the Fe-GAC composite. Addition of 100 ppb vanadate to the NSFI-53
challenge water had almost no effect on As(V) removal. Addition of 20 ppm
silicate to the NSFI-53 challenge water decreased As(V) adsorption at both pH
6.5 and 8.5, compared against the NSFI-53 challenge water prepared without
silicate added. Thus silicate appears to be a major foulant/competitor for As(V)
adsorption sites on Fe-GAC.
• In batch and column tests, decreasing the ionic strength of a groundwater through
blending with nanopure water resulted in improved As(V) adsorption. These
experiments maintained the same ratio of ions, but demonstrate that the
concentration of competing ions affect As(V) adsorption.
• Rapid small scale column tests (RSSCTs) were conducted in continuous flow
operation with empty bed contact times (EBCTs) of 2.5 to 10 minutes. Fe-GAC
composites that performed best in batch testing, also had the best performance in
RSSCTs. The ASU and Solmetex Inc. Fe-GAC composites achieved the longest
run length (i.e., bed volumes treated). Lengthening the EBCT resulted in a larger
number of bed volumes treated. The minimum EBCT recommended EBCT is 4
minutes; at 2.5 minute EBCT the mass transfer zone was not captured within the
column.
• RSSCTs operated for 50,000 to over 100,000 bed volumes before complete As(V)
breakthrough.
• RSSCTs were also conducted with a surface water containing dissolved organic
carbon (DOC) and model groundwaters spiked with MTBE and benzene. Both
waters also contained As(V). As(V) was removed effectively, and did not seem
to be impacted by the presence of DOC, MTBE or benzene. Fe-GAC removed
not only the As(V), presumably due to adsorption onto iron (hydr)oxides, but also
removed the DOC and synthetic organic chemicals (MTBE and benzene).
Organics were presumably removed by adsorption onto the GAC. Therefore, iron
impregnation does not appear to hinder organic removal by GAC, but adds the
capability to remove arsenic.
• GAC is a relatively inexpensive substrate to impregnate with iron. For water
treatment situations were co-removal of arsenic and other contaminants suitable
for removal by GAC is desired (e.g., SOCs, radionuclides, DOC) Fe-GAC
composites are a feasible option.

70
RECOMMENDATIONS FOR FUTURE RESEARCH

• Optimization of the selection of base GAC substrate and iron deposition process
needs to be conducted. This should follow the basic technique which resulted in
the best performing Fe-GAC composites (produced by ASU and Solmetex Inc.).
• Batch equilibrium tests proved to be excellent indicators of more labor-intensive
RSSCT performance. Future work should use batch tests to screen new
composites. However, optimization of Fe-GAC composites must also consider
that deeper iron penetration into the GAC base material may impact arsenate mass
transfer. Therefore, RSSCTs should continue to be conducted to assess mass
transport limitations.
• Because Fe-GAC composites are new materials, there is a need to compare
arsenic breakthrough achieved from pilot tests against those achieved by RSSCTs
to validate the scaling approach.
• Improved spectroscopic analysis of Fe-GAC composites is warranted to
understand the nano-structure of the iron deposits (coated surface versus crystals
crossing through the pore space). This could help optimize the iron deposition
process (e.g., duration of iron treatment). Ultimately the goal should be to
maximize iron content for the best iron mineralogy to result in the most arsenic
removal.
• Forensic analysis of spent media would provide additional insight into the role of
foulants and changes in mineralogy over time.
• A detailed economic analysis of iron-oxide impregnated GAC composites should
be conducted. Without full-scale production of these products the economic
analysis could be conducted based upon costs of raw materials, costs for similar
processes applied for other products, and lifecycle cost of the media.

71
72
 REFERENCES

Badruzzaman, M. (2002). Scaling laboratory arsenic removal data for porous sorbents to
the pilot scale. Civil and Environmental Engineering. Tempe, AZ, Arizona State
University: 146.

Badruzzaman, M., P. Westerhoff, et al. (2004). “Intraparticle diffusion and adsorption of

arsenate onto Granular Ferric Hydroxide (GFH).” Water Research 38(18): 4002-
4012.

Bang, S., X. Meng, et al. (2002). “Removal of arsenic from water by zero-valent iron.”
Abstracts of Papers of the American Chemical Society 224: 129-ENVR.

Chen, A. S. C., K. A. Fields, et al. (2002). “Field evaluation of As removal by

conventional plants.” Journal American Water Works Association 94(9): 64-77.

Cheng, R. C., H. C. Wang, et al. (1994). “Enhanced Coagulation For Arsenic Removal.”
Journal American Water Works Association 86(9): 79-90.

Coleman, S. J., P. R. Coronado, et al. (2003). “Granular Activated Carbon Modified with
Hydrophobic Silica Aerogel - Potential Composite Materials for the Removal of
Uranium from Aqueous Solutions.” Environmental Science & Technology 37(10):
2286-2290.

Driehaus, W., M. Jekel, et al. (1998). “Granular ferric hydroxide - a new adsorbent for
the removal of arsenic from natural water.” Journal of Water Services Research
and Technology-AQUA 47(1): 30-35.

Farrell, J., J. P. Wang, et al. (2001). “Electrochemical and spectroscopic study of arsenate
removal from water using zero-valent iron media.” Environmental Science &
Technology 35(10): 2026-2032.

Faust, S. D. and O. M. Aly (1987). Adsorption Processes for Water Treatment.

Stoneham, Butterworth Publishers.

Hem, J. (1992). Study and Interpretation of the chemical characteristics of natural water
(Third edition). Washington, DC, US Geological Society: 263.

Hering, J. G., P. Y. Chen, et al. (1997). “Arsenic removal from drinking water during
coagulation.” Journal of Environmental Engineering-ASCE 123(8): 800-807.

73
Highfield, D. E. (2002). Arsenic occurrence in metro Phoenix groundwater and treatment
by Granular Ferric Hydroxide. Civil and Environmental Eng. Arizona State
University. Tempe, ASU.

Jekel, M. and R. Seith (2000). Comparison of Conventional and New Techniques for the
Removal of Arsenic in a Full Scale Water Treatment Plant. AWWA Inorganics
Conference (Feb. 28-29), Albuquerque, NM.

Lackovic, J. A., N. P. Nikolaidis, et al. (2000). “Inorganic arsenic removal by zero-valent

iron.” Environmental Engineering Science 17(1): 29-39.

Lin, T. F. and J. K. Wu (2001). “Adsorption of arsenite and arsenate within activated

alumina grains: Equilibrium and kinetics.” Water Research 35(8): 2049-2057.

Manning, B. A., M. L. Hunt, et al. (2002). “Arsenic(III) and Arsenic(V) reactions with
zero-valent iron corrosion products.” Environmental Science & Technology
36(24): 5455-5461.

Matis, K. A., A. L. Zouboulis, et al. (1997). “Flotation removal of As(V) onto goethite.”
Environmental Pollution 97: 239-245.

McNeill, L. S. and M. Edwards (1995). “Soluble Arsenic Removal At Water-Treatment

Plants.” Journal American Water Works Association 87(4): 105-113.

McNeill, L. S. and M. Edwards (1997). “Arsenic removal during precipitative softening.”

Journal of Environmental Engineering-ASCE 123(5): 453-460.

Meng, X. G., S. Bang, et al. (2002). “Chemical reactions between arsenic and zero-valent
iron in water.” Abstracts of Papers of the American Chemical Society 224: 052-
ENVR.

Nriagu, J. O. and J. S. Lee (2002). Arsenic carbonate complexes in aqueous systems.

Biogeochemistry of Environmentally Important Trace Elements, American
Chemical Society.

Pierce, M. L. and C. B. Moore (1982). “Adsorption of arsenite and arsenate on

amorphous iron hydroxide.” Water Research 16: 1247-1253.

Pontius, F. W., K. G. Brown, et al. (1994). “Health Implications of Arsenic in Drinking-

Water.” Journal of American Water Works Association 86(9): 52-63.

Raven, K. P., A. Jain, et al. (1998). “Arsenite and arsenate adsorption on ferrihydrite:
Kinetics, equilibrium, and adsorption envelopes.” Environmental Science &
Technology 32(3): 344-349.

74
Reed, B., E., R. Vaughan, et al. (2000). “As(III), As(V), Hg, and Pb removal by Fe-oxide
impregnated activated carbon.” Journal of Environmental Engineering - ASCE
126(9): 869-873.

Reynolds, T. D. and P. A. Richards (1996). Unit Operations and Processes in

Environmental Engineering. Boston, PWS Publishing Company

Rook, J. J. (1976). “Haloforms in drinking water.” Journal of American Water Works

Association 68(3): 168.

Scott, K. N., J. F. Green, et al. (1995). “Arsenic Removal By Coagulation.” Journal

American Water Works Association 87(4): 114-126.

Selvin, N., G. Messham, et al. (2000). The Development of Granular Ferric Media-
Arsenic Removal and Additional Uses in Water Treatment. AWWA Water Quality
Technology Conference, Salt Lake City.

Sinha, S., N. Lee, et al. (2002). Innovative technologies for arsenic removal. WQTC.

Smedley, P. L. and D. G. Kinniburgh (2002). “A review of the source, behaviour and

distribution of arsenic in natural waters.” Applied Geochemistry 17(5): 517-568.

Su, C. M. and R. W. Puls (2001). “Arsenate and arsenite removal by zerovalent iron:
Kinetics, redox transformation, and implications for in situ groundwater
remediation.” Environmental Science & Technology 35(7): 1487-1492.

Wang, L., A. S. C. Chen, et al. (2002). “Field evaluation of As removal by IX and AA.”
Journal American Water Works Association 94(4): 161-173.

75
76
APPENDIX A –ISOTHERMS FOR COMPARISON OF FE-GAC COMPLETED
AFTER COMPOSITE SELECTION FOR RSSCTS

77
After the completion of batch and column tests presented in Chapters 3, 4, and 5 new Fe-
GAC composites continued to be produced in an attempt to optimize their properties to
maximize As(V) removal. Two batches of Fe-GAC composites were produced and
evaluated in batch experiments for As(V) removal.

Figure A-1 summarizes the isotherms obtained for Fe-GAC composites prepared by
Clemson University following a similar approach as employed for C-II4B-3. The results
indicate a slight improvement in As(V) removal occurred for two of the new materials
(C-II4C-3 and C-II4C-4). However, when these two new materials were compared
against materials synthesized using the oxidant/ferrous process employed at ASU and
Solmetex Inc., the sample C-II4C-3 and C-II4C-4 had >50% less As(V) capacity of the
best performing materials (A-3B or S-4) (Figure A-2). A-3B-11 and S-4-11 composites
were prepared with the same approach for iron deposition and contain similar iron
content (11%Fe and 10%Fe, respectively).

10 0.57
q =0.04Ce ; R2 =0.8
A-1 0.73
q =0.05Ce ; R2 = 0.98
A-2 0.42
q =0.17Ce ; R2 =0.98
A-3 0.65
q = 0.13Ce ; R2 =0.98
A-4
q (µgAs/mg adsorbent)

0.55
q =0.07Ce ; R2 =0.99
1 C-II4B-3

0.1
As(V) ~ 100ppb
pH = 7.0 (5mM
NaHCO3)
Mesh size: 100x140
Contact Time = 7 Days
0.01
1 10 100
Ce( µgAs/L)

Figure A-1 Isotherm test in model solution for arsenate removal showing 4 new
versions of C-II4B in parallel with old version. These new media were prepared by
iron hydroxide precipitation and have same iron content than C-II4B-3.

78
 10
A-3-7 A-J1-6
C-II4C-4 A-3B-11
S-4-11
q (µgA s/mg adsorbent)

0.65
C-II4C-4: q =0.12Ce ; R2 =0.98
0.1 0.55
S-4: q =0.33Ce ; R2 =0.95
As(V) ~ 100ppb
0.67
pH = 7.0 (5mM NaHCO 3) A-3B : q =0.25Ce ; R2 =0.89
0.56
Mesh size: 100x140 A-3: q =0.22Ce ;R2 =0.92

Contact Time = 7 Day s A-J1: q =0.13Ce

0.58
; R2 =0.99
0.01
1 10 100
Ce( µgA s/L)

Figure A-2 Isotherm test in model solution using Fe-GACs from ASU, Solmetex and
Clemson University. October 2005

79
80
 ABBREVIATIONS

As(III) Arsenite
As(V) Arsenate
ASU Arizona State University
AZ Arizona

BV Bed volume

DOC Dissolved organic carbon

EBCTLC Empty bed contact time of large scale column

EBCTSC Empty bed contact time of RSSCT column

Fe Iron
Fe-GAC Iron (hydr)oxide impregnated granular activated carbon
ft Foot

g gram
GAC Granular activated carbon
GC Gas chromatography
GFAA Graphite furnace atomic absorption spectroscopy
gpm Gallons per minute
GW Groundwater

K Freundlich isotherm adsorption coefficient

L Liter
LLNL Lawerence Livermore National Lab

M Meter
mL Milliliter
MTBE Methyl tert-butyl ether

NSFI National Sanitation Foundation International

pHZPC pH of zero point of charge

ppb parts per billion
ppm parts per million

q Adsorption capacity

RSSCT Rapid small scale column test

81
SEM Scanning electron microscopy
Si Silica
SOC Synthetic organic chemical

TM Trademark

UVA254 Ultraviolet absorbance at 254 nm

V Vanadate
VOC Volatile organic chemical

82