www.epa.gov/aaa
United States
Environmental Protection
Agency
Identification and Characterization
Methods for Reactive Minerals
Responsible for Natural
Attenuation of Chlorinated Organic
Compounds in Ground Water
Office of Research and Development
National Risk Management Research Laboratory, Ada, Oklahoma 74820
-------
-------
Identification and Characterization
Methods for Reactive Minerals
Responsible for Natural
Attenuation of Chlorinated Organic
Compounds in Ground Water
Yongtian He
National Research Council Research Associate
Chunming Su
John Wilson
RickWilkin
Cherri Adair
Tony Lee
U.S. EPA/ORD Ground Water and Ecosystems
Restoration Division
Paul Bradley
U.S.G.S. South Carolina Water Science Center in
Columbia, SC
Mark Ferrey
Minnesota Pollution Control Agency
Office of Research and Development
National Risk Management Research Laboratory, Ada, Oklahoma 74820
-------
Notice
The U.S. Environmental Protection Agency through its Office of Research and
Development funded portions of the research described here. Mention of trade
names and commercial products does not constitute endorsement or recommendation
for use. All research projects making conclusions and recommendations based on
environmentally related measurements and funded by the Environmental Protection
Agency are required to participate in the Agency Quality Assurance Program. This
project was conducted under a Quality Assurance Project Plan for Task 3674 and
19421. Work performed by U.S. EPA employees or by the U.S. EPAon-site analytical
contractor followed procedures specified in these plans without exception. Information
on the plans and documentation of the quality assurance activities and results are
available from Cherri Adair or John Wilson.
-------
Foreword
The U.S. Environmental Protection Agency (EPA) is charged by Congress with protecting the Nation's land, air, and water
resources. Under a mandate of national environmental laws, the Agency strives to formulate and implement actions leading
to a compatible balance between human activities and the ability of natural systems to support and nurture life. To meet
this mandate, EPA's research program is providing data and technical support for solving environmental problems today
and building a science knowledge base necessary to manage our ecological resources wisely, understand how pollutants
affect our health, and prevent or reduce environmental risks in the future.
The National Risk Management Research Laboratory (NRMRL) is the Agency's center for investigation of technologi-
cal and management approaches for preventing and reducing risks from pollution that threatens human health and the
environment. The focus of the Laboratory's research program is on methods and their cost-effectiveness for prevention
and control of pollution to air, land, water, and subsurface resources; protection of water quality in public water systems;
remediation of contaminated sites, sediments and ground water; prevention and control of indoor air pollution; and resto-
ration of ecosystems. NRMRL collaborates with both public and private sector partners to foster technologies that reduce
the cost of compliance and to anticipate emerging problems. NRMRL's research provides solutions to environmental
problems by: developing and promoting technologies that protect and improve the environment; advancing scientific and
engineering information to support regulatory and policy decisions; and providing the technical support and information
transfer to ensure implementation of environmental regulations and strategies at the national, state, and community levels.
Monitored Natural Attenuation is widely used by U.S. EPA to manage risk associated with hazardous organic contami-
nants in soils and ground water. The Agency prefers attenuation mechanisms that destroy contaminants. As a result,
site characterization at hazardous waste sites has focused on biological degradation of organic contaminants because
biodegradation was the only mechanism that was widely acknowledged to destroy organic contaminants. In recent years
it has become increasingly apparent that abiotic degradation mechanisms can make a substantial contribution to natural
attenuation of a variety of halogenated organic compounds in soil, sediment and ground water.
This report provides a technical basis to evaluate the contribution of abiotic processes to MNA of halogenated organic
compounds. The report reviews the current knowledge of the rate of transformation of halogenated organic compounds
that is associated with reactive mineral phases in soil and aquifer sediment. The report also reviews the known trans-
formation products.
David G. Jewett, \cting Director
Ground Water and
Ecosystems Restoration Division
National Risk Management Research Laboratory
-------
-------
Contents
Notice ii
Foreword iii
Acknowledgements xiii
Executive Summary xv
Evolution of Applications for Abiotic Processes xvi
Important Reactive Iron Minerals xvii
Iron Sulfides xviii
Magnetite xix
Green Rust xx
Adsorbed Fe(II) xx
Phyllosilicate Clays xxi
Iron Carbonates xxi
1.0 Formation and Stability of Reactive Iron and Sulfur Minerals Under Various
Geochemical Conditions 1
1.1 Formation Processes of Reactive Minerals 3
1.1.1 Formation of Pyrite and Mackinawite 4
1.1.2 Formation of Siderite 6
1.1.3 Formation of Green Rusts 6
1.1.4 Formation and Transformation of Goethite and Ferrihydrite 7
1.2 Geochemical Modeling 7
2.0 Iron Sulfide Minerals 11
2.1 Applications of Iron Sulfide Minerals 11
2.2 Abiotic Transformation of Chlorinated Ethylenes 12
2.2.1 Transformation Products 12
2.2.2 Rates of Degradation on Iron Sulfide Minerals 15
2.2.3 Factors Affecting Degradation 15
2.2.3.1 Effect of Freeze Drying FeS on Rate Constants for Tranformation of TCE 15
2.2.3.2 Effect of pH on the Rate of Transformation of TCE on FeS 16
2.2.3.3 Organic Molecules 17
2.2.3.4 Metal Ions 17
2.2.3.5 Dissolved Oxygen (DO) 17
2.2.4 Degradation Mechanisms 17
2.3 Abiotic Transformation of Chlorinated Alkanes 18
2.3.1 Degradation Products 18
2.3.2 Rates of Degradation on Iron Sulfide Minerals 19
2.3.3 Conditions Affecting Degradation 20
2.3.3.1 pH 20
2.3.3.2 Metals 21
2.3.3.3 Sulfide 21
2.3.3.4 Organic Molecules 21
2.4 Abiotic Transformation of Chlorinated Methanes 22
2.4.1 Degradation Products 22
2.4.2 Rate Constants 22
2.4.3 Conditions Affecting Reactivity 24
-------
2.4.3.1 pH 24
2.4.3.2 Sulfide 24
2.4.3.3 Organic Compounds 25
2.5 Abiotic Transformation of the Brominated Ethane EDB 25
2.5.1 Degradation Products 25
2.5.2 Rate Constants 25
2.5.3 Conditions That Affect Degradation 26
2.5.3.1 Sulfate and Sulfide 26
2.5.3.2 pH 26
2.6 Formation, Transformations, and Reactivity of Iron Sulfides 26
2.6.1 FeS Mineralogy and Reactivity 26
2.6.2 Formation of FeS Minerals 27
2.6.3 Transformation of FeS Minerals 28
2.6.4 Effect of Transformations of Iron Sulfides on Reactivity 29
2.6.5 Effect of Transition Metals on Degradation 30
2.7 Environmental Implications 30
2.7.1 Geochemical Parameters 30
2.7.2 Differentiating Biotic and Abiotic Degradation 31
2.7.3 Geochemical Modeling 31
3.0 Magnetite 33
3.1 Structure of Magnetite 33
3.2 Transformation Products of Chlorinated Organic Compounds on Magnetite 33
3.3 Effect of Properties of Magnetite on Rate of Transformation 34
3.3.1 Effect of Particle Size and Surface Area 34
3.3.2 Effect of pH 35
3.3.3 Effect of Fe2+ Sorbed to the Surface of Magnetite 35
3.4 Rates of Transformation of Chlorinated Organic Compounds on Magnetite 36
3.4.1 Degradation of cis-DCE and Vinyl Chloride on Magnetite 36
3.4.2 Degradation of PCE and TCE on Magnetite 40
3.4.3 Degradation of Carbon Tetrachloride on Magnetite 42
3.5 Rates of Transformation Normalized to Magnetic Susceptibility 45
4.0 Transformation of Chlorinated Hydrocarbons by Green Rusts 47
4.1 Structure and Reactivity of Green Rusts 47
4.1.1 Chemical Composition and Crystal Structure of Green Rusts 47
4.1.2 Chemical Reactivity of Green Rusts 48
4.1.3 Abiotic Degradation of PCE and TCE by Green Rusts 49
.3.1 Reports of Degradation of PCE and TCE in the Literature 49
.3.2 Degradation of TCE in Studies Performed at Kerr Center (U.S. EPA) 49
4.1.4 Abiotic Degradation of Chlorinated Methanes and Alkanes by Green Rusts 51
.4.1 Reports of Degradation of Chlorinated Methanes and Alkanes in the Literature 51
4. .4.2 Degradation of Carbon Tetrachloride in the EPA Study 52
.4.3 Reports of Degradation of Chlorinated Alkanes in the Literature 53
4.2 Occurrence and Determination of Green Rusts 55
4.2.1 Geochemical Conditions that Favor Green Rust Formation 55
4.2.2 Structural Stability of Carbonate Green Rust 56
5.0 Phyllosilicate Clays 59
5.1 Structural Iron in Phyllosilicate Clays 59
5.2 Degradation Processes on Phyllosilicate Clays 59
5.3 Rate Constants for Degradation of Chlorinated Alkenes 60
5.4 Rate Constants for Degradation of Carbon Tetrachloride and Chlorinated Alkanes 60
-------
5.5 Effect of Sorbed Iron (II) on the Rate of Degradation 61
5.6 Extrapolation of Rate Constants to the Field 61
6.0 Methods to Characterize Sediments and Organic Contaminants 67
6.1. Isotopic Fractionation to Characterize Degradation of Organic Compounds 67
6.1.1 Isotopic Fractionation of Carbon Tetrachloride on Various Iron Minerals 67
6.1.2 Isotopic Fractionation of PCE and TCE on Iron(II) Monosulfide 68
6.1.3 Isotopic Fractionation of EDB on Iron(II) Monosulfide 68
6.1.4 Isotopic Fractionation of TCE on Green Rust 69
6.1.5 Isotopic Fractionation of TCE and cis-DCE on Magnetite 69
6.1.6 Isotopic Fractionation of Chlorinated Alkanes on Phyllosilicate Clay 70
6.2 Methods to Estimate the Quantity of Reactive Minerals 70
6.2.1 Characterization by Chemical Extraction Methods 71
6.2.1.1 Sulfur Minerals in Soil and Sediments 71
6.2.1.2 Interferences with Chemical Extraction Methods for Sulfur Minerals 72
6.2.2 Characterization of Chemical Identify by Spectroscopy or Diffraction 73
6.2.2.1 Characterization by X-ray Diffraction (XRD) 73
6.2.2.2 Characterization of Green Rusts by X-ray Diffraction (XRD) and Atomic Force
Microscopy (AFM) 74
6.2.2.3 Characterization of Green Rusts by Infrared Spectroscopy (IR) 76
6.2.3 Estimating Magnetite in Sediment from Magnetic Susceptibility 76
6.3 Methods to Determine Oxidation Status 78
6.3.1 Characterization of Green Rusts by Mossbauer Spectroscopy 78
6.3.2 Characterization of Green Rust by X-ray Adsorption Near Edge Spectroscopy (XANES) 80
6.4 Methods to Estimate the Specific Surface Area of Minerals 82
6.4.1 Surface Area from Peak Broadening during X-ray Diffraction 82
6.4.2 Surface Area from Electron Microscopy 82
6.4.3 Characterization of Green Rusts by Scanning/Transmission Electron Microscopy 83
6.5 Collection and Handling of Samples 84
7.0 Recommendations for Future Research 85
7.1 Further Explore the Role of Phyllosilicate Clays 85
7.2 Reexamine the Role of Pyrite 85
7.3 Determine if Geochemical Models Can Be a Useful Surrogate to Predict the Rates of
Abiotic Degradation at Field Scale 85
7.4 Use Geochemical Models to Improve Estimates of Degradation on Mineral Surfaces 85
7.5 Characterize Isotopic Fractionation of cis-DCE and Vinyl Chloride on Reactive Minerals ... 85
7.6 Characterize the Role of Manganese Oxides 86
8.0 References 87
AppendixA Iron Sulfides 101
A. 1 Experimental Details Related to TCE Degradation by FeS 101
A.2 Experimental Details Related to EDB Degradation by FeS 102
Appendix B Magnetite 103
B.I Building 102 Site and Site A on the TCAAP, North of St. Paul, Minnesota 103
B.2 Baytown Superfund Site, Minnesota 108
B.3 Thermo-Chem Site, East of Muskegon, Michigan 109
B.4 Products of Degradation of cis-DCE in Sediment from TCAAP Minnesota Site Ill
B.5 Removal of Vinyl Chloride in Sediment from TCAAP Minnesota Site 112
Appendix C Materials and Methods for Laboratory Studies of Abiotic Transformation of TCE and
Carbon Tetrachloride by Green Rusts 113
-------
Appendix D Quality Assurance Documentation 115
D.I Analysis of Halogenated Organic Compounds in Water Samples 115
D.2 Analysis of Ratio of Stable Isotopes of Carbon in TCE and cis-DCE 119
D.3 Analysis of Total Iron in Sediment 122
D.4 Analysis of Sodium, Potassium, Calcium, Magnesium, and Iron in Water Samples 122
D.5 Analysis of Sulfate and Chloride in Water Samples 124
D.6 Analysis of Sulfide in Water Samples 124
D.7 Determination of Magnetic Susceptibility 124
D.8 Scintillation Counting 125
-------
Figures
Figure 1.1. Mineral stability diagrams: a) Eh versus log a H2S showing stability fields for
goethite, magnetite, carbonate green rust, siderite, mackinawite, and pyrite (Fe=10~4,
C=10~3, pH =7.5, hematite suppressed); b) Eh versus log a HCO3~ showing stability
fields for goethite, magnetite, carbonate green rust, and siderite (Fe=10~4, pH =7.5,
hematite suppressed) 4
Figure 1.2. Ternary diagram showing the importance of controlling factors on iron sulfide
formation in natural aquifers and biowalls 5
Figure 1.3. Saturation indices (SI) for a) disordered mackinawite and b) crystalline mackinawite
as a function of pH and time (ground water from the biowall at OU-1 on Altus AFB, OK). . 10
Figure 2.1. TCE degradation by chemically synthesized FeS atpH 7.2 16
Figure 2.2. Effect of pH on the rate of TCE degradation on chemically synthesized FeS. The data
series Not freeze dried [2] and Freeze dried [2] repeat data from Figure 2.1 16
Figure 2.3. EDB degradation by FeS at pH 7.2 26
Figure 2.4. Effect of pH on the first order rate constant for EDB degradation by FeS. The rate
constant is normalized to the concentration of FeS in suspension 26
Figure 3.1. Effect of solution pH on the rate constant for degradation of carbon tetrachloride on
the surface of magnetite 35
Figure 3.2. Effect of pH on the solution concentrations of Fe2+ in equilibrium with magnetite 36
Figure 3.3. Sorption isotherms for Fe2+ on the surface of magnetite 36
Figure 3.4. Removal of cis-DCE in aquifer sediment containing magnetite 37
Figure 3.5. The surface area specific rate of degradation of cis-DCE on magnetite can explain the
removal of cis-DCE in sediment from the TCAAP. 38
Figure 4.1. The crystal structure of green rust compounds consists of layers of Fe(II)(OH)6 in
which some of the Fe(II) is replaced by Fe(III) 47
Figure 4.2. X-ray diffraction pattern of freshly synthesized carbonate green rust and sulfate green
rust (less than 24 hours after synthesis) scanned as glycerol smears 48
Figure 4.3. Removal of TCE in the presence of carbonate green rust as a function of reaction
time, pH, and excess dissolved Fe(II) 50
Figure 4.4. Removal of TCE in the presence of sulfate green rust as a function of reaction time
with and without a 0.05 M pH 8 Trizma buffer 50
Figure 4.5. Removal of TCE in the presence of sulfate green rust as a function of reaction time,
concentration of added CuSO4, and presence or absence of 0.05 M Trizma buffer 51
Figure 4.6. Removal of carbon tetrachloride in the presence of sulfate green rust as a function of
reaction time and pH 53
Figure 4.7. Proposed pathways for the reduction of chlorinated ethanes in aqueous suspensions
of green rusts and in green rust suspensions spiked with Ag(I) (AgGR) or Cu(II)
(CuGR); however, some elements of the pathways shown are not relevant to all
experimental systems 54
Figure 4.8. Adsorption of the phosphate on the lateral {1010} face of the GR(CO32) crystal that
stabilizes it (From Bocher et al., 2004) 57
-------
Figure 6.1. Fractionation of stable isotopes of carbon during abiotic degradation of TCE in aquifer
sediment containing magnetite (site at Building 102 at TCAAP) 69
Figure 6.2. Fractionation of stable isotopes of carbon during abiotic degradation of cis-DCE in
aquifer sediment containing magnetite 70
Figure 6.3. Powder X-ray diffraction scans of mixtures of magnetite and quartz 74
Figure 6.4. Blow-up of the 2-theta region from 44 to 48°, showing the most intense diffraction
peak for magnetite at 45.2° 74
Figure 6.5. Relationship between the mass magnetic susceptibility of a sediment sample and the
content of magnetic materials 77
Figure 6.6. Mossbauer spectra of products obtained during reduction of y-FeOOH (80 mM) with
formate (75 mM) in the presence of AQDS (100 |o,M) in bacterial cultures (initially 8 x
109 cells m/L) 79
Figure 6.7. Total electron yield XANES measurements of the Fe-LII, III absorption edges for green
rust and reference samples FeCl2 (11+) and goethite (III+) 81
Figure 6.8. TEM image and corresponding electron diffraction pattern of carbonate green rust
hexagonal crystals mixed with a minor fine-grained phase (y-FeOOH) obtained after
6 days of bacterial reduction of lepidocrocite (initially 80 mM y-FeOOH, 75 mM
formate, and 100 |^M AQDS) 83
Figure B.I. Orientation of monitoring wells at the Building 102 site on the TCAAP. 104
Figure B.2. Orientation of monitoring wells at Site A on the TCAAP. 105
Figure B.3 Removal of PCE in autoclaved sediment from the Building 102 Site and Site A on the
former TCAAP, north of St. Paul, MN 107
Figure B.4. Removal of TCE in autoclaved sediment from the Building 102 Site and Site A on the
former TCAAP, north of St. Paul, MN 107
Figure B.5. Removal of cis-DCE in autoclaved sediment from the Building 102 Site and Site A on
the former TCAAP, north of St. Paul, MN 108
Figure B.6. Relationship between monitoring wells in Table B.3 and TCE contamination in the
underlying Prairie Du Chien Aquifer at Baytown Township, MN 108
Figure B.7. Removal of TCE in sediment from the Baytown Site, north of St. Paul, Minnesota 109
Figure B. 8. Location of sediment used for microcosms in the plume of contamination at the
Thermo-Chem site near Muskegon, Michigan 110
Figure B.9. Removal of cis-DCE in sediment from the Thermo-Chem site, east of Muskegon,
Michigan Ill
Figure B. 10.Removal of cis-DCE in autoclaved sediment from the former TCAAP, north of St.
Paul, MN 112
Figure B. 11 .Removal of vinyl chloride in sediment from the former TCAAP, north of St. Paul, MN. ... 112
-------
Tables
Table 1. Degradation of carbon tetrachloride on reactive iron and sulfur minerals in laboratory
experiments xviii
Table 1.1. Mineralogical data for iron- and sulfur-bearing phases of interest 1
Table 1.2. Field data and results of geochemical speciation modeling of abiowall at the OU-1 site
on Altus AFB, OK 9
Table 2.1. Rate Constants and products of abiotic degradation of chlorinated ethylenes by FeS
minerals 13
Table 2.2. Rates and products of abiotic degradation of chlorinated alkanes on FeS minerals 19
Table 2.3. Rates and products of abiotic degradation of chlorinated methanes on FeS minerals 23
Table 3.1. Surface area specific rate constants for removal of cis-DCE and vinyl chloride on
magnetite 39
Table 3.2. Surface area specific rates of removal of PCE and TCE on magnetite 41
Table 3.3. Surface area specific rate constants for removal of carbon tetrachloride on magnetite 43
Table 3.4. Relationship between the rate of removal of PCE, TCE, cis-DCE or vinyl chloride and
the content of total iron, and the magnetic susceptibility of aquifer sediment 44
Table 4.1. Specific surface area values reported in the literature for synthetic green rusts 48
Table 4.2. Rate constants for transformation of chlorinated hydrocarbons by green rusts 55
Table 5.1. Rate constants for degradation of chlorinated alkenes on phyllosilicate clay minerals 63
Table 5.2. Rate constants for degradation of carbon tetrachloride and chlorinated alkanes on
phyllosilicate clay minerals 65
Table 5.3. Maximum quantity (C°RC) of chlorinated hydrocarbon that can be reduced by reactive
iron in representative phyllosilicate clays 66
Table B.I. Distribution of chlorinated organic compounds and geochemical parameters in ground
water at the Building 102 site on the TCAAP in 2005 104
Table B.2. Distribution of chlorinated organic compounds and geochemical parameters in ground
water at Site A on the TCAAP in 1998 105
Table B.3. Distribution of chlorinated organics and dissolved oxygen in the ground water in the
Baytown TCE plume 109
Table B.4. Distribution of chlorinated organic compounds and geochemical parameters in ground
water on the bank of Black Creek at the Thermo-Chem site in Michigan 110
Table B.5. Distribution of 14C from 14C-c/s-DCE in microcosms and container controls 112
Table D.I. Quality of Data on Concentrations of Halogenated Organic Compounds as Determined
in Experiments Described in Section 2 116
Table D.2. Quality of Data on Concentrations of Acetylene as Determined in Experiments
Described in Section 2 117
Table D.3. Quality of Data on Concentrations of Halogenated Organic Compounds as Determined
in Experiments Described in Section 3 and Appendix B 118
Table D.4. Quality of Data on Concentrations of Halogenated Organic Compounds as Determined
in Experiments Described in Section 4 119
-------
Table D.5. Quality parameters for analysis of 513C in TCE and cis-DCE 121
Table D.6. Quality of Data on Concentrations of Total Iron in Sediment as Determined in
Experiments Described in Section 3 122
Table D.7. Quality of Data on Concentrations of Sodium, Potassium, Calcium, Magnesium and
Iron used in the Geochemical Modeling in Section 1 123
Table D.8. Quality of Data on Concentrations of Sulfate and Chloride used in the Geochemical
Modeling in Section 1 124
-------
Acknowledgements
Peer reviews for this document were provided by Bill Batchelor (Department of
Civil Engineering, Texas A&M University), Dick Brown (a technical director for
ERM, Inc. based in Ewing, New Jersey), James Henderson (a project director for the
DuPont Corporate Remediation Group based in Charlotte, North Carolina), Romona
Darlington (Research Scientist at Battelle Memorial Institute, Columbus, Ohio),
J. Gillette (Air Force Center for Engineering and the Environment, San Antonio,
Texas), Michael Barcelona (Department of Chemistry, Western Michigan University,
Kalamazoo, Michigan) and Robert Ford (U.S. EPA National Risk Management
Research Laboratory, Cincinnati, Ohio).
Significant technical support was provided by Tracy Pardue, Lisa Hudson, Sean
Beach, Steve Markham, Vanessa Scroggins, Ying Wang, and John Cox (Shaw
Environmental), Lynda Callaway and Kristie Hargrove (U.S. EPA), and Kevin Smith
and John Skender (Student Contractors).
Pat Bush (an Information Coordinator with the Senior Environmental Employee
Program, a grantee with U.S. EPA at the R.S. Kerr Environmental Research Center,
Ada, Oklahoma) is acknowledged for her technical editing to provide consistency in
formatting and grammar. Martha Williams (a Publication Editor for SRA, a contractor
to U.S. EPA at the R.S. Kerr Environmental Research Center in Ada, Oklahoma)
assisted with final editing and formatting for publication. Kathy Tynsky (a Graphics
Designer for SRA, a contractor to U. S. EPA at the R.S. Kerr Environmental Research
Center in Ada, Oklahoma) assisted with photos and proofing for this publication.
-------
-------
Executive Summary
This report is intended to facilitate the
application of abiotic processes to remediate
contamination from halogenated hydrocarbons
in ground water. It is intended for scientists
or engineers who design remedies for con-
taminated ground water, or who review the
remedial proposals of others. It reviews the
literature on the rate of degradation of particu-
lar halogenated hydrocarbons on particular
reactive minerals, provides information on
currently available techniques to characterize
the reactive minerals that may be present in
aquifer material, and it evaluates the prospects
for applying the available analytical techniques
to make a quantitative prediction of the rate or
extent of degradation.
Ideally this report would identify appropriate
analytical techniques that are appropriate and
sensitive for each class of reactive mineral,
and would provide equations that would relate
the quantity or surface area of each mineral to
the rate of degradation that might be expected.
For iron(II) monosulfides, the current state
of science approaches this expectation. For
magnetite, the available technique is sensitive,
but not particularly specific, and the predic-
tions allowed by the current state of practice
are only semi-qualitative. For green rusts,
the kinetics of degradation are understood in
simple laboratory systems, but the current state
of science is not adequate to allow predictions
of the rate of reaction in complex aquifer
materials. For other reactive minerals, even
less information is available.
This summary provides a brief review of
previous research activity and engineering
practice dealing with abiotic processes. It
reviews the structure and reactivity of the
major classes of reactive iron and sulfur miner-
als in ground water, and considers the efficacy
of current approaches to characterize the
reactive minerals.
Section 1 of this report reviews the structure
and composition of the major classes of
minerals that contribute to abiotic degradation
of halogenated hydrocarbons in ground water.
Section 1 also identifies the geochemical envi-
ronments in which the various minerals are
stable, and illustrates the use of geochemical
modeling to predict the occurrence of reactive
minerals in aquifer material from an analysis
of relevant parameters in ground water.
Sections 2, 3, 4, and 5 deal in detail with the
most common and well-studied reactive miner-
als; Section 2 deals with iron sulfides, Section
3 with magnetite, Section 4 with green rusts,
and Section 5 with phyllosilicate clays. Each
section in turn reviews the available literature
on the rates of reaction of various halogenated
hydrocarbons, the influence of geochemical
parameters (such as pH) on the rate of reac-
tion, describes the degradation products that
can be expected from various halogenated
hydrocarbons, and describes in detail the most
appropriate analytical techniques to character-
ize the reactive mineral.
Section 6 describes the present state of practice
for analysis of reactive minerals. The section
discusses the sensitivity of and specificity
of analytical methods that are available to
determine reactive iron sulfur minerals in
aquifer sediment. It describes precautions that
are necessary to preserve the integrity of core
samples that are collected for the analysis of
reactive minerals. Section 6 also evaluates
the use of stable isotope analyses to monitor
the extent of degradation of the chlorinated
hydrocarbons at field scale.
Section 7 provides recommendations for future
research.
-------
Evolution of Applications for Abiotic
Processes
For many years, the only non-biological reac-
tions that were considered to have application
to the restoration of ground water contaminated
with halogenated organic compounds were the
reactions of 1,1,1-TCAto form 1,1-DCE or
acetate (Vogel and McCarty, 1987; Vogel et
al., 1987), and reactions that destroyed EDB
in the presence of HS~ in solution (Barbash
and Reinhard, 1989). Catalytic reactions on
surfaces were generally ignored.
This situation changed with the successful
application of permeable reactive barriers
(PRBs) containing zero-valent iron to treat TCE
in ground water (Gillham and O'Hannesin,
1994; Wilkin and Puls, 2003). Interest in
the research community and the remediation
community expanded out of a narrow focus on
biological processes to include abiotic reac-
tions on surfaces. Detailed and comprehensive
studies compared the rates of reaction of
chlorinated hydrocarbons on iron(II) mono-
sulfides (Butler and Hayes, 1998, 1999, 2000,
2001), on magnetite (Lee and Batchelor, 2002a)
and on green rusts (Lee and Batchelor, 2002b;
O'Loughlin et al., 2003).
Kennedy et al. (2006b) applied the laboratory
studies of Butler and Hayes (1999, 2001) on
TCE degradation by iron(II) monosulfides to
explain the removal of TCE that was observed
in a PRB at Altus AFB, OK. The PRB was
constructed with sand and shredded plant
mulch. The sand contained iron in coatings on
the quartz grains. Ambient concentrations of
sulfate in the ground water were high, and large
amounts of iron(II) monosulfide accumulated in
the PRB. Kennedy et al. (2006b) attempted to
treat TCE in ground water by injecting magne-
sium sulfate and lactate into a plume. Sulfate
reduction produced large quantities of iron
sulfides in the aquifer matrix.
Iron and sulfur minerals are common corrosion
products in zero-valent iron reactive barriers,
and many of these minerals can transform or
degrade chlorinated organic compounds. Until
recently, the formation of secondary minerals
has generally been viewed as an undesirable
outcome that limits the long-term performance
of reactive barriers. However, some of the cor-
rosion products that are associated with granu-
lar or nano-sized iron particles may contribute
to the overall treatment effectiveness of reac-
tive barriers (Wilkin and Puls, 2003; Nooten et
al., 2008).
If the ambient concentrations of sulfate are
high, reactive iron and sulfur minerals can
be expected to form during in situ anaerobic
bioremediation of chlorinated hydrocarbons.
Bioremediation involves the addition of a
biodegradable carbon source to support biologi-
cal reductive dechlorination. The same carbon
source will support sulfate reduction to produce
sulfide.
Abiotic processes also have a role in the natural
attenuation of contaminants. Abiotic degrada-
tion of PCE and TCE tends to favor dichloro-
elimination reactions to produce acetylene over
a sequential hydrogenolysis reaction to produce
cis-DCE and vinyl chloride (Butler and Hayes,
2000). Abiotic processes can also degrade
chlorinated ethylenes to glycolate, acetate,
formate, and carbon dioxide (Darlington et al.,
2008). Anaerobic biological reactions follow
the hydrogenolysis pathway exclusively. The
removal or treatment of chlorinated hydrocar-
bons through abiotic reactions may avoid the
production of toxic daughter products such as
vinyl chloride.
Ferrey et al. (2004) applied the laboratory
studies of Lee and Batchelor (2002a) with
magnetite to explain the natural attenuation of
cis-DCE in a plume of contaminated ground
water. Despite the disappearance of cis-DCE.,
vinyl chloride or ethylene did not accumulate in
the ground water.
-------
Brown et al. (2007) suggested that a role for
abiotic processes in natural attenuation could
be demonstrated in four ways. First of all,
abiotic processes are probably important in
plumes of PCE or TCE when the concentra-
tions of PCE and TCE decline with distance
along the flow path, and there is no evidence
of the accumulation of their chlorinated
transformation products cis-DCE and vinyl
chloride. Second, mineralogical analyses can
be performed on aquifer sediments to recognize
reactive minerals such as magnetite or iron
monosulfides. Third, the ground water can
be monitored for reaction products that are
unique to abiotic reactions, such as acetylene.
This line of evidence can be expanded to look
at the fractionation of stable isotopes in the
chlorinated hydrocarbons. Finally, protocols
for microcosm studies should be modified or
expanded to specifically examine abiotic reac-
tions. Removal of contaminants in autoclaved
controls should not be discounted as the trivial
loss of material from the container.
Important Reactive Iron Minerals
A variety of iron-bearing soil minerals can
degrade chlorinated hydrocarbons. Iron
sulfides (disordered mackinawite, mackinawite,
and pyrite), iron oxides (magnetite), green rust,
and iron-bearing clays have been shown to
support complete or nearly complete transfor-
mation of PCE, TCE, and carbon tetrachloride.
These minerals have been identified in aquatic
environments, typically in iron-reducing and/
or sulfate-reducing environments. Mineral
surfaces act as electron donors and/or reaction
mediators to increase the rate of reductive
dechlorination. Laboratory evidence suggests
that the more thermodynamically stable phases
(pyrite and magnetite) support comparatively
slower rates of TCE reduction than metastable
phases, such as mackinawite.
Table 1 compares the rates of degradation of
carbon tetrachloride on mackinawite, magne-
tite, sulfate green rust, goethite, lepidocrocite,
pyrite, siderite, and hematite in laboratory
studies. The reactions occur at the surfaces of
minerals, and the rates of reaction should be
proportional to the surface areas of the minerals
presented to water. This information is easily
attained for specific samples of pure minerals,
but it is difficult to estimate the specific surface
area of a particular mineral in soils and aquifer
materials. As a consequence, Table 1 ranks the
reactions on the minerals by the magnitude of
their rate constants that have been normalized
by the mass concentration of the mineral. Data
on surface areas and surface-area normalized
rate constants are presented for comparison.
Mackinawite was the most reactive mineral;
magnetite was approximately tenfold less reac-
tive than mackinawite, and sulfate green rust
was approximately tenfold less reactive than
magnetite.
Iron(II) sorbed to the surface of minerals
contributes to the degradation of carbon
tetrachloride. The rates of degradation for
mackinawite, magnetite, goethite, lepidocrocite,
siderite, and hematite were determined in the
presence of 1 mM dissolved Fe2+. The degra-
dation of carbon tetrachloride by reactions on
goethite, lepidocrocite, siderite, and hematite
are probably carried out by iron(II) sorbed to
the mineral surfaces. The rates on goethite and
lepidocrocite are roughly comparable to the
rate of degradation on magnetite. This might
be expected for hematite. Hematite in the
presense of iron(II) is transformed to magnetite
(Behrends and Van Cappellen, 2007). The rates
on siderite and hematite are much slower.
Table 1 lists the specific surface areas of the
minerals used in the experiments. With the
exception of pyrite, the minerals were synthe-
sized by precipitation from solution and the
particle size and specific surface area of the
laboratory preparations were loosely compa-
rable to the particle size and surface area of
natural minerals in aquifer material. In con-
trast, pyrite was prepared by grinding a large
specimen to a powder, and the specific surface
-------
Table 1. Degradation of carbon tetrachloride on reactive iron and sulfur minerals in laboratory experi-
ments. In some cases, Fe2+ was added to the reaction medium. The pseudo first order rate
constant is normalized to the concentration of the mineral (units of L g~1 day1) or to the surface
area of the mineral presented to water (L rrv2 day1).
Mineral
mackinawite
with Fe2+
magnetite
with Fe2+
sulfate green
rust
goethite with
Fe2+
lepidocrocite
with Fe2+
pyrite
siderite with
Fe2+
hematite
with Fe2+
First Order
Rate Constant
L g"1 day1
15
2.3
0.30
0.18
0.059
0.014
0.0074
0.0067
L m~2
day1
1.1
0.12
0.021
0.010
0.0033
0.16
0.00019
0.00048
Specific
Surface
m2/g
13
19
14.1
17
18
0.088
38
14
Cone.
g/L
4
2.6
5
3.1
2.8
14.8
1.3
3.6
pH
7.2
7.3
7.6
7.2
7.2
6.5
7.1
7.2
Reference
Zwank et al. (2005)
Zwank et al. (2005)
O'Loughlin et al. (2003)
Zwank et al. (2005)
Zwank et al. (2005)
Kriegman-King and
Reinhard (1994)
Zwank et al. (2005)
Zwank et al. (2005)
area of the pyrite was nearly one hundredfold
lower than the other minerals. When the rate
of degradation on pyrite is expressed on a mass
basis, the rate is very low. When the rates are
compared on a surface area basis, the degrada-
tion of carbon tetrachloride on pyrite is compa-
rable to the rate on magnetite.
Because they are more reactive, mackinawite,
magnetite, and green rusts have received the
most attention, and more information is avail-
able in the literature on their behavior. This
report will emphasize the degradation of chlori-
nated hydrocarbons on iron sulfides (including
mackinawite), on magnetite, and on green rusts.
Iron Sulfides
Transition metal sulfides (e.g., FeS, MS, CuS,
ZnS, CdS) have exceedingly low solubility
products and might be expected to form in
sulfate-reducing environments. However,
transition metals other than iron are typically
present in trace amounts in aquifer materi-
als, which does not allow for any significant
accumulation of sulfide minerals other than
those of iron. Due in part to the fact that iron
is the most abundant of these transition metals,
iron sulfides are the only metal sulfides com-
monly recognized in soils and sediments. In
contaminated systems with high metal load-
ings, sulfides of Hg, Zn, Cd, and Cu have been
reported. Studies on FeS amended with other
metals (e.g., Cr, Mn, Ni, Cu, Zn, Cd, and Hg)
suggest that increased rates of transformation of
chlorinated hydrocarbons may be achieved in
systems containing FeS that have been sub-
stituted with Co, Ni, or Hg (Jeong and Hayes,
2003; 2007).
Several iron sulfide phases have been syn-
thesized in the laboratory, either as transient
intermediates or as stable end products, and
-------
are therefore likely to form in subsurface
environments. These phases are: disordered
mackinawite, FeS; mackinawite, FeS; cubic
iron sulfide, FeS; hexagonal pyrrhotite, Fe^S;
greigite, Fe3S4; smythite, Fe9Sn; marcasite,
orthorhombic FeS2; and, pyrite, cubic FeS2.
Pyrrhotite and pyrite represent the thermody-
namically stable phases at the temperatures
and pressures characteristic of shallow ground
water aquifers. Disordered mackinawite,
mackinawite, and greigite are metastable with
respect to pyrite and/or stoichiometric pyr-
rhotite but are considered to be the principal
precursor phases to pyrite (Schoonen, 2004).
Laboratory studies have primarily focused on
the reductive dechlorination ability of pre-
cipitated FeS (e.g., Butler and Hayes, 1998;
1999; 2000). Compounds investigated include
hexachloroethane, PCE, TCE, 1,1-DCE, penta-
chloroethanes, tetrachloroethanes, trichloro-
ethanes, dichloroethanes, carbon tetrachlor-ide,
and tribromomethane. No appreciable reac-
tion was observed for 1,1-dichloroethane,
1-2-dichloroethane, and 1,1,2-trichloroethane,
but all other compounds were transformed with
half-lives of hours to days. There are several
key observations from these lab-based studies.
First, the degradation reactions occur at the
mineral surface and not in aqueous solution.
Second, the rate of FeS-mediated reductive
dechlorination is a strong function of pH, with
the rate increasing with increasing pH. Third,
there is some evidence to suggest that the
reaction mechanism may be associated with
surface-bound Fe(II) and not surface-bound
S(-II). Fourth, acetylene is a major reaction
product for both PCE and TCE transformation.
This contrasts with the sequential hydrogenoly-
sis products such as cis-DCE and vinyl chlo-
ride that are produced in the microbiological
transformation of these compounds. Finally,
the oxidation or transformation of FeS in the
environment will have a strong influence on
the potential for sustained intrinsic remediation
(Butler and Hayes, 2001; He et al., 2008).
Laboratory investigations with pyrite, an iron
disulfide, are more limited compared to FeS,
but include examinations of carbon tetra-
chloride transformation (Kriegman-King and
Reinhard, 1994; Lipczynska-Kochany et al.,
1994) and TCE transformation (Weerasooriya
and Dharmasena, 2001; Lee and Batchelor,
2004; Pham et al., 2008). Kreigman-King
and Reinhard (1994) proposed that electron
transfer at the pyrite surface occurs at sulfur
sites because carbon disulfide was detected as
a reaction product and because the surfaces
of the pyrite grains used in batch experiments
were depleted in iron (based on X-ray photo-
electron spectroscopy measurements). Pham
et al. (2008) recently found that in the presence
of pyrite and dissolved oxygen, TCE could be
oxidized to organic acids, including dichlo-
roacetic acid, glyoxylic acid, oxalic acid, and
formic acid. It seems unlikely that this type of
oxidative transformation could be sustained in
natural systems where pyrite would be present
in aquifer sediments. Extensive studies on
acid mine drainage have documented the rapid
oxidation of pyrite in oxidizing environments.
Nevertheless, the results of Pham et al. (2008)
are somewhat intriguing and indicate that mul-
tiple transformation pathways may be possible
for trichloroethylene in the presence of iron
sulfides like pyrite, and possibly mackinawite.
Consult Section 2 for more details on iron
sulfides.
Magnetite
Magnetite, a mixed Fe(II)-Fe(III) iron oxide,
is a widespread but typically minor compo-
nent of aquifer sediments. The structure of
magnetite is that of an inverse spinel, which
can be expressed by rewriting the formula as
Fe3+(Fe2+,Fe3+)O4. It is usually present as a
detrital phase derived from bedrock weathering.
However, very fine-grained authigenic magne-
tite can be produced by iron-reducing bacteria
or via the transformation of Fe(III) hydroxides
under reducing conditions. Under sulfate-
reducing conditions, magnetite will slowly
-------
dissolve and react to form pyrite. This trans-
formation of magnetite to pyrite is believed to
be important under high sulfide concentrations
(>1 mM) over several decades (Canfield and
Berner, 1987). At lower concentrations of
dissolved sulfide, magnetite dissolution without
pyrite replacement is observed. Magnetite is
also a common corrosion product in passive
reactive barriers (PRBs) constructed with
zero-valent iron. In PRBs, green rusts form as
a corrosion product of zero-valent iron, then
magnetite forms as a transformation product of
the green rust.
Consult Section 3 for more details on
magnetite.
Green Rust
Green rust compounds are compositionally
variable, mixed valence Fe(II)/Fe(III) layered
hydroxides (Hansen, 1989; Taylor, 1973).
Green rust compounds have been identified
in reducing soils and are sometimes corrosion
products in iron-based reactive barriers. Green
rust structural units consist of alternating
positively charged tri-octahedral metal hydrox-
ide sheets and negatively charged interlayers
of anions. Anions present in the interlayer
positions typically are Cl~, CO32~, or SO42~.
Laboratory experiments have revealed mixed
results regarding the efficiency of dechlorina-
tion in the presence of green rust (Erbs et al.,
1999; Lee and Batchelor, 2003; 2004). The
reactivity of green rust can be enhanced by
adding Cu2+ or Ag+ ions during green rust
synthesis (O'Loughlin et al., 2003). Addition
of these metals may lead to a galvanic couple
involving a zerovalent metal (i.e., Cu or Ag)
and green rust, with contaminant reduction
occurring on the surface of the metal and green
rust providing the source of electrons for the
electrochemical cell.
Consult Section 4 for more details on green
rusts.
Adsorbed Fed I)
The presence of mineral surfaces has been
shown to dramatically increase rates of con-
taminant reduction by Fe(II). Amonette et
al. (2000), Charlet et al. (2002), Pecher et al.
(2002), Szecsody et al. (2004), and Eisner et
al. (2004a) show that sorption of Fe(II) on the
iron oxyhydroxide goethite led to high rates of
carbon tetrachloride degradation.
Importantly, no reduction of carbon tetra-
chloride occurred in systems with Fe(II) but
no goethite, indicating that the degradation
processes were controlled at the mineral
surface. In their review of reactions involving
sorbed Fe(II) as an electron donor, Schoonen
and Strongin (2005) postulate that this increase
in reducing power on surfaces results from the
formation of a Me-O-Fe11 species, where Me
is structural iron in the mineral crystal. The
bond to the structural iron makes the sorbed
Fe(II) a stronger electron donor, because of the
electron density the structural iron is transferred
through the bridging oxygen atom to the sorbed
Fe(II). However, the mineral surface may also
contribute by providing a sink for Fe(III) that
otherwise would be in solution, thus chang-
ing the ratio of Fe(II) to Fe(III), which would
affect the thermodynamics of the Fe(II)/Fe(III)
couple and make Fe(II) in solution a stron-
ger reductant (Bill Batchelor, Texas A&M
University, College Station, Texas, personal
communication).
In general, the rate of contaminant degradation
determined in experimental studies is propor-
tional to the concentration of Fe(II) adsorbed at
the mineral surface as determined by chemical
extraction or from the measured disappearance
of Fe(II) from solution. Amonette et al. (2000)
suggested that a simultaneous transfer of two
electrons takes place from two adjacent Fe(II)
surface sites that react with a sorbed carbon
tetrachloride molecule, thus leading to a rate
dependence that is second-order with respect
to the concentration of carbon tetrachloride.
Although the exact controls and mechanisms
-------
of contaminant reduction are topics of continu-
ing research, it is clear that the primary role of
the goethite surface is to catalyze the reac-
tion by fixing the position of the two charged
reactants in a geometry that is suitable for
reaction with carbon tetrachloride (Amonette
et al., 2000). Mossbauer studies with 57Fe and
56Fe compounds have revealed interesting and
unanticipated complexity; sorbed Fe(II) leads
to electron transfer and reduction of structural
Fe(III) near the mineral surface (Williams and
Scherer, 2004).
Phyllosilicate Clays
Another class of iron-bearing minerals that are
common components of aquifer sediments but
have received less attention includes the iron-
bearing clay minerals biotite, montmorillonite,
and vermiculite (Lee and Batchelor, 2004) or
smectite (Neumann et al., 2009). Because iron-
bearing clays are often abundant soil compo-
nents, these minerals could play a significant
role in affecting the transport and fate of chlo-
rinated organic compounds. Lee and Batchelor
(2003) found that the reductive capacities of
iron-bearing phyllosilicates were from one to
three orders of magnitude less than those of
other iron sulfides and oxides. Neumann et al.
(2009) concluded that structural Fe2+ in smec-
tites will become important in the subsurface
only when the iron(hydr)oxides are reductively
dissolved. Brown et al. (2007) presented data
from a site (their Site A), where ground water
in an aquifer containing glauconitic marine clay
was contaminated with 1,1,1-TCA, 1,1-DCE,
and 1,1-DC A. The sediments were naturally
reducing. The concentrations of chlorinated
hydrocarbons were reduced ten thousand fold
as ground water moved 200 meters along the
flow path.
Iron Carbonates
The iron carbonate siderite (FeCO3) is closely
related to the minerals rhodochrosite (MnCO3),
magnesite (MgCO3), and calcite (CaCO3).
These minerals often have some substitution
of their metal ions. Siderite is a common
mineral associated with sedimentary rocks, bog
deposits, and coal fields, where it occurs as a
reddish-brown mineral in shale layers, nodules,
concretions, and fossil burrows. Although
little work has been carried out on evaluat-
ing the reactivity of siderite with chlorinated
compounds, it is included in this discussion
because it is expected to form in reducing,
nonsulfidic ground-water systems. Note that an
iron hydroxycarbonate mineral appears to be a
major species of carbonate precipitate within
granular zerovalent iron PRBs (Wilkin and
Puls, 2003). Studies to evaluate the controls
on precipitation of this phase are underway at
the Kerr Center (U.S. EPA/ORD) and it appears
that elevated pH (-10) in zerovalent iron PRBs
is an important factor.
-------
1.0
Formation and Stability of Reactive Iron and Sulfur
Minerals Under Various Geochemical Conditions
The minerals which support abiotic transforma-
tions of chlorinated compounds include macki-
nawite, pyrite, siderite, magnetite, green rust,
and goethite with adsorbed Fe(II). Table 1.1
describes the structure of these minerals.
Table 1.1 Mineralogical data for iron- and sulfur-bearing phases of interest.
Reactive
Mineral
General Description
Structural Image
Mackinawite
FeS (Tetragonal)
Specific gravity = 4.17
Each Fe(II) atom is bonded to four
sulfur atoms. Tetrahedra share edges
to form layers that are stacked and
interact by van der Waals forces.
Litharge (PbO) structure type.
Pyrite
FeS2 (Cubic)
Specific gravity = 4.95-5.10
Each Fe(II) atom is in octa-
hedral coordination with the
center of a disulfide unit.
NaCl structure type.
-------
Table 1.1. (cont.) Mineralogical data for iron- and sulfur-bearing phases of interest.
Reactive
Mineral
General Description
Structural Image
Siderite
FeCO3 (Trigonal)
Specific gravity = 3.96
Each Fe(II) atom is in octahedral
coordination with (CO3)2~ ions. The
unit cell is slightly distorted to
accommodate large planar (CO3)
groups which contain carbon
atoms at the center of an equilat-
eral triangle of oxygen atoms.
NaCl structure type.
Magnetite
Fe3O,4 (Cubic)
Specific gravity = 5.2
One third of iron atoms are in
tetrahedral coordination with
oxygen; two thirds of the iron
atoms are in octahedral coordina-
tion. One half of the six-fold
coordination iron atoms are Fe(II).
Spinel (inverse) structure type.
Green Rust
(Fe2+, Mg)6 Fe23+(OH)8-4H2O
(Fougerite; Trigonal)
Specific gravity = 3.5 (calc)
Fe(OH)2 layers alternate with inter-
layers of anions (chloride, sulfate,
or carbonate) and water molecules.
-------
Table 1.1. (cont.) Mineralogical data for iron- and sulfur-bearing phases of interest.
Reactive
Mineral
General Description
Structural Image
Goethite
FeOOH (Orthorhombic)
Specific gravity = 4.27-4.29
Oxygen atoms are arranged in a
sequence of hexagonal close-packed
arrays with Fe(III) in octahedral
interstices. Fe(II) adsorbed at
the mineral-water interface.
Diaspore structure type.
1.1 Formation Processes of Reactive
Minerals
This section discusses the relative stability and
conditions that are associated with the forma-
tion of the principal reactive iron minerals.
Because iron exhibits two oxidation states, Fe2+
and Fe3+, the stability of iron-bearing minerals
is a function of the redox state of the system,
which can be expressed in terms of Eh. In
addition, inclusion of several different elements
in the structure of these minerals requires the
consideration of other geochemical parameters,
such as pH, dissolved CO2, and dissolved H2S.
Figure 1.1 presents mineral stability diagrams
plotted using the geochemical parameters that
were identified as being of importance. Mineral
stability diagrams presented in Figure 1.1
were constructed with the EQ/36 thermody-
namic database, modified to include data for
green rust and iron sulfides. It is important
to recognize that mineral stability diagrams,
such as the one presented in Figure 1.1, are
constructed assuming that equilibrium is
established. Natural systems are typically not
at thermodynamic equilibrium, yet these types
of diagrams are highly useful in linking specific
geochemical conditions (e.g., pH, concentra-
tion of bicarbonate) with minerals that might
be expected to precipitate from solution.
Two diagrams (Figure 1.1 a and Figure Lib)
are plotted, one for sulfidic environments where
the activity of the HCO3~ anion is taken to be
10~3, and the other for non-sulfidic environments
where the concentrations of dissolved sulfide
are so low that pyrite and mackinawite do
not plot on the Eh-log HCO3~ diagram. These
diagrams represent a theoretical prediction of
natural settings, yet they are in overall qualita-
tive agreement with the results of experimental
and natural measurements of mineral forma-
tion. In aerobic, high Eh environments, only
Fe(III) oxyhydroxides (goethite) are expected
to form. Siderite is expected to form only at
low Eh, low H2S, and high HCO3-. Stability
fields for the carbonate form of green rust and
for magnetite are present in systems containing
low concentrations of inorganic carbon and
low Eh. At the specified conditions, green rust
-------
appears to be a stable phase. Carbonate green
rust occupies a redox region intermediate to
siderite and magnetite. Pyrite occupies a wide
stability field encompassing low Eh and low to
moderate concentrations of dissolved sulfide.
1.1.1 Formation of Pyrite and Mackinawite
Examination of Figure 1.1 a shows that pyrite
is the stable iron mineral under low Eh and
moderate to high concentrations of dissolved
sulfide. Such conditions are characteristic of
anaerobic ground-water systems where sulfate
reduction is on-going. The occurrence of
pyrite in aquifer sediments is nearly always
due to precipitation in place (authigenic forma-
tion) because detrital pyrite is chemically and
physically unstable. Discussion of iron sulfide
formation may be divided into issues relating to
the sources of iron and sulfur, the factors limit-
ing formation, and the mechanism of formation.
The principal source of iron for pyrite forma-
tion in subsurface environments is detrital
iron minerals. Surface waters contain almost
no dissolved iron and biogenic materials are
typically low in iron content. Detrital iron
minerals can dissolve by bacterial or abiotic
processes. The most common source of dis-
solved iron is from the chemical or biological
reduction of iron(III) hydroxides to produce
dissolved iron(II) (Poulton, 2003). However,
not all detrital iron minerals react completely
to form pyrite. The most reactive iron-bearing
phases are those that are very fine grained,
such as hydrous iron(III) oxide, fine-grained
goethite, and iron-bearing clay minerals
(e.g., Raiswell and Canfield, 1996; Poulton
et al., 2004). These fine-grained minerals
often occur as adsorbed coatings on detrital
silicates. In contrast, sand- and silt-sized
grains of magnetite and ferruginous silicates
are comparatively unreactive and can often
a)
-0.1
W -0.2
"o
c
LU
-0.3
-0.4
v
_Goethite\
\
A AltusAFB
_ Magnetite \ blOWQll
Greenrust-CO3 \
\ !
Siderite n-sMatikinawite
| J£Ss, HiPJ
25°C
2 -11 -10 -9 -8 -7 -8 -« -4 -
log a H2S(aq)
b)
OJJ
0.6
V) 0.4
"5
LU
0
-0.2
-0.4
3
- 25°C
o. .
H20
Goethite
-
~.
Magnetite r — """"
Greenrust-CO__<**"*"''^ Siderite
, H,
5-4-3-2
log a HCO3
Figure 1.1. Mineral stability diagrams: a) Eh versus log a H2S showing stability fields for goethite, magne-
tite, carbonate green rust, siderite, mackinawite, and pyrite (Fe=10'4, C=1&3, pH =7.5, hematite
suppressed); b) Eh versus log a HCO3 showing stability fields for goethite, magnetite, carbon-
ate green rust, and siderite (Fe=10-4, pH =7.5, hematite suppressed). The log a H2S and log a
HCO3 refer to the common logarithm of the chemical activity of H2S or HCO3, as opposed to their
concentration.
-------
co-exist, in a metastable state in conditions
that should support the precipitation of pyrite.
Pyrite forms as a result of a series of processes
stemming from the initial reaction between
dissolved sulfide and the iron released from
detrital iron minerals. The major source of
dissolved sulfide is bacterial reduction of
sulfate. Sulfate in ground water is usually
from dissolution of pyrite or other sulfides in
aerobic environments or from the leaching of
sulfate minerals that are components of certain
sedimentary rocks. Consequently, the primary
factors that limit the amount of pyrite, or any
iron sulfide that forms, are: the concentration
and reactivity of iron compounds, the availabil-
ity of dissolved sulfate, and the concentration of
organic carbon that acts as a carbon source for
sulfate-reducing bacteria to produce dissolved
sulfide. In natural systems, the concentration of
reactive organic carbon is oftentimes found to
be a limiting factor on the amount of bacterial
sulfate reduction that occurs. In some organic-
carbon-rich environments, where dissolved
sulfide is abundant, the amount of iron sulfide
precipitation is limited by the amount of reac-
tive iron minerals. In ground-water systems,
the concentration of dissolved sulfate, reactive
carbon, or reactive iron can all be limiting fac-
tors for iron sulfide precipitation (Figure 1.2).
In biowalls, the formation of iron sulfides
is not limited by the abundance of organic
carbon but can be limited by the availability
of either sulfate or iron (Figure 1.2). Thus,
understanding the primary controls on reactive
mineral formation is essential for designing and
implementing remedial systems in the field.
Carbon Limitation
Iron Limitation
Sulfur Limitation
Figure 1.2. Ternary diagram showing the impor-
tance of controlling factors on iron
sulfide formation in natural aquifers
and biowalls.
It is now well known that pyrite formation
proceeds through the formation of metastable
iron monosulfide precursors. This has been
shown by laboratory experimentation conducted
to simulate a range of natural environments.
Minerals that form initially at room tem-
perature by the reaction of H2S and HS~ with
fine-grained goethite or dissolved ferrous iron
are mackinawite (tetragonal FeS) and greigite
(cubic Fe3S4). Both minerals are black and
soluble in concentrated hydrochloric acid.
Mackinawite and greigite are thermodynami-
cally metastable relative to pyrite (and other
iron sulfides such as troilite and pyrrhotite).
Although pyrite has a wide stability field
on Figure l.la, its formation is impeded by
relatively slow nucleation kinetics, unlike
mackinawite which readily precipitates when
solutions reach saturation. This fundamental
difference in nucleation kinetics between
mackinawite and pyrite represents one of
the limitations of equilibrium diagrams such
as Figure l.la to accurately depict mineral
formation at low temperatures. The stability
field of the thermodynamically stable phase,
pyrite, is necessarily expanded over the meta-
stable phase, mackinawite. The mechanism
of the mackinawite-to-pyrite transformation
has been the subject of many laboratory
and field investigations (Schoonen, 2004;
-------
Rickard and Luther, 2007). He et al. (2008)
discuss the implication of this transformation
process relative to dechlorination of TCE on
iron sulfides. Pyrite is generally considered
to be much less reactive toward chlorinated
hydrocarbons than are iron(II) monosulfides.
He et al. (2008) measured Eh and dissolved sul-
fide concentrations in ground water from a plant
mulch based passive reactive barrier (a biowall)
at the OU-1 site at Altus AFB, OK. They also
sampled and characterized iron and sulfur
mineral phases that were accumulating in the
biowall matrix. Iron and sulfur minerals were
dominated by pyrite and mackinawite. The
range of values for Eh and the chemical activity
of H2S in ground water in the OU-1 biowall are
presented in Figure 1. la. Consistent with the
observed accumulation of pyrite and macki-
nawite, the Eh and chemical activity of H2S
in water in the biowall fall within the stability
fields of pyrite and mackinawite in Figure 1.1 a.
1.1.2 Formation of Siderite
The conditions under which siderite is stable
are limited. Figure 1.1 a and Lib show that Eh
must be low (<0 mV) and dissolved sulfide con-
centrations must be low as well (<10~73 molar),
a combination that would not be expected for
reducing ground water systems that contain any
appreciable concentrations of sulfate. Thus,
siderite formation is expected in anaerobic
environments only where sulfate is absent. In
addition to waters containing dissolved sulfide,
siderite is not expected from water rich in
dissolved calcium. Following the reaction:
CaCO3 + Fe2+ = FeCO3 + Ca2+
So that
= {Ca2+}/{Fe2+}
For siderite to be stable relative to calcite the
concentration of iron must be greater than
-1% of the concentration of calcium (Wilkin
and Puls, 2003). Substantial concentrations
of calcium ion can prevent the precipita-
tion of siderite in ground-water systems.
1.1.3 Formation of Green Rusts
Green rust is a mixed Fe(II)-Fe(III) hydroxide
consisting of positively charged trioctahe-
dral metal-hydroxide layers and negatively
charged interlayer anions and water molecules
(Table 1.1). It is believed to be an intermediate
reaction product as reduced minerals oxidize
to mixed ferric/ferrous or ferric minerals
such as magnetite and goethite (Figure 1.1 a).
Green rust minerals have been prepared in
the laboratory, using various chemical pro-
cedures, from solutions containing iron(II)
and chloride ions, iron(II) and sulfate ions,
and iron(II) and carbonate ions (e.g., Taylor
et al., 1980; Hansen, 1989; Drissi et al.,
1995; Ayala-Luis et al., 2008). Studies have
shown that green rust minerals form from the
bioreduction of hydrous iron(III) oxides by
dissimilatory iron-reducing bacteria. Green
rust has been identified using transmission
electron microscopy as a corrosion product in
zerovalent iron PRBs (Furukawa et al., 2002).
Two types of green rust (GR) minerals are
distinguishable based upon X-ray diffraction
analysis: GR1 in which the distance between
hydroxide sheets is between about 0.75 and
0.80 nm (e.g., carbonate green rust) and GR2
in which the distance between sheets is about
1.1 nm (e.g., sulfate green rust). Solid-phase
characterization and geochemical modeling
studies of iron corrosion in zerovalent iron
passive reactive barriers (PRBs) indicate
that if the ground water is not impacted by
mine waste, the carbonate forms of green
rust will precipitate preferentially over the
sulfate form in zerovalent iron systems
(Wilkin and Puls, 2003). These observa-
tions can be understood by considering the
following exchange equilibrium based upon
the anhydrous components of green rust:
Fe6(OH)12S04 + C032 = Fe6(OH)12CO3 + SO42
So that
^={S02-}/{C02-}
-------
Assuming ideal mixing relations in the solids
(i.e., the activities of the solids are taken to
be unity) and using available thermodynamic
data from Bourrie et al., 1999, we estimate
A: = 103.1. Typically the [SO42- / CO32] ratio
in ground water that has not been impacted
by mine waste is less than 103.1, which is
consistent with the predominance in aqui-
fers and most reactive barriers of carbon-
ate green rusts over sulfate green rusts.
1.1.4 Formation and Transformation of
Goethite and Ferrihydrite
Goethite (FeOOH) and ferrihydrite (Fe(OH)3)
are the most common Fe(III) (oxy)hydroxides in
aquifers. Note that if the occurrence of goethite
were to be suppressed from Figure Lib then
Fe(OH)3 would occupy nearly the same region
as goethite. These iron(III) minerals often have
high surface area and intrinsic reactivity toward
various metals and nutrients. Ferrihydrite is
generally considered to be the most labile and
most bioavailable Fe(III) hydroxide or Fe(III)
oxide for dissimilatory iron-reducing bacteria.
Ferrihydrite can undergo a number of trans-
formation pathways leading to other minerals,
some of which are of interest in the context of
abiotic reductive dechlorination. Reaction of
ferrihydrite with dissolved Fe(II), derived from
the reductive dissolution of Fe(III) minerals,
converts ferrihydrite to Fe phases such as goe-
thite, lepidocrocite, green rust, and magnetite.
The mineralization pathways are influenced by
the concentration of dissolved Fe(II), pH, and
concentration of ligands in solution. Magnetite
formation from ferrihydrite is favored at high
loadings of Fe(II) (>1 mmol Fe(II) per gram)
and in the absence of bicarbonate. Magnetite
precipitation is also favored at higher pH. At
lower surface loadings of Fe(II), ferrihydrite
converts to the oxyhydroxides, goethite and lep-
idocrocite depending on the presence or absence
of ligands in solution (Hansel et al., 2005).
A geochemical environment of particular
interest is the transitional redox regime
where Fe(II) production occurs, but Fe(III)
oxyhydroxides still remain. The widespread
occurrence of iron-reducing bacteria may
provide a continuous supply of Fe(II) ions at
the surface of minerals that may enhance the
natural attenuation of ground-water contaminant
plumes (Williams and Scherer, 2004; Tobler
et al., 2007). This coupling of microbial
respiration and abiotic reduction of organic
compounds may allow for simultaneous natural
attenuation of different chemical classes.
1.2 Geochemical Modeling
Reactive minerals can be unstable in the pres-
ence of oxygen in the atmosphere, and they may
be present in aquifer sediments at low concen-
trations. As a consequence, the direct identi-
fication of reactive minerals in raw samples
collected from the field is a difficult task. Most
of the successful approaches to identify reactive
minerals in aquifer material include a process
to concentrate the reactive minerals through the
use of density or sizing separations, followed
by microscopy or other solid-phase analyses.
In some cases the presence or absence of
reactive minerals can be predicted using
thermodynamic calculations. Thermodynamic
calculations based on measured concentra-
tions of chemicals in ground water can provide
clues to the potential occurrence of specific
minerals in the aquifer sediment (Billon et al.,
2003). Geochemical speciation modeling is a
tool that can be used to constrain whether or
not specific minerals that could support abiotic
transformations are present in a system. In
essence geochemical speciation models use
water chemistry to constrain the mineralogical
composition of the mineral phases present.
It is important to keep in mind that the results of
geochemical modeling exercises are not defini-
tive, but they can be used as supporting lines of
evidence and provide important insight about
the geochemical environment in the subsurface.
The approach is more robust when the modeling
encompasses water samples collected through
time. Use of modeling is most successful
-------
when specific questions are posed, i.e., is it
possible that an iron sulfide is precipitating in
a reactive barrier? It is imperative that models
be supported with the appropriate field data
and thermodynamic data. A specific example
of modeling is presented here to explore the
formation of reactive iron minerals in a plant
mulch based passive reactive barrier (a biowall)
that was installed at the OU-1 site on Altus
AFB, Oklahoma, to treat TCE in ground water.
In this example, the equilibrium geochemical
modeling software Geochemist's Workbench
(Release 6.0, Rockware, Golden, CO) was
used to model ground water chemical specia-
tion. The "React" module was used to model
chemical speciation in aqueous solutions and
determine the saturation states of solutions with
respect to selected iron sulfides and iron car-
bonates. The modeling provides a quantitative
means of assessing whether or not solutions are
undersaturated, oversaturated, or saturated with
respect to specific mineral phases. Solutions
that are at saturation (equilibrium) with respect
to a specific mineral provide a case that the
mineral could be present within the aquifer
material (in this case the biowall) that is in con-
tact with that water. Solutions that are under-
saturated with respect to a particular mineral
can be expected to dissolve that mineral into
solution. Solutions that are oversaturated tend
to precipitate, assuming that kinetic barriers of
nucleation do not prevent mineral precipitation.
The standard database (thermo.dat) of
Geochemist's Workbench was applied with
important modifications to include ther-
modynamic data from Rickard (2006) for
the solubility of crystalline mackinawite
[FeS(c)] and Benning et al. (2000) for
the solubility of disordered mackinawite
[FeS(d)] following the reactions:
FeS(c) + 2H+= Fe2++ H2S log K= 3.50
and
FeS(d) + 2FT = Fe2+ + H S log K= 3.98
By examining both of these reactions, it is
possible to compare the composition of ground
water with solutions that are saturated against
well-crystalline and freshly precipitated FeS.
As shown in Table 1 .2, inputs to the model
were pH, and the concentrations of Na, K, Ca,
Mg, Fe, Cl, SO42 H2S(aq), and HCO3-. It was
found that the high concentrations of sulfide
in the biowall necessarily drove the concentra-
tion of iron to very low levels. Non-detects
cannot be used as model inputs. The concen-
trations in the water sample must be above
the quantitation limit for the method used to
analyze the samples. It was necessary to use
inductively coupled plasma mass spectrometry
(ICP-MS) to determine the concentrations of
iron in ground water samples from the bio-
wall. Note that the iron concentrations reported
in Table 1.2 have units of parts per billion.
These measurements are well below typical
concentration ranges possible using spectro-
photometric or emission-based ICP methods.
Based on the distribution of the measured
concentrations of the major anions and cations,
the charge balance was calculated and found
to be generally better than 10% (which is
suitable for the modeling). In this case, since
we are interested in understanding the solu-
tion saturation states of the iron monosulfides
and iron carbonate, the redox state was not
fixed using the Eh value measured in the field.
Instead, sulfate and sulfide were decoupled,
i.e., not linked by redox equilibrium restric-
tions. In this way the field-determined Eh
could be compared to a calculated Eh based
on the measured concentrations of sulfide and
sulfate and the pH following the reaction:
SO42
9FT + 8e = HS + 4HO
Comparing results for the measured and
calculated Eh values shows fairly good agree-
ment, particularly for samples containing high
dissolved sulfide concentrations (>100 mg/L;
Table 1.2). The platinum electrode response
in this system is likely reflecting the sulfide/
elemental sulfur pair (Berner, 1963).
-------
Table 1.2. Field data and results of geochemical speciation modeling of a biowall at the OU-1 site on
Altus AFB, OK.
Field Data
Well
EPA110
EPA 109
EPA 108
EPA 107
EPA 106
EPA 105
EPA 104
EPA 103
EPA 102
EPA101
pH
SU
6.58
6.77
6.81
6.60
6.64
6.73
6.77
6.75
6.70
6.98
Eh
m
mV
-171
-206
-209
-194
-185
-185
-230
-196
-195
-70
Na
ppm
428
389
396
366
383
346
383
370
379
398
K
ppm
32
9.0
6.8
16.2
19.9
3.9
21.2
11.9
27.6
4.9
Ca
ppm
386
333
353
373
407
396
444
406
447
395
Mg
ppm
119
112
107
93.9
145
138
152
158
156
155
Fe
ppb
48.7
11.3
27.5
4.60
170
19.2
291
53.0
68.2
11.2
ci-
ppm
289
299
298
264
308
293
289
253
269
301
so/
ppm
27
89
355
67
686
1030
389
806
319
1600
DIG
ppm
523
398
336
414
325
229
352
311
493
96
IH2S
ppm
164
344
208
188
88
121
210
190
178
1
Model Outputs
Well
EPA110
EPA 109
EPA 108
EPA 107
EPA 106
EPA 105
EPA 104
EPA 103
EPA 102
EPA101
Ehc
mV
-198
-210
-207
-197
-190
-196
-202
-200
-199
-199
Charge
Balance
%
-8.4
-11.5
-6.8
-7.1
-3.3
-6.4
-5.9
-5.8
-7.5
0.1
Pyrite
Mackinawite
(crystalline)
Mackinawite
(disordered)
Gypsum
Siderite
Saturation Index (SI)
9.42
9.03
9.37
8.53
10.0
9.19
10.5
9.75
9.77
7.03
-0.39
-1.00
-0.62
-1.41
0.13
-0.79
0.40
-0.34
-0.23
-1.45
-0.87
-1.48
-1.09
-1.89
-0.35
-1.27
-0.08
-0.82
-0.72
-1.93
-2.09
-1.60
-1.00
-1.67
-0.71
-0.55
-0.69
-0.65
-1.01
-0.40
-2.46
-3.57
-3.06
-3.65
-1.91
-2.99
-2.02
-2.76
-2.61
-2.39
Notes: All samples collected on 11/19 and 11/20, 2008. Samples for cations and anions were filtered through 0.45 ^m disc-filters in the
field. Samples for analysis of cations were acidified with HNO3 for preservation. All samples were stored on ice or in a refrigerator
prior to analysis. The saturation index (SI) is used to quantify the degree of saturation of a mineral and is defined as follows: SI = log
(lAP/Ksp). Where IAP is the ion activity product and Ksp is the solubility product constant for the phase in question. For phases at
saturation, IAP=Ksp and SI = 0. A negative SI indicates that the phase is undersaturated (IAPKsp)
indicates the phase is oversaturated. The analyses were checked for internal consistency using charge balance. The charge
balance was calculated as follows: (L (meq/L positive charge) -1, (meq/L negative charge))/ (L (meq/L positive charge) + E (meq/L
negative charge)) x 100. Ehm and Eh? are the field determined Eh and calculated Eh values, respectively. Calculated Eh values are
based on the sulfate/sulfide redox pair.
-------
The saturation states of the iron minerals were
evaluated to determine if iron sulfides were
likely to be forming within the biowall. We
focused on the iron sulfides pyrite and macki-
nawite. Other iron sulfides, such as marcasite,
troilite, and pyrrhotite are not likely candidates
in this system. Marcasite often precipitates
from low pH solutions. Troilite and pyrrho-
tite are simply not encountered as formation
products in low-temperature systems. The
results of the modeling for the iron minerals
are provided in Table 1.2. Ground water within
the biowall is nearly saturated with respect to
iron monosulfide. Mean saturation indices for
crystalline mackinawite and disordered macki-
nawite are -0.14 and -0.82, respectively (n=28).
Figure 1.3 shows the SI values for crystalline
mackinawite and disordered mackinawite as a
function of pH. Notice that nearly half of the
SI values calculated for crystalline mackinawite
are >0, suggesting over saturation and that
SI=0 is a practical cap for values reflecting the
saturation state of disordered mackinawite.
s
~3T
'5
10
c
Ix:
o
TO
CO
"o"
"i
ro
,E
V
o
ffl
•
CO
1.5-
1.0-
0.5-
0.0-
-0.5-
-1.0-
-1.5-
-20-
1 5
1.2-
0.9-
0.6-
0.3-
0.0-
-0.3-
-0.6-
-0.9-
-1.2-
-1 5 -
a) * April 2006
» June 2007
^ November 2007
oversi'duii-iled m ^
imderaaliiraled % \
* *
* ^ * * * *
'* *
»
* *
1 r t i ' i • i '
b)
• £
% * * %
ov,ra,Ur*d . »" *
i vere* ur*n - ^
undersoSuroied * • •
•
9
•
* *
6.0
6.4 6.6 6.8
pH
7.0
Figure 1.3. Saturation indices (SI) for a) disor-
dered mackinawite and b) crystalline
mackinawite as a function of pH and
time (ground water from the biowall at
OU-1 on Altus AFB, OK).
These features are consistent with the formation
and aging of mackinawite in the biowall. The
saturation state of freshly precipitated macki-
nawite (disordered) should not be exceeded
because rapid precipitation kinetics are expected
to drive solution concentrations down to the
solubility limit. Clustering of SI values around
SI=0 for crystalline mackinawite suggests that
freshly precipitated and aged mackinawite may
control the dissolved sulfide and iron concentra-
tions in the biowall. Note that the disulfide
pyrite is highly oversaturated in the biowall
(mean SI= 9.6). This is expected because of
kinetic barriers to direct precipitation of pyrite
in aqueous solutions (Schoonen and Barnes,
1991). Even though the precipitation of pyrite
does not regulate concentrations of sulfide and
iron, there is a thermodynamic driving force
for the mackinawite-to-pyrite transformation.
These interpretations reinforce the results
of solid-phase investigations in the biowall
reported by He et al. (2008). The important
implication is that geochemical speciation
modeling, based on water chemistry param-
eters, can be useful in evaluating whether or
not potentially reactive phases are present
in situations where solids cannot be col-
lected and analyzed for mineral content.
The saturation states of the gypsum (calcium
sulfate) and siderite phases were also evalu-
ated in order to determine what reactions likely
control the sulfate concentrations within the
biowalls and whether iron carbonate formation
was likely to compete with iron sulfide precipi-
tation. Sulfate is important as an electron accep-
tor for sulfate-reducing bacteria (SRB). Ground
water within the biowall is undersaturated
in both gypsum and siderite. Consequently,
precipitation of a calcium sulfate is not a sink
for dissolved sulfate in the biowall; this may
be linked to the fact that ground-water is likely
supporting the precipitation of calcium carbon-
ate (calcite or aragonite) which is oversaturated
in the biowall. Siderite is highly undersaturated
so the primary precipitation reactions involv-
ing iron are those that form iron sulfides.
-------
2.0
Iron Sulfide Minerals
2.1 Applications of Iron Sulfide Minerals
Iron sulfides are important redox active minerals
that can form under anaerobic conditions. They
are found in engineered systems such as zero
valent iron passive reactive barriers, at contami-
nated sites that contain mixed wastes, and in
natural sediments. Engineering systems where
iron sulfide minerals are expected to play an
important role in the abiotic degradation of chlo-
rinated organic contaminants include permeable
reactive barriers (PRBs), landfills, waste disposal
facilities, and sites where chlorinated organic
compounds are treated by anaerobic bioreme-
diation. In addition, iron sulfide minerals may
contribute to abiotic degradation of chlorinated
organic contaminants in anaerobic sediments in
natural wetlands and salt marsh environments.
In natural sediments, mixed plumes, and engi-
neered systems, FeS is generally formed from
Fe2+ and HS" that is produced in these systems.
Under anaerobic conditions, sulfate reducing
bacteria (SRB) generate HS". The Fe2+canbe
produced by dissimilatory iron reducing bacteria
or through reductive dissolution of hematite and
other iron oxides by hydrogen sulfide (Pyzik and
Sommer 1981). The production of dissolved sul-
fide species in the presence of dissolved metals
can result in precipitation of metal sulfide min-
eral phases as shown in the following reaction:
H2S + Fe2+ = FeS + 2FT
Under ambient conditions, disordered macki-
nawite or FeSam precipitate is usually the first
iron sulfide phase formed (Rickard and Luther
2007; Rickard and Morse 2005). Amorphous
FeS and poorly crystalline mackinawite (tetrago-
nal structure) has been observed in PRB systems
constructed with zero valent iron (ZVI) and in
biowall systems constructed with plant mulch
(Benner et al. 1999; Phillips et al. 2000). In nat-
ural settings, iron monosulfide (FeS) formation
is commonly observed in anoxic sediment envi-
ronments. The HS" produced by sulfate reduc-
ing bacteria may also directly react with iron
oxides (e.g. hematite, goethite, ferrihydrite) and
produce iron sulfides on the mineral surfaces.
Iron sulfide minerals mediate reductive trans-
formation of a variety of halogenated organic
compounds (Butler and Hayes 1998; Butler and
Hayes 2000; Jeong and Hayes 2003; Jeong et
al. 2007; Jeong et al. 2008). The transformation
of chlorinated hydrocarbons such as TCE, PCE,
carbon tetrachloride (CT), and trichloroethane
(TCA) by a variety of FeS and FeS2 minerals
is well documented in the literature (Butler and
Hayes 1999; Butler and Hayes 2001; Liang et
al. 2007; Shen and Wilson 2007; Butler and
Hayes 1998; Butler and Hayes 1999; Butler and
Hayes 2000; Butler and Hayes 2001; Gander et
al. 2002; Hansson et al. 2008). Degradation of
hexachloroethane (HCA), tetrachloroethylene
(PCE), and trichloroethylene (TCE) by nickel
and copper sulfides has also been observed under
certain conditions (Butler and Hayes 1998).
Biotite and vermiculite (Kriegman-King and
Reinhard 1992), and kaolinite (Hanoch et al.
2006) have been shown to enhance dechlorina-
tion by forming new FeS phases in the system.
This section reviews the abiotic degradation
of different groups of halogenated aliphatic
hydrocarbons by iron sulfide minerals, includ-
ing discussion of the rate of degradation, the
products of degradation, and the conditions
associated with the degradation processes.
The compounds will be discussed in four
separate groups: (1) the chlorinated ethylenes
-------
including PCE, TCE, cis-DCE , and vinyl
chloride, (2) chlorinated alkanes including
1,1,1-TCA, 1,1,2-TCA, 1,1-DCA, and 1,2-DCA,
(3) chlorinated methanes including carbon
tetrachloride, chloroform, and dichlorometh-
ane and (4) EDB (1,2-dibromoethane).
2.2 Abiotic Transformation of
Chlorinated Ethylenes
2.2.1 Transformation Products
Abiotic transformation kinetics are usually
expressed in the literature as a first order rate
constant or a surface area normalized first order
rate constant. The first order rate constant or
pseudo first order rate constant is obtained from
fitting the experimental data to a first order rate
law. A surface area normalized rate constant is
calculated by dividing the first order rate con-
stant by the surface area of the mineral (m2/L)
in suspension in the solution. The reactions
occur on surfaces of the minerals, and the rate
of reaction should be proportional to the surface
area of the mineral presented to the water, and
not to the bulk concentration of the mineral
suspended in the water. The surface area
normalized rate constants may provide a better
comparison of rate constants, especially when
the minerals vary widely in their specific surface
area (m2/g). However, information on specific
surface area is not always readily available.
Table 2.1 lists rates reported in the literature
for abiotic degradation of chlorinated ethylenes
by FeS phases. Most studies investigated the
degradation of chlorinated ethylenes by FeS
and pyrite; very few studies included other FeS
phases such as troilite, or greigite. Most of
these studies focus on PCE and TCE degrada-
tion; very few studies included DCE and vinyl
chloride degradation by iron sulfide minerals.
Vogel et al. (1987) identified a variety of abiotic
and biotic processes that could degrade chlo-
rinated aliphatics, including hydrogenolysis,
dihalo-elimination (loss of two adjacent chlo-
rines forming a C-C bond), and coupling (loss
of chlorines on two separate molecules forming
a C-C bond, joining the two molecules). They
also identified other abiotic processes such as
dehydrohalogenation and hydrolysis. For TCE,
transformation by FeS or FeS2 seems to occur
via parallel reaction pathways. TCE is trans-
formed by reductive elimination to acetylene
via chloroacetylene as an intermediate, or by
hydrogenolysis to cis-DCE (Butler and Hayes
1999; Jeong et al. 2007; Lee and Batchelor
2002a). Acetylene was reported to be the major
reaction product in the reductive transformation
of TCE by FeS (accounting for 70 to 80% of
removed TCE), followed by cis-DCE (account-
ing for 7%), and ethylene, ethane, and other
C2 to C6 hydrocarbons (accounting for 15 to
20%). The C4 and C6 minor products were
attributed to coupling of radical intermediates
formed in the TCE transformation process.
Sivavec and Horney (1997) reported a similar
product distribution for transformation of TCE
by the FeS mineral phase troilite. Acetylene
was also detected as a product in the reac-
tion of TCE and PCE with nickel and copper
sulfides (Butler and Hayes 1999). Lee and
Batchelor (2002a) found the main transforma-
tion product for TCE dechlorination by pyrite
was acetylene, accounting for 86% of TCE that
was removed. Minor products were cis-DCE
(6.6% of TCE) and ethylene (4.4% of TCE).
Butler and Hayes (1999) proposed that PCE
was transformed by FeS via parallel reac-
tion pathways to acetylene, TCE, and cis-
DCE. PCE was transformed to acetylene
eight times faster than it was to TCE. No
vinyl chloride was detected over the course
of the experiment, which is likely due to
only minimal accumulation of cis-DCE.
Lee and Batchelor (2002a) showed that the
main transformation product of PCE on pyrite
is acetylene, which accounted for 71.1% of
PCE that was degraded. Minor products were
ethylene (13.3%) and TCE (14.9%). Neither
cis-DCE nor vinyl chloride was observed.
PCE could be transformed to TCE, followed
by the parallel transformation to acetylene
and cis-DCE. The reductive dechlorination of
-------
cis-DCE in pyrite suspension produced acety-
lene and ethylene, which accounted for 44.3%
and 17.5% of removed cis-DCE, respectively.
Neither vinyl chloride nor ethane was observed.
However, a trace amount of vinyl chloride was
detected as a transformation product of cis-
DCE and trans-DCE by zerovalent Fe° and Zn°
(Arnold and Roberts 2000). To our knowledge,
there is no investigation of DCE degrada-
tion by FeS that has been reported to date.
Lee and Batchelor (2002a) proposed that
dechlorination of vinyl chloride by pyrite
followed a hydrogenolysis pathway to produce
ethylene and ethane, with ethylene account-
ing for 77.3% of vinyl chloride that was
removed. Ethylene was further reduced to
ethane in the suspension, accounting for 11.4%
of the target organic removed. Acetylene was
not observed throughout the experiment.
No report on vinyl chloride degradation
by FeS has been published.
Table 2.1. Rate Constants and products of abiotic degradation of chlorinated ethylenes by FeS minerals.
Chlorinated
solvents
PCE
PCE
PCE
PCE
PCE
TCE
TCE
FeS mineral
phases
Fe2+ and HS~
in solution
FeS (freeze
dried)
Mackinawite
(FeS)
FeS
Pyrite
FeS (freeze
dried)
FeS (freeze
dried)
Degradation
product
None detected
TCE, acetylene
via parallel reac-
tion pathways,
minor cis-DCE
Acetylene, TCE,
c/5-DCE,
1,1 -DCE
cis-DCE, TCE,
ethylene
TCE, acetylene,
ethylene
Acetylene,
cis-DCE via
parallel reaction
pathways, minor
amount of
vinyl chloride
Acetylene,
cis-DCE, vinyl
chloride
Conditions
pH 7.0, O.lm
NaHCO3
1.85 mM FeCl2
2.06 mM Na2S
pH 8.3, 10
g/L, 0.5m2/L
amended 10~3
M cysteine
pH8.3, lOg/LFeS
0. 1M tris buffer
w/O.OlM FeCl2
w/O.OlM CoCl2
w/O.OlM NiCl2
w/O.OlM HgCl2
pH7, lOg/L
pH8, lOg/L
pH9, lOg/L
pH8, 84 g/L
2340 m2!/1
pH 8.3, 10
g/L, 0.5m2/L
amended 10"3M
cysteine
pH 7.3, lOg/L
pH 9.3, lOg/L
ImM MnCl2
2,2'-bipyridine
1% 2-propanol
Rate constant
Not detected
<10% in 33 days
<10% in 33 days
(1.37+0.02)xlO-2 d'1
2.74 xlO'2 L nr2 d'1
(3.67+0.08)xlO-3 d'1
7.34 xlO'2 L nr3 d'1
(1.82+0.24)xlO-2 d'1
(LlS+O^^xlO-2^1
(2.13+0.18)xlO-1d-1
(1.05+0.22)xlO-2 d'1
(3.69+0.38)xlO-2 d'1
(6.3+1. 6)xlO'5 L nr2 d'1
(5.3+0.5 l)xlO-4 L nr2 d'1
(1.21+0.12)xlO-3Lm-2d-1
(1.01+0.02) d'1
1.97xlO-5 L m-2 d'1
(3.58+0.34)xlO-2d-1
7.16xlO-2Lm-2d-1
(1.03+0. 12)xlO-2d-'
2.06 xlO'2 L m-2 d'1
(l^O+O^^xlO'M-1
(4.80+1.03)xlO-2d-1
(3.38+0.29)xlO-2d-1
(3.02+0.21)xlO-2d-1
(2.81+0.62)xlO-2d-1
Reference
Doong and
Wu 1992
Butler and
Hayes 1999
Jeong and
Hayes 2007
Liang et
al. 2007
Lee and
Batchelor 2002a
Butler and
Hayes 1999
Butler and
Hayes 2001
-------
Table 2.1. Continued.
Chlorinated
solvents
TCE
TCE
TCE
TCE
TCE
TCE
TCE
cis-DCE
1,1 -DCE
cis-DCE
1,1 -DCE
Vinyl
chloride
FeS mineral
phases
FeS
Mackinawite
Pyrite
Pyrite
Pyrite
FeS, not
freeze dried
FeS, freeze
dried
FeS
FeS
Pyrite
FeS (freeze
dried)
Pyrite
Degradation
product
cis-DCE,
vinyl chloride,
ethylene
Acetylene, cis-
DCE, 1,1 -DCE
acetylene,
ethylene
Organic acid,
CO,, Cl
2.
cis-DCE,
acetylene,
ethylene
Acetylene,
DCE
Acetylene,
DCE
none detected
none detected
acetylene,
ethylene
none detected
C2H4, C2H6
Conditions
pH8, lOg/L FeS,
HEPES buffer
pH9, lOg/L,
CHES buffer
pH 8.3, lOg/L
FeS, O.lMtris
buffer
w/0.01MFeCl2
w/0.01MCoCl2
w/0.01MNiCl2
w/0.01MHgC!2
2g/L
ionic strength
0.01 Mwith
NaClO4, pH
adjusted with
NaOH or TRIS
or PIPES buffer
lOOg/Lpyrite,
DO 0.54 to
8.6 mg/L
pH 8, 84 g L-1
2340 m2 L-1
pH7.2, 18g/LFeS
pH7.2, 18g/LFeS
pH 8.3, lOg/L
FeS, 0.1M
tris buffer
pH 8.3, lOg/L
FeS, 0.1M
tris buffer
pH 8, 84 g L-1
2340 m2 I/1
pH8.3, 10
g/L, 0.5m2/L
pH 8, 84 g L-1
2340 m2 L-1
Rate constant
(1.61±0.19)xlO-4Lm-2d-1
(6A±0.8\)xW-4Lm-2d-1
(5.09±0.24)xlO-2d-1
(2.01±0.17)xlO-2d-1
(1.01±0.09)xlO-2d-1
(3A3±O.U)xW-2d-1
(2.71±0.17)xlO-2d-1
pH 4.3, 0.182^,
pH 5.3, 0.182 d'1,
pH 6.4, 0.293 d-1
pH 7.4, 0.552 d-1
pH 8.2, 0.866 d-1
pH8.9, 1.21 d-1
0.096100.312^
(1.6010.02)^
2.53 x 10-5 Lm-2^1
(1.52±0.10)xlO-1d-1
(7.18±3.69)xlO-3d-1
Degradation
not detected
Degradation
not detected
(0.98±0.02) d-1
1.32xlO-5Lm-2d-1
Degradation
not detected
(1.7110.06)^
2.27xlO-5Lm-2d-1
Reference
Liang et
al. 2007
Jeong and
Hayes 2007
Weerasooriya
and
Dharmasena
2001
Pham et
al. 2008
Lee and
Batchelor
2002a
Heetal. 2010
Heetal. 2010
Jeong et
al. 2007
Jeong et
al. 2007
Lee and
Batchelor
2002a
Butler and
Hayes 1999
Lee and
Batchelor
2002a
-------
2.2.2 Rates of Degradation on Iron Sulfide
Minerals
Table 2.1 lists the rate constants for degrada-
tion of PCE, TCE, DCE, and vinyl chloride by
common iron sulfide minerals under different
conditions. The FeS used in most of these stud-
ies was chemically synthesized in the laboratory,
while all FeS2 used in the experiments was
hydrothermal pyrite from natural deposits.
Previous studies established some general
trends with regard to degradation of different
compounds by iron sulfide minerals. TCE
seems to be generally degraded faster than PCE.
Butler and Hayes (1999) reported a corrected
pseudo first order rate constant (kobj ) for PCE
degradation by FeS of (5.7 + l.O/x' 1O4 tr1.
This value is lower than the kobj value for
TCE of (1.49 + 0.14) x 1O3 tr1." Pseudo first
order rate constants for degradation of TCE
and PCE, normalized for surface area of FeS
(mackinawite) in suspension, were 7.16 xlO~2
and 2.74 xlO~2 L m~2 d"1 respectively. Similar
results for relative rates of degradation of
chlorinated ethylenes have been reported on
mackinawite (Butler and Hayes 1999) and
zerovalent metals (Arnold and Roberts 1998).
Degradation of PCE and TCE on pyrite is
slower than degradation by the other iron sulfide
minerals. Lee and Batchelor (2002a) investi-
gated abiotic dechlorination of chlorinated ethyl-
enes by pyrite, and they found that rate constants
for TCE were the highest, followed by those
for vinyl chloride, PCE, and cis-DCE. Pseudo
first order rate constants for degradation of TCE
and PCE, normalized for surface area of FeS2
(pyrite) in suspension, were 2.53 xlO~5 L m~2 d"1
and 1.97 xlO~5 L m~2 d"1. Rate constants for cis-
DCE and vinyl chloride were 1.32 xlO~5 L m~2 d"1
and 2.27 xlO~5 L m~2 d"1. The rates of degrada-
tion by pyrite (Lee and Batchelor, 2002a) are
from 1390 to 2900 times lower than those
reported for degradation on mackinawite and
troilite (Butler and Hayes 1999; Sivavec and
Horney 1997). These results indicate that
pyrite is less reactive than meta-stable iron
sulfides for the reductive dechlorination of
chlorinated ethylenes in natural and engi-
neered systems (Lee and Batchelor 2002a).
2.2.3 Factors Affecting Degradation
To understand how different experimental
conditions affect degradation, earlier studies
investigated the effect of a number of environ-
mental parameters such as pH, the concentration
of sulfide in solution, and the concentration
of organic matter on the rates of degradation
of chlorinated ethylenes. These observations
may help identify and optimize conditions to
attain efficient dechlorination at field scale.
2.2.3.1 Effect of Freeze Drying FeS on Rate
Constants for Tranformation of TCE.
Most of the studies in the literature were con-
ducted with FeS that was freeze dried after it
was prepared. Figure 2.1 presents data from
He et al. (2010) that compares the degradation
of TCE by freeze dried FeS (FD-TCE) and
FeS that was not freeze dried (for experimental
details, see Appendix A. 1). In the presence of
FeS that was not freeze dried, TCE degraded
quickly and after 28 days almost all the TCE
was gone. With freeze dried FeS, concentrations
of TCE decreased steadily over time, but at a
much slower rate than non-freeze dried FeS.
Even after 56 days of incubation, over 60% of
the TCE remained. Control experiments with
5 mg/L and 50 mg/L hydrogen sulfide but no
FeS showed no evidence of TCE degradation.
The calculated pseudo first order rate constant
for degradation of TCE by FeS that was not
freeze dried, normalized to the concentration
of FeS in suspension (moles per liter), was
(0.75 + 0.042) M-1 d-1. This is much faster
than the corresponding rate constant for freeze
dried FeS [(0.036 + 0.018) M'1 d'1]. The rate
of TCE degradation on freeze dried FeS in
Figure 2.1 is generally consistent with rate
constants obtained by Butler and Hayes (2001)
and Jeong et al. (2007) using freeze dried FeS.
-------
-TCE on FeS not freeze dried
TCE on FeS freeze dried
-TCE Control no FeS
-Acetylene from TCE on FeS not freeze dried
- cis-DCE from TCE on FeS not freeze dried
o
S
o
O
10
20 30 40
Time (d)
50 60
Figure 2.1. TCE degradation by chemically synthe-
sized FeS at pH 7.2. Error bars are one
standard deviation.
Shen and Wilson (2007) conducted column
experiments to simulate TCE degradation in
a mulch biowall, and the pseudo first order
rate constant for TCE removal, normalized
to the amount of biogenic FeS that formed in
the system (0.53 to 2.3 M"1 d"1) was roughly
equivalent to the rate constant obtained from
non-freeze dried chemically synthesized
FeS as presented in Figure 2.1. The rate of
TCE degradation on FeS that is not freeze
dried is from 20 to 50 times greater than the
rate of degradation on freeze dried FeS.
2.2.3.2 Effect of pH on the Rate of Transformation
of TCE on FeS
Figure 2.2 compares data from Butler and
Hayes (2001) for the rate constant of TCE
degradation on freeze dried FeS to data from
He et al. (2010) for TCE degradation on both
freeze dried FeS and FeS that was not freeze
dried. The pseudo first order rate constants
are normalized to the amount of FeS in the
suspensions. Regardless of how the FeS
was prepared, the rate of abiotic reaction of
TCE on FeS increased with increasing pH.
T3
"7
S
.2
to
c
0
O
to
o:
Figure
0.
0.
-«- Not feeze dried [1] (He et al., 2010)
a Not freeze dried [2] (He et al., 2010)
A Freeze dried
-«-- Freeze dried
3_,
7 -
0.6-
0.5-
0.4-
0.3-
0.2-
0.
2.
1 -
-i .
[2] (He et al
(Butler and
,2010)
Hayes, 2001)
T
^ _«- o>
5678
pH
2. Effect of pH
9
10
on the rate of TCE deg
dation on chemically synthesized FeS.
The data series Not freeze dried [2]
and freeze dried [2] repeat data from
Figure 2.1.
Butler and Hayes (2001) showed that solu-
tion pH had a significant influence on both the
overall rate constants for TCE transformation by
FeS (k, ') as well as the distribution of reaction
v obs '
products. A similar dependence of rate constants
on pH was also observed in the transforma-
tion of TCE by pyrite (Lee and Batchelor
2002a; Weerasooriya and Dharmasena 2001).
The reason for the pH effect on TCE degradation
by FeS remains unclear. One plausible explana-
tion is that the reactive chemical species on
the surface of the FeS changes as a function of
pH. Bebie et al. (1998) have proposed that the
surface chemistry of metal sulfides is dominated
by sulfhydryl functional groups. An increase in
pH would result in an increased concentration of
deprotonated surface species. The difference in
reactivity between the protonated and deproton-
ated forms of surface acid/conjugate base pairs
can be explained by the greater driving force
for electron donation by the more deproton-
ated ligands, which could cause an increased
rate of reductive dechlorination at higher pH.
Butler and Hayes (2001) suggested surface func-
tional groups that are more deprotonated would
-------
also be better nucleophiles, which could increase
the rate of formation of chloroacetylene and its
degradation product (acetylene) by a nucleo-
phile-induced dichloroelimination pathway.
Similarly, Weerasooriya and Dharmasena (2001)
explained the influence of pH on the transforma-
tion of TCE by pyrite on pH-dependent equi-
libria between pyrite surface functional group.
Alternatively, the pH effect on TCE degradation
can be related to faster electron transfer at higher
pH. Conway et al. (1980) use cyclic voltamme-
try to measure peak currents in FeS2 (pyrite) and
F6j xS (pyrrhotite). Because the peak current is
proportional to the rate of electron transfer at the
electrode-solution interface, an increase in peak
current with increasing pH implies a faster rate
of electron transfer at higher pH values. Conway
et al. (1980) suggested this pH-dependence of
the peak current was greater as deprotonation of
sulfide species increases in the interfacial region
with increasing pH. As a result, electron transfer
by FeS should increase at higher pH, leading
to an increase in the rate of TCE degradation.
2.2.3.3 Organic Molecules
Organic molecules with strong affinities for the
FeS surface may affect rate constants through
surface electron transfer reactions. Butler and
Hayes (1998) show that addition of cysteine
significantly inhibits the rate of TCE reductive
dechlorination by FeS. This rate reduction may
be due to adsorption of cysteine or methionine
to surface iron atoms in competition with HS~ or
OH", causing an energetic barrier or steric barrier
to electron transfer (Butler and Hayes 2002). On
the other hand, addition of 2,2-bipyridine and
2-propanol did not significantly affect pseudo
first order rate constants or branching ratios for
TCE transformation by FeS (Butler and Hayes
2001). However, unlike what they observed in
the TCE system, Butler and Hayes (1998) found
that 2,2'-bipyridine increased hexachloroethane
degradation by FeS by an order of magnitude.
2.2.3.4 Metal Ions
Recently, Jeong and Hayes (2007) reported
the impact of transition metals [Fe(II), Co(II),
Ni(II), and Hg(II)] on the dechlorination rates
of PCE and TCE. The impact of Fe(II) on
the dechlorination rates varied with its con-
centration, due to the formation of several
Fe sorption phases. Both Co(II) and Hg(II)
enhanced the reductive dechlorination of
PCE and TCE, but Ni(II) slowed the rate of
dechlorination. Jeong and Hayes (2007) attrib-
uted their results to different dechlorination
reactivities of new sulfide phases formed by
interaction of the transition metals with FeS
(See Section 2.3.3.2. for more discussion).
2.2.3.5 Dissolved Oxygen (DO)
Most of the past studies on degradation of
chlorinated compounds by pyrite have been
conducted under anaerobic conditions. The
degradation of chlorinated compounds in pyrite
suspension in the presence of O2 has been less
studied. Kriegman-King and Reinhard (1994)
reported that oxygen inhibited carbon tetrachlo-
ride transformation by pyrite. In a more recent
study, Pham et al. (2008) show that the rate
of transformation of TCE on pyrite depended
strongly on dissolved oxygen. The transforma-
tion rate increased with an increase in DO. The
rate constant for degradation increased from
0.004 hr1 up to 0.013 hr1 as the DO increased
from 0.017 to 0.268 mmol/L. In the more aero-
bic systems, TCE transformed to several organic
acids including dichloroacetic acid, glyoxylic
acid, oxalic acid, formic acid, and finally to
CO2 and chloride ion. Dichloroacetic acid was
the only chlorinated degradation product of
TCE that was observed in their experiments.
2.2.4 Degradation Mechanisms
The reductive dehalogenation of chlorinated
compounds by the transfer of electrons from
the mineral surface is similar to the mechanism
of TCE dehalogenation on zerovalent iron
(ZVI) (Liu et al. 2005; Pham et al. 2008). Two
major pathways of reductive transformation of
-------
TCE have previously been proposed (Butler
and Hayes 2001; Lipczynska-Kochany et al.
1994). The first pathway, sometimes called
|3-elimination or dichloroelimination, results in
the formation of an additional carbon-carbon
bond with loss of two halogens on adjacent
carbon atoms. TCE undergoes dichloroelimina-
tion to acetylene via the transient intermediate
chloroacetylene (Arnold and Roberts 2000;
Burris et al. 1996; Campbell et al. 1997; Roberts
et al. 1996). Acetylene may undergo subse-
quent hydrogenation to ethylene and/or ethane
under certain conditions (Arnold and Roberts
2000; Campbell et al. 1997), and it can also
be hydrolyzed. Dichloroelimination results in
the formation of relatively benign products.
In the second pathway TCE undergoes sequen-
tial hydrogenolysis, forming cis-DCE, vinyl
chloride, and ethylene (Butler and Hayes
2001). Hydrogenolysis (replacement of a
halogen by hydrogen) of both chloroethylenes
leads to the formation of more toxic and recal-
citrant chlorinated byproducts including the
dichloroethylenes and vinyl chloride. Each
of these reactions involves a net transfer of
two electrons. Jeong et al. (2007) observed
that acetylene was subsequently transformed
in the presence of FeS, but little transforma-
tion of cis-DCE and 1,1-DCE was observed.
A third mechanism proposed by Pham et al.
(2008) suggested that the transformation of
TCE in aerobic pyrite suspension follows a
Fenton-like mechanism, in which the reduction
of O2 at the pyrite surface can induce hydroxyl
radical formation (Cohn et al. 2006; Pham
et al. 2008; Rimstidt and Vaughan 2003).
The transformation of PCE and other con-
taminants by iron sulfide minerals is similar
to the degradation of TCE. Jeong and Hayes
(2007) proposed that PCE transforms to
TCE via hydrogenolysis, to acetylene via
P-elimination and to 1,1-DCE via a-elimination.
The process of P -elimination results in loss
of two halogens on the same carbon atom.
Lee and Batchelor (2002a) suggested that the
main pathway for the reductive dechlorination
of cis-DCE by pyrite is the reductive elimina-
tion pathway. They also proposed a sequential
hydrogenolysis pathway to explain the reduc-
tive transformation of vinyl chloride in pyrite
suspension (vinyl chloride —» ethylene —»
ethane, which all involves 2 electron reductions).
2.3 Abiotic Transformation of
Chlorinated Alkanes
2.3.1 Degradation Products
Based on the reaction pathway of 1,1,1-TCA
proposed by Fennelly and Roberts (1998),
1,1,1-TCA can be initially transformed to
1,1-DCA through hydrogenolysis, to 1,1-dichlo-
roethylene through dehydrochlorination, and to
2,2,3,3-tetrachlorobutane by coupling of two
dichloroethyl radicals (H3C-C -C12) that were
produced by a one electron reduction of 1,1,1-
TCA. In addition, it has been reported that
carbene intermediates (H3C- C-C1) formed by
two-electron reduction to 1,1,1-TCA triggered
the formation of acetaldehyde and vinyl chloride
(Butler and Hayes 2000; Choi et al. 2009).
Butler and Hayes (2002) showed that 1,1,1-TCA
degraded on FeS, and that 1,1-DCA accounted
for approximately 2% of 1,1,1-TCA that was
degraded, however, 1,1-DCA, 1,2-DCA, and
1,1,2-TCA showed no appreciable transforma-
tion by FeS over approximately 120 days.
Gander et al. (2002) investigated 1,1,1-TCA
transformation by iron sulfide and a methano-
genic microbial consortium. In reactors contain-
ing FeS alone, minor amounts of 1,1-DCA (2%)
and 2-butyne (4%) were observed as products,
resulting in a carbon mass recovery of ~6% of
the 1,1,1 -TCA that was degraded. Other poten-
tial degradation products (1,2-dichloroethylene,
acetic acid, acetylene, ethane, ethylene, acetalde-
hyde, c/s-2-butene, and vinyl chloride) were not
detected. In reactors containing LEG (a metha-
nogenic culture of bacteria enriched on lactate)
alone, the only identified product was 1,1-DCA,
which accounted for 46 + 8% of the 1,1,1-TCA
that was transformed (Gander et al. 2002).
-------
2.3.2 Rates of Degradation on Iron Sulfide
Minerals
Table 2.2 lists rates reported in the literature
for degradation of chlorinated alkanes by FeS
minerals. The rate constant for degradation of
1,1,1-TCA transformation by 10 g/L chemi-
cally synthesized FeS was (0.28 ± 0.03) d'1,
corresponding to a half-life of 2.5 days (Gander
et al. 2002). Choi et al. (2009) show that 82%
of initial 1,1,1-TCA(0.5 mM) was removed in
contact with 33 g/L of FeS within 48 h, pro-
ducing 1,1-DCA as the major transformation
product. The rate constant for the reductive
dechlorination of 1,1,1-TCA was (0.0375 +
0.0018) tr1. Butler and Hayes (2000) reported
a rate constant for degradation of 1,1,1-TCA
on 10 g/L FeS of 0.235 d"1, corresponding to
a pseudo first order rate constant normalized
to the surface area of FeS of 0.63 L nr2 d"1.
This is approximately tenfold faster than the
pseudo first order rate constant for degradation
of TCE on FeS, which was 0.0716 L nr2 d'1
(Butler and Hayes, 1999, see Table 2.1).
Wilson et al. (2008) investigated 1,2-DCA
degradation by biogenic FeS. They extracted
Table 2.2. Rates and products of abiotic degradation of chlorinated alkanes on FeS minerals.
Chlorinated
solvents
1,1 -DCA
1,2-DCA
1,2-DCA
1,1,1-TCA
1,1,1-TCA
1,1,1-TCA
1,1,1-TCA
1,1,2-TCA
1,1,1,2-TeCA
1,1,2,2-TeCA
PCA
Pentachloroethane
FeS
mineral
phases
FeS
FeS
Biogenic FeS
Fe2+ and
HS-
in solution
FeS, freeze
dried
FeS
FeS
FeS
FeS
FeS
FeS
Degradation
product
None detected
None detected
Not monitored
None detected
1,1 -DCA,
2-butyne
1,1 -DCA
1,1 -DCA,
ethylene
1,1-DCE, vinyl
chloride
1,1-DCE
TCE, c/s-DCE,
trans-DCE,
acetylene
PCE, TCE
Conditions
Same as above
Same as above
pH 7, 26 g/L FeS
pH7.0, 0.1M
NaHCO3and
1.85mMFeCl2
2.06 mM Na2S
pH7.5, 10 g/L FeS,
100 mM NaHCO3
pH8.3, 10 g/L FeS,
0.05 m2 /g, ionic
strength 0.1M
pH 7.5, 33 g/L
FeS, Tris buffer
pH8.3, 10 g/L FeS,
0.05 m2 /g, ionic
strength 0.1M
Same as above
Same as above
pH8.3, 10 g/L FeS,
0.05 m2 /g, ionic
strength 0.1M
Rate constant
Rate not significant
at 95% confidence
Rate not significant
at 95% confidence
2.8xlO-2d-1
<10% in 33 days
<10% in 33 days
0.28+0.03 d'1
(2.35+0.34)xlO-1d-1
(0.63+0.09) L m-2 d'1
(9.0+0.043)xlO-1d-1
Rate not significant
at 95% confidence
(0.73+0.12)^'
(1. 46+0.24) Lm-2d-'
(4.56+9.6)x lO'2 d'1
(9.12+19)xlO-2Lm-2d-1
(2.28+1.03) d'1
(4.6+2.1) Lm-2d-'
Reference
Butler and
Hayes 2000
Butler and
Hayes 2000
Wilson et al. 2008
Doong and
Wu 1992
Gander et al. 2002
Butler and
Hayes 2000
Choi et al. 2009
Butler and
Hayes 2000
Butler and
Hayes 2000
Butler and
Hayes 2000
Butler and
Hayes 2000
-------
Table 2.2. Continued.
Chlorinated
solvents
HCA
Hexachloroethane
HCA
HCA
HCA
HCA
Y-HCH
hexachloro-
cyclohexane
FeS
mineral
phases
FeS, freeze
dried
FeS, freeze
dried
FeS, freeze
dried
FeS
FeS2
pyrite
FeS, freeze
dried
Degradation
product
PCE, PCA,
TCE, cis-DCE,
acetylene
PCE, PCA,
TCE, cis-DCE,
acetylene
PCE
Not monitored
Not monitored
pentachloro-
cyclohexene,
tetrachloro-
cyclohexene,
dichlorocyclo-
hexadiene,
dichloro-
benzenes,
chlorobenzene
Conditions
pH7.8, lOg/LFeS,
0.05 m2 /g, ionic
strength 0.1M
Same as above
pH8.3, lOg/LFeS,
0.05 m2 /g, ionic
strength 0.1M
pH 7.2, 2.0 g/L ,
25 m2/L, 1 mM Fe2+
pH 7.2, 28 g/L ,
25 m2 IL, 1 mM Fe2+
pH 8.3, tris-HCl,
10 g/L, 10 m2/g
ionic strength 0.1M
Rate constant
(1.74+0.19) d-1
(3.5+0.38) Lm2^1
(2.97+0. 18) d'1
(5.9+0.36) L m-2 d'1
(1.80+0.08) d'1
(3.61+0.16)Lm-2d-1
0.38 L m-2 d'1
0.38 L m-2 d'1
(l.lxlO-2)^1
(l.lxlO-OLm2^1
Reference
Butler and
Hayes 1998
Butler and
Hayes 2000
Jeong and
Hayes 2003
Eisner et al. 2004b
Eisner et al. 2004b
Liu et al. 2003
a rate constant of 0.028 d"1 in contact with
26 g/L FeS. This rate constant was an order of
magnitude lower than rate constants for TCE
and EDB degradation under similar conditions.
Butler and Hayes (2000) reported first order rate
constants for degradation of hexachloroethane,
pentachloroethane, 1,1,2,2-tetrachloroethane,
1,1,1,2-tetrachloroethane, and 1,1,1-trichloro-
ethane in water containing 0.5 m2 L"1 FeS as
poorly crystalline mackinawite. Table 2.2
expresses their data as a first order rate con-
stant, and as a first order rate normalized to
the surface area of FeS exposed to water.
The surface area specific rates varied from a
high of (28 + 1.5) m2!/1 d'1 for pentachloro-
ethane to a low of (5.9 + 0.36) m2!/1 d'1
for hexachloroethane.
2.3.3 Conditions Affecting Degradation
Similar to other groups of halogenated
aliphatic compounds, TCA degrada-
tion is affected by solution composi-
tion and environmental conditions.
2.3.3.1 pH
Previous studies demonstrated that the reduc-
tive dechlorination of TCE by FeS was
strongly dependent on the suspension pH and
generally increased with increasing pH (Butler
and Hayes 1998; Butler and Hayes 2001).
There is no report available in the literature
concerning the effect of pH on TCA degrada-
tion by FeS. However, an effect similar to
the effect on degradation of TCE would be
expected because the FeS surface will be
affected by changes in pH in a similar way.
-------
Butler and Hayes (1998) investigated the
influence of pH on the reductive dechlorina-
tion of hexachloroethane (HCA) by FeS
occurring as poorly crystalline mackinawite.
Increasing the solution pH from 7.1 to 9.5
caused a significant increase in the rate of
reductive dechlorination of HCA by FeS. They
proposed that a pH-dependent equilibrium
between the protonated and deprotonated
forms of FeS surface species is responsible
for the pH dependence of the rate of reduc-
tive dechlorination of HCA by FeS, with the
deprotonated species having greater reactivity.
2.3.3.2 Metals
Metal ions may be characterized as "hard" or
"soft" depending on the Lewis acid base prin-
ciple. Metal ions in water impact the dechlo-
rination rates of hexachloroethane depending
on their relative hardness. While hard metals
decrease the dechlorination rates, intermediate
to soft metals enhance the rates (Butler and
Hayes 1998; Jeong and Hayes 2007). Choi
et al. (2009) show that the reduction rate of
1,1,1-TCAby FeS increased with increasing
concentration of transition metals. The rate
constants in the presence of 10 mM Co(II) and
Ni(II) were 0.06 and 0.11 h"1, approximately 1.3
and 3.0 times greater than those by FeS alone.
Jeong and Hayes (2003) investigated the impact
of various transition metals on the reductive
dechlorination by mackinawite using a read-
ily degradable chlorinated organic compound,
hexachloroethane (HCA). They found that
10-3 M Cr(III) and Mn(II) (hard metal ions)
decreased the dechlorination rates. These hard
metal ions, due to their weak affinity for sul-
fides, are thought to form surface precipitates of
hydroxides around FeS under the experimental
conditions. These hydroxide precipitates
should hinder the electron transfer between
FeS and HCA (Butler and Hayes 1998; Jeong
and Hayes 2003). In contrast, Co(II), Ni(II),
Cu(II), Zn(II), Cd(II), and Hg(II) (intermedi-
ate/soft metal ions) at concentrations of 10~4,
10"3, and 10"2 M increased the rates of HCA
degradation on FeS (Jeong and Hayes 2003).
2.3.3.3 Sulfide
Choi et al. (2009) evaluated the effect of HS in
solution on the rate of degradation of 1,1,1- TCA
by FeS. Although 1,1,1- TCA was reductively
degraded by HS" in homogeneous solution
to a certain extent, the rate of reaction was
enhanced in the presence of FeS. Reductive
dechlorination of 1,1,1-TCAby FeS increases
with HS" concentrations. The rate of 1,1,1- TCA
degradation on FeS was 0.0375 + 0.0018 h"1.
The rate in the presence of 5 mM HS" was three
times greater, and the rate in the presence of
20 mM HS" was ten times greater than degrada-
tion by FeS without HS" (Choi et al. 2009). The
HS" added to the suspension of FeS may bind
to the surface of FeS and enhance the reaction
rate of 1,1,1-TCAby facilitating the electron
transfer or by forming reactive iron sulfide
species on the surface of FeS (Choi et al. 2009).
Sulfide may also enhance dechlorination through
formation of new FeS phases. Choi et al. (2009)
show that iron sulfide can be deposited on the
surface of iron metal and facilitate the electron
transfer from iron metal to chlorinated organic
compounds. Kriegman-King and Reinhard
(1992) found that the rates of transformation
of HCA by biotite or vermiculite increased
significantly when HS" was added to the
experimental system. The increase in reactiv-
ity may have been due to the formation of a
secondary iron sulfide phase upon reaction of
HS" with Fe2+ associated with the clay miner-
als (Kriegman-King and Reinhard 1992).
2.3.3.4 Organic Molecules
No experiment has been conducted to inves-
tigate the effect of natural organic matter
(NOM) on 1,1,1-TCA degradation. However,
several studies investigated the effect of NOM
on HCA degradation. These studies indicate
that organic compounds with strong metal
chelating properties and delocalized TT* elec-
trons significantly increase the rate of HCA
dechlorination (Jeong and Hayes 2003).
Butler and Hayes (1998) found that compounds
with a bidentate chelating functional group,
-------
including 2,2'-bipyridine, 1,10-phenanthrolene,
and EDTA, could significantly increase the rate
of HC A reductive dechlorination. The presence
of this functional group likely increased the
driving force for adsorption of these compounds
to the mineral surface (Butler and Hayes 1998).
On the other hand, oxalate, succinate, and hydro-
quinone, which contain phenolic and carboxylic
functional groups, had no significant effect on
the HCA dechlorination rate (Butler and Hayes
1998). Addition of 1 mM cysteine, a thiol-con-
taining amino acid, to the FeS reaction mixture
reduced the rate of reductive dechlorination of
HCA by almost one-half, and the addition of 1
mM methionine, an alkylthio amino acid, only
slightly depressed the rate. This rate reduction
may be due to adsorption of cysteine or methio-
nine to surface iron atoms in competition with
HS" or OH", causing an energetic or steric barrier
to electron transfer (Butler and Hayes 1998).
2.4 Abiotic Transformation of
Chlorinated Methanes
2.4.1 Degradation Products
Hanoch et al. (2006) has reviewed the capac-
ity of common soil minerals to degrade carbon
tetrachloride. Carbon tetrachloride can be
reductively dechlorinated by pyrite (FeS2) and
FeS (Assaf-Anid and Lin 2002; Kenneke and
Webber 2003; Lee and Batchelor 2002a; Pham et
al. 2008) and by marcasite, which is also a form
of FeS2 (Butler and Hayes 1998; Kriegman-King
and Reinhard 1994). There is one study on the
degradation of chloroform on FeS (Kenneke
and Webber 2003). To our knowledge, no
information is available on degradation of
dichloromethane by FeS or FeS2 minerals.
Laboratory investigations have demonstrated
that carbon tetrachloride can be degraded along
two competing pathways: through hydrogenoly-
sis to chloroform, and through hydrolysis to
carbon dioxide (Criddle and McCarty 1991).
The iron sulfide minerals pyrite and marcasite
were shown to reduce carbon tetrachloride
to chloroform (Kriegman-King and Reinhard
1994). Kenneke and Webber (2003) reported
69% conversion of carbon tetrachloride to
chloroform in the presence of FeS. Devlin and
Miiller (1999) reported that the principal product
of carbon tetrachloride reduction by FeS was
chloroform (48%) followed by carbon dioxide
(10%), formate (5%), and carbon disulfide
(2%). In addition, formation of non-chlorinated
products such as carbon monoxide, methane,
acetylene, ethylene, and ethane was also
observed during abiotic degradation of carbon
tetrachloride by various minerals (Criddle and
McCarty 1991; Devlin and Muller 1999; Eisner
2004a; Kenneke and Webber 2003; Kriegman-
King and Reinhard 1992; Maithreepala and
Doong 2005; O'Loughlin et al. 2003).
Kriegman-King and Reinhard (1994) show that
the product distribution of carbon tetrachloride
degradation varies greatly depending on the
reaction conditions. Under aerobic conditions,
the major product was carbon dioxide which
could account for 70% to 80% of the carbon tet-
rachloride transformed. In contrast, the anaero-
bic system forms approximately 50% chloroform
and only 10% to 20% carbon dioxide.
2.4.2 Rate Constants
Hanoch et al. (2006) reviewed carbon tetra-
chloride dechlorination by some common soil
minerals. In general, the rate of degradation
on FeS minerals was the fastest among all the
minerals that were tested. The rate of degrada-
tion on pyrite was slower than on FeS, but still
faster than the other Fe(II) containing minerals.
Table 2.3 lists rate constants reported in the
literature for degradation of chlorinated meth-
anes by FeS minerals. Butler and Hayes (2000)
reported rapid degradation of carbon tetrachlo-
ride in the presence of FeS, with an observed
first order rate constant of (6.39 + 0.79) x 10~2
h"1. Table 2.3 expresses their data as this first
order rate, and as a first order rate normalized
to the surface area of FeS exposed to water.
The surface area specific rate of degradation of
carbon tetrachloride was (5.2 + 0.6) m2L"1 d"1.
-------
Table 2.3. Rates and products of abiotic degradation of chlorinated methanes on FeS minerals
Chlorinated
solvents
CC14
CC14
Carbon
Tetrachloride
CC14
CC14
CC14
CC14
CC14
CC14
CC14
CC14
FeS min-
eral phases
Fe2+ or HS
in solution
Mackinawite 1
Mackinawite 2
FeS
FeS
FeS2
FeS, fresh
FeS, aged
FeS
FeS, freeze
dried
FeS coat-
ing on Iron
oxides
formed by
treating
with HS
Pyrite
Degradation
product
CHC13
CHC13
Chloroform
CHC13,
possibly CH4,
C2H4, C2H6
Not
monitored
Not
monitored
CHC13, CS2
CHC13, CS2
CHC13,
CH2C12
CHC13
CHC13
CHC13, CO2,
CS,, formate
z
Conditions
pH7.0, 0.1M
NaHCO3and
1.85mMFeCl2
2.06 mM Na2S
4 g/L FeS,
13 m2/g
pH 7.2,
0.6 g/L FeS,
77 m2/g, pH 7.2
FeS 33 g/L, Tris
buffer, pH 7.5
200g/L FeS,
0.05M I, pH 6.5
200g/L FeS2,
0.05M I, pH 6.5
0.73 g/L FeS
0.73 g/L FeS
pH8, 18 g/L FeS
lOg/L FeS, 0.05
m2/g, pH 8.3,
O.lMTris
pH8, O.lMtris
0.13 g/L FeS
on goethite
0.20 g/L FeS
on hematite
pH 6.5, fresh
ground pyrite,
14.8 g/L, 0.01
m2/g, anaerobic
conditions
Rate constant
90% removal in 10 days
0.23 d'1
<10% removal in 33 days
(5.84±0.31)xl01d-1
(1.2±0.06)Lm-2d-1
(1.5210.11)^
(3.0±0.22)xlO-2Lm-2d-1
(2.98±0.22)xl01 d'1
(4.15±0.12)xl01d-1
(4.15±0.19)xl01d-1
66% removed in one day
l.OTd'1
72% removed in one day
1.2yd-1
80% removed in 2.5 hours
9.7X101 d-1
(2.610.31)^
(5.2±0.62)Lm-2d-1
(0.2810.14)^
(0.2210.12)^
zero order kinetics
0.053 molnvM-1
first order kinetics
O.nd-1
O.ieLm-2^1
Reference
Doong and
Wu 1992
Zwank et
al. 2005
Choi et
al. 2009
Lipczynska-
Kochany et
al. 1994
Lipczynska-
Kochany et
al. 1994
Devlin and
Miiller 1999
Assaf-Anid
and Lin
2002
Butler and
Hayes 2000
Hanoch et
al. 2006
Kriegman-
King and
Reinhard
1994
-------
Table 2.3. Continued.
Chlorinated
solvents
CC14
CC14
CC13
Chloroform
CHC13
FeS min-
eral phases
Pyrite
Pyrrhotite
FeS coat-
ing on Iron
oxides
formed by
treating
with HS
FeS
Degradation
product
CHC13
CHC13
Not
monitored
Conditions
6 g/L FeS
6 g/L FeS
pH8, O.lMtris
0.13 g/L FeS
on goethite
0.20 g/L FeS
on hematite
pH 7.8, 0.044
g/L, 0.14 m2/g
Rate constant
20% removed in one day
0.22 d-1
60% removed in one day
0.91 d'1
(2.23±0.29) x 10-3 d-1
(9.37±0.98) x 10-3 d-1
0.717 m^min-1 in 0.059 L
6.1 xlO'Lm-'d-1
Reference
Devlin and
Miiller 1999
Devlin and
Miiller 1999
Hanoch et
al. 2006
Kenneke
and Weber
2003
2.4.3 Conditions Affecting Reactivity
2.4.3.1 pH
Similar to PCE and TCE degradation, pH could
have important effects on carbon tetrachloride
degradation. This is because pH may affect the
surface properties of the minerals, the aque-
ous speciation of sulfide, and probably the
surface speciation of sulfide (Kriegman-King
and Reinhard 1992). However, Kriegman-
King and Reinhard (1992) didn't find a strong
effect of solution pH on the degradation of
carbon tetrachloride by biotite and vermiculite
over a pH range of 6 to 10 standard units.
We are not aware of any studies investigat-
ing the effect of pH on carbon tetrachloride
degradation in the presence of iron sulfide
minerals. Based on the discussion in section
2.2.3.2, we would expect pH to have a similar
effect on carbon tetrachloride degradation
by FeS as on degradation of TCE by FeS.
2.4.3.2 Sulfide
Sulfide was expected to play an important role
in the kinetics because it could act as either
an electron donor or a nucleophile. Another
major role sulfide could play is to form new
FeS phases that could significantly enhance
dechlorination. The reductive dechlorination
of carbon tetrachloride to chloroform by dis-
solved hydrogen sulfide was reported to occur
more rapidly when biotite and vermiculite were
present, and the reaction was suggested to take
place at the mineral surface (Kriegman-King and
Reinhard 1992). Carbon tetrachloride was also
shown to be reduced to chloroform by a sodium
sulfide solution containing hematin (Assaf-
Anid and Lin, 2002) and in water containing
free Fe2+ and HS ( Doong and Wu 1992). The
reductive dechlorination of carbon tetrachloride
to chloroform by dissolved hydrogen sulfide
in the presence of the clay minerals could be
due to the fact that sulfide can react with dis-
solved iron (released to solution from mineral
dissolution) to form iron sulfide minerals,
which can then react with carbon tetrachlo-
ride (Kriegman-King and Reinhard 1992).
-------
2.4.3.3 Organic Compounds
Organic compounds representative of natural
organic matter (NOM) affect the FeS-mediated
reductive dechlorination of carbon tetrachloride
to a greater or lesser extent, depending on their
structure (Hanoch et al., 2006; Assaf-Anid and
Lin, 2002; Butler and Hayes, 1998). These
effects have been attributed to the formation
of electron donating or electron withdrawing
surface complexes (Butler and Hayes 1998).
Devlin and Miiller (1999) found that in the pres-
ence of humic acid, which acted as an electron
mediator, the reduction of carbon tetrachloride
was accelerated by a factor often. Addition of
vitamin B12 to aqueous FeS systems not only
enhanced the rate of degradation of carbon
tetrachloride, but also increased the yields
of the hydrogenolysis products chloroform
and dichloromethane (Assaf-Anid and Lin
2002). This indicates that organic co-solutes
can influence the carbon tetrachloride product
distribution, perhaps by acting as hydrogen
atom donors. Hashsham et al. (1995) found
that the addition of cobalamine to their micro-
cosms increased the rate of transformation of
14C-labelled carbon tetrachloride one hundred-
fold and resulted in 73% of carbon tetrachloride
conversion to 14C-labelled carbon disulfide.
2.5 Abiotic Transformation of the
Brominated Ethane EDB
2.5.1 Degradation Products
There is little information about ethylene
dibromide (EDB) degradation in the literature.
Several earlier studies investigated biological
degradation of EDB (Bouwer and McCarty
1985; Henderson et al. 2008; Tandoi et al. 1994).
These studies showed that EDB was biodegrad-
able under aerobic and anaerobic conditions.
EDB can be degraded through three potential
pathways: through hydrolytic debromination to
form 2-bromoethanol; through hydrogenolysis
(reductive dehalogenation) to form bromoethane;
and through dibromoelimination (dihaloelimina-
tion) to form ethylene (Wilson et al. 2008).
The only study of abiotic degradation of EDB
by FeS to date was conducted by Wilson et
al. (2008). In experiments on EDB degrada-
tion by laboratory synthesized FeS, ethylene
was the major degradation product. Methane
was detected but only at background levels,
and ethane and acetylene were not detected
(personal communication Yongtian He
and John T Wilson, unpublished data).
2.5.2 Rate Constants
The microcosm study by Wilson et al. (2008)
shows that the removal of EDB followed first
order kinetics without a lag, indicating that
removal was an abiotic process that did not
require acclimation of an active biological
process. However, the study was conducted
with biogenic FeS isolated from laboratory
columns that simulated a plant mulch based
passive reactive barrier. The biogenic FeS
used in the study still contained small amounts
of residual plant material, and was actively
methanogenic and sulfate reducing. To confirm
that the removal was truly an abiotic process,
the experiment was repeated with chemically
synthesized FeS. The FeS was not freeze
dried. Details of the procedure are provided in
Appendix A.2. The rate constant for degradation
of EDB in the microcosm study of Wilson et al.
(2008), normalized to the amount of biogenic
FeS exposed to the porewater (moles FeS per
liter porewater), was 0.781 ± 0.249 M'1 d'1.
Figure 2.3 depicts the abiotic degradation
of EDB by chemically synthesized FeS at
pH 7.2. The rate constant of EDB degra-
dation by synthesized FeS at pH 7.2 was
0.833 ± 0.004 M-1 d'1. There was good agree-
ment between the rate of degradation on bio-
genic FeS and chemically synthesized FeS.
-------
10000
1000 -
05
m
Q
LU
100 -
10
Rate constant normalized to
concentration of FeS in suspension
0.833 ± 0.004 M"1 d"1
10 15 20
Time (days)
25
Figure 2.3. EDB degradation by FeS at pH 7.2.
2.5.3 Conditions That Affect Degradation
Wilson et al. (2008) summarized major fac-
tors that could be associated with EDB
degradation in gasoline release sites.
2.5.3.1 Sulfate and Sulfide
The consumption of sulfate during sulfate
reduction produces reactive HS~ that can destroy
EDB, or can react with iron(III) minerals to
produce FeS which also reacts to degrade EDB.
The rate of reactions with HS~ in solution only
becomes important to the abiotic degradation of
EDB at HS" concentrations above 0.2 mg/L. At
lower concentrations of HS" in solution, aque-
ous hydrolysis is faster than reaction with HS".
EDB can also react with H2S and HS" to
produce a variety of thiols and thio-
ethers (Schwarzenbach et al. 1985).
2.5.3.2 pH
Figure 2.4 depicts the effect of pH on the
rate of EDB degradation by chemically syn-
thesized FeS. There was a linear increase in
the rate of degradation with pH in the range
from pH 5 to 9 (Kuder et al., In prep.).
As discussed in Section 2.2.3.2, the rate of
abiotic transformation of TCE by iron(II) sulfide
also increases as the pH increases. Based on
the data for TCE and EDB, pH seems to have a
1.6 -i
-~ 1.4 -
T3
V 1.2 -
jS 0.8 -
I 0.6-
« °-4 "
I 0.2-
0
7
PH
Figure 2.4. Effect ofpH on the first order rate
constant for EDB degradation by FeS.
The rate constant is normalized to the
concentration of FeS in suspension.
greater impact on rate of EDB degradation than
on TCE degradation. The greater impact may be
a composite of the direct effect of pH and the
effect of pH on the concentration of sulfide in
solution. As pH increases, the deprotonation of
the surface of the FeS increases, which should
increase its reactivity toward EDB. In experi-
ments with chemically synthesized FeS, the
measured concentration of sulfide also increases
with pH (2, 9, 26, 40 mg/L at pH 5.4, 6.4, 7.2,
8.4, respectively). Previous studies by Wilson et
al. (2008) showed that the presence of sulfide in
solution can contribute to EDB degradation. In
contrast to EDB, sulfide doesn't seem to have
significant impact on TCE degradation rate.
2.6 Formation, Transformations, and
Reactivity of Iron Sulfides
2.6.1 FeS Mineralogy and Reactivity
Mackinawite belongs to the P4/nmm space
group (Taylor and Finger 1970). Mackinawite
possesses a tetragonal layer structure (Wolthers
et al. 2003), where the iron atoms are linked in
a tetrahedral coordination to four equidistant
sulfur atoms. These tetrahedrons share edges
to form a layer structure. These layers are
stacked along the c-axis and held together by
weak van der Waals forces (Vaughan and Craig
1978). Within the layers, each Fe atom is in a
-------
square planar coordination with four Fe atoms
with an Fe-Fe distance of 2.5967 A (Lennie et
al. 1995). This is similar to the Fe-Fe distance
in a-iron. The Fe-Fe bonding in mackinawite
allows it to conduct electricity like a metal
(Jeong et al. 2008; Taylor and Finger 1970;
Vaughan and Ridout 1971; Wolthers et al. 2003).
The high reactivity of iron sulfide minerals in
degrading chlorinated solvents may be related
to the following unique properties of FeS:
(1) The high electron transfer capability
of mackinawite. As discussed above,
mackinawite is a metallic conductor
with the delocalized electrons in
the plane of the mackinawite layers
(Vaughan and Ridout 1971).
(2) The hydrophobic character of
mackinawite enhances sorption of
TCE and other halogenated aliphatic
compounds. Weerasooriya and
Dharmasena (2001) suggest that
TCE may be readily attracted to
pyrite by hydrophobic interactions,
thus facilitating the surface mediated
electron transfer process. Park
et al. (2006) suggest that FeS is
more hydrophobic than FeS2 and
that the hydrophobicity of the
particle surface is a critical factor
in surface-mediated dehalogenation
of chlorinated compounds.
(3) FeS is a hydrate with an expanded
lattice. Morse and Cornwell
(1987) proposed that disordered
mackinawite may be a hydrate.
Mackinawite has an expanded
lattice, in particular along the c axis,
with intercalated water molecules
between the tetrahedral sheets and
possibly intercalated hydroxyl groups
with associated protons, cations,
or sulfide vacancies to balance
charge (Wolthers et al. 2003).
(4) FeS has a high reduction potential
and chalcophilic (sulfur-loving)
nature. Both Fe2+ and HS~ are
strong reductants, which contributes
to the high reduction potential of
FeS phases. Several experimental
studies have shown that iron sulfides
are more reactive and effective
in degradation of the chlorinated
compounds than other Fe(II)
containing minerals and Fe° in PRB
treatment systems. Eisner et al.
(2004b) showed a general trend of
reaction rates for hexacholoroethane
and 4-chloronitrobenzene where the
rate of degradation increases in the
order of Fe(II)+siderite < Fe(II)+
iron oxides < Fe(II)+iron sulfides.
Newly formed FeS phases have been
reported to be significantly more
reactive than iron metal in TCE
transformation (Butler and Hayes
2001). Jeong and Hayes (2007)
showed that in the presence of
metals, the chalcophilic nature of FeS
facilitated the formation of secondary
sulfide phases such as CoS, which
increases TCE degradation.
2.6.2 Formation of FeS Minerals
The formation of iron monosulfide (FeS) in
a natural environment typically begins with
biologically produced Fe2+ and HS~. Another
possible mechanism is to proceed with reductive
dissolution of Fe3+ oxide by HS'that is generated
by sulfate reducing bacteria. Reductive dissolu-
tion of goethite by HS~ can be described with the
following reaction (Pyzik and Sommer 1981):
2FeOOH(s) + HS + H2O = 2Fe2+ + S° + 5OH
The Fe2+ formed via reductive dissolution
then precipitates with aqueous H2S or HS" to
form FeS (Devlin and Muller 1999; Neal et al.
2001; Pyzik and Sommer 1981). This FeS may
take the form of mackinawite (FeSj x, where
x = 0.025) (Berner 1970; Taylor and*Finger
-------
1970), amorphous phases (Berner 1970; Neal
et al. 2001), and a phase tentatively identified
by Neal et al. (2001) as hexagonal pyrrhotite
(Fe^S). While H2S and HS are generated by
sulfate reducing bacteria in natural systems,
FeS can also form under strictly abiotic con-
ditions (Benning et al. 2000; Hanoch et al.
2006; Herbert et al. 1998; Pyzik and Sommer
1981). In typical anoxic aquatic and sedimen-
tary systems, the FeS concentration ranges
from 1-100 mMol /kg (Rickard, 1997).
Sulfide mediated reductive dissolution of
iron oxides can also lead to formation of
elemental sulfur (Pyzik and Sommer 1981),
polysulfides (S42~ or S52) (Pyzik and Sommer
1981), thiosulfate (S2O32-) (dos Santos Afonso
and Stumm 1992; Pyzik and Sommer 1981),
sulfite (SO32~) (dos Santos Afonso and Stumm
1992), and sulfate (SO42-) (dos Santos Afonso
and Stumm 1992; Hanoch et al. 2006). These
intermediate sulfur species could play an
important role in FeS transformation.
2.6.3 Transformation of FeS Minerals
The most abundant iron sulfide phases in
natural environments are amorphous iron (II)
monosulfide (FeSam), mackinawite (tetragonal
FeS), greigite (Fe3S4), and pyrite (FeS2). Most
of these iron sulfide phases are metastable at
ambient temperatures. FeS can transform to
Fe3S4 and FeS2 through biotic and/or abiotic
pathways (Butler and Hayes 2001; Hanoch et al.
2006; Jeong and Hayes 2003). It is generally
accepted that the reaction sequence FeSam —»
mackinawite —» greigite —» pyrite is involved
in pyrite formation (Lennie et al. 1997).
Greigite is the sulfur analogue of magnetite
and has a similar inverse spinel structure. The
greigite structure can be regarded as a cubic,
close-packed array of sulfur atoms linked by
smaller iron atoms (Rickard and Luther 2007).
The mackinawite to greigite transformation
occurs through an oxidation process (Boursiquot
et al. 2001). Posfai et al. (1998) showed that
some bacteria form intracellular nanometer-scale
crystals of greigite (Fe3S4) from nonmagnetic
mackinawite (tetragonal FeS) and possibly from
cubic FeS. Rickard (1997) proposed that the
oxidation of FeS to Fe3S4 is a solid state reac-
tion, which is supported by several other stud-
ies (Boursiquot et al. 2001; Jeong et al. 2008;
Lennie et al. 1997; Posfai et al. 1998a,b). In
this process, two thirds of Fe(II) are oxidized
to Fe(III) and the S(-II) remains unoxidized.
Transformation of mackinawite to greigite can
be generally described by the following reaction:
3FeS + S°(s) = Fe3S
34
FeS may be transformed to FeS2 based
on three different mechanisms.
(1) Apolysulfide pathway is proposed
to be the primary pathway for
FeS2(s) formation. In this process,
iron sulfide precursors react
with S°, polysulfides, or other
S intermediates to form pyrite
through the following reactions:
FeS (s) + S°^FeS2(s)
FeS (s) + Sn2-^FeS2(s) + S^2-
Production of intermediate sulfur
species [for example, elemental
sulfur, polysulfides, and thiosulfate
(S2O32-)] is initiated through H2S
oxidation by oxidants such as O2,
NO3~, Fe oxides, and manganese
oxides (Butler and Hayes 2000;
Butler and Rickard 2000; Henneke et
al. 1997; Schoonen and Barnes 1991).
(2) Another potential transformation
mechanism is through greigite
formation. Hunger and Benning
(2007) show strong evidence for
greigite being an intermediate on
the formation pathway of pyrite,
and that pyrite did not form directly
from mackinawite. Their data show
mackinawite disappeared as greigite
was produced, and that pyrite was
-------
produced as greigite disappeared.
Benning et al. (2000) conducted
experiments with ex-situ, fireeze-
dried mackinawite that show that
at 25°C, 78% of mackinawite was
transformed to pyrite, via greigite,
within 24 h. Wang and Morse (1996)
provided experimental evidence
that the formation of pyrite is most
rapid in the presence of greigite.
The transformation of greigite
to pyrite can be described as:
Fe3S4
2S°(s) = 3FeS2
(3) Under strictly anoxic conditions,
H2S can act as an oxidant. FeS2(s)
has been shown to form via the
"H2S oxidation pathway" (Rickard
1997; Rickard and Luther 1997):
FeS(s) + H2S(aq) = FeS2(s) + H2(g)
Rickard (1997) showed that the H2S pathway
proceeded more rapidly than the polysulfide
pathway and all other pathways involving
intermediate sulfur species. Hurtgen et al.
(1999) suggested that the accumulation of
H2S in pore waters may promote pyrite forma-
tion. In an anoxic system with excess sulfide,
sulfide could mediate the partial oxidation of
mackinawite to greigite (Jeong et al. 2008;
Rickard 1997; Vaughan and Ridout 1971). In
addition, water has also been proposed as an
oxidizing agent for this transformation process
(Jeong et al. 2008; Rickard and Morse 2005).
For the first two pathways, intermediate sulfur
species are critical in pyrite formation. Thus the
availability of an oxidant to produce sulfur spe-
cies with intermediate oxidation states controls
the transformation of FeS to pyrite. Pyrite is
frequently found at the interface between oxidiz-
ing and reducing conditions in the environment
and the occurrence of pyrite is closely related to
the oxidation of sulfide (Hurtgen et al. 1999). In
a PRB field study, Herbert et al. (2000) observed
a greater accumulation of mildly oxidized sulfur
species (pyrite, S0) at the front end of the barrier.
This observation suggests that oxidizing agents
may be present at low concentrations in the
ground water that enters the PRB from up gradi-
ent, but the oxidizing agents are consumed as
groundwater passes through the reactive barrier.
2.6.4 Effect of Transformations of Iron
Sulfides on Reactivity
Butler and Hayes (2001) suggested that the
oxidation of mackinawite to greigite is faster at
acidic versus alkaline pH, and that mineral aging
in low pH groundwaters may eventually result in
the deactivation of iron sulfide mineral surfaces
with respect to reductive dechlorination. Butler
and Hayes (2001) have observed that aged FeS
at 76°C becomes completely unreactive to TCE
over 6 months. They detected greigite forma-
tion by XRD in aged FeS. Lee and Batchelor
(2002a) showed that the rate constant for
degradation of TCE on pyrite was 1390 to 2900
times slower than rate constants of degradation
reported for mackinawite and troilite. Devlin
and Miiller (1999) showed that FeS is more reac-
tive than FeS2 in degrading carbon tetrachloride.
Shao and Butler (2009) had a similar observation
in their microcosm study, which showed the
decline in the rate of degradation of carbon tetra-
chloride corresponded to a decrease in the abun-
dance of FeS and an increase in the abundance
of pyrite. Kriegman-King and Reinhard (1994)
showed that the condition of the pyrite surface
was found to be an important factor in determin-
ing both the rates of reaction and the product
distributions of carbon tetrachloride degradation.
Pseudo zero order kinetics was observed with
the fastest rates recorded for acid-washed and
freshly exposed pyrite surfaces in anaerobic
solutions. The slowest rates corresponded to
reactions on pyrite in aerobic solutions (Devlin
and Muller 1999; Hansson et al. 2008).
These observations show that different mineral
phases have different reactivity. The transfor-
mation of iron sulfide minerals is expected
to play a significant role in determining their
reactivity. Because the iron sulfide minerals
-------
are metastable at ambient conditions in the
environment, understanding the reactivity of
different iron sulfide phases and their transfor-
mation under different geochemical conditions
is important in understanding their role in
engineering remediation and natural attenuation.
2.6.5 Effect of Transition Metals on
Degradation
The presence of transition metals is common
as a constituent in mixed wastes with
chlorinated organic solvents. In addition,
transition metals can also be released under
reducing conditions where FeS typically
forms. Through their various interactions
with FeS, transition metals can change the
efficiency of FeS in the reductive dechlorina-
tion of chlorinated organic compounds.
Metal ions are characterized as hard or soft
depending on the Lewis acid base principle.
Due to their high affinity for sulfides, however,
intermediate to soft metal ions can react with
FeS in various ways: precipitation of pure metal
sulfides (MS), formation of metal-substituted
FeS by lattice exchange, and coprecipitation of
the mixed sulfides in a Fe-M-S system (Jeong
and Hayes 2003). Evidence of the incorporation
of metal impurities into iron sulfides has been
previously reported. For example, chemical
analyses of natural mackinawite by Ostwald
(1978) show that this mineral commonly con-
tains nickel, cobalt, and copper in its lattice.
Di Toro et al. (1990) found that metals with
lower sulfide solubility in general show a
greater tendency to replace iron atoms. In
contrast, Mn(II), whose sulfide is more soluble
than mackinawite, was not found to form
any sulfide precipitates in a suspension of
mackinawite (Arakaki and Morse 1993).
Choi et al. (2009) showed that the addition of
the transition metal Ni(II) and HS caused a
noticeable morphologic change of the surface
of FeS. The transition metal was substituted
for structural iron resulting in a decrease in
the iron content of the FeS (from 52.6% down
to 46.9%). They found that one third of the
transition metal in a suspension of FeS existed
in the zerovalent form. The zerovalent form
acts as a catalyst to accelerate degradation
of carbon tetrachloride and 1,1,1-TCA.
Depending on how metal ions interact with
FeS, transition metals can either inhibit or
stimulate the degradation of chlorinated organic
compounds. Fe(II) released as a result of
the interaction of FeS with intermediate/soft
metals, enhances the dechlorination of HCA
at concentrations of 10~4 and 10~3 M through
sorbed or dissolved Fe(II) species, while
Fe(OH)2(s) formed at a higher concentration
(10~2 M) also enhances reductive dechlorina-
tion. Co(II) and Hg(II) enhanced the reductive
dechlorination of PCE and TCE, but Ni(II)
slowed dechlorination (Jeong and Hayes 2007).
Choi et al. (2009) showed that the reduction
rate of 1,1,1-TCA increased with increasing
concentration of transition metals. The rate
constants with 10 mM Co(II) and Ni(II) were
0.06 and 0.11 h"1, approximately 1.3 and 3.0
times greater than rates with FeS alone. Jeong
and Hayes (2003) showed that metal ions
impacted the dechlorination rate of HCA by FeS
as a result of formation of secondary phases.
Rate increases observed in systems amended
with Co(II), Ni(II), and Hg(II) were not simply
explained by the formation of pure metal sulfide.
Instead, metal-substituted FeS or coprecipi-
tated sulfides are thought to be responsible for
the significantly increased rates observed in
these systems (Jeong and Hayes 2003).
2.7 Environmental Implications
2.7.1 Geochemical Parameters
Previous studies have investigated a series
of environmental factors that could affect the
abiotic degradation of halogenated aliphatic
compounds. It may be possible to use these
geochemical parameters to identify environments
where processes that degrade these organic
compounds are favored. Parameters such as
-------
transition metals, sulfide concentration, and natu-
ral organic matter have different effects on the
degradation of halogenated aliphatic compounds
depending on the specific compounds and
specific conditions involved. At the current state
of knowledge, it is difficult to use these param-
eters as reliable indicators of the degradation of
chlorinated organic compounds. However, pH
and the presence of anaerobic conditions seem to
affect the transformation of halogenated com-
pounds in a more consistent and predictable way.
Under iron and sulfur reducing conditions, the
formation of FeS can be expected. An increase
in pH will generally increase the rate of degrada-
tion of chlorinated organic compounds on FeS.
2.7.2 Differentiating Biotic and Abiotic
Degradation
Several approaches have been proposed to
differentiate abiotic degradation from bio-
logical degradation of chlorinated hydrocar-
bon compounds. Because biotic and abiotic
degradation follow different mechanisms and
have different degradation products, it may be
possible to differentiate biological degrada-
tion from abiotic degradation based on the
distribution of products. Acetylene has been
proposed as a product from abiotic degrada-
tion pathways (Butler and Hayes 1999).
Devlin and Miiller (1999) noticed that when
carbon tetrachloride reacted with freshly
precipitated, amorphous FeS at near neutral
pH, the ratio of the concentrations of chlo-
roform and carbon disulfide produced from
the reaction was 2:1 + 0.4. They proposed
that the 2:1 ratio may be a useful tool for
distinguishing abiotic transformations from
biodegradation in sulfate reducing environ-
ments where FeS is actively precipitated.
Abiotic processes may fractionate stable carbon
isotopes of carbon more strongly than biodegra-
dation. An analysis of stable isotope ratios can
be used as a tool to differentiate biotic vs. abiotic
degradation of TCE and PCE (Liang et al. 2007;
Zwank et al. 2005). More importantly, stable
isotope ratios can be used to recognize trans-
formation of chlorinated organic compounds
when it is difficult or impossible to characterize
degradation from the production or accumula-
tion of degradation products. As an example,
acetylene can be biologically degraded under
anaerobic conditions, and may not accumulate
in ground water where it is being produced.
2.7.3 Geochemical Modeling
When comprehensive data is available on water
chemistry, geochemical modeling may be used
as a technique to predict the formation of FeS
and other sulfides through the calculation of sat-
uration indices. Geochemical models, based on
water chemistry parameters, are useful in evalu-
ating whether or not potentially reactive phases
are present in situations where solids cannot be
collected and analyzed for their mineral content.
-------
-------
3.0
Magnetite
3.1 Structure of Magnetite
Magnetite (Fe3O4) is a mixed valence iron
mineral, containing roughly twice as much
Fe3+ as Fe2+. The formula can be expressed
as (Fe3+(Fe2+,Fe3+)O4 or as FeOFe2O3. See
Table 1.1 for a depiction of the three dimen-
sional structure. Halogenated organic com-
pounds can react with structural Fe2+ in magne-
tite, or with Fe2+ species sorbed to the surface
of magnetite. One third of the iron is in a
tetrahedral coordination with oxygen. All of the
iron in the tetrahedral layer is Fe3+. Two thirds
of the iron is in an octahedral coordination with
oxygen. Half of the iron in the octahedral layer
is Fe3+ and half is Fe2+. Danielsen and Hayes
(2004) and McCormick and Adriaens (2004)
note that electrons can "hop" between atoms of
iron in the octahedral layer, making iron in this
layer, or Fe2+ sorbed to this layer, more reactive.
3.2 Transformation Products of
Chlorinated Organic Compounds on
Magnetite
Vikesland et al. (2007) described the reac-
tion of carbon tetrachloride (CC14) with
magnetite as the transfer of one electron,
with simultaneous cleavage of a C-C1 bond.
The transfer of one electron produces the
trichloromethyl radical (-CC13) which can
react with hydrogen donors to produce chlo-
roform (HCC14) in the following reaction:
3Fe304 + CC14 + H+
Fe2+ +HCC13 + Cl"
The trichloromethyl radical can also react to
form other intermediates that finally decay
or react to form carbon monoxide, meth-
ane, or formate (McCormick and Adriaens,
2004; Vikesland et al., 2007). The final
distribution of products depends on pH
and availability of hydrogen donors.
Less is known of the reactions of alkenes with
magnetite. At this writing, the authors are
not aware of any published study that identi-
fies transformation products of PCE, TCE,
cis-DCE or vinyl chloride on magnetite. The
only transformation product identified by Lee
and Batchelor (2002a) on magnetite was chlo-
ride. Ferrey et al. (2004) analyzed pore water
in microcosms for reductive dechlorination
products of cis-DCE, such as vinyl chloride and
ethylene and ethane, and did not find them.
Darlington et al. (2008) followed the evolu-
tion of labeled carbon from 14C-TCE that was
incubated with samples of material from a
sandstone aquifer that had been contaminated
with chlorinated solvents. The aquifer matrix
contained a variety of potentially reactive miner-
als including iron sulfides, pyrite, fougerite,
magnetite, biotite, and vermiculite. The ground
water was predominantly anaerobic. In some
of the living microcosms, there was extensive
transformation of the TCE to cis-DCE. Within
15 to 16 months, up to 10% of the label accumu-
lated as 14C-CO2. In sterile samples (autoclaved
controls), the majority of label accumulated in
a fraction that was water soluble, but was not
stripped from the water by gas sparging after the
pH had been adjusted with HC1. Further analy-
sis indicated that this non-strippable fraction
included glycolate, formate, and acetate. The
only degradation products that were identified
are carbon dioxide and organic compounds that
can be readily biodegraded to carbon dioxide.
Ferrey et al. (2004) described the degradation of
cis-DCE in sediment from the site of the former
Twin Cities Army Ammunition Plant (TCAAP)
in Minnesota. The primary reactive mineral in
the sediment was magnetite. The surface area of
magnetite in the sediment, together with the sur-
face area specific rates of cis-DCE degradation
-------
on magnetite published by Lee and Batchelor
(2002a), could explain the removal of cis-DCE
in the sediment. Ferrey, Wilson, Lee, and
Bradley (Section B.4 of Appendix B) compared
the evolution of label from l4C-cis-DCE that
was incubated in sediment from TCAAP that
had been autoclaved. The only transformation
product that could be identified was 14C-CO2.
Danielsen and Hayes (2004) and McCormick
and Adriaens (2004) speculated that Fe2+ sorbed
to the octahedral layer of magnetite would
stabilize the trichlorocarbene ion, and shift
the products of carbon tetrachloride degrada-
tion away from chloroform toward carbon
monoxide, formate, and methane. If a similar
mechanism controls degradation of alkenes,
then carbon dioxide would be the ultimate
degradation product of cis-DCE on magnetite.
Little is known of reactions of chlorinated
alkanes on magnetite. Eisner et al. (2004b) eval-
uated the degradation of hexachloroethane by
Fe2+ sorbed to the surface of a series of minerals
including magnetite. The primary transforma-
tion product recovered was PCE. Degradation
of hexachloroethane on magnetite proceeded
through beta elimination, also called dihaloelimi-
nation, where two chlorine atoms were removed
from adjacent carbon atoms. This pathway
is also an important pathway for reaction of
alkenes on pyrite (Lee and Batchelor, 2002a)
and on iron sulfides (Butler and Hayes, 1998).
3.3 Effect of Properties of Magnetite
on Rate of Transformation
3.3.1 Effect of Particle Size and Surface Area
The rate of degradation of a chlorinated
organic compound on magnetite depends
on a number of factors. Because the reac-
tion occurs at the solid surface, the rate is
proportional to the surface area of magnetite
presented to the pore water in the aquifer
sediment, and rates in laboratory studies are
conventionally normalized to the surface
area of magnetite in the experimental system
(Lee and Batchelor, 2002a). Because volume
varies as the cube of the linear dimension,
while surface area varies as the square,
finely divided magnetite has much greater
surface area for a given mass of magnetite.
Vikesland et al. (2007) compared the rate of
degradation of carbon tetrachloride on magnetite
with an average particle size of 9 nm to deg-
radation on magnetite with an average particle
size of 80 nm. The rate of carbon tetrachloride
degradation in the 9 nm preparation was 1.4
L g-1 day1 atpH = 7.8, while the rate in the 80
nm preparation was 0.014 L g-1 day1. When
the rates were normalized to the surface area
of the magnetite, the rate of degradation of
carbon tetrachloride was still significantly faster
in the 9 nm preparation. The specific surface
areas of the 9 nm and 80 nm preparations were
63.5 and 14.5 m2 g-1 respectively. The surface
area normalized rate constants were 2.2 x
10-2 and 0.099 x 1Q-2 L nr2 day1, respectively.
Vikesland et al. (2007) attributed the faster
degradation in the preparation having an aver-
age particle size of 9 nm and specific surface
area of 63.5 m2 g-1 to quantum confinement
effects. These effects become important at
particle sizes less than approximately 20 nm.
The most detailed laboratory evaluations of the
degradation of chlorinated organic compounds
on magnetite are provided in McCormick et
al. (2002) and Lee and Batchelor (2002a). For
the laboratory studies to extrapolate to field
conditions, the particle size distribution of
magnetite should be similar. McCormick et
al. (2002) prepared biogenic magnetite from
a culture of iron reducing bacteria. Their
preparation of magnetite had a specific surface
area of 77.5 m2 g'1. Lee and Batchelor (2002a)
synthesized magnetite by mixing solutions
of iron(II) chloride and iron(III) nitrate, then
precipitating magnetite by adjusting the pH
with NaOH. The specific surface area of the
magnetite was 57.2 m2 g'1. As discussed in
Section 3.4.1, the average specific surface area
of magnetite isolated from aquifer sediment at
the TCAAP in Minnesota was 24 m2 g-1. The
-------
specific surface of these laboratory preparations
of magnetite compare well with the average
specific surface area of magnetite isolated from
aquifer sediment at the TCAAP in Minnesota.
3.3.2 Effect of pH
The rate of degradation of a variety of organic
compounds on magnetite increases as the pH
increases. This effect has been shown for carbon
tetrachloride by Danielsen and Hayes (2004) and
Vikesland et al. (2007), and for RDX (hexa-
hydro-l,3,5-trinitro-l,3,5-triazine) by Gregory
et al. (2004). Danielsen and Hayes (2004)
attributed the increase in reactivity to the degree
of protonation of sites on the mineral surface.
The rate of carbon tetrachloride degradation
on deprotonated sites in their preparation of
magnetite was approximately ten times greater
than the rate of degradation on protonated sites.
As the solution pH increased, the fraction
of sites which were deprotonated increased.
Ferrous iron sorbed to the surface of iron
minerals is postulated to form a =Fe-O-Fen
species where = Fe refers to iron in the struc-
ture of the mineral (Schoonen and Strongin,
2005). Danielsen and Hayes (2004) describe the
protonation reaction with the following equation:
= FemOFeu+ + H2O O = Fe1L'OFeL'OH" + H+
Figure 3.1 presents data from Danielsen and
Hayes (2004) on the effect of pH on the rate of
degradation of carbon tetrachloride on magne-
tite. They did not add additional Fe2+ to their
preparation of magnetite, and their preparation
of magnetite did not contain other iron minerals
such as siderite [Iron(II) Carbonate] that would
poise the concentration of Fe2+ in solution. As
a result, Figure 3.1 presents the effect of solu-
tion pH on carbon tetrachloride degradation
by structural iron in magnetite. There was a
steady increase in the rate constant for degrada-
tion with increased pH. In general, the rate
constant increases approximately 2.7 fold for
each ten-fold increase in [OH ], up to pH 8.0.
0.01 -i
0.006
0.003
ro
0
o
0.001
0.0006
0.0003
0.0001
8
PH
10
Figure 3.1. Effect of solution pH on the rate con-
stant for degradation of carbon tetra-
chloride on the surface of magnetite.
Redrawn from Danielsen and Hayes
(2004).
3.3.3 Effect of Fe2+ Sorbed to the Surface of
Magnetite
Lee and Batchelor (2002a) compared the
rates of degradation of cis-DCE and vinyl
chloride on magnetite as prepared, and in
the presence of 2.4 g/L Fe2+. In the presence
of a high concentration of Fe2+ in solution
(2,400 mg/L), the rates of degradation of cis-
DCE and vinyl chloride were tenfold higher.
Presumably, high concentrations of Fe2+ in
solution increased the concentration of Fe2+
that was sorbed to the surface of the magnetite
and was available to provide electrons for
reaction with cis-DCE and vinyl chloride.
Amonette et al. (2000) compared the effect of
sorbed Fe2+ on the rate of carbon tetrachloride
degradation on goethite. When the pH was
fixed at 7.0, the rate of carbon tetrachloride
degradation was directly proportional to the
density of Fe2+ sorbed to the goethite (mmol
g"1). Klausen et al. (1995) compared the effect
of sorbed Fe2+ on the rate of nitrobenzene
degradation on magnetite. At pH 7.0, the rate
increased concomitant with the amount of
sorbed Fe2+. Gregory et al. (2004) compared
the effect of sorbed Fe2+ on the rate of RDX
degradation on magnetite. The rate increased
-------
with the amount of sorbed Fe2+; however, the
increased rate was not directly proportional to
the increase in sorbed Fe2+. The apparent order
of the reaction with RDX with respect to the
concentration of sorbed Fe2+ was 1.8 + 0.3.
The concentration of sorbed Fe2+ on the surface
of magnetite should be related to the con-
centration of Fe2+ in solution. Danielsen and
Hayes (2004) determined the concentration of
iron in solution that would be in equilibrium
with their preparation of magnetite. The iron
was provided by dissolution of structural iron
in magnetite. Their data are presented in
Figure 3.2. If the concentration of total iron
exceeded the equilibrium concentration, some
other process such as biological iron reduction
must have contributed to iron in solution. If the
concentration of total iron was lower than the
values in Figure 3.2, some other process must
have removed the iron from solution, such as
sorption of iron to the surface of the magnetite.
100 -i
D)
LJ_
•a
a>
.a
o
(/5
.Q
0.1 -
0.05 -
Gregory et al. (2004)
Klausenetal. (1995)
*-
100 200 300 400
Dissolved Fe2+ (mg/L)
50
Figure 3.3. Sorption isotherms for Fe2+ on the
surface of magnetite. Redrawn from
Klausen et al. (1995) and Gregory et al.
(2004).
3.4 Rates of Transformation of
Chlorinated Organic Compounds on
Magnetite
3.4.1 Degradation ofcis-DCE and Vinyl
Chloride on Magnetite
At the Twin Cities Army Ammunition Plant
(TCAAP) north of St. Paul, Minnesota, a large
plume of TCE had moved beyond the plant
boundary and impacted municipal water supply
wells down gradient of the plant. In response,
the Army installed a series of pump-and-treat
wells that intercepted the plume near the plume
boundary. Based on a ground water transport
and fate model, the concentrations of TCE in
the plume were more than an order of magni-
tude lower than concentrations that would be
expected if there were no degradation of TCE.
Because of this discrepancy between measured
-------
and expected concentrations of TCE, the site
was proposed for a retrospective evaluation of
the U.S. EPA Technical protocol for evaluat-
ing natural attenuation of chlorinated solvents
in ground water (Wiedemeier et al., 1998).
Although TCE disappeared in the down gradi-
ent portion of the plume, there was very little
cis-DCE, and no vinyl chloride or ethylene
in the ground water. Low concentrations of
cis-DCE would be expected if cis-DCE were
being removed by natural biodegradation.
However, the retrospective evaluation found that
the geochemical environment in the plume was
not consistent with natural anaerobic biological
degradation of cis-DCE (Wilson et al., 2001).
In an attempt to identity the processes that were
removing chlorinated solvents, a microcosm
study was constructed with sediment from the
site (Ferrey et al., 2004). Ferrey et al. (2004)
selected sediment from an anaerobic, heavily
contaminated zone in the middle of the plume
with measureable Fe2+ in ground water, a less
contaminated zone at the water table, and a
marginally contaminated zone in oxygenated
ground water below the plume. The microcosm
study documented rates of removal of cis-DCE
in the sediment (Figure 3.4) that were consistent
with and would explain the low concentra-
tions of cis-DCE in the field scale plume.
Further, the microcosm study showed that
the removal process was indifferent to steril-
izing the sediment in an autoclave. Because
sterilizing the sediment did not affect removal,
natural biodegradation could not explain the
removal of cis-DCE. The rates of removal in
the microcosm study were also consistent with
the rate of decline in concentrations of cis-DCE
over time in monitoring wells at the site.
Working with magnetite that was chemi-
cally synthesized in their laboratory, Lee and
Batchelor (2002a) showed that magnetite could
degrade a variety of chlorinated alkenes includ-
ing PCE, TCE, cis-DCE, and vinyl chloride.
To determine whether magnetite in the aquifer
sediment could explain the removal of cis-DCE
in ground water at the TCAAP, Ferrey et al.
(2004) isolated the magnetic fraction from the
sediment used for their microcosm study, used
electron microscopy to estimate the distribution
of particle sizes, and used X-ray diffraction anal-
ysis to estimate the mean particle size for mag-
netite. They used the magnetic susceptibility of
the sediment to estimate the content of magnetite
in the sediment, and from the content and
specific surface area of magnetite they estimated
the surface area of magnetite exposed to water in
the sediment from TCAAP. Lee and Batchelor
(2002a) had published two surface area specific
rate constants for degradation of cis-DCE on
magnetite, one for magnetite by itself, and a
faster rate for magnetite in the presence of high
concentrations of Fe2+ in solution. A model
assuming a specific surface area for magnetite
of 100 m2 g"1, and the higher rate of degradation
most closely matched the experimental data.
10000-,
1000 -
O)
LU
o
Q
100 -
10 -
Live Microcosms
Sterile Microcosms
Container Controls
0 200 400 600 800 1000
Time of Incubation (days)
Figure 3.4. Removal of cis-DCE in aquifer sediment
containing magnetite. Redrawn from
Ferrey et al. (2004).
-------
Fitted line is a model based on magnetite
particle size and magnetic susceptibility of
the sediment, and the surface area
specific rate of c/s-DCE degradation in
Lee and Batchelor (2002a).
10000
1000 -
§> 100 -
O
Q
10 -
live microcosms
sterile microcosms
-200 0 200 400 600 800 1000
Time of Incubation (days)
Figure 3.5. The surface area specific rate of deg-
radation of c/s-DCE on magnetite can
explain the removal of c/s-DCE in sedi-
ment from the TCAAP. Redrawn from
Ferrey et al. (2004).
and the experimental data from the microcosm
study were in good agreement (Figure 3.5).
Following the assumption in the model above,
the mass magnetic susceptibility of the sedi-
ment used in the experimental treatments in
Ferrey et al. (2004) was used to estimate the
concentration of magnetite in the sediments.
The concentration of magnetite in the laboratory
experiments of Lee and Batchelor (2002a) were
similar to the natural concentration of magne-
tite in the sediment from the TCAAP, and the
specific surface area of the magnetite that was
chemically synthesized by Lee and Batchelor
(2002a) was similar to the specific surface area
of magnetite isolated from the TCAAP sediment.
Figure 3.5 presents a further elaboration on
the model of cis-DCE degradation presented
in Ferrey et al. (2004). Ferrey et al. (2004)
used the peak broadening of the strongest
magnetite reflections at d spacings of 2.97,
2.53, and 1.48 A to estimate the mean particle
size. They reported a mean magnetite crystal
size of 100 nm. The actual best estimate of
the particle size was 46 nm (Robert Ford, U.S.
EPA, personal communication). Assuming a
crystal of magnetite is a cube, this would cor-
respond to specific surface area of 24 m2 g"1. In
the model presented in Figure 3.5, the amount
of magnetite in the sediment was estimated
from the magnetic susceptibility of sediment
(1.6 x 10'6 m3 kg"1) assuming that the quantity
of magnetite is proportional to the mass mag-
netic susceptibility, and that a mass magnetic
susceptibility of 2.7 x 10~6 m3 kg"1 is equivalent
to 1% by weight magnetite in silica sand. The
relationship between mass magnetic susceptibil-
ity and the content of magnetite is described
in more detail in Section 3.5.1. The model
further assumed that the water filled porosity of
the sediment was 0.25. The model projection
-------
Table 3.1. Surface area specific rate constants for removal of c/s-DCE and vinyl chloride on magnetite.
Reference
Source of
Magnetite
First-Order Rate
Constant
± 95%
Confidence
Interval
Microcosm
Container Control
year1
Magnetite
exposed to
pore water
gL-1
Specific
Surface
Area
Magnetite
m2g-1
Surface Area
Specific
Rate Constant for
Transformation
L nr2 day1
Removal of cis-DCE
Lee and
Batchelor
(2002a)
Lee and
Batchelor
(2002a)
Ferrey et al.
(2004)
Ferrey et al.
(2004)
Ferrey et al.
(2004)
This study
This study
This study
Chemically
synthesized
Chemically
synthesized,
2400 mg/L Fe2+
Natural in
TCAAP shallow
Fe2+ < 2.5 mg/L
Natural in
TCAAP
Intermediate
Fe2+ < 2.5 mg/L
Natural in
TCAAP Deep
oxic
Natural in
TCAAP T5
shallow
Natural in
TCAAP Site A
anoxic
Natural in
TCAAP Site 102
oxic
0.58 ± 0.090
0.077 ±0.082
2.29 ± 0.25
0.077 ±0.052
0.31 ± 0.08
0.077 ±0.052
0.82 ± 0.39
0.26 ±0.08
0.73 ± 0.18
0.27 ±0.04
0.65 ± 0.20
0.27 ±0.04
63
63
27
47
41
25
29
27
57.2
57.2
24
24
24
24
24
24
0.560 x 10-6
5.74 x IQ-6
1.32 x 10-6
2.8 x IQ-6
0.44 x 10-6
1.97 x IQ-6
1.47 x IQ-6
1.46 x 10-6
Removal of vinyl chloride
Lee and
Batchelor
(2002a)
Lee and
Batchelor
(2002a)
Ferrey et al.
(2004) and
This study
Chemically
synthesized
Chemically
synthesized,
2400 mg/L Fe2+
Natural in
TCAAP Deep
sediment,
oxidized
0.311 ± 0.12
63
63
41
57.2
57.2
24
0.564 x 10-6
5.78 xlQ-6
0.44 x 10-6
-------
To extract a surface area specific rate constant
from the data published by Ferrey et al. (2004),
the rate constant for removal of the chlorinated
organic contaminants in the sediment was
divided by the surface area of magnetite exposed
to ground water in the sediment. Table 3.1
compares the surface-area specific rate constants
for removal of cis-DCE and vinyl chloride in
the experiments of Lee and Batchelor (2002a) to
rate constants calculated from the data on aquifer
sediment presented in Ferrey et al. (2004).
As mentioned above, Lee and Batchelor (2002a)
determined rate constants for cis-DCE removal
for their magnetite as synthesized, and for
magnetite that was incubated with high concen-
trations of Fe2+ in solution to maximize sorption
of Fe2+ to the surface of their magnetite prepara-
tion. The rate constant for cis-DCE degradation
was tenfold higher on magnetite in the Fe2+
solution. As mentioned previously, Ferrey et
al. (2004) selected sediment from an anaerobic,
heavily contaminated zone in the middle of the
plume with measureable Fe2+ in ground water,
a less contaminated zone at the water table,
and a marginally contaminated zone in ground
water from the bottom of the plume. There was
a six-fold difference in the rates of cis-DCE
degradation. Degradation was faster in the more
contaminated, reduced sediment. The rate of
cis-DCE degradation in sediment from the con-
taminated intermediate depth was comparable to
the rate of removal on chemically synthesized
magnetite incubated with high concentrations
of Fe2+, while the rate of degradation in the oxic
sediment was comparable to removal on chemi-
cally synthesized magnetite without added Fe2+.
Table 3.1 compares rate constants from pub-
lished microcosm studies to rate constants from
microcosm studies that were conducted as part
of this study. Details of the microcosm studies
are presented in Appendix B. The microcosm
study described as TCAAP T5 in Table 3.1
used material from the same location as used
for Ferrey et al. (2004). The experimental
procedures followed Ferrey et al. (2004). In
contrast, the microcosm studies described as
TCAAP Site A and TCAAP Site 102 are from
two other sites on the TCAAP. These studies
were designed to strictly exclude any biological
activity from either Dehalococcoides strains
that could use cis-DCE as an electron acceptor,
or iron-reducing bacteria that can use cis-DCE
as an electron donor or carbon source (Bradley,
2003). The sediment was dried to a powder,
autoclaved, and made into microcosms with
sterile oxygenated water. The microcosms
were incubated in a dark cabinet exposed to
air. The sediment or microcosms were never
exposed to hydrogen in an anaerobic glove
box. The surface area specific rates of removal
of cis-DCE in these microcosms were com-
parable to the rates of removal of cis-DCE in
the microcosm studies that were constructed
and incubated in an anaerobic glove box.
Table 3.1 also compares the rate constants for
removal of vinyl chloride in the experiments
of Lee and Batchelor (2002a) to unpublished
data from the experiment described in Ferrey
et al. (2004). The data and experimental
details are provided in Appendix B. As was
the case for cis-DCE, the rate of degrada-
tion of vinyl chloride was ten times faster on
the chemically synthesized magnetite incu-
bated with Fe2+, and the rate of removal of
vinyl chloride in the oxic sediment was very
similar to the rate on the chemically synthe-
sized magnetite without additional Fe2+.
3.4.2 Degradation of PCE and TCE on
Magnetite
Surface area specific rate constants for removal
of PCE and TCE are presented in Table 3.2. The
rate constants for PCE and TCE degradation
reported by Lee and Batchelor (2002a) on chem-
ically synthesized magnetite were very similar
to their rates for degradation of cis-DCE and
vinyl chloride. The rates of degradation of PCE
or TCE in oxidized natural sediment from the
TCAAP, or from the nearby Baytown Superfund
site in Minnesota, are fivefold to tenfold higher
than the rates on the chemically synthesized
-------
magnetite. Lee and Batchelor (2002a) did not
report experiments on the effects of Fe2+ on rates
of degradation of PCE and TCE on magnetite.
If the rates in the presence of Fe2+ are enhanced
tenfold, as was the case with cis-DCE and vinyl
chloride, the rates of degradation of PCE or TCE
on the chemically synthesized magnetite would
compare closely with the rates of PCE or TCE
degradation in the sediment from the TCAAP.
The rates of degradation of PCE, TCE,
cis-DCE, and vinyl chloride on chemically
synthesized magnetite as reported by Lee and
Batchelor (2002a), on aquifer sediment from
the TCAAP as reported by Ferrey et al. (2004),
and this study all fall into a fairly narrow range
(0.56 to 8.6 x IQ-6 L nr1 day1). If the surface
area of magnetite in a sediment sample can be
determined, it may be possible to estimate the
rate of degradation of PCE, TCE, cis-DCE., or
vinyl chloride within an order of magnitude.
Table 3.2. Surface area specific rates of removal of PCE and TCE on magnetite.
Reference
Source of
Magnetite
First-Order Rate
Constant
± 95%
Confidence
Interval
Microcosm
Container Control
yr1
Magnetite
exposed to
pore water
gL-1
Specific
Surface
Area
Magnetite
n^g-1
Surface Area
Specific
Rate Constant for
Transformation
L m~2 day"1
Removal of PCE
Lee and
Batchelor
(2002a)
This study
This study
Chemically
synthesized
Component
TCAAP Site A
sediment
Component
TCAAP Site 102
sediment
0.20 ± 0.15
0.83 ±0.23
1.32 ± 0.45
0.83 ±0.23
63
51
57.2
24
0.84 x 10-6
5.6 x 10-6
Removal of TCE
Lee and
Batchelor
(2002a)
This study
This study
This study
Chemically
synthesized
Component
TCAAP Site A
sediment
Component
TCAAP Site 102
sediment
Component
Baytown sediment
0.32 ± 0.11
0.51 ±0.13
0.95 ± 0.31
0.51 ±0.13
0.98 ± 0.49
0.45 ±0.11
63
51
21
57.2
24
24
0.72 x 10-6
4.1 x 10-6
9.3 x 10-6
-------
As mentioned above, the sediment from
TCAAP Site A and Site 102 received spe-
cial treatment to preclude any possibility
of biodegradation of chlorinated organic
compounds. The sediments were dried to
a powder before they were autoclaved, and
were resuspended in oxygenated water. The
removal of TCE and PCE in microcosms
constructed with sediment from Site A and
Site 102 was compared to loss from container
controls that were filled with sterile water
but no sediment (Table 3.2). Both sediments
removed cis-DCE at the same rate (Table 3.1),
but the sediment from Site A did not remove
TCE or PCE any faster than the loss from the
container controls, while the sediment from
Site 102 removed TCE and PCE at rates that
were equivalent to the removal of cis-DCE.
It is difficult to explain the behavior of PCE
and TCE in the sediment from Site A. There
was no treatment in the study with a "living"
microcosm. As a result, there is no way to
determine whether the lack of activity was
a property of the sediment as collected, or
whether the magnetite was altered when the
sediment was dried and autoclaved. The
aquifer at Site A was anoxic with undetect-
able concentrations of dissolved oxygen
and measurable concentrations of Fe2+and
Mn2+, while the aquifer at Site 102 often
had measureable concentrations of dissolved
oxygen in addition to Fe2+ and Mn2+.
Notice that the rate of removal of PCE and
TCE in the microcosms constructed with
sediment from Site A was significantly slower
than the removal in the container controls that
were only filled with sterile water. The sedi-
ment in the microcosms may have restricted
exchange of water in the microcosms with the
face of the Teflon-faced septum that sealed the
microcosm, and thus imposed a mass transfer
limitation on escape of PCE or TCE from the
container. The removal of PCE or TCE in the
Site A microcosms is probably a better control
on losses of PCE and TCE from the Site 102
microcosms or the Baytown microcosms
than is the removal in the container controls.
Because the removal in the Site A microcosms
was slow (0.2 to 0.3 per year) compared to
the removal in the other microcosms, the
removal in the other microcosms was not
corrected for removal in the container controls.
3.4.3 Degradation of Carbon Tetrachloride
on Magnetite
The rate of carbon tetrachloride degradation on
magnetite was at least one hundredfold faster
than the rate of degradation of the chlorinated
alkenes (Compare Table 3.2 and 3.3). In the
presence of Fe2+, the rates of degradation of
carbon tetrachloride were roughly similar on
magnetite produced by a culture of iron reduc-
ing bacteria (McCormick et al., 2002) or mag-
netite synthesized in the laboratory (Danielsen
and Hayes, 2004; Zwank et al., 2005).
Vikesland et al. (2007 and Supporting
Information) compared the rate of degrada-
tion of carbon tetrachloride on chemically
synthesized nano-iron (particle size 9 nm)
with a specific surface area of 63.5 m2 g"1
and "conventional" iron (particle size 80 nm)
with a specific surface area of 14.5 m2 g"1.
The surface area specific rate constant for
degradation of carbon tetrachloride was 22
times faster with nano-iron. Vikesland et al.
(2007) attributed the faster rate to quantum
confinement in the nano-scale particles. They
also showed that aggregation of the magnetite
particles relieved the quantum confinement,
and produced lower rates of degradation. The
biogenic magnetite particles produced in the
iron-reducing culture of McCormick et al.
(2002) had a specific surface area of 68.9 m2
g"1, indicating that they were also nano-scale,
yet the rate of carbon tetrachloride degrada-
tion was similar to conventional magnetite.
Perhaps aggregation reduced the rate of
degradation in their experimental system.
-------
Table 3.3. Surface area specific rate constants for removal of carbon tetrachloride on magnetite.
Reference
Source of Magnetite
Magnetite
exposed to
pore water
gL-1
Specific
Surface Area
Magnetite
m2g-!
Surface Area Specific
Rate Constant for
Transformation
L m~2 day"1
Removal of carbon tetrachloride
McCormick et
al. (2002)
Danielsen
and Hayes
(2004)
Zwank et al.
(2005)
Vikesland et al.
(2007) and SI
Vikesland et al.
(2007) and SI
Separated from
a culture of iron
reducing bacteria
32 to 37 mg/L Fe2+
Chemically
synthesized
pH 7.03
0.82 mg/L Fe2+
Chemically
synthesized
pH 7.2,
57 mg/L Fe 2+
Chemically
synthesized
pH7.8, 9nm
Chemically
synthesized
pH 7.8, 80 nm
7.3 to 26
25
2.6
5
5
77.5 or 68.9
18
19
63.5
14.5
8.9 x 10-4
4.8 x IQ-4
1.2 x 10-1
2.2 x 10-2
9.9 x 10-4
-------
Table 3.4. Relationship between the rate of removal of PCE, TCE, c/s-DCE or vinyl chloride and the content of
total iron, and the magnetic susceptibility of aquifer sediment.
Source of Sediment
First-Order
Rate Constant
for Removal in
Sediment
± 95%
Confidence
Interval
Microcosm
Container
Control
yr-1
Mass Magnetic
Susceptibility
SI Units
m3 kg -1
First-Order
Rate Constant
Normalized by
Mass Magnetic
Susceptibility
yr -1 m ~3 kg
Iron
Expected
From
magnetite
mg kg -1
Total
Iron
Nitric
Acid
extract
mg kg -1
Removal of PCE in sediment
TCAAP Site A
TCAAP 102
0.20 ± 0.15
0.83 ±0.23
1.32 ± 0.45
0.83 ±0.23
1.0 x 10-6
0.91 x 10-6
1.4 x 106
5,090
4,600
7,780
5,520
Removal of TCE in sediment
TCAAP Site A
TCAAP 102
Baytown
0.32 ± 0.11
0.51 ±0.13
0.95 ± 0.31
0.51 ±0.13
0.98 ± 0.49
0.45 ±0.11
1.0 x 10-6
0.91 x 10-6
0.41 x 10-6
1.1 x 106
2.4 x 106
5,090
4,600
1,940
7,780
5,520
12,100
Removal of cis-DCE in sediment
TCAAP shallow
(Ferrey et al., 2004)
TCAAP intermediate
(Ferrey et al., 2004)
TCAAP deep
(Ferrey et al., 2004)
TCAAP shallow
T5
TCAAP Site A
TCAAP 102
Thermo-Chem
0.58 ± 0.090
0.077 ±0.082
2.29 ± 0.25
0.077 ±0.052
0.31 ± 0.08
0.077 ±0.052
0.82 ± 0.39
0.26 ±0.08
0.73 ± 0.18
0.27 ±0.04
0.65 ± 0.20
0.27 ±0.04
0.58 ± 0.39
0.26 ±0.08
0.9 x 10-6
1.6 x 10-6
1.4 x 10-6
0.86 x 10-6
1.0 x 10-6
0.91 x 10-6
0.15 x 10-6
0.64 x 106
1.4 x 106
0.22 x 106
0.95 x 106
0.73 x 106
0.71 x 106
3.9 x 106
4,540
8,460
7,320
4,309
5,090
4,600
656
6,520
10,300
9,160
Not done
7,780
5,520
1,500
Removal of vinyl chloride in sediment
TCAAP deep
(Ferrey et al., 2004)
0.311 ±0.12
1.4 x 10-6
0.22 x 106
7,320
9,160
-------
3.5 Rates of Transformation
Normalized to Magnetic
Susceptibility
A determination of mass magnetic susceptibility
is a useful tool for screening sediment for the
presence of magnetite, and thus for the possibil-
ity of abiotic degradation of chlorinated organic
compounds on magnetite. Table 3.4 compares
the abiotic rate of removal of PCE, TCE, cis-
DCE and vinyl chloride normalized to the mass
magnetic susceptibility. The normalized rate of
removal should be influenced by the effect of
pH, the specific surface area of the magnetite,
the presence or absence of sorbed Fe2+, and
variations in the mass magnetic susceptibility of
the magnetite in the sediment samples. Despite
all these influences, and the uncertainly associ-
ated with the estimate of the rate of removal and
the mass magnetic susceptibility, the normal-
ized rates extend over little more than an order
of magnitude. The rates were higher in the
sediments that were anoxic based on measured
concentrations of Fe2+ in ground water (TCAAP
intermediate, or Thermo-Chem). Presumably
the magnetite in sediment in contact with Fe2+
in ground water had higher concentrations of
Fe2+ sorbed or precipitated on the surface.
Note in Table 3.4 that no rates are reported
for removal of PCE or TCE in sediment from
the TCAAP Site A. Although a mass mag-
netic susceptibility in the range of 1 x 10~6
m3 kg"1 may indicate that degradation of PCE
or TCE at rates near 0.3 to 2 per year is pos-
sible, but should not be taken as proof that
abiotic degradation will occur at those rates.
-------
-------
4.0
Transformation of
Chlorinated Hydrocarbons by Green Rusts
4.1 Structure and Reactivity of Green
Rusts
4.1.1 Chemical Composition and Crystal
Structure of Green Rusts
Green rusts are Fe(II)-Fe(III) hydroxysalts
belonging to the general class of layered
double hydroxides (Hansen, 2001). They
consist of ordered layers of Fe(II)/Fe(III)
hydroxides carrying positive charges, and
interlayers containing anions and water
molecules (Figure 4.1). Green rusts have
a general formula [?Q(ll\ xFe(III)x(OH)2]x+
[(x/n)An~ •mH2O]x" where An~is the intercalated
anions (An- = Cl~, SO42-, CO/ ...) and x is the
Fe(III) molar fraction (x = 0.25 to 0.33) (Genin
et al., 1998; Hansen, 2001). The type of anion
(charge, shape, and size) defines the crystal
structure, which gives a distinct X-ray diffraction
pattern for green rusts (Figure 4.2).
Green rusts can exchange counter ions, so if
a sulfate green rust were prepared and added
to a solution of bicarbonate, it might well be
converted to carbonate green rust, depending
on the relative concentrations of the two anions.
The carbonate green rust would be favored over
sulfate green rust when [SO42-]/CO32-] < 1259
(Su and Puls, 2004) based on thermodynamic
calculations. Such a conversion may not be
assured due to possible slow kinetics, but it is a
possibility. It is expected that carbonate green
rust should be the dominant form of green rusts
that could be formed in non-mining impacted
sites. Green rusts can be readily synthesized in
a laboratory. Green rusts have been reportedly
found in soils and sediments under suboxic and
anoxic conditions (Genin et al., 1998; Refait et
al., 2001; Bearcock et al., 2006).
Figure 4.1. The crystal structure of green rust com-
pounds consists of layers of Fe(ll)(OH)6
in which some of the Fe(ll) is replaced
by Fe(lll). These alternate with layers
of anions (SO/-, CO/-, and Ch) and
water which bind the Fe(OH)6 layers via
hydrogen bonding (From Randall et al.,
2001). If the interlayer is dominated by
sulfate, carbonate, or chloride, then the
material is described as sulfate green
rust, carbonate green rust, or chloride
green rust.
-------
c
3
o
o
IUUUU •
18000
16000
14000
12000
10000
8000
6000
4000
2000
n .
IO
8
10 03
I GR(S042-)
_J\ ,X
O)
S £ GR(C032-)
'"I O
0
LJLjLjL_^^JUL^___.
— j\ —
0 10 20 30 40 50 60 70 80 90 100
Degree 26 / Fe Ka
Figure 4.2. X-ray diffraction pattern of freshly syn-
thesized carbonate green rust and sul-
fate
uieziiz.eu uaiuuiiate yieen /uii c«//u iu;-
fate green rust (less than 24 hours after
synthesis) scanned as glycerol smears.
Only d-spacings for basal reflections
are shown (nm).
4.1.2 Chemical Reactivity of Green Rusts
Green rusts are very reactive minerals involved
in the redox cycling of iron in both aquatic and
terrestrial environments. Green rusts react read-
ily with redox sensitive contaminants due to the
iron in the layered structure. Their large external
and internal surface area promotes the adsorption
of contaminants and facilitates the association
of contaminants with the reactive sites. They
have been shown to chemically reduce a range
of organic and inorganic contaminants such as
nitrates (Hansen et al., 2001), Cr(VI) (Loyaux-
Lawniczak et al., 2000), Se(IV) (Myneni et
al., 1997), Ag(I), Au(III), Cu(II), and U(VI)
(O'Loughlin et al., 2003; Suzuki et al., 2008),
as well as chlorinated hydrocarbons (Erbs et al.,
1999; Lee and Batchelor, 2002b; O'Loughlin
and Burris, 2004; Scherer et al., 2007; Chun
et al., 2007; Choi and Lee, 2008). Green rusts
also adsorb contaminants such as As(V) and
As(III) via surface complexations (Randall et al.,
2001; Su and Wilkin, 2005). Carbonate green
rust partially oxidizes As(III) to form As(V) (Su
and Wilkin, 2005).
The rates of degradation of chlorinated hydro-
carbons on the surface of green rusts should be
a function of their surface area that is presented
to the chlorinated hydrocarbons in water solu-
tion. Table 4.1 presents data on the specific
surface area of a variety of forms of green rust.
The specific surface area varies by almost two
orders of magnitude. Bond and Fendorf (2003)
reported values of 30.1 m2^1 for carbonate green
rusts GR(CO32), 19.0 m2^1 for chloride green
rusts GR(C1") and 3.6 m2^1 for sulfate green
rusts GR(SO42") using the BET N2 adsorption
method on a Coulter SA 3100 surface area ana-
lyzer. Samples were dried in the glove box by
vacuum desiccation for 1 week prior to analysis.
They claim that exposure to air during transfer to
the instrument was minimal as evidenced by the
lack of color change even in the most sensitive
green rust sample, GR(CO32~). The highest
specific surface area (86.3 m2^1) was reported
by Lee and Batchelor (2002b) for sulfate green
rust using EGME adsorption. To date, to the
best of our knowledge, there is no report in the
literature on the specific surface area of naturally
occurring green rusts.
Table 4.1. Specific surface area values reported in
the literature for synthetic green rusts
Green
Rust
GR(C1)
GR(CO32)
GR(CO32)
GR(CO32)
GR(S042)
GR(SO42)
GR(SO42)
GR(SO42)
Surface
area
m'g1
19.0
30.1
37.1 ±9.4
47 ±7
(n = 5)
3.6
14.1
28.4
86.3
Method
BETN2
BETN2
BETN2
BETN2
BETN2
BETN2
BETN2
EGME
Reference
Bond and
Fendorf (2003)
Bond and
Fendorf (2003)
Chun et al.
(2007)
Williams and
Scherer (2001)
Bond and
Fendorf (2003)
O'Loughlin et al.
(2003)
This study
Lee and
Batchelor
(2002b)
-------
4.1.3 Abiotic Degradation ofPCE and TCE by
Green Rusts
4.1.3.1 Reports of Degradation of PCE and TCE
in the Literature
Earlier reports show that abiotic transformation
of PCE and TCE by sulfate green rusts occur
via reductive p-elimination to produce acetylene
and other completely dechlorinated products
(Lee and Batchelor, 2002b; Scherer et al., 2007).
In contrast, microbial reductive dechlorination of
PCE takes place via sequential hydrogenolysis to
yield less chlorinated ethenes along the sequence
TCE, c/5-DCE, vinyl chloride, ethene, and
ethane.
In contrast to the reports of Lee and Batchelor
(2002b, 2003), Scherer et al. (2007) reported
little reduction of 17 mg/L PCE or 13 mg/L TCE
by chloride green rust, sulfate green rust, or
carbonate green rust at concentrations of 5 or 7
g/L at pH 7.0 to 8.3. However, they did report
substantial TCE degradation after 60 days of
incubation when TCE was reacted with 10 g/L
sulfate green rust at pH 8.2 in 0.02 M NaHCO3.
Choi and Lee (2008) did not detect PCE degra-
dation when PCE was reacted with sulfate green
rust for 2 days. PCE only degraded if platinum
chloride (0.5 mM) was added to the reaction
mixture. This seems to be in contradiction to the
data presented in Figure 1 of Lee and Batchelor
(2002b), where approximately 20% of PCE
disappeared after 2 days of reaction. But that
decrease in concentration was only partly due to
degradation. As indicated by the control, much
of the reduction was due to rapid partitioning of
PCE from the water to the container. The con-
centration in the control was decreased to about
90% of the original value in the first sample
taken. Therefore, it would be more reasonable
to say that 5 to 10% of PCE was degraded in 2
days. A difference in experimental procedures
might explain the difference in behavior of green
rust in these experiments; the green rust used in
Lee and Batchelor (2002b) was dried before it
was used in the experiment, but the green rust
used in Choi and Lee (2008) was never dried.
Data from the unpublished dissertation of Choi
(2005) supports the lack of reactivity of sulfate
green rust with PCE when sulfate green rust is
not dried (See Table 4.2, p 50 in Choi, 2005). It
is unclear why drying of sulfate green rust would
enhance its reactivity.
Reactive metal ions may be important to under-
standing the degradation of chlorinated hydro-
carbons by green rusts. Maithreepala and Doong
(2005), found that Cu(II) in chloride green rust
enhanced the degradation of PCE, TCE, and
carbon tetrachloride.
4.1.3.2 Degradation of TCE in Studies Performed
at Kerr Center (U.S. EPA)
In an effort to verify the earlier findings,
researchers in EPA's Ground Water and
Ecosystems Research Division (GWERD)
conducted batch tests on the abiotic degradation
of TCE and carbon tetrachloride by synthetic
green rusts. The concentrations of chlorinated
compounds (2-10 mg/L) used in this study are
comparable to those found in the plume areas of
contaminated sites. Freshly synthesized green
rusts were used to ensure material integrity. The
concentration of green rusts in water (4.17 g/L)
would correspond to a concentration of iron
in aquifer sediment of approximately 0.03%.
Calculations assume a water filled porosity of
30% and a density of 2.65 g/cm3for aquifer
sediment. Experimental details are provided in
Appendix C.
Figure 4.3 depicts the degradation of TCE at a
nominal concentration of 10 mg/L in suspen-
sions containing 4.17 g/L of carbonate green
rust. Over 31 days of incubation, there was
no significant degradation of TCE at pH 10.3,
or pH 8.8 or pH 7.0. None of the expected
reaction products of TCE (c/s-DCE, 1,1-DCE,
trans-DCE, vinyl chloride, acetylene, ethene,
ethane) were detected at quantifiable concentra-
tions. Because this type of green rust is likely
the most common green rust in ground waters
contaminated with chlorinated hydrocarbons,
these results would suggest that the potential for
-------
degradation of TCE by green rusts may not be as
great as previously believed.
0.5
O
o
c
0.5
0.0
-0.5
-1.0
-1.5
TCE C0 = 10 mg/L,
•
GR(CO32') = 4.17 g/L
• • •
• pH10.3(a)
• pH 8.8 (b)
A pH 7.0, 35
Fittob, R2
Fit toe, R2
mM Fe(ll) (c)
= 0.065
= 0.115
= 0.012
200
400
Time (hours)
600
800
Figure 4.3. Removal of TCE in the presence of
carbonate green rust as a function of
reaction time, pH, and excess dissolved
Fe(ll). TCE was reacted with carbon-
ate green rust in 0.037 M Na2SO4 +
0.030 M Na2CO3.
Lee and Batchelor (2002b) reported a surface-
area-normalized pseudo first order initial rate
constant (&sa) value of 8.55 x 10~5 L m~2 day1
for TCE degradation by sulfate green rust at
pH 7. Over a pH range from 6.8 to 10.1, Lee
and Batchelor (2002b) showed that as the pH
increased from 6.8, the initial ksa values were
1.64 fold higher at 7.3; 2.16 fold higher at pH
8.1; 3.05 fold higher at pH 9.2; and 3.96 fold
higher at pH 10.1.
In the experiments at Kerr Center, TCE was not
degraded in a suspension of sulfate green rust at
pH 8.2 or at pH 8.0 (Figure 4.4). No detectable
products were found.
0.0'
o
o
O -0.5
-1.0
-1.5
TCE C0 = 10 mg/L, GR(S042-) = 4.17 g/L
pH 8.2 (a)
pH 8.0 Trizma buffer (b)
Fit to a, R2 = 0.078
Fittob, R^ = 0.136
200
400
600
800
Time (hours)
Figure 4.4. Removal of TCE in the presence of
sulfate green rust as a function of reac-
tion time with and without a 0.05 M pH
8 Trizma buffer. TCE was reacted with
sulfate green rust in 0.074 M Na2SO4.
In the EPA study, TCE degraded when 0.5 mM
CuSO4 was present in the suspensions of sulfate
green rust. The average final equilibrium pH
was 7.9. After 30 days of incubation, 42%
of the TCE disappeared, yielding a value of
1.36 x 10"4 L m~2 day1 for the k . This value
J sa
is slightly larger than the initial ksa value of
8.55 x 10'5 L m^day1 reported by Lee and
Batchelor (2002b) for TCE degradation by
sulfate green rust at pH 7 without a catalyst such
as Cu(II).
Reaction products included c/s-DCE, 1,1-DCE,
tmns-DCE, methylene chloride, ethene, and
ethane. Addition of both 0.5 mM Cu(II) and
0.05 M Trizma buffer (pH 8.0) did not produce
detectable TCE transformation. Addition of
2.0 mM Cu(II) without a buffer also did not pro-
duce measurable TCE degradation (Figure 4.5);
however, trace amounts of ethene and ethane
were detected in these two treatments. This
implies that there is an optimal concentration of
Cu(II) at which the reaction is catalyzed. Even
in the presence of the optimal concentration of
Cu(II), TCE degradation was not great and there
was more scatter in the data with Cu(II) than
without Cu(II).
-------
0.5
-1.0
-1.5
-2.0
GR(SO42-) = 4.17g/L
5 mg/L TCE, 0.5 mM Cu(ll|, pH 7.9 (a|
10 mg/L TCE, 0.5 mM Cu(ll|, pH 8.0 Trizma buffer (b|
10 mg/L TCE, 2 mM Cu(ll), pH 7.7 (c)
Fit to a, R2 = 0.521
Fit to b, R2 = 0.043
Fit to c, R2 = 0.001
200
400
600
800
Time (hours)
Figure 4.5. Removal of TCE in the presence of sul-
fate green rust as a function of reaction
time, concentration of added CuSO4,
and presence or absence of 0.05 M
Trizma buffer. TCE was reacted with
sulfate green rust in 0.074 M Na2SO4.
The differences in results between the EPA study
and the study of Lee and Batchelor (2002b) may
be explained by differences in experimental
procedures. The ratios of the mass of sulfate
green rust to the mass of TCE in both studies
were similar; however, the preparation of sulfate
green rust used by Lee and Batchelor (2002b)
had a specific surface area that was three times
greater than the sulfate green rust used in the
EPA study (Table 1). The higher surface area
may have resulted in enhanced dechlorination.
Also, drying of green rust may have had an
effect. Further work is needed to provide a
useful understanding of the role of green rusts
in the abiotic degradation of PCE and TCE in
contaminated ground water.
4.1.4 Abiotic Degradation of Chlorinated
Methanes andAlkanes by Green Rusts
4.1.4.1 Reports of Degradation of Chlorinated
Methanes and Alkanes in the Literature
Chun et al. (2007) studied the kinetics and
pathways of the degradation of selected
halogenated disinfection byproducts (DBFs)
in the presence of carbonate green rust.
Trichloronitromethane was rapidly degraded
to methylamine via sequential hydrogenolysis
followed by nitroreduction. Haloacetic acids
reacted solely via sequential hydrogenolysis.
Trichloroacetonitrile, 1,1,1 -trichloropropanone,
and trichloroacetaldehyde hydrate were trans-
formed via hydrolysis and hydrogenolysis.
In contrast, chloroform was unreactive over
300 hours of incubation.
The chemical nature of the pH buffer affected
the rates of reductive dehalogenation of disin-
fection byproducts, with faster rates in MOPS
buffer than in carbonate buffer, the latter being
representative of the buffer in drinking water
systems. Carbonate green rust was unstable in
both buffers and transformed to magnetite within
48 hours.
O'Loughlin et al. (2003) reported a ksa value of
2.07 x 10-2L nr2 day1 at pH 7.6 for the reduc-
tion of carbon tetrachloride by sulfate green rust
within the first 6 hours of the reaction. Overall,
the reaction was not well described by first-order
kinetics. The authors demonstrated the catalytic
activity of transition metal species in the reduc-
tion of chlorinated hydrocarbons, suggesting
the potential for enhanced reduction by green
rusts in the presence of an appropriate transition
metal catalyst. The rates of reduction of carbon
tetrachloride were greatly increased in systems
amended with Cu(II), Au(III), and Ag(I) (listed
in order of increasing rates) relative to green rust
alone.
Erbs et al. (1999) reported values of 0.41 to
1.88 day1 for degradation of carbon tetrachloride
by sulfate green rust at pH « 8. They provided
no information on the concentration of sulfate
green rust or the total surface area of the sulfate
green rust.
Various degradation products have been reported
for carbon tetrachloride. Erbs et al. (1999)
showed that reduction of carbon tetrachloride by
sulfate green rust at pH 8 produced chloroform
(CHC13) and hexachloroethane (C2C16) as the
main chloroaliphatic products, while the sulfate
green rust was oxidized to magnetite (Fe3O4).
The formation of C2C16 indicates a coupling
reaction between trichloromethyl radicals in the
suspension.
-------
The first-order rate constants for transformation
of carbon tetrachloride with zerovalent iron and
with sulfate green rust were in the same range.
Thus, green rusts formed during corrosion of
iron(O) under nonacid conditions may make a
substantial contribution to the total reduction
of carbon tetrachloride measured in iron(O)
systems.
Chloroform was much less susceptible than
carbon tetrachloride to reductive dechlorination
by sulfate green rust. Rates of degradation for
chloroform were approximately 100 times less
than rates for carbon tetrachloride.
O'Loughlin et al. (2003) observed a variety
of intermediates and products of carbon tet-
rachloride degradation, including chloroform,
dichloromethane, chloromethane, methane,
acetylene, ethene, ethane, carbon monoxide,
tetrachloroethene, and various nonchlorinated C3
and C4 compounds. The distribution of products
during the reductive dechlorination of carbon
tetrachloride was highly dependent on the transi-
tion metal used.
Recently, X-ray absorption fine structure analy-
sis of aqueous green rust suspensions amended
with Ag(I), Au(III), or Cu(II) showed that the
metals were reduced to their zerovalent forms.
A possible mechanism for carbon tetrachloride
reduction is the formation of a galvanic couple
involving the zerovalent metal and green rust,
with reduction of carbon tetrachloride occurring
on the surface of the metal and green rust serv-
ing as the bulk electron source. The enhanced
reduction of carbon tetrachloride by suspensions
of green rust amended with Ag(I), Au(III), or
Cu(II) may prove useful in the development of
improved materials for remediation of chlori-
nated organic contaminants.
No evidence in the study of O'Loughlin et al.
(2003) showed that the reduction of carbon
tetrachloride and chloroform to methane can be
envisioned as resulting from a series of sequen-
tial hydrogenolysis reactions (i.e., the stepwise
replacement of hydrogen for chlorine), such that
carbon tetrachloride —» chloroform —» dichloro-
methane —» chloromethane —» methane. Results
suggest that the reduction of carbon tetrachloride
by green rust, silver green rust, gold green rust,
and copper green rust involves processes other
than sequential hydrogenolysis. In this reaction
sequence, carbon tetrachloride is reduced primar-
ily to methane and other non-chlorinated end
products, largely through a series of one-electron
reductions forming radicals and carbenes or
carbenoids.
4.1.4.2 Degradation of Carbon Tetrachloride in
the EPA Study
The degradation of carbon tetrachloride by
sulfate green rust was pH dependent (Figure
4.6). At low pH, only limited degradation
of carbon tetrachloride was observed; the ks^
was 1.58 x 10-6 L nrMay1 at pH 4.45 and
3.20 x 10-4 L m-2 day1 at pH 6.05. The £ a values
ranged from 2.02 x 10~3 L nr2 day1 at pH 6.64
to 3.87 x 10-3L nvMay1 at pH 9.80, suggest-
ing that in this pH range the reaction rates are
not strongly influenced by changes in pH. The
highest &sa value of 5.61 x 10~3L m~2 day1 was
observed at pH 11.16. In the EPA study, chloro-
form, methylene chloride, methane, and ethylene
were detected as reaction products.
-------
-2-I-4"
O
O -4
c
-6
-8
CT C0 = 2 mg/L, GR(S042-) = 4.1 7 g/L
100 200 300
Time (hours)
400
pH 4.45 (a)
pH 6.05 (b)
pH 6.64 (c|
pH 7.92 (d|
pH 8.86 (e)
pH 9.80 If)
Fit
Fit
Fit
Fit
Fit
o a, Rz = 0.161
o b, Rz = 0.169
o c, Rz = 0.970
o f, Rz = 0.721
o g, Rz = 0.792
Figure 4.6. Removal of carbon tetrachloride in
the presence of sulfate green rust as
a function of reaction time and pH.
Carbon tetrachloride was reacted with
sulfate green rust in 0.074 M Na2SO4.
reduced than the isomer with greater symmetry
(e.g., 1,1,1-TCA> 1,1,2-TCA). The addition
of Ag(I) or Cu(II) to suspensions of green rust
resulted in a substantial increase in the rate of
reduction of halogenated ethanes as well as sig-
nificant differences in the product distributions
with respect to green rust alone.
Dechlorination of most chlorinated ethanes
was through reductive |3-elimination to pro-
duce vinyl chloride, ethene, and ethane, and/or
through dehydrochlorination to produce DCE
isomers (Figure 4.7). The reaction mechanism
for dechlorination of 1,1,1-TCAby sulfate green
rust is uncertain.
4.1.4.3 Reports of Degradation of Chlorinated
Alkanes in the Literature
O'Loughlin and Burris (2004) examined reduc-
tion of halogenated ethanes in aqueous suspen-
sions of sulfate green rusts, both alone and with
the addition of Ag(I)(AgGR) and Cu(II) (CuGR).
Hexachloroethane (HCA), pentachloroethane
(PCA), 1,1,1,2-tetrachloroethane (1,1,1,2-TeCA),
1,1,2,2-tetrachloroethane (1,1,2,2-TeCA),
1,1,1 -trichloroethane (1,1,1 -TCA), 1,1,2-tri-
chloroethane (1,1,2-TCA), 1,1-dichloroethane
(1,1-DCA), and 1,2-dibromoethane were reduced
in the presence of green rust alone, AgGR, or
CuGR. Only 1,2-dichloroethane and chloro-
ethane were nonreactive. The reduction was
generally more rapid for more highly substituted
ethanes than for ethanes having fewer halogen
groups (HCA > PCA > l,l,l,2-TeCA> 1,1,1-
TCA> l,l,2,2-TeCA> 1,1,2-TCA> 1,1-DCA),
and isomers with the more asymmetric distri-
butions of halogen groups were more rapidly
-------
Cl
ci-;:/
Cl
HCA
pi
CI
Cl
\ / dehydrochlorination \
Cl Cl
PCE
,CI
Cl
c=c—ci
DCAc
(not observed)
01
Cl ^ ' Cl
\ / dehydrochlorination
C=C -*---
\
Cl
Cl
Figure 4.7. Proposed pathways for the reduction of chlorinated ethanes in aqueous suspensions of green
rusts and in green rust suspensions spiked with Ag(l) (AgGR) or Cu(ll) (CuGR); however, some
elements of the pathways shown are not relevant to all experimental systems. Pathways leading
to the formation of the C4 hydrocarbons observed as products of the reduction of 1,1,1-trichloro-
ethane (1,1,1-TCA) by AgGR or 1,1,2-trichloroethane (1,1,2-TCA) and pentachloroethane (PCA)
by CuGR are not shown. HCA = hexachloroethane; 1,1,1,2-TeCA = 1,1,1,2-tetrachloroethane;
1,1,2,2-TeCA = 1,1,2,2-tetrachloroethane;1,1-DCA = 1,1-dichloroethane; 1,2-DCA = 1,2-dichlo-
roethane; CA = chloroethane; PCE = perchloroethene; TCE = trichloroethene; 1,1-DCE = 1,1-di-
chloroethene; cis-1,2-DCE = cis-1,2-dichloroethene; trans-1,2-DCE = trans-1,2-dichloroethene;
VC = vinyl chloride; DCAc = dichloroacetylene; CAc = chloroacetylene; Ac = acetylene (From
O'Loughlin and Burns, 2004).
-------
Table 4.2 summarizes the available information chlorinated hydrocarbons by green rusts.
on the rate constants for abiotic degradation of
Table 4.2. Rate constants for transformation of chlorinated hydrocarbons by green rusts. Rate constants are
quoted or converted to surface-area-normalized pseudo-first-order initial rate constant (/csa) when
information is available.
Compound
PCE
TCE
TCE
TCE
TCE
TCE
TCE
TCE
TCE
TCE
CT
CT
CT
CT
CT
CT
CT
CT
cis-DCE
VC
Mineral
GR(S042)
GR(S042)
GR(CO32)
GR(CO32)
GR(CO32)
GR(S042)
GR(S042)
GR(S042)
GR(S042)
GR(S042)
GR(S042)
GR(S042)
GR(S042)
GR(S042)
GR(S042)
GR(S042)
GR(S042)
GR(S042)
GR(S042)
GR(S042)
pH
7.0
7.0
10.3
8.8
7.0
8.0
8.2 Trizma
7.9
8.0 Trizma
7.7
7.6
4.45
6.05
6.64
7.92
8.86
9.80
11.16
7.0
7.0
Rate Constant &sa
L m 2 day x
1.62 x 10-4
8.55 x 10-5
No reaction
No reaction
No reaction
No reaction
No reaction
1.36 x 10-4
No reaction
No reaction
2.07 x 10-2
1.58 x 10-6
3.20 x IQ-4
2.56 x 10-3
2.02 x lO-3
3.87 x 10-3
3.65 x 10-3
5.61 x 10-3
5.19 x 10-5
7.77 x 10-5
Reference
Lee and Batchelor (2002b)
Lee and Batchelor (2002b)
This Report
This Report
This Report (35 mM Fe(II))
This Report
This Report
This Report (0.5 mM Cu(II))
This Report (0.5 mM Cu(II))
This Report (2.0 mM Cu(II))
O'Loughlin et al. (2003)*
This Report
This Report
This Report
This Report
This Report
This Report
This Report
Lee and Batchelor (2002b
Lee and Batchelor (2002b)
"for the data of first 6 hours of reaction only.
4.2 Occurrence and Determination of
Green Rusts
Our current knowledge on the occurrence and
distribution of green rusts in contaminated sites
is not adequate to evaluate their contribution to
the abiotic transformation of chlorinated hydro-
carbons in ground water. More work is needed
to attain a practical capability to predict how
and when green rust may form at a site and to
quantitatively predict their capacity to degrade
chlorinated hydrocarbons.
4.2.1 Geochemical Conditions that Favor
Green Rust Formation
Green rusts can form from partial oxida-
tion of Fe2+ in the groundwater or directly
from corrosion of metallic iron that is used
in reactive barriers and for reinforcement of
underground radioactive waste repositories.
The name "fougerite" has been formally
approved by the Commission on New Minerals
and Mineral Names of the International
Mineralogical Association (IMA) (number
-------
2003-057), on January 29, 2004 (Feder et al.,
2005). The mixed valence compound foug-
erite (IMA 2003-057), a layered Fe(II)-Fe(III)
hydroxysalt, is found in transitionally oxic
and anoxic environments like hydromorphic
soils (Genin et al. 1998; Trolard et al. 2007).
Fougerite (IMA 2003-057) is a mixed M(II)-
M(III) hydroxysalt of the green rust group,
where M(II) can be Fe or Mg, and M(III)
is Fe. The general structural formula is:
[Fe1_x2+Fe/+Mgy(OH)?+2j]+ljc/«^-n. wH20]- where
A is the interlayer anion andn its valency, with
1/4 < x/(l +y)< 1/3 and m<(\-x+y). The
mineral forms by partial oxidation and hydro-
lysis of aqueous Fe2+, to give small crystals
(400-500 nm) in the form of hexagonal plates.
The mineral is unstable in air and transforms to
lepidocrocite or goethite. The name is for the
locality of the occurrence, a forested Gleysol
near Fougeres, Brittany, France. Its character-
istic blue-green color (5BG6/1 in the Munsell
system) has long been used as a universal crite-
rion in soil classification to identify Gleysols.
Dissolved silicon affects phase formation during
neutralization and partial oxidation of iron(II)
chloride solutions. In the absence of Si, the
initial phase further reacts to form green rust,
presumably a double hydroxide of Fe2+ and Fe3+
iron which ultimately forms lepidocrocite. In
the presence of Si, the formation of green rust
structure from the initial phase is inhibited, as
evidenced by symmetrical basal and asymmetri-
cal hk x-ray diffraction (XRD) lines, and ferri-
hydrite is the ultimate oxidation product (Karim,
1986).
Biogenic formation of carbonate green rust was
observed in well controlled laboratory experi-
ments (Ona-Nguema et al., 2002a; Ona-Nguema
et al., 2002b). They confirmed that carbonate
green rust is metastable with respect to magne-
tite in the presence of y-FeOOH.
The competitive formation of carbonate green
rust and sulfate green rust in bacterial cultures is
dependent on the relative ratio (R) of bicarbonate
and sulfate concentrations (Ona-Nguema et al.,
2004). When R > 0.17, only carbonate green
rust was formed whereas when R < 0.17, a
mixture of sulfate green rust and carbonate green
rust was obtained. Ona-Nguema et al. (2004)
used Shewanellaputrefaciens, a dissimilatory
iron-reducing bacterium (DIRB) to anaerobically
catalyze the transformation of a ferric oxyhy-
droxide, lepidocrocite (y-FeOOH), into Fe(II) in
the presence of various sulfate concentrations.
These results demonstrated that the sulfate green
rust can originate from the microbial reduc-
tion of y-FeOOH and confirmed the preference
for carbonate over sulfate during green rust
precipitation.
Zegeye et al. (2005) investigated the formation
of only sulfate green rust during the reduction of
y-FeOOH by a DIRB. Their experimental study
demonstrated that, under a H2 atmosphere, the
biogenic solid that was produced was a sulfate
green rust and it was the sole iron(II-III) bearing
mineral, regardless of the initial lepidocrocite
concentration. The crystals of the biotically
formed sulfate green rust were significantly
larger than those observed for sulfate green
rust obtained through abiotic preparation. The
formation of green rust by Shewanella species
isolated from a wide range of habitats and pos-
sessing varied metabolic capabilities suggests
that under favorable conditions biogenic green
rusts may be formed by a diverse array of DIRB
(O'Loughlin et al., 2007; Zegeye et al., 2007).
4.2.2 Structural Stability of Carbonate Green
Rust
Synthetic carbonate green rust is not stable; it
quickly changes to magnetite and Fe(OH)2. No
information is available on the potential for
degradation of chlorinated solvents by Fe(OH)2.
Studies have shown that the presence of certain
anions help stabilize the crystal structure from
degradation. These anions include phosphate
(Bocher et al., 2004) and arsenate or arsenite (Su
and Wilkin, 2005). Bocher et al. (2004) showed
that in the presence of phosphate as low as
{[POJmm0 /[Fe]} K 1%, carbonate green rust is
-------
stable. No P was detected by TEM-EDX when
the basal (0001) crystal faces were analyzed,
showing that no P was inserted in the GR inter-
layers. In contrast, very low quantities of P were
detected when the analyses were performed on
the lateral faces of the green rust crystals. The
preference of phosphate species for lateral faces
is due to the fact that the {1010} faces contain
mono- and di-coordinated OH-surface groups.
The PO43 adsorbed on the lateral faces may act
as a barrier that slows down the release in solu-
tion of the CO32 species preventing carbonate
green rust from transforming into a mixture of
Fe3O4 and Fe(OH)2. See Figure 4.8.
Since carbonate green rust has been found to
form in the natural environment by coprecipita-
tion, some kind of stabilization agents (phos-
phate, silicate, etc.) may play a role.
0
o o e
o o o ©
o o o 0,
o "
O O
© o o o
e o o
€>
o
Fe'"
Fe"
Figure 4.8. Adsorption of the phosphate on the
lateral {1010} face of the GR(CO/~)
crystal that stabilizes it (From Bocher et
ai, 2004).
-------
-------
5.0
Phyllosilicate Clays
5.1 Structural Iron in Phyllosilicate
Clays
Phyllosilicate clays are composed of sheets of
SiO4 and A1O4 tetrahedra with an approximate
chemical formula of (Al,Si)3O4. Other elements
such as iron can replace silicon or aluminum
in the sheets. This iron is incorporated into the
structure of the clay with co-valent bonds, and is
referred to as structural iron. The structural iron
can be iron(II) or iron(III), and its oxidation state
can change as the clay interacts with oxidiz-
ing or reducing agents. The structural iron can
also interact through shared oxygen atoms with
iron(II) sorbed to the surface of the clay.
In experimental systems, the rate law for the
degradation of the organic compound will
frequently change over time. Lee and Batchelor
(2004) described the kinetics of degradation
of PCE, TCE, cis-DCE, and vinyl chloride on
biotite, vermiculite, and montmorillonite with
a model that contained two reactive sites. One
type of reactive site was present at lower con-
centrations than the other type of site, but was
much more reactive than the second type of
site. Over time, the more reactive type of site
is consumed, and the overall rate of degradation
slows. Neumann et al. (2008) used nitroaro-
matic compounds as probes to characterize and
compare two preparations of montmorillonite
clay and a ferruginous smectite that had different
proportions of iron(II) and iron(III) and differ-
ent concentrations of total iron. They could
explain the contrasting behavior of the probes on
the clays by associating more reactive and less
reactive sites with Fe(III)-O-Fe(II) entities and
Fe(II)-O-Fe(II) entities in the crystal structure of
the clays.
5.2 Degradation Processes on
Phyllosilicate Clays
Kriegman-King and Reinhard (1992) examined
the degradation products of carbon tetrachloride
on biotite in the presence of HS~. The primary
products were chloroform, carbon disulfide, and
carbon dioxide. The pH of the water had little
effect on the rate of degradation, or the relative
concentration of the degradation products.
Cervini-Silva et al. (2001) found that
1,1,2,2-tetrachloroethane degraded to TCE
on ferruginous smectite. Neumann et al.
(2009) also compared degradation of carbon
tetrachloride, 1,1,1,2-tetrachloroethane,
1,1,2,2-tetrachloroethane, and several model
compounds on ferruginous smectite. The
only product detected from the degradation of
carbon tetrachloride was chloroform. There
was no effect of pH on the rate of degradation.
Transformation of 1,1,1,2-tetrachloroethane was
primarily through reductive |3-elimination to
form 1,1-dichloroethylene , while transformation
of 1,1,2,2-tetrachloroethane was through
dehydrochlorination to produce TCE.
Nzengung et al. (2001) showed that PCE
degraded on smectites to form TCE, and then
cis-DCE, vinyl chloride, acetylene, ethylene, and
ethane. In contrast to the behavior of carbon
tetrachloride, Lee and Batchelor (2004) found
that the rate constant for degradation of PCE
on biotite increased with increasing pH. As
pH increased from 5.5 to 8.1, the rate constant
increased threefold, then the rate constant
doubled between pH 8.1 and 8.5.
-------
5.3 Rate Constants for Degradation of
Chlorinated Alkenes
Table 5.1 provides rate constants available from
the literature for the degradation of chlorinated
alkenes on the phyllosilicate clays: biotite,
montmorillonite, and vermiculite. The observed
pseudo first-order rate constants are normalized
to the surface area of the mineral presented to
water in the suspension, and to the amount of
structural iron in the minerals. The clays had
been pretreated with dithionite to reduce the
structural iron to iron(II). Lee and Batchelor
(2004) found that the kinetics of degradation
were biphasic, with an initial rapid phase that
was followed by a transition to a slower phase.
The kinetic parameters reported by Lee and
Batchelor (2004) were used to calculate pseudo
first-order rate constants for the initial rapid
phase, which are reported in Table 5.1.
The first order rate constants normalized to
the concentration of structural iron fell into a
fairly narrow range, extending between 0.020
and 0.65 M^d"1. The rates of degradation of
PCE, TCE, cis-DCE, and vinyl chloride were
comparable on a particular preparation of clay,
but varied from clay to clay. On montmoril-
lonite, the rate constants for degradation of the
four chlorinated alkenes varied from 0.65 M^d"1
for PCE to 0.33 M^d'1 for vinyl chloride. On
vermiculite, the rates varied from 0.020 M^d"1
for PCE to 0.18 M-'d'1 for vinyl chloride.
The rate constants for degradation of the chlo-
rinated alkenes on the phyllosilicate clays (after
reduction of structural iron with dithionite) were
comparable to the rate constant for degradation
on pyrite, and exceeded the rate constant for
degradation on magnetite (Table 5.1).
When the rate constant was normalized to
the surface area of the clay mineral, the rate
constants varied widely, depending on the clay
mineral. The preparations of biotite, vermiculite,
and montmorillonite used by Lee and Batchelor
(2004) have specific surface areas of 1.9 m2/g,
26.7 m2/g and 488 m2/g respectively. The
reactivity of the phyllosilicate clays was more
influenced by the concentration of structural iron
than by the surface area presented to water. As
the specific surface area increased, the normal-
ized rate of degradation decreased. The rate
constants for the four chlorinated alkenes varied
from 9.4 L m2 d"1 to 13.5 L m2 d"1 on biotite, from
0.062 L m2 d'1 to 0.67 L m2 d'1 on vermiculite,
and from 0.0090 L m2 d'1 to 0.018 L m2 d'1 on
montmorillonite.
5.4 Rate Constants for Degradation
of Carbon Tetrachloride and
Chlorinated Alkanes
Table 5.2 provides rate constants available from
the literature for the degradation of carbon tet-
rachloride, 1,1,1-TCA, and 1,1,2,2-tetrachloro-
ethane on phyllosilicate clays, including biotite,
a ferruginous smectite, and vermiculite. When
expressed on a surface area basis, the reactiv-
ity of the phyllosilicate clays was intermediate
between pyrite and magnetite. Rate constants
reported by Kriegman-King and Reinhard (1992)
and Neumann et al. (2009), when normalized
to the concentration of structural iron in their
experiments, also indicate that reactivity of
structural iron in biotite and vermiculite is
comparable to the reactivity of structural iron in
pyrite or magnetite. The rate constants for deg-
radation of carbon tetrachloride on these clays
were approximately ten to one-hundred times
larger than the rate constants for the chlorinated
alkenes (compare Table 5.2 and 5.1).
The rate constants for degradation of carbon
tetrachloride on a preparation of ferruginous
smectite was approximately one-hundred fold
higher than rate constants for degradation on
biotite and vermiculite. It is unclear whether this
result comes from some difference in experimen-
tal procedure or approach, or represents a real
difference in the behavior of carbon tetrachloride
on these clays.
The rate constant for degradation of
1,1,1,2-tetrachloroethane was equivalent to
the rate constants for degradation of carbon
-------
tetrachloride, and the rate constant for
degradation of 1,1,2,2-tetrachloroethane was
three to fourfold faster.
5.5 Effect of Sorbed Iron (II) on the
Rate of Degradation
The effect of sorbed iron(II) is modest. Lee and
Batchelor (2004) compared the effect of sorbed
iron on the kinetics of degradation of chlorinated
alkenes on biotite. As sorbed iron(II) increased
from zero to 0.15 mM g"1, the rate of degradation
on the clay surface increased 1.5 fold. Across
a variety of experimental conditions, including
a range of pH from 5.5 to 8.1, and for PCE,
TCE, c/s-DCE, and vinyl chloride reacting on
biotite, vermiculite, and montmorillonite, the
rate constant for reaction on the clay surface was
generally less than twofold higher in the pres-
ence of 4.28 mM Fe(II).
5.6 Extrapolation of Rate Constants to
the Field
Nzengung et al. (2001) reported that the rate
constants for degradation of PCE on the clay
minerals was not a linear function of the concen-
tration of clay minerals. The rate of degradation
in a 1% (weight/volume) suspension of clay
was actually slower than the rate of degradation
in a 0.5% suspension. Nzengung et al. (2001)
explained this effect as being due to aggregation
of the clay particles that reduced the surface area
available to react with chlorinated organic com-
pounds in water. Aggregation will be important
for natural clays in aquifer sediment.
If the rate constants in Tables 5.1 and 5.2 are
realized in aquifer material at field scale, they
can make a substantial contribution to the
destruction of these contaminants (Lee and
Batchelor, 2004). However, the rate constants
in Tables 5.1 and 5.2 are likely to be an upper
boundary on the rates that might be achieved
at field scale. This is particularly true because
most of the experiments were done with clays
that had been artificially reduced with dithionite.
The rate constants in Tables 5.1 and 5.2 are
probably not appropriate for aquifers that are
naturally aerobic. They may be appropriate for
aquifers with ongoing sulfate reduction.
The rate constants presented in this section were
developed for homogeneous samples of clays in
laboratory experiments. To our knowledge, no
one has published an estimate of structural iron
in clay minerals in a sample of aquifer material,
and related the concentration of structural iron
to the rate of degradation of chlorinated organic
contaminants. Because the reactivity of phyl-
losilicate clays is more directly related to the
concentration of reactive structural iron in clay
than to the total weight of clay or the surface
area of clay, the total content of structural iron
in clay, or perhaps the total content of structural
iron(II) in clay, would be an appropriate param-
eter to predict abiotic degradation.
An estimate of total structural iron in the clay
fraction of aquifer sediment is theoretically pos-
sible at the present state of practice, but making
that estimate would require a complex series of
extractions and digestions. To determine the
concentration of structural iron in phyllosilicate
clays, it would first be necessary to perform
extractions to remove the exchangeable iron
and iron oxide minerals. Depending on the
redox status of the aquifer, it might be necessary
to remove the chromium reducible sulfides to
remove iron associated with mackinawite and
pyrite. Presumably, the major portion of iron
remaining in the sample would be structural iron
associated with phyllosilicate clays. The final
step will require digestion of the sample with
48% HF and 3.6 M H2SO4 and determination of
iron using the 1,10-phenanthroline colorimetric
method (Stucki, 1981; Komadel and Stucki,
1988).
The biphasic kinetics of degradation seen in
laboratory experiments poses another impedi-
ment to extrapolation of the rate constants in
Tables 5.1 and 5.2 to predict behavior in the
field. Lee and Batchelor (2004) estimated the
reductive capacity of individual samples of phyl-
losilicate clays to PCE, TCE, c/s-DCE, and vinyl
-------
chloride. Table 5.3 uses their data to estimate
the total concentration of these chemicals that
could be degraded by the clays when present
at 10% of the total dry weight of aquifer sedi-
ment. This would be a high concentration of
clay for a sediment that would readily transmit
water and would function as a useful aquifer.
Approximately 20 to 50 mg/L of chlorinated
alkenes would be destroyed. Phyllosilicate clays
can only be expected to make a significant con-
tribution to the natural attenuation of relatively
dilute plumes of chlorinated solvents.
The reactive iron in the phyllosilicate clays is
a small fraction of the total structural iron in
the clays. Lee and Batchelor (2003) found that
only 1% of Fe(II) in the clay they examined
could reduce PCE. Table 5.3 also expresses the
specific initial reductive capacity of the clays
examined by Lee and Batchelor (2004) on the
basis of moles of chlorinated solvent destroyed
per mole of total iron. Reductive dechlorination
requires two electrons to remove one chlorine
from the organic compound. Two moles of iron
would be required to dechlorinate PCE, TCE,
cis-DCE, or vinyl chloride to the next degrada-
tion product. A very small fraction of the total
iron in the clays was available to reduce the
chlorinated organic compounds.
-------
Table 5.1 Rate constants for degradation of chlorinated alkenes on phyllosilicate clay minerals. Data for pyrite
and magnetite are provided for comparison.
Compound
PCE
PCE
PCE
PCE
PCE
PCE
TCE
TCE
TCE
cis-DCE
cis-DCE
cis-DCE
VC
Mineral
Pyrite
Magnetite
Biotite
reduced with
dithionite
Montmorillonite
reduced with
dithionite
Montmorillonite
reduced with
dithionite
Vermiculite
reduced with
dithionite
Biotite
reduced with
dithionite
Montmorillonite
Vermiculite
reduced with
dithionite
Biotite
reduced with
dithionite
Montmorillonite
reduced with
dithionite
Vermiculite
reduced with
dithionite
Biotite
reduced with
dithionite
pH
8
7
7
7
8.5
7
7
7
7
7
7
7
7
Other
Conditions
22 °C, 10 mM
NaHCO3
22 °C, 10 mM
NaHCO3
22 °C, 10 mM
NaHCO3
22 °C, 10 mM
NaHCO3
25 °C, 34.5
mM dithionite
22 °C, 10 mM
NaHCO3
22 °C, 10 mM
NaHCO3
22 °C, 10 mM
NaHCO3
22 °C, 10 mM
NaHCO3
22 °C, 10 mM
NaHCO3
22 °C, 10 mM
NaHCO3
22 °C, 10 mM
NaHCO3
22 °C, 10 mM
NaHCO3
Observed
first order
rate constant
normalized to
surface area
of mineral
L rrr2 d-1
2.4 x 10 5
0.082 x 10 5
13.5 x lO'5
0.018 x 10-5
1.5 x 10-5
0.074 x 10-5
10 x lO'5
0.018 x 10-5
0.062 x lO'5
9.4 x 10-5
0.010 x 10-5
0.30 x lO'5
12 x lO'5
Observed
first order
rate constant
normalized to
concentration
structural
iron
M-1 day1
0.17
0.0036
0.12
0.65
19
0.020
0.092
0.64
0.016
0.086
0.37
0.079
0.111
Reference
Lee and Batchelor
(2002a)
Lee and Batchelor
(2002a)
Lee and Batchelor
(2004)
Lee and Batchelor
(2004)
Nzengung et al.
(2001)
Lee and Batchelor
(2004)
Lee and Batchelor
(2004)
Lee and Batchelor
(2004)
Lee and Batchelor
(2004)
Lee and Batchelor
(2004)
Lee and Batchelor
(2004)
Lee and Batchelor
(2004)
Lee and Batchelor
(2004)
-------
Table 5.1 Continued.
Compound
vc
vc
Mineral
Montmorillonite
reduced with
dithionite
Vermiculite
pH
7
7
Other
Conditions
22 °C, 10 mM
NaHCO3
22 °C, 10 mM
NaHCO3
Observed
first order
rate constant
normalized to
surface area
of mineral
0.0090 x 10-5
0.67 x 10-5
Observed
first order
rate constant
normalized to
concentration
structural
iron
0.33
0.18
Reference
Lee and Batchelor
(2004)
Lee and Batchelor
(2004)
-------
Table 5.2 Rate constants for degradation of carbon tetrachloride and chlorinated alkanes on phyllosilicate clay
minerals. Data for pyrite and magnetite are provided for comparison.
Compound
carbon
tetrachloride
carbon
tetrachloride
carbon
tetrachloride
carbon
tetrachloride
carbon
tetrachloride
carbon
tetrachloride
carbon
tetrachloride
carbon
tetrachloride
1,1,1,2-
tetrachloro-
ethane
1,1,2,2-tetra-
chloroethane
Mineral
Pyrite
Magnetite
Biotite
Biotite
Ferruginous
Smectite
Ferruginous
Smectite
Reduced with
dithionite
Ferruginous
Smectite
reduced with
dithionite
Vermiculite
Ferruginous
Smectite
Ferruginous
Smectite
pH
6.5
7.0
8.6
8.8
NR
8.4
7.5
8.3
NR
NR
Other
Conditions
25 °C, 1 M NaCl
25 °C, 1 M NaCl
50 °C, 1 mM
HS-
50 °C, 1 mM
HS-
5 mM NaCl
5 mM NaCl
25 °C, 0.5 M
NaCl
50 °C, 1 mM
HS-
5 mM NaCl
5 mM NaCl
Observed
first order
rate constant
normalized to
surface area
of mineral
L nr2 d'1
1620 x 10 4
4.80 x 10-4
33.2 x lO'4
23.1 x 10-4
not reported
not reported
not reported
8.95 x lO'4
not reported
not reported
Observed
first order
rate constant
normalized to
concentration
structural iron
M-1 day1
1.7
0.67
4.4
3.1
232
483
1.75
7.0
380
1,040
Reference
Kriegman-King and
Reinhard (1994)
Danielsen and
Hayes (2004)
Kriegman-King and
Reinhard (1992)
Kriegman-King and
Reinhard (1992)
Cervini-Silva et al.
(2001)
Cervini-Silva et al.
(2001)
Neumann et al.
(2009) from panel
A, Figure S2
Kriegman-King and
Reinhard (1992)
Cervini-Silva et al.
(2001)
Cervini-Silva et al.
(2001)
-------
Table 5.3. Maximum quantity (C°RC) of chlorinated hydrocarbon that can be reduced by reactive iron in repre-
sentative phyllosilicate clays. Calculated by multiplying the concentration of clay exposed to pore
water (g/L) by the specific initial reductive capacity of the clay (SR) as published in Lee and Batch-
elor (2004). Calculations assume a water filled porosity of 25% and a clay content of 10% of the
total dry weight of sediment.
Clay
Biotite
Biotite
Biotite
Biotite
Vermiculite
Vermiculite
Vermiculite
Vermiculite
Montmorillonite
Montmorillonite
Montmorillonite
Montmorillonite
Chlorinated
hydrocarbon
PCE
TCE
cis-DCE
VC
PCE
TCE
cis-DCE
VC
PCE
TCE
cis-DCE
VC
SR
laMg'1
0.4
0.482
0.612
1.06
0.177
0.188
0.635
1.01
0.271
0.282
0.318
0.353
C°
^ RC
mgl/1
52.7
50.4
47.2
52.7
23.3
19.6
48.9
50.2
35.7
29.5
24.5
17.5
Iron
content
of clay
mg/g
117
117
117
117
56.7
56.7
56.7
56.7
7.5
7.5
7.5
7.5
c°
^ RC
normalized
to content of
total iron
mg(CH)
g(Total Iron)'1
0.57
0.54
0.51
0.57
0.52
0.44
1.09
1.11
5.99
4.94
4.11
2.94
C°
^ RC
normalized
to content of
total iron
M(CH)
M(Total Iron)'1
1.9 x 10-4
2.3 x 10-4
2.9 x 10-4
5.1 x 10-4
1.7 x 10-4
1.9 x 10-4
6.2 x 10-4
9.9 x 10-4
20 x 10-4
21 x 10-4
24 x 10-4
26 x 10-4
-------
6.0
Methods to Characterize Sediments and
Organic Contaminants
This section discusses methods that are
currently available to predict or under-
stand abiotic transformations of organic
compounds in aquifer sediments.
6.1 Isotopic Fractionation to
Characterize Degradation of
Organic Compounds
Degradation of organic compounds either
through abiotic processes or through biological
processes can change the ratio of stable iso-
topes in the material that remains behind and
has not been degraded. This process is called
isotopic fractionation. In many cases it is
possible to recognize degradation of organic
contamination in field scale plumes from a
change in the ratio of stable isotopes of carbon
or hydrogen (Hunkeler et al., 2008). In some
cases, it is possible to infer the extent of deg-
radation from isotopic fractionation, and to
separate the reduction in concentration caused
by degradation from reductions in concentra-
tion caused by dilution and dispersion.
Commercial laboratories can determine the
ratio of stable isotopes of carbon in chlorinated
organic compounds such as PCE, TCE, and cis-
DCE in water samples, even when the concen-
trations are as low as 5 to 10 ug/L. At the time
of this writing, the cost is near $300 per sample.
Reactions in which bond cleavage is the rate-
limiting step can result in the enrichment of the
heavier isotope (13C) in the remaining parent
compound (Bloom et al. 2000; Eisner et al.
2005; Zwank et al. 2005). The magnitude of
isotope fractionation can be described by the
bulk enrichment factor, s, „, derived from the
' bulk'
Rayleigh model (Liang et al. 2007). The more
negative the value of s, „, the greater the change
° bulk' ° O
in ratio of stable isotopes for a given amount of
degradation. For abiotic transformations, the
magnitude depends on the mineral surface pre-
sented for reaction, and on the reaction pathway.
For biological transformations, the magnitude
of isotope fractionation depends on various
factors including temperature, bacterial spe-
cies, structure of the organic compounds, and
availability of nutrients (Bloom et al. 2000).
Readers who are not familiar with the use
of stable isotope analysis to document the
extent of degradation of organic compounds
are directed to Hunkeler et al. (2008) for the
definition of units, a discussion of analyti-
cal techniques, and example calculations.
6.1.1 Isotopic Fractionation of Carbon
Tetrachloride on Various Iron Minerals
Zwank et al. (2005) compared the rate of
degradation and the isotopic fractionation of
carbon tetrachloride on a variety of iron minerals
under comparable conditions. Minerals included
magnetite (Fe3O4), mackinawite (FeS), goethite
(a-FeOOH), lepidocrocite (y-FeOOH), hematite
(Fe2O3), and siderite (FeCO3). The pH was
buffered to between 7.1 and 7.3. The minerals
were suspended in a solution of 1 to 2.3 mM
Fe2+, and the surface area of the minerals in
the suspension was approximately 50 m21/1.
Zwank et al. (2005) found that the abiotic
reductive dehalogenation of carbon tetrachloride
was associated with substantial carbon isotopic
enrichment. They found that the observed
bulk enrichment factors (&, „) correlated
v bulk'
neither with the surface-normalized reaction
rate constants nor with the type of products
formed. However, the bulk enrichment factors
fell into two distinctly different ranges for the
two principal groups of minerals studied. With
-------
iron (hydr)oxide minerals (goethite, hematite,
lepidocrocite, and magnetite) and with siderite,
the s-values for carbon tetrachloride dehalogena-
tion were remarkably similar (the range was
-29 + 3%o). Because this value matched well
with the theoretical estimates for the cleav-
age of an aliphatic C-C1 bond, they suggested
that dissociative electron transfer to carbon
tetrachloride controls the reaction rates for this
group of iron minerals. Conversely, carbon
tetrachloride transformation by different prepara-
tions of mackinawite was accompanied by a
significantly lower carbon isotopic fractionation
(sbulk = -15.9 + 0.3%o), possibly due to the
presence of nonfractionating rate-determining
steps or a significantly different transition state
structure of the reaction (Zwank et al. 2005).
Neumann et al. (2009) compared degradation
of carbon tetrachloride on ferruginous smectite.
The only product detected from the degrada-
tion of carbon tetrachloride was chloroform.
The value of SK ,. was -13.3%o to -10.9%o.
bulk
6.1.2 Isotopic Fractionation ofPCE and TCE
on Iron(II) Monosulfide
Liang et al. (2007) compared the enrichment
of stable carbon isotopes during biological
and nonbiological degradation of PCE and
TCE. Significant carbon isotope fractionation
was observed during FeS-mediated reductive
dechlorination of tetrachloroethylene (PCE)
and trichloroethylene (TCE). Bulk enrichment
factors for PCE were -30.2 + 4.3%o (pH 7),
-29.54 + 0.83%o (pH 8), and -24.6 + l.l%o
(pH 9). For TCE, sbulk values were -33.4 +
1.5%o (pH 8) and -27.9 + 1.3%o (pH 9).
The bulk enrichment factors for abiotic
degradation are substantial. Based on the
uncertainty in the measurement, isotopic ratios
in two different samples of PCE or TCE can
usually be distinguished if the ratios differ
by more than 2%o. If the bulk enrichment
factor is near -30%o, and as little as one-half
the original material is degraded, the isotopic
shift would be 21%o. If degradation of PCE or
TCE occurs through an abiotic reaction with
FeS, then it should be possible to recognize
degradation and predict the extent of degrada-
tion using stable carbon isotope analyses.
Bulk enrichment factors are also available
for microbial reductive dechlorination by two
isolated pure cultures \Desulfuromonas michi-
ganensis strain BB1 (BB1) and Sulfurospirillum
multivorans (Sm)] and a bacterial consortium
[BioDechlor Inoculum (BDI)]. The sbulk values
for biological PCE microbial dechlorination
were -1.39 + 0.21%o (BB1), -1.33 + 0.13%o
(Sm), and -7.12 + 0.72%o (BDI), while those for
TCE were -4.07 + 0.48%o (BB1), -12.8 + 1.6%o
(Sm), and -15.27 + 0.79%o (BDI) (Liang et al.
2007). In each case, the fractionation during
abiotic transformation was substantially greater
than the fractionation during biodegradation.
Because s, „ values for microbial reductive
bulk
dechlorination of PCE and TCE are gener-
ally smaller in magnitude (less negative)
than those for abiotic reductants, Zwank et
al. (2005) suggested that differences in sbulk
values could be used to distinguish abiotic and
biotic reductive dechlorination of PCE, but
not TCE, in model sulfate-reducing systems.
6.1.3 Isotopic Fractionation ofEDB on
Iron(II) Monosulfide
The isotopic enrichment factor for degrada-
tion of EDB on biogenic FeS as determined
by Wilson et al. (2008) was -20.2 + 2.23%o
at 95% confidence. The enrichment of
carbon isotopes during abiotic degradation
of EDB during reaction with FeS (-20.2%o)
is significantly greater than the enrichment
during biodegradation (-5.7%o) (Wilson et al.
2008). This relationship has been observed for
other halogenated organic compounds. The
contrast in behavior of stable carbon isotopes
in EDB during biodegradation and abiotic
transformation parallels the behavior of PCE
and TCE as observed by Liang et al. (2007).
-------
6.1.4 Isotopic Fractionation ofTCE on Green
Rust
At this writing, there is only one value for sbulk
associated with the degradation of a chlorinated
hydrocarbon on green rust. Liang et al. (2009)
reported a value of-23.00 + 1.8%o for degrada-
tion of TCE on chloride green rust. This value
is very similar to the fractionation of TCE on
magnetite (Section 3) and on FeS (Liang et al.
2007; Zwank et al. 2005). Because the TCE
is strongly fractionated, it should be possible
to recognize abiotic transformation of TCE on
sulfate green rust in the field (when it occurs)
by an analysis of stable carbon isotopes.
6.1.5 Isotopic Fractionation of TCE and cis-
DCE on Magnetite
Section 3 and Appendix B describe experiments
conducted by staff at the Kerr Center on abiotic
transformation of TCE and cis-DCE in sedi-
ments from two sites on the Twin Cities Army
Ammunition Plant (TCAAP) north of St Paul,
Minnesota. Sediment containing magnetite was
collected from Site A and the Building 102 site.
As mentioned in Section 3 and Appendix B,
there was no evidence of removal of PCE or
TCE in sediment from Site A at the TCAAP;
however, PCE and TCE were degraded in
material from the site at Building 102. Figure
6.1 depicts the fractionation of stable carbon
isotopes of TCE in pore water in the micro-
cosms that were constructed with sediment
from the site at Building 102. The fractionation
of TCE was strong (s = -39 %o + 12 %o at
95% confidence). This strong fractionation
should make it possible to use stable carbon
isotope analyses of TCE in ground water to
recognize degradation of TCE at field scale.
In contrast to TCE, the fractionation of cis-DCE
on magnetite is much weaker (Figure 6.2).
Panel A of Figure 6.2 presents the data with
an expanded y-axis, to facilitate comparison
of differences. Panel B of Figure 6.2 plots the
data on the same scale as Figure 6.1. The bulk
isotopic enrichment factor (s) extracted from
combined data from both Site A and the site
at Building 102 is only -0.60 %o. The value is
not different from zero at 90% confidence. The
value of & would have to be more negative than
-1.4 %o to be different from zero at 95% con-
fidence. Based on this limited data set, it will
be difficult to use stable isotope fractionation to
recognize degradation of cis-DCE at field scale.
The strong contrast in the isotopic enrichment
factor between TCE and cis-DCE suggests that
there are important differences in the mecha-
nisms or pathways of degradation of TCE and
cis-DCE on the surface of magnetite. This might
explain why PCE and TCE did not degrade in
microcosms constructed from sediment from Site
A at TCAAP, but cis-DCE did degrade in micro-
cosms made from sediment from Site A. The
bulk isotopic enrichment factors for degradation
of PCE and TCE in sediment from the Building
102 site are similar to the values reported by
Liang et al. (2007) for degradation of PCE and
TCE on FeS [-29.54 + 0.83%o ( for PCE at
pH 8), and -33.4 + 1.5%o (for TCE at pH 8)].
LU
O
O
150 -i
125 -
100 -
75-
50-
25 -
0 -
-25
-50
y = -39.359X - 35.261
0 -1 -2 -3 -4
Natural Log of Fraction Remaining (C/Co)
Figure 6.1. Fractionation of stable isotopes of car-
bon during abiotic degradation of TCE
in aquifer sediment containing mag-
netite (site at Building 102 at TCAAP).
The uncertainty in 813C is indicated by
the vertical bars through the symbols.
-------
LU
o
Q
CO
G
O
-20 i
-21 -
-22 -
-23 -
-24 -
-25 -
-26 -
-27 -
-28
.Site 102
o site A
° Container Controls
0 -1 -2 -3 -4
Natural Log of Fraction Remaining (C/Co)
150
LU125
o 100
co 75
50
25
G
o
-25
p
D
Site 102
Site A
Container Controls
.»«*»«.
0 -1 -2 -3 -4
Natural Log of Fraction Remaining (C/Co)
Figure 6.2. Fractionation of stable isotopes of
carbon during abiotic degradation of
cis-DCE in aquifer sediment containing
magnetite.
6.1.6 Isotopic Fractionation of Chlorinated
Alkanes on Phyttosilicate Clay
Neumann et al. (2009) also compared deg-
radation of 1,1,1,2-tetrachloroethane and
1,1,2,2-tetrachloroethane on ferruginous smec-
tite. Transformation of 1,1,1,2-tetrachloroethane
was primarily through reductive |3-elimination
to form 1,1-dichloroethylene (sbulk,-11.4%o +
2.0%o and -14.9%o + 1.7%o), while transforma-
tion of 1,1,2,2-tetrachloroethane was through
dehydrochlorination to produce TCE (sbulk,
-24.0%o + 0.8%o and -20.8%o + 2.2%o). "
6.2 Methods to Estimate the Quantity
of Reactive Minerals
The essential components of a study to char-
acterize reactive iron and sulfur minerals will
include sample collection, sample fractionation
(e.g., size or magnetic fractionation), determina-
tion of bulk elemental composition (e.g., total
Fe and S), determination of chemical speciation
of Fe and S, and determination of mineralogy.
There are two particular challenges related to
the detection and measurement of the reactive
iron-bearing minerals. The first challenge is that
they are often sensitive to changes in the redox
environment. Exposure to molecular oxygen
from the atmosphere quickly destroys green
rusts and iron(II) monosulfides. The second
challenge is that the reactive iron and sulfur
minerals are usually present in soil and aquifer
sediment at low concentrations. Concentrations
of mackinawite or green rust as low as 0.1%
by weight, or concentrations of magnetite as
low as 1% by weight, can transform chlorinated
hydrocarbons at environmentally significant
rates (Section 2 and Section 4 of this Report).
These low, but environmentally relevant, con-
centrations can be difficult to characterize.
Amonette (2002) provides a recent description
of the methods used to identify and quantify
minerals in soils. As will be discussed below,
established methods with adequate sensitivity
and resolution are available to determine the
bulk concentration of mackinawite and other
iron(II) monosulfides, as well as, pyrite and
other iron(II) disulfides. Established methods
are not readily available to directly determine
magnetite or green rusts at natural concentra-
tions in environmental samples. However,
magnetite can be easily concentrated by
magnetic separation, and the bulk concentra-
tion of magnetite can be estimated from the
magnetic susceptibility of the sample.
Established methods are not available to
estimate the specific surface area of miner-
als in soils and aquifer materials. If the
-------
minerals can be concentrated, the particle
size can be estimated from the broadening of
peaks in the X-ray diffraction spectrum, or
by examination by scanning or transmission
electron microscopy, and surface area can be
calculated from the particle size distribution.
6.2.1 Characterization by Chemical Extraction
Methods
Chemical extraction methods are perhaps the
most frequently used techniques for establish-
ing iron and sulfur partitioning in sediments
and soils. Methods designed to differentiate
iron minerals typically involve complexation
and protonation reactions. The amount of
iron extracted by hydrochloric acid, oxalic acid,
ascorbic acid, and buffered sodium dithionite
solutions is often used to distinguish mineral
phases (Heron et al., 1994; Kostka and Luther,
1994). Methods for partitioning total sulfur in
a sample among organic sulfur, acid-volatile
sulfides (e.g., mackinawite), and chromium-
reducible sulfur (e.g., pyrite and sulfur) are
well established and sensitivity to 0.01% by
weight is achievable for most sample types
(e.g., Burton et al. 2008; Tack et al., 1997).
6.2.1.1 Sulfur Minerals in Soil and Sediments
Sulfur can generally be fractioned into sulfate,
elemental sulfur, acid volatile sulfide (AVS),
and chromium reducible sulfur (CrRS). In
sediments, AVS and CrRS are the two most
important fractions in regulating the behavior
of contaminants. AVS is commonly used as
a general estimate for FeS, and CrRS is an
estimate for FeS2 in sediment. Analysis of AVS
and CrRS is commercially available from labs
that support the AMIB A protocol (Aqueous and
Mineral Intrinsic Bioremediation Assessment)
as described in Kennedy et al. (2004). The
labs can be located by using AMIBA or
Aqueous and Mineral Intrinsic Bioremediation
Assessment in a search engine on the internet.
AVS is operationally defined as sulfide that can
be extracted with 1M HC1. The AVS extraction
procedure is intended to digest labile sulfide
compounds, including iron monosulfides such
as mackinawite (FeS). AVS measurements are
generally considered to represent hydrogen
sulfide released during the acid dissolution of
iron monosulfides. AVS can also include sul-
fide associated with other acid-soluble sulfide
minerals (e.g., poorly crystalline ZnS), residual
pore-water sulfide (Morse and Rickard 2004),
sulfide adsorbed to mineral or organic surfaces
released at low pH, or sulfide associated with
metal sulfide nanoparticles or clusters (Rickard
and Morse 2005; Wilkin and Bischoff 2006).
CrRS is selective for inorganic S with an
oxidation state less than +6 (Canfield et al.
1986). CrRS measurements are commonly
used to determine the pyrite-sulfur content of
sediments. CrRS is extracted with 0.5M CrCl2
in 1M HC1 which is usually prepared using
a Jones reductor. CrRS can be analyzed in
sequence after extraction and removal of AVS
and elemental sulfur; or CrRS can be deter-
mined before removal of AVS and elemental
sulfur. When this is done, the measured CrRS
includes both AVS and elemental sulfur.
Qualitatively, the most common methods to
determine the presence of iron sulfide compo-
nents are X-ray diffraction (XRD), X-ray fluores-
cence (XRF), and scanning electron microscopy-
energy dispersive spectroscopy (SEM-EDS).
XRD can be used to detect certain minerals in a
mixture without regard to their abundance; XRF
can be used to determine the total elemental
composition of a sample; and SEM-EDS can be
used to visualize ultrastructures with subsequent
estimation of their elemental composition. None
of these methods however, are appropriate for
quantitative analysis (Popa and Kinkle 2000).
There are several procedures available in the
literature for quantitative determination of AVS
and CrRS in sediment. Generally there are two
ways to extract sulfide from sediment samples:
diffusion (Hsieh et al. 2002; Ulrich et al. 1997)
and purge and trap (Morse and Cornwell 1987).
After the sulfide is extracted, the concentration
-------
of sulfide can be measured several ways.
These include coulometry (Oita 1983; Wilkin
and Bischoff 2006), colorimetry (Allen et al.
1993; Cornwell and Morse 1987), gravimetric
analysis (Di Toro et al. 1990; Leonard et al.
1993), sulfide ion-specific electrodes (Pesch et
al. 1995), or gas chromatography with photo-
ionization detection (Casas and Crecelius 1994;
Lasorsa and Casas 1996; Leonard et al. 1999).
Lasorsa and Casas (1996) provide a detailed
comparison of the advantages and disadvantages
of some common analytical methods for AVS
and CrRS measurements. The colorimetric
method is relatively easy to use and is less
time-consuming than the gravimetric method,
and is cost effective in terms of consumables. It
is also the most forgiving when extremely high
or low levels of AVS are encountered, though
dilution of the trapping solution is required to
analyze high concentrations of AVS. Instrument
failure and maintenance were seldom a problem.
However, this method produced a relatively
large volume of acid waste (approximately
280 mL per sample), and the mixed diamine
reagent used in this method is the most hazard-
ous of all of the reagents used in the available
methods in terms of corrosivity and toxicity.
The gravimetric method was more time-
intensive and complicated, and more expensive
on a per-sample basis than the other methods.
This is due to the expense of the silver nitrate
reagent, as much as $10 per sample. The cost
for equipment is lower than the methods using
photoionization or coulometric detectors. Only
glassware and an accurate balance sensitive
to 0.001 g are required. The method was not
as forgiving when extremely high or low AVS
concentrations were encountered, either due to
uncertainty in weighing extremely small masses
or the possibility of exhausting the silver reagent
at very high concentrations of AVS. The volume
of waste produced is significantly higher than
that of the colorimetric method (approximately
500 mL) because of an extra trap to remove HC1
and the larger volumes of trapping reagent.
The ion selective electrode method was the
simplest to carry out and relatively inexpen-
sive in terms of both consumable items and
instrumentation. This method can be used
over a wide range of concentrations without
dilution of the trapping solution. It produced
a volume of waste similar to that of the colo-
rimetric method (approximately 280 mL).
The photoionization detection method is rela-
tively simple, most cost effective in terms of
consumables, and accurate at low concentra-
tions. Difficulty may be experienced when
extremely high AVS concentrations were
encountered. The method works well when the
NaOH-trapping step is eliminated, but since the
photoionization detection method is linear to
only about 1 pmole of H2S, the NaOH-trapping
step must be used with samples that have
sulfide concentrations greater than 10 pmole/g.
A coulometric method was introduced by Wilkin
and Bischoff (2006) to determine sulfur frac-
tions in environmental samples. This method
is simple, fast and sensitive, with a detection
limit of 5 |j,g S g"1 and average precision of
91% (Wilkin and Bischoff 2006). Only a small
sample is required. The method can be applied
to samples of sediments, soil, aquifer solids,
and waste products. Another major advantage
is the fact that the evolution of hydrogen sulfide
gas is detected in-line. Chemical extraction
end points can be precisely determined without
need to set arbitrary time limits as in other
methods. This method generates the smallest
volume of waste (10 mL), but the coulometer
is expensive. The anode and cathode solutions
used in the measurement are highly toxic.
6.2.1.2 Interferences with Chemical Extraction
Methods for Sulfur Minerals
Occasionally, when sediment samples are
analyzed for sulfides by different methods, there
are discrepancies between the concentrations
that are reported. Some of the discrepancies
are probably caused by inadequate treatment
or preservation of the samples. Problems can
-------
occur during collection and processing under
N2 or with precipitation of sulfides with zinc
acetate, or oxidation after collection during
shipment, or during steps involved in process-
ing the samples such as washing or distillation.
Alternatively, variations in the composition of
the sample may be important (e.g., ferric iron
interference, crystal size, and surface properties).
In the presence of acid-soluble ferric miner-
als, AVS can be underestimated using the hot
distillation procedure (Pruden and Bloomfield,
1968; Berner, 1974; Cornwell and Morse, 1987).
This underestimation is caused by the dissolution
of ferric minerals in strong acid, presumably
allowing soluble Fe(III) to oxidize sulfide (Hsieh
et al. 2002). To minimize oxidation of AVS
by Fe(III) during the hot distillation procedure,
SnCl2 has been added to the sample (Cornwell
and Morse, 1987). However, Hsieh et al., (2002)
found that the addition of 1 g SnCl2 powder to
their samples did not prevent the ferric interfer-
ence on AVS. They also found that addition
of 2 mL 1 M ascorbic acid to the sample
effectively protects AVS from oxidation in the
presence of 0.39 mmol acid-soluble Fe(III).
Another major interference is elemental sulfur
which could interfere during the chromium
reduction step in the determination of CrRS.
This would result in an overestimation of FeS2
(Popa and Kinkle 2000). Elemental sulfur (S°g)
can be extracted by shaking the sediment with
20 mL of acetone for 24 h (Wieder et al., 1985),
followed by a further 10 mL acetone rinse. The
S (S°8) content of the acetone phase can then
be determined by the Cr-reduction method.
6.2.2 Characterization of Chemical Identity by
Spectroscopy or Diffraction
6.2.2.1 Characterization by X-ray Diffraction
(XRD)
X-ray diffraction is generally the preferred
method for mineral identification. X-ray diffrac-
tion patterns are diagnostic features that can be
searched and matched with available software
to confidently identify constituent minerals
contained in natural samples. One drawback
with the method relates to the practical detection
limit for particular minerals. In most situations,
the mineral phases responsible for driving abiotic
dechlorination reactions are typically expected
to be less than one percent by weight of the total
sediment. Detection limits using conventional
powder X-ray diffraction techniques will vary
depending on the sample matrix and on the
degree of crystallinity of the component miner-
als, but quantitation limits of about 1 to 5 weight
percent may be achieved for a given phase with-
out pre-concentration. In natural samples, quartz
usually overwhelms other minerals because it is
abundant and because it diffracts more strongly
than other minerals. For equivalent amounts of
quartz and a mixture of other minerals, the stron-
gest peaks will often be from quartz. Detection
limit problems are consequently confounded
by the fact that quartz is usually the major
mineral component in ground-water aquifers.
As an example, mechanical mixtures of magne-
tite and quartz were prepared at three different
dilutions ranging from 2% to 10% magnetite
by weight. The mixed components were of
comparable grain size (-325 mesh). For each
sample, X-ray diffraction scans were collected
in overnight runs to maximize sensitivity and
increase the signal to noise ratio. The raw
diffraction scans are shown on Figure 6.3 along
with Powder Diffraction Files for magnetite
(PDF 19-0629) and quartz (PDF 46-1045).
-------
o
o
ex
-------
from the time the water was taken from the flow-
ing stream, the finished sample had been stored
under a nitrogen atmosphere. The samples
were transported to the laboratory in a nitrogen
atmosphere and stored in an anaerobic chamber
(4% H2 in N2). They imaged the samples with
AFM in the glovebox. Reproducibility tests
made with synthetic green rust proved that if it
is dry, green rust fixed on Si-bearing substrates
is stable in air for more than 24 hours. In this
case, the powder XRD data were collected in air.
Green rusts form in soils and sediments when
the level of oxygen is very low; in air they will
transform to normal brown rust within a few
minutes. This has prevented researchers from
finding out exactly how and under what condi-
tions they form. To overcome these problems,
Christiansen and Stipp (2003) recreated the
conditions conducive to green rust formation
within the beamline of an X-ray machine. Using
a complex chemical reaction cell, and with the
unique power and intensity of the X-ray beam,
they were able to gain a unique insight into
the atomic scale formation and crystallization
of green rust under conditions close to which
it forms in contaminated land environments.
Specifically, they used an AFM mounted inside a
glove box to observe the formation and trans-
formation of sulfate green rust and developed a
method for obtaining XRD patterns from sam-
ples in an inert atmosphere. This avoids the use
of glycerol and other treatments that may alter
the crystal structure, making it possible to collect
clearer patterns while avoiding any possibility
of alteration. Some previously undefined XRD
peaks from untreated sulfate green rust suggest
internal ordering of sulfate and water. AFM
evidence indicates that sulfate green rust can
form from topotactic transformation of Fe(OH)2
as the starting phase (Bernal et al., 1959).
Carbonate green rust was identified as a major
Fe(0) corrosion product by XRD, even in simu-
lated groundwater containing added sulfate and
no added carbonate. Carbon dioxide diffusion
from air provided carbonate for the formation of
carbonate green rust (Su and Puls, 2004). This
suggests that carbonate green rust is either kinet-
ically favored or thermodynamically more stable
than sulfate green rust in the studied system.
Bearcock et al. (2006) used XRD to character-
ize green rust present in a 45-60 mm thick
band which lay just below the surface (& 4
mm) of an ochreous deposit at an abandoned
coal mine site. SEM imaging of the sample
(presence of |j,m-sized hexagonal crystals)
and chemical analyses (Fe(II)/Fe(III) = 2:1)
support the XRD identification of green rust
either as pyroaurite (Powder Data File 01-070-
2150) or GR2 (00-052-0163) (sulfate green
rust). The sample contained 14.1 wt.% FeO.
In a recent study, Sumoondur et al. (2008)
obtained direct evidence for the formation of
magnetite via a green rust intermediate. The
Fe(II) induced transformation of ferrihydrite,
was quantified in situ and under O2-free condi-
tions using synchrotron-based time-resolved
energy dispersive X-ray diffraction. At pH
9 and Fe(II)/Fe(III) ratios of 0.5 and 1, rapid
growth (6 min) of sulfate green rust and its
subsequent transformation to magnetite was
observed. Electron microscopy confirmed these
results, showing the initial rapid formation of
hexagonal sulfate green rust particles, followed
by the corrosion of the green rust as magnetite
growth occurred, indicating that the reaction
proceeds via a dissolution-reprecipitation
mechanism. At pH 7 and a Fe(II)/Fe(III)
ratio of 0.5, sulfate green rust was the stable
phase, with no transformation to magnetite.
Information is lacking about the detection limit
of green rusts using XRD for natural soil and
sediment samples. Successful identification
of green rusts with XRD is usually associated
with a dark bluish green color in the sample.
XRD analysis is commercially avail-
able and the cost is reasonable.
-------
6.2.2.3 Characterization of Green Rusts by
Infrared Spectroscopy (IR)
IR can be used to determine the presence of
anions in the green rust interlayers. For example,
Ona-Nguema et al. (2004) used diffuse reflec-
tance infrared Fourier transform spectroscopy to
confirm the presence of intercalated carbonate
and sulfate in the structure of biogenic green
rust, in combination with XRD, TEM, and SEM.
Su and Puls (2004) used FTIR-photoacoustic
spectroscopy (interlayer carbonate stretching
mode at 1352 - 1365 cm"1) to identify carbonate
green rust along with XRD and SEM as a major
iron corrosion product in zerovalent iron-packed
columns constructed for arsenic remedia-
tion. Commercial IR services are available
and most research labs are equipped with IR.
6.2.3 Estimating Magnetite in Sediment from
Magnetic Susceptibility
There is no direct chemical "test" that is specific
for magnetite at its natural abundance in sedi-
ment. However, the magnetic susceptibility
of magnetite can provide a useful estimate of
the quantity of magnetite in aquifer sediment.
Magnetite is the most abundant mineral in
natural sediments that exhibits magnetic behav-
ior. Magnetism is an expression of the magnetic
moments associated with the spin of unpaired
electrons that are involved in chemical bonds.
In magnetite the moments are highly aligned,
but exist in two sets of opposing but unequal
forces constrained by the crystal lattice of the
mineral. As a result, when an external magnetic
field is imposed on a sample of magnetite, the
magnetic moments will align with the imposed
field and add to the total magnetic field. The
extent to which an imposed magnetic field is
strengthened by the sample is termed the mag-
netization of the sample. The volume magnetic
susceptibility of the sample is the ratio of the
magnetization of the sample (the magnetic
dipole moment per unit volume) measured in
amperes per meter divided by the magnetic
field strength, measured in amperes per meter.
As defined above, volume magnetic susceptibil-
ity is dimensionless. However, the actual value
for volume magnetic susceptibility depends
on how the magnetic permeability of free
space is defined in the system of measurement.
Unfortunately, two separate systems are in
common use, the COS system (centimeter gram
second) with values reported as emu or e.m.u.
for electromagnetic unit, and the International
System of Units (abbreviated SI from the French
Le Systeme International d'Unites) based on
the meter, kilogram and second. This Report
will use the SI system of measurement. A value
for volume magnetic susceptibility in the COS
system can be converted to the SI system by
multiplying by 4n. We will express measure-
ments of magnetic susceptibility of aquifer
sediment as the mass molecular susceptibility.
The mass magnetic susceptibility is the volume
magnetic susceptibility divided by the density of
the sample in kg nr3, and the unit for mass mag-
netic susceptibility in the SI system is m3 kg"1.
Canfield and Berner (1987) studied the dissolu-
tion and pyritization of magnetite in marine
sediments. Horneman et al. (2004) studied
dissolution of arsenic from iron(II) minerals
in anoxic aquifers in Bangladesh. As part of
their studies they compared the mass magnetic
susceptibility of their sediments to the mass
of materials that could be separated from the
sediment with a magnet. Their results are
presented in Figure 6.5. The correlation of mass
magnetic susceptibility with weight content of
geological materials was roughly linear from
less than 100 mg/kg to more than 10,000 mg/
kg. However, there was significant scatter in
the data. Individual measurements could vary
by a factor of three from the regression line.
Figure 6.5 also presents the relationship between
mass and mass magnetic susceptibility for four
known materials. Magnetite Sample 1 was pro-
vided by Robert C. Thomas of Athens Georgia;
its specific surface area is 18m2 g"1. Magnetite
Sample 2 was purchased from Cerac Specialty
Inorganics (particle size 5 um or less, 99% pure).
Magnetite Sample 3 was ground from magnetite
-------
crystals. Finally, the Magnetite Calibration
Standard in Figure 6.5 was the standard provided
with a MS2 Magnetic Susceptibility System
(Bartington Instruments Ltd., Oxford, England).
The standard contained 1% by weight magne-
tite in silica. It was contained in a cylindrical
sample pot with a capacity of 10 mL. As was
the case with the natural sediment samples, the
magnetic susceptibility of the magnetite stan-
dards varied from each other by a factor of three.
100000 -,
=| 10000-
-§
| 1000 -
"CD
5 100 -
o
1
§5 -m -
m.
«•
x Horneman et al.
o Canfield and Berner
• Magnetite Sample 1
A Magnetite Sample 2
» Magnetite Sample 3
• Calibration Standard
0.01 0.1 1 10
Magnetic Susceptibility (10~6 m3 kg"1
Figure 6.5. fte/af/onsft/p between the mass mag-
netic susceptibility of a sediment
sample and the content of magnetic
materials.
All the data on magnetic susceptibility in
Section 3 and Appendix B were acquired with
a MS2 Magnetic Susceptibility System using
the MS2B detector (Bartington Instruments
Ltd., Oxford, England). This unit is widely
used by geologists and geochemists to char-
acterize samples of rock and sediment. The
Bartington Magnetic Susceptibility System
claims good reproducibility down to 0.001 x
10"6 m3 kg"1. If the relationship in Figure 6.5
is extrapolated to this value, the content of
magnetic materials would be near 1 mg/kg.
This concentration is negligible. The estima-
tion of magnetite in sediment is more likely to
be limited by a spurious signal from another
mineral, than by the sensitivity of the instrument.
The combined data of Horneman et al. (2004)
and Canfield and Berner (1987) were used
to extract a relationship between the mass
magnetic susceptibility of a sediment sample
and the expected concentration of magnetic
materials. A linear regression of the logarithm
of the concentration of magnetic materials
on the logarithm of the magnetic susceptibil-
ity of the sample produces equation 6.1.
and equation 6.2,
Equation 6.1
Equation 6.2
where MM is the concentration of magnetically
separable materials (mg/kg) and MS is the mass
magnetic susceptibility of the sample (m3 kg"1).
Equation 6.2 was used to calculate the mass
of magnetic material (assumed to be magne-
tite) exposed to pore water in the sediment
samples in Table 3.1 and 3.2 in Section 3.
The reproducibility and linearity of magnetic
susceptibility instruments such as the Bartington
MS2 Magnetic Susceptibility System is good,
generally much less than ±10% of the expected
value. Error in predicting the content of mag-
netic materials from mass magnetic suscepti-
bility will come from the natural variation in
mass magnetic susceptibility of magnetite from
one sample to another, and from the magnetic
susceptibility of other minerals that might be in
the sample. Hematite (a-Fe2O3) and goethite
(a-FeOOH) are the only common iron minerals
that have large enough mass magnetic suscep-
tibilities to be confused with magnetite. The
mass magnetic susceptibility of finely divided
magnetite (particle diameter 12 nm to 69 nm)
ranges from 500 to 1116 x 1Q-6 m3 kg"1, while
the mass magnetic susceptibility of hematite and
goethite range from 0.27 to 1.69 x 1Q-6 m3 kg"1
(Table 2.2, Bearing, 1999). The mass magnetic
susceptibility of magnetite is one thousand-
fold greater. If the iron content represented
by the predicted concentration of magnetite
is at least 1% of the total extractable iron, it
is unlikely that the magnetic susceptibility of
other iron minerals will cause significant error.
-------
If it is necessary to do so, the chance of con-
fusing another mineral with magnetite can be
evaluated by comparing the total extractable iron
concentration in a sample to the concentration of
iron in magnetite that would be predicted from
the mass magnetic susceptibility of the sample.
For an example of this evaluation, compare data
in Table 3.4 in Section 3. Equation 6.2 was
used to calculate the concentration of magnetite
that would be necessary to produce the mass
magnetic susceptibility of the sediment. On a
mass basis, magnetite is 72% iron by weight.
To estimate the iron associated with magnetite,
the concentration of magnetite in Table 3.4 was
multiplied by 0.72. In sediments with detectable
rates of abiotic removal of chlorinated organics,
the amounts of iron that would be in the amounts
of magnetite that were predicted from the mass
magnetic susceptibility of the sediment varied
from 16% to 83% of the total extractable iron.
6.3 Methods to Determine Oxidation
Status
6.3.1 Characterization of Green Rusts by
Mossbauer Spectroscopy
Mossbauer spectroscopy (MS) has often been
used to characterize green rusts and to follow
their Fe(II)/Fe(III) ratio during the process
of oxidation of Fe(OH)2 in the presence of
aggressive anions such as Cl", SO42", CO32.
For example, Genin et al. (2002) studied
chloride pitting of concrete reinforcing bars
and the bacterial corrosion of cast iron in
water pipes or steel sheet piles in harbors.
Refait et al. (2001) studied a green rust mineral
extracted from hydromorphic soils in Fougeres
(France) by X-ray absorption spectroscopy
(XAS) and transmission Mossbauer spectros-
copy (TMS). The Mossbauer spectrum of the
mineral, measured at 77 K, was composed
of four quadrupole doublets: D\ and D2 due
to Fe2+ [5 = 1.26 mm/s and A£Q = 2.5 and
2.9 mm/s, respectively] and/)3 and/)4 due
to Fe3+ [5 = 0.46 mm/s and A£Q = 0.5 and
1.0 mm/s, respectively]. Finally, synthetic
Mg2+-Fe2+-Fe3+ hydroxycarbonates could be pre-
pared by coprecipitation from Mg and Fe salts
and lead to Mossbauer spectra similar to that of
the mineral. In particular, the partial substitu-
tion of Fe2+ by Mg2+ proved to be consistent
with the existence of the unusual doublet D4.
Ona-Nguema et al. (2002b) obtained Mossbauer
spectra measured at 77 K of precipitates that
were sampled after 1 and 6 days of bioreduc-
tion. The spectra exhibited four quadrupole
doublets (Figure 6.6, Panel a and b), the dif-
ference concerning only abundance areas.
Doublets Dv D2, and D3 are typical of a green
rust spectrum at 77 K. D3 is due to high-spin
Fe3+ in octahedral sites with small values
of isomer shift 5 and quadrupole splitting
A, whereas Dl and D2 with larger 5 and A
values are due to high-spin Fe2+ in octahedral
sites. Doublet/) corresponds to Fe(III) with
small 5 and A values, as small as those found
for ferric oxyhydroxide paramagnetic at 77
K, e.g., y-FeOOH with A = 0.57 mm s'1.
Phases present after 6 days of bioreduc-
tion were also analyzed by TMS at 12 K
(Figure 6.6, Panel c). The spectrum consisted
of magnetically split components for y-FeOOH
and paramagnetic doublets for green rusts.
Asymmetrically broadened absorption lines were
fitted with three sextets; the main outer sextet S
had a field of 449 kOe. Doublet £>y (Figure 6.6,Y
Panel b) and sextet S (Figure 6.6, Panel c) were
both due to unreduced lepidocrocite. D dis-
played a larger abundance after 1 day than after
6 days (Figure 6.6, Panel a and b) confirming the
decrease of y-FeOOH by reduction. Information
is provided from spectra of Figure 6.6, Panel a
and b concerning the Fe(II)/Fe(III) ratio
inside the green rust. Assuming equal Lamb-
Mossbauer factors/for all sites, the abundance
of each iron site was proportional to the area
under the peaks, and the Fe(II)/Fe(III) ratio
was equal to ratio {Dl + D2}/D3. The ratio was
32/33 and 44/41 for the samples obtained after
1 day and 6 days of bioreduction, respectively.
-------
(c)
12K
-202
Velocity (mm s"1)
100
-------
that oxidize relatively easily. The presence of
such components makes identification of green
rust very difficult. However, in contrast to
the ferric oxides, which are the final oxidation
products of green rusts, most ferric components
in layer-silicates may be expected to remain
paramagnetic at temperatures above 12 K.
Another important factor facilitating detection
of green rust in the ochre sludge aggregate
was its relatively high concentration in the
sample (Koch and M0rup, 1991). From the 80
K spectrum of the green core they estimated
that Fe2+ amounts to ~6% of the total iron, and
assuming a Fe2+ to Fe3+ ratio of 1.5 in the green
rust, it was estimated that ~10% of the Fe in
the sample is present in the green rust. The
morphology of the green core, the sharp color
transition, and the absence of Fe2+ in the freshly
precipitated ochre indicate that the anaero-
bic transformation of organic matter plays a
major role in the reduction of Fe3+ compounds
necessary for the formation of green rust.
Feder et al. (2005) used a miniaturized
Mossbauer spectrometer, adapted to the Earth's
conditions from the instrument developed for
Mars space missions, to study in situ varia-
tions of iron minerals in a gleysol. Mossbauer
spectra were obtained over a depth interval from
15 to 106 cm. Measurements were repeated
at the same depth at different times to follow
mineralogical transformations over time. X-ray
diffraction (XRD) and selective extraction
techniques were performed on soil samples
from the gleysol. The level of the water table
was measured and the composition of the soil
solution was monitored continuously in situ
with an automatic multi-parameter probe.
All the Mossbauer spectra obtained are charac-
teristic of the Fe(II)-Fe(III) green rust-fougerite,
a natural mineral of the meixnerite group. The
structural formula of this group is: [Fel _xn
Fe/1 (OH)2+2yp*[xA • mH2O;r, where x is the
ratio Fe3+/Fe t and A is the intercalated anion.
tot
No other iron phases were detected from the
Mossbauer spectra or by XRD. About 90% of
total iron was extractible by dithionite-citrate-
bicarbonate, and 60% by citrate bicarbonate. In
the horizons showing oximorphic properties that
were in the upper part of the gleysol, the x ratio
in fougerite, as deduced from Mossbauer spectra,
was approximately 2/3. In the deepest horizons
that show reductomorphic properties, the x ratio
was only 1/3. Rapid mineralogical transforma-
tions were observed at well-defined points in
the soil profile, as evidenced by variations in
the x observed when Mossbauer spectra were
acquired at different times at the same depth.
Mossbauer spectroscopy can provide informa-
tion on the Fe bonding environment and the
Fe(II)/Fe(III) ratio; however, the spectra are not
unique for green rust. Common soil minerals
such as chlorite and several Fe-bearing clays
generate Mossbauer parameters at tempera-
tures around 15 °C (Ballet et al., 1985; Feder
et al., 2005) that are very similar to those of
green rust (Gancedo et al., 1976; Cuttler et
al., 1990). Definitive identification cannot
be made by Mossbauer spectroscopy alone.
Analysis of samples by Mossbauer spectros-
copy is commercially available from university
laboratories. Billing is often on a per hour
basis instead of a per sample basis. Fees in the
range of $500 to $1000 per sample are typical.
6.3.2 Characterization of Green Rust by X-ray
Adsorption Near Edge Spectroscopy
(XANES)
X-ray adsorption near edge spectroscopy
(XANES) is well-suited to identifying different
oxidation states of an element. Because of the
rich fine structure in the L-edge transition metal
spectrum originating in crystal-field splitting
multiple effects, the distribution of oxidation
states in an inhomogeneous sample can be
distinguished in a single spectrum. Two dif-
ferent methods are available for measuring the
XANES; a total-electron yield method which
detects the composition of the "near-surface"
region, and a transmission method which
-------
averages over the composition of the bulk of
a thin section of material. Both methods lend
themselves to spectroscopic imaging, using
X-ray photoelectron emission microscopy
(XPEEM) for near-surface spectro-microscopy,
and scanning transmission x-ray microscopy
(STXM) for the bulk measurements.
Kneedler et al. (1997) used a BL-7 Spectro-
Microscopy Facility STXM and photoemission
apparatus in their work. To produce green
rust, a reaction was initiated between goethite
(oc-FeOOH) and a solution of FeCl2. By halting
the reaction prematurely, precipitates of green
rust were obtained. After drying, the precipitates
were mounted on a sample puck. Powders of
two reference compounds were also mounted
on the puck, FeCl2 (Fe(II)) and goethite (Fe(III))
before insertion into an ultra-high vacuum
chamber on the BL-7 instrument. To prevent
contamination of the green rust sample, all
stages of preparation and introduction into the
analysis equipment were performed anaerobi-
cally, using an oxygen-free glove box and a
sealed transport unit. To obtain XANES spectra
for each specimen, the sample current to ground
was measured as a function of incident photon
energy through the Fe LII, III adsorption edge.
Figure 6.7 shows the XANES spectra for green
rusts and two reference compounds. The
distinction between Fe(II) and Fe(III) oxidation
states is obvious from a comparison of the two
reference compounds FeCl2 and goethite (top and
bottom, respectively). The green rust sample
(second from top) has a distinct signature, which
resembles most closely the Fe(II) spectrum,
but contains a Fe(III) component; the charac-
teristic peak of the Fe(III) signature is seen as
a shoulder to the right of the main peak. Thus,
the green rust can be regarded to first approxi-
mation as a superposition of Fe(II) and Fe(III)
character. To establish the effect of oxygen
contamination, the green rust was later exposed
to air for 50 minutes, and reintroduced for a
final XANES spectrum (third from top). While
there is still a mixed Fe(II)/Fe(III) character
evident in the spectrum, the Fe(III) component
now dominates, indicating a conversion of
iron at the Fe(II) oxidation state to Fe(III).
I
3
32
0)
c
o
fc
JU
LU
"ro
•s
Fe(IK) ref.
Green rust
Green rust,
exp. to air
Fe(lll+) ref.
700.0
710.0 720.0
Photon Energy (eV)
730.0
Figure 6.7. Total electron yield XANES measure-
ments of the Fe-LII, III absorption
edges for green rust and reference
samples FeCI2 (11+) and goethite (III+).
Spectra are normalized to constant
background increase through the edges
(From Kneedler et al., 1997).
Refait et al. (2001) extracted a green rust mineral
from hydromorphic soils in Fougeres (France)
and characterized it by X-ray absorption spec-
troscopy (XAS) and transmission Mossbauer
spectroscopy (TMS). The XAS spectrum at the
Fe K absorption edge of this mineral proved
to be very similar to that of synthetic green
rusts. However, the radial distribution func-
tion obtained for the green rust mineral proved
to be intermediate between those of carbonate
green rust and pyroaurite, that is between the
Fe2+-Fe3+ and Mg2+-Fe3+ hydroxycarbonates.
Consequently, a partial substitution of Fe2+by
Mg2+ occurs, leading to the general formula of
[Fe^+Mg^+Fe^OH^^F [x/n A-. m H2O]-
where An~ is the interlayer anion. The Mossbauer
spectrum of the green rust mineral, measured at
77 K, was composed of four quadrupole dou-
blets: D and D due to Fe2+ [5 « 1.26 mm/s and
-------
AEQ a; 2.5 and 2.9 mm/s, respectively] and D3
and D4 due to Fe3+ [5 « 0.46 mm/s and AEQ ^0.5
and 1.0 mm/s, respectively]. Finally, synthetic
Mg2+-Fe2+-Fe3+ hydroxycarbonates could be
prepared by coprecipitation from Mg and Fe
salts which produced Mossbauer spectra similar
to that of the mineral. In particular, the partial
substitution of Fe2+ by Mg2+ proved to be consis-
tent with the existence of the unusual doublet D4.
A variety of analytical instruments is usually
required to study and identify green rusts. Suzuki
et al. (2008) used XANES, the extended X-ray
absorption fine structure (EXAFS) and XRD
measurements to characterize the effect of the
addition of copper sulfate ions on the chemical
state and local structure of sulfate green rust.
The Fe K edge XANES spectra showed that
Fe(II) in green rust was partially oxidized by
the addition of the copper sulfate solution. The
Cu K edge XANES spectra showed that the
copper sulfate ions in the green rust suspen-
sion were reduced to zerovalent copper. Radial
structural functions indicated that the green rust
was composed of edge sharing FeO6 octahedral
units, and that the structure was changed by the
oxidation of Fe(II). In addition, it was found
that the green rust was partially oxidized to
a-FeOOH by the addition of copper ions.
A limitation of the XANES technique is that the
XAS spectra of various green rusts proved to
be independent of the interlayer anion, and the
nature of the anions present in the mineral green
rust could not be determined by the XANES
method. XAS analyses are conducted in national
labs and are not available in commercial labs.
6.4 Methods to Estimate the Specific
Surface Area of Minerals
Abiotic degradation of organic compounds on
reactive minerals is a heterologous reaction. As
a consequence, the rate of reaction is strongly
influenced by the surface area of the mineral pre-
sented to the organic compound. Finely divided
minerals with high specific surface area will be
more reactive. As a practical matter, information
on the specific surface area is rarely available for
specific minerals in aquifer sediments. When
this information is available, it was obtained at
considerable cost using specialized equipment.
6.4.1 Surface Area from Peak Broadening
during X-ray Diffraction
Laboratory studies of chemically synthesized
reactive minerals usually report the specific
surface area of the mineral preparation as
determined by Brunauer-Emmett Teller (BET)
analysis. This approach determines a sorption
isotherm for nitrogen gas, and thus measures
the entire surface in a sample, regardless of
its mineralogical composition. This approach
is not applicable for reactive minerals that are
present as a minor component of sediments.
If magnetite is separated and enriched from
sediment by magnetic separation, it may be
possible to estimate magnetite particle size
from the broadening of characteristic peaks
during XRD analysis, as was done by Vikesland
et al. (2007) and Ferrey et al. (2004).
6.4.2 Surface Area from Electron Microscopy
If the magnetite is separated and enriched, it
may also be possible to estimate particle size
distribution using electron microscopy, as was
done by Vikesland et al. (2007) and Ferrey
et al. (2004). He et al. (2008) used Electron
Microscopy to estimate the particle size dis-
tribution of pyrite and FeS in complex natural
materials. The dimensions and geometry of the
particles as revealed from electron microscopy
are used to calculate the ratio of surface area
to volume (m2/m3) for typical particles of the
reactive mineral. Then the concentration of
the mineral in the sample (g/g) is determined
by XRD or chemical extraction methods, or
from magnetic susceptibility. Then the specific
surface area for the mineral in the bulk sedi-
ment (m2/g) is calculated by multiplying the
concentration of the mineral in the sediment by
the ratio of surface area to volume, then divid-
ing by the bulk density of the mineral (g/m3).
-------
Jeong et al. (2008) used electron microscopy to
characterize the specific surface area of synthetic
mackinawite, and compared the specific surface
area as estimated from crystal size and shape
to surface area as estimated by peak broaden-
ing during XRD analysis, and a method based
on the weight associated with a mono-layer of
ethylene glycol monoethyl ether. The method
based on electron microscopy estimated lower
values for specific surface, probably because
of aggregation of the crystals of mackinawite.
6.4.3 Characterization of Green Rusts by
Scanning/Transmission Electron
Microscopy
Both scanning electron microscopy (SEM) and
transmission electron microscopy (TEM) have
been used to help identify green rusts. The
SEM image of a naturally occurring green rust
(fougerite) sample showed hexagonal crystals
about 0.5 |j,m size (Trolard et al., 2007). The
SEM image of synthetic carbonate green rust
showed a particle size range from 0.1 to 0.3
|j,m (Su and Wilkin, 2005). Ona-Nguema et al.
(2002b) showed that after incubation of lepi-
docrocite in an iron reducing bacterial culture
for 6 days, the TEM image of the particles
(Figure 6.8) showed large hexagonal crystals of
carbonate green rust, measuring about 10 |j,m.
The electron diffraction pattern of [001] zone
(caption to Figure 6.8), indexed in the hexagonal
representation ofR3m space group, yielded the
same parameter a as that obtained by XRD.
F i g u re 6.8. TEM image and corresponding electron
diffraction pattern of carbonate green
rust hexagonal crystals mixed with a
minor fine-grained phase (j-FeOOH)
obtained after 6 days of bacterial reduc-
tion of lepidocrocite (initially 80 mM
j-FeOOH, 75 mM formate, and WO ^.M
AQDS). Electron diffraction pattern
along [001] zone axis (From Ona-
Nguema et al., 2002b).
The limitation of microscopy in detecting green
rusts is that the hexagonal morphology is not
unique to green rusts; many clay-sized minerals
show similar hexagonal features. Thus, cau-
tion is needed in image interpretation. SEM
and TEM analyses can be performed via com-
mercial labs; however, the analysts should be
well familiar with the nature of the samples
because the morphology of crystals can change
when the samples are prepared for examina-
tion. Imaging of samples by SEM and TEM
is commercially available from university
laboratories. Billing is often on a per hour
basis instead of a per sample basis. Fees in the
range of $500 to $1000 per sample are typical.
-------
6.5 Collection and Handling of
Samples
Because many iron and sulfur minerals are
sensitive to oxidation, extreme caution should be
taken in handling these samples to ensure that
they reflect the actual conditions in contami-
nated subsurface environments. Upon exposure
to air, iron(II) monosulfides, as measured by
acid volatile sulfide (AVS), are destroyed by
chemical oxidation. When this happens, the
concentration of AVS that is determined in
the sample can grossly underestimate the true
concentration of AVS in the material in the
environment. The process of oxidation begins
when the sediments are brought up from depth
during sampling. Because the process continues
during handling, storage, and analysis, it is
crucial for the accurate measurement of AVS
in anoxic sediment to use techniques for sam-
pling, handling and analysis that minimize the
effects of oxidation (Lasorsa and Casas 1996).
Duan et al. (1997) found that AVS in their
samples was largely oxidized to the level
of elemental sulfur unless they protected
their samples by in situ treatment with
zinc acetate. However, Lasorsa and Casas
(1996) used zinc acetate for sample pres-
ervation, and found it to be ineffective.
Several studies (Di Toro et al. 1990; Lasorsa
and Casas 1996; Ulrich et al. 1997) show that
sulfide levels are best maintained when samples
are handled under a nitrogen atmosphere, stored
at 4 °C or frozen at -20 °C and analyzed within
2 weeks of collection. When sampling condi-
tions make it impossible to handle the samples
under an inert atmosphere, it is recommended
that the samples be handled as quickly as
possible and then frozen until analysis within
2 weeks of collection. Wilkin and Bischoff
(2006) found that freezing sediment samples in
the field immediately after collection, shipping
the samples frozen on dry ice, and thawing
and processing the samples in an anaerobic
glove box protected reduced iron and arsenic
species in the sample from oxidation. The
process of collecting, storing, and handling
materials prior to characterization is critical.
Minerals containing ferrous iron and sulfide are
typically sensitive to air exposure. Freezing of
materials and storage under oxygen-free condi-
tions will generally minimize the oxidation of
redox-sensitive minerals (Wilkin, 2006). Core
samples should be frozen with liquid nitrogen
or dry ice in the field immediately after they are
acquired. In some cases, it is possible to freeze
the sample in-situ with liquid nitrogen, and then
retrieve the sample in the frozen state (He et al.
2008). Samples should be kept frozen during
shipment and storage, and should be thawed
for analysis under an anaerobic atmosphere.
Size separation and magnetic separation
can be particularly useful to isolate reac-
tive minerals in aquifer materials; however,
it must be emphasized that fractionation
procedures should be carried out under an
anaerobic atmosphere to prevent oxidation.
-------
7.0
Recommendations for Future Research
We offer a few specific suggestions and
recommendations for future research.
7.1 Further Explore the Role of
Phyllosilicate Clays
Despite promising laboratory experi-
ments, we are not aware of any field work
that has attempted to associate removal of
PCE, TCE, cis-DCE, or vinyl chloride in
ground water to reaction of these contami-
nants with phyllosilicate clays. A good place
to start would be contamination in aqui-
fers containing glauconitic clays (green-
sand) as described in Brown et al. (2007).
7.2 Reexamine the Role of Pyrite
Pyrite as it develops in anaerobic ground water
aquifers may be more reactive than is currently
acknowledged in conventional wisdom. The
available laboratory work on reactions with
pyrite was done with specimens of hydrothermal
pyrite that was ground to an appropriate particle
size, instead of using biogenic pyrite. Biogenic
pyrite that accumulates in contaminated aquifers
has a complex crystal shape (framboids, see He
et al., 2008) that may have a greater specific
surface area, or present a more reactive surface,
than the pyrite used in laboratory experiments.
7.3 Determine if Geochemical Models
Can Be a Useful Surrogate to
Predict the Rates of Abiotic
Degradation at Field Scale
The acquisition of aquifer material and char-
acterization of the material for reactive miner-
als can be expensive and time consuming. It
may be possible to empirically associate rate
constants for abiotic degradation in field scale
plumes to hydrologic parameters of the aqui-
fer and geochemical parameters of the ground
water. This could be done by characterizing
field scale plumes to extract a rate constant
for abiotic degradation, then using statistical
methods to attempt to correlate the field scale
rate to pertinent geochemical parameters.
7.4 Use Geochemical Models to
Improve Estimates of Degradation
on Mineral Surfaces
At the present state of practice, geochemi-
cal models are used to predict the mineral
phases that might occur in contact with ground
water, but are not used to refine the estimates
of rate of reaction on the surfaces of mineral
phase that are recognized and determined by
chemical extraction techniques, or XRD, or
magnetic susceptibility. It may be possible to
refine an estimate of the field scale rate deg-
radation by assigning a rate constant to an
important mineral phase that is appropriate
for the particular geochemical environment
in that ground water under consideration.
7.5 Characterize Isotopic Fractionation
of cis-DCE and Vinyl Chloride on
Reactive Minerals
In many chlorinated solvent plumes, there
is evidence of natural anaerobic biodegrada-
tion of PCE and TCE based on accumula-
tion of cis-DCE; however, vinyl chloride, the
expected transformation product of cis-DCE,
does not accumulate to an appreciable extent.
In the past, this has been interpreted as a fail-
ure of cis-DCE to further degrade (cis stall)
even though transport and fate models of the
evolution of the plume indicate that a sink for
cis-DCE is necessary to model the disposition
of cis-DCE in the plume. There is increas-
ing success in using stable isotopes to explain
the behavior of cis-DCE and vinyl chloride
in ground water. Work is needed to deter-
mine the carbon isotopic enrichment factor
for degradation of cis-DCE on iron(II) mono-
-------
sulfides, and enrichment factors of all the
chlorinated alkenes on phyllosilicate clays.
7.6 Characterize the Role of
Manganese Oxides
Manganese oxides can be an important com-
ponent of aquifer matrix material. A variety of
organic compounds have been shown to react
with manganese oxide (MnO2). Cheney et al.
(1998) reported a rapid reaction of atrazine with
synthetic 5-MnO2 to produce N-dealkylated
forms of atrazine. The rate of dealkylation was
rapid compared to rates of biological degradation
of atrazine. Barrett and McBride (2005) showed
that glyphosate and its degradation product
aminomethylphosphonate reacted with birnes-
site (a manganese oxide mineral) to produce
ortho-phosphate. Mihelcic and Luthy (1988)
showed that a-naphthol was rapidly degraded by
manganese oxide. Ko et al. (2007) found that
4-chlorophenol degraded on manganese oxide.
The rate of degradation was at a maximum at pH
near the zero point of charge for the manganese
dioxide (pH 4 to 5). Petrie et al. (1995) showed
that addition of manganese oxide to soil poi-
soned with mercuric chloride and sodium azide
showed significant reductions in the concentra-
tion of pentachlorophenol compared to poisoned
soil without addition of manganese oxide.
Despite literature that documents rapid reac-
tion of chlorinated aromatic organic com-
pounds with manganese oxide, to our knowl-
edge, nothing has been published concerning
the rate of abiotic degradation of chlorinated
alkenes or alkanes on manganese oxides.
-------
Allen, H.E., G. Fu, and B. Deng. Analysis of
acid-volatile sulfide (AVS) and simultane-
ously extracted metals (SEM) for estima-
tion of potential toxicity in sediments.
Environmental Toxicology Chemistry
12: 1441-1453 (1993).
Amonette, I.E. Methods for determination of
mineralogy and environmental availabil-
ity. In Soil Mineralogy with Environmental
Applications (eds. J. Dixon and D. Schulze),
Soil Science Society of America, Madison,
WI (2002).
Amonette, I.E., DJ. Workman, D.W.
Kennedy, J.S. Fruchter, and Y.A. Gorby.
Dechlorination of carbon tetrachlo-
ride by Fe(II) associated with goethite.
Environmental Science & Technology
34: 4606-4613 (2000).
Arakaki, T. and J.W. Morse. Coprecipitation
and adsorption of Mn(II) with mackinawite
(FeS) under conditions similar to those
found in anoxic sediments. Geochimica et
Cosmochimica Acta 57: 9-14 (1993).
Arnold, W.A. and A.L. Roberts. Pathways of
chlorinated ethylene and chlorinated acety-
lene reaction with Zn(0). Environmental
Science & Technology 32: 3017-3025 (1998).
Arnold, W. A. and A.L. Roberts. Pathways and
kinetics of chlorinated ethylene and chlori-
nated acetylene reaction with Fe(0) particles.
Environmental Science & Technology
34: 1794-1805 (2000).
Assaf-Anid, N. and K.Y. Lin. Carbon tetra-
chloride reduction by Fe2+, S2~, and FeS with
vitamin B12 as organic amendment. Journal
of Environmental Engineering 128: 94-99
(2002).
8.0
References
Ayala-Luis, K.B., C.B. Koch, andH.C.B.
Hansen. The standard Gibbs energy of
formation of Fe(II)Fe(III) hydroxide sulfate
green rust. Clays and Clay Minerals 56: 633-
644 (2008).
Ballet, O., J.M.D. Coey, KJ. Burke. Magnetic
properties of sheet silicates; 2:1:1 layer min-
erals. Physics and Chemistry of Minerals
12: 370-378 (1985).
Barbash, I.E. and M. Reinhard. Abiotic deha-
logenation of 1,2-dichloroethane and
1,2-dibromoethane in aqueous solution
containing hydrogen sulfide. Environmental
Science & Technology 23: 1349-1358 (1989).
Barrett, K.A. and M.B. McBride. Oxidative
degradation of glyphosate and amino-
methylphosphonate by manganese oxide.
Environmental Science & Technology 39:
9223-9228 (2005).
Bearcock, J.M., W.T. Perkins, E. Dinelli, and
S.C. Wade. Fe(II)/Fe(III) 'green rust' devel-
oped within ochreous coal mine drainage
sediment in South Wales, UK. Mineralogical
Magazine 70: 731-741 (2006).
Bebie, J., M.A.A. Schoonen, M. Fuhrmann and
D.R. Strongin. Surface charge development
on transition metal sulfides: An electrokinetic
study. Geochimica et Cosmochimica Acta
62: 633-642(1998).
Behrends, T. and P. Van Cappellen.
Transformation of hematite into magnetite
during dissimilatory iron reduction - condi-
tions and mechanisms. Geomicrobiology
Journal 24: 403-416 (2007).
Benner, S.G., D.W. Blowes, W.D. Gould, R.B.
Herbert, Jr., and CJ. Ptacek. Geochemistry
of a permeable reactive barrier for metals
-------
and acid mine drainage. Environmental
Science & Technology 33: 2793-2799 (1999).
Benning, L.G., R.T. Wilkin and H.L. Barnes.
Reaction pathways in the Fe-S system
below 100°C. Chemical Geology 167: 25-51
(2000).
Bernal, J.D., D.R. Dasgupta, and A.L. Mackay.
The oxides and hydroxides of iron and their
structural interrelationships. Clay Mineral
Bulletin^. 15-30(1959).
Berner, R. A. Electrode studies of hydrogen
sulfide in marine sediments. Geochimica et
CosmochimicaActa21: 563-575 (1963).
Berner, R.A. Iron sulfides in Pleistocene deep
Black Sea sediments and their paleo-ocean-
ographic significance. In, The Black Sea—
Geology, Chemistry and Biology., eds. E. T.
Degens, D. A. Ross. Am. Assoc. Pet. Geol.,
Mem., 20: 524-531 (1974).
Berner, R.A. Sedimentary pyrite formation.
American Journal of Science 268: 1-23
(1970).
Billon, G., B. Ouddane, J. Laureyns and A.
Boughriet. Analytical and thermodynamic
approaches to the mineralogical and compo-
sitional studies on anoxic sediments. Journal
of Soils and Sediments 3:180-187 (2003).
Bloom, Y, R. Aravena, D. Hunkeler, E. Edwards
and S.K. Frape. Carbon isotope fractionation
during microbial dechlorination of trichlo-
roethene, c/5-l,2-dichloroethene, and vinyl
chloride: Implications for assessment of
natural attenuation. Environmental Science &
Technology 34: 2768-2772 (2000).
Bocher, R, A. Gehin, C. Ruby, J. Ghanbaja, M.
Abdelmoula, J-M.R. Genin. Coprecipitation
of Fe(II-III) hydroxycarbonate green rust
stabilised by phosphate adsorption. Solid
State Sciences 6: 117-124 (2004).
Bond, D.L. and S. Fendorf Kinetics and struc-
tural constraints of chromate reduction
by green rusts. Environmental Science &
Technology 37: 2750-2757 (2003).
Bourrie, G., F. Trolard, J. Genin, A. Jaffrezic, V.
MaTtre and M. Abdelmoula. Iron control by
equilibria between hydroxy-green rusts and
solutions in hydromorphic soils. Geochimica
Cosmochimica Acta 63: 3417-3427 (1999).
Boursiquot, S., M. Mullet, M. Abdelmoula, J.-M.
Genin and J.-J. Ehrhardt. The dry oxidation
of tetragonal FeSj 4 mackinawite. Physics
and Chemistry of Minerals 28: 600-611
(2001).
Bouwer, EJ. and PL. McCarty. Ethylene dibro-
mide transformation under methanogenic
conditions. Applied and Environmental
Microbiology 50: 527-528 (1985).
Bradley, P.M. History and ecology of chloroeth-
ene biodegradation: a review. Bioremediation
Journal 7: 81-109 (2003).
Brown, R.A., J.T. Wilson and M. Ferrey.
Monitored natural attenuation forum: the
case for abiotic MNA. Remediation Journal
17: 127-137 (2007).
Burris, D.R., C.A. Delcomyn, M.H. Smith and
A.L. Roberts. Reductive dechlorination of
tetrachloroethylene and trichloroethylene
catalyzed by vitamin B12 in homogeneous
and heterogeneous systems. Environmental
Science & Technology 30: 3047-3052 (1996).
Burton, E.D., L.A. Sullivan, R.T. Bush, S.G.
Johnston and A.F. Keene. A simple and
inexpensive chromium-reducible sulfur
method for acid-sulfate soils. Applied
Geochemistry 23: 2759-2766 (2008).
Butler, E.G. and K.F. Hayes. Effects of solution
composition and pH on the reductive dechlo-
rination of hexachloroethane by iron sulfide.
Environmental Science & Technology
32: 1276-1284 (1998).
Butler, E.G. and K.F. Hayes. Factors influenc-
ing rates and products in the transformation
-------
of trichloroethylene by iron sulfide and iron
metal. Environmental Science & Technology
35: 3884-3891 (2001).
Butler, E.G. and K.F. Hayes. Kinetics of the
transformation of halogenated aliphatic
compounds by iron sulfide. Environmental
Science & Technology 34: 422-429 (2000).
Butler, E.G. and K.F. Hayes. Kinetics of
the transformation of trichloroethylene
and tetrachloroethylene by iron sulfide.
Environmental Science & Technology
33:2021-2027(1999).
Butler, E.G. and K.F. Hayes. Reductive
Transformation of Halogenated Aliphatic
Pollutants by Iron Sulfide. Chapter 6,
In: Chemicals in the Environment: Fate,
Impacts, and Remediation, eds., R. L.
Lipnick, R. P. Mason, M. L. Phillips, and
C. U. Pittman, American Chemical Society
Symposium Series No. 806, Washington, DC
(2002).
Butler, IB. and D. Rickard. Framboidal pyrite
formation via the oxidation of iron(II) mono-
sulfide by hydrogen sulfide. Geochimica et
Cosmochimica Acta 64: 2665-2672 (2000).
Campbell, T.J., D.R. Burris, A.L. Roberts and
J.R. Wells. Trichloroethylene and tetra-
chloroethylene reduction in a metallic iron-
water-vapor batch system. Environmental
Toxicology and Chemistry 16: 625-630
(1997).
Canfield, D.E. and R.A. Berner. Dissolution and
pyritization of magnetite in anoxie marine
sediments. Geochimica et Cosmochimica
Acta 51: 645-659(1987).
Canfield, D.E., R. Raiswell, J.T. Westrich, C.M.
Reaves and R. A. Berner. The use of chro-
mium reduction in the analysis of reduced
inorganic sulfur in sediments and shales.
Chemical Geology 54: 149-155 (1986).
Casas, A. and E.A. Crecelius. Relationship
between acid volatile sulfide and the toxicity
of zinc, lead and copper in marine sediments.
Environmental Toxicology and Chemistry
13: 529-536(1994).
Cervini-Silva, J., R.A. Larson, J. Wu and J.W.
Stucki. Transformation of chlorinated ali-
phatic compounds by ferruginous smectite.
Environmental Science & Technology
35: 805-809(2001).
Charlet, L., D. Bosbach and T. Peretyashko.
Natural attenuation of TCE, As, Hg linked
to heterogeneous oxidation of Fe(II): an
AFM study. Chemical Geology 190: 303-319
(2002).
Cheney, M.A., J.Y. Shin, D.E. Crowley, S.
Alvey, N. Malengreau and G. Sposito.
Atrazine dealkylation on a manganese
oxide surface. Colloids and Surfaces A:
Physiochemical and Engineering Aspects
137: 267-273 (1998).
Choi, J.Y. Degradation of perchloroethylene and
nitrate by high-activity modified green rusts.
Doctoral Dissertation. 150 pp. Texas A&M
University (2005).
Choi, J., K. Choi and W. Lee. Effects of tran-
sition metal and sulfide on the reductive
dechlorination of carbon tetrachloride and
1,1,1-trichloroethane by FeS. Journal of
Hazardous Materials 162: 1151-1158 (2009).
Choi, J., and W. Lee. Enhanced degradation of
tetrachloroethylene by green rusts with plati-
num. Environmental Science & Technology
42: 3356-3362 (2008).
Christiansen, B.C. and S.L.S. Stipp. Structure
and reactivity of green rust. Geophysical
Research Abstracts, Vol. 5, 00644, European
Geophysical Society (2003).
Christiansen, B.C., T. Balic-Zunic, K.
Dideriksen, and S.L.S. Stipp. Identification
of green rust in groundwater. Environmental
-------
Science & Technology 43: 3436-3441
(2009).
Chun, C.L., R.M. Hozalski, and W.A. Arnold.
Degradation of disinfection byproducts by
carbonate green rust. Environmental Science
& Technology 41: 1615-1621 (2007).
Clesceri, L.S., A.E. Greenberg and A.D. Eaton
(editors). Standard methods for the exami-
nation of water and wastewater, American
Public Health Association, American Water
Works Association, Water Environment
Federation, 20th edition, Washington DC,
USA (1999).
Cohn, C.A., S. Mueller, E. Wimmer, N. Leifer,
S. Greenbaum, D.R. Strongin and M.A.A.
Schoonen. Pyrite induced hydroxyl radi-
cal formation and its effect on nucleic acids.
Geochemical Transactions 7: 3. Published
online 2006 April 4. doi: 10.1186/1467-4866-
7-3 (2006).
Conway, B.E., J.C.H. Ku and EC. Ho. The elec-
trochemical surface reactivity of iron sulfide,
FeS2. Journal of Colloid Interface Science
75: 357-372 (1980).
Cornwell, J.C. and J.W. Morse. The charac-
terization of iron sulfide minerals in anoxic
marine sediments. Marine Chemistry
22: 193-206 (1987).
Criddle, C.S. and PL. McCarty. Electrolytic
model system for reductive dehalogenation
in aqueous environments. Environmental
Science & Technology 25: 973-978 (1991).
Cuttler, A.H., V. Man, T.E. Cranshaw, and G.
Longworth. A Mossbauer study of green
rust precipitates: I. Preparation from sulphate
solutions. Clay Minerals 25: 289-301 (1990).
Danielsen, K., and K.F. Hayes. pH dependence
of carbon tetrachloride reductive dechlorina-
tion by magnetite. Environmental Science &
Technology 38: 4745-4752 (2004).
Darlington, R., L. Lehmicke, R.G. Andrachek
and D.L. Freedman. Biotic and abiotic
anaerobic transformations of trichloroethene
and c/5-l,2-dichloroethene in fractured sand-
stone. Environmental Science & Technology
42(12): 4323-4330 (2008).
Dearing, J. Environmental Magnetic
Susceptibility, Using the Bartington MS2
System. Chi Publishing, Kenilworth, England
ISBN 09523409 09 (1999).
Devlin, J.F. and D. Miiller. Field and laboratory
studies of carbon tetrachloride transforma-
tion in a sandy aquifer under sulfate reduc-
ing conditions. Environmental Science &
Technology 33: 1021-1027 (1999).
Di Toro, D.M., J.D. Mahony, DJ. Hansen,
K.J. Scott, M.B. Hicks, S M. Mayr and
M.S. Redmond. Toxicity of cadmium in
sediments: the role of acid volatile sulfide.
Environmental Toxicology and Chemistry 9:
1487-1502 (1990).
Doong, R.-A. and S.-C.Wu. Reductive dechlori-
nation of chlorinated hydrocarbons in aque-
ous solutions containing ferrous and sulfide
ions. Chemosphere 24: 1063-1075 (1992).
dos Santos Afonso, M. and W. Stumm.
Reductive dissolution of iron (III) (hydr)
oxides by hydrogen sulfide. Langmuir 8:
1671-1675 (1992).
Drissi, S.H., P. Refait, M. Abdelmoula and
J.M.R. Genin. The preparation and thermo-
dynamic properties of Fe(II)-Fe(III) hydrox-
ide-carbonate (green rust 1); Pourbaix dia-
gram of iron in carbonate-containing aqueous
media. Corrosion Science 37: 2025-2041
(1995).
Duan, W.-M., M.L. Coleman, and K. Pye.
Determination of reduced sulphur species
in sediments-an evaluation and modified
technique. Chemical Geology 141: 185-194
(1997).
-------
Eisner, M., S.B. Haderlein, T. Kellerhals,
S. Luzi, L. Zwank, W. Angst and R.P.
Schwarzenbach. Mechanisms and products
of surface-mediated reductive dehalogenation
of carbon tetrachloride by Fe(II) on goethite.
Environmental Science & Technology
38: 2058-2066 (2004a).
Eisner, M., R.P. Schwarzenbach, and S.B.
Haderlein. Reactivity of Fe(II)-bearing
minerals toward reductive transformation of
organic contaminants. Environmental Science
& Technology 38: 799-807 (2004b).
Eisner, M., L. Zwank, D. Hunkeler and R.P.
Schwarzenbach. A new concept linking
observable stable isotope fractionation to
transformation pathways of organic pollut-
ants. Environmental Science & Technology
39: 6896-6916 (2005).
Erbs, M., H.C.B. Hansen and C.E. Olsen.
Reductive dechlorination of carbon tetra-
chloride using iron(II) iron(III) hydroxide
sulfate (green rust). Environmental Science
& Technology 33: 307-311 (1999).
Feder, R, F. Trolard, G. Klingelhofer, and G.
Bourrie. In situ Mossbauer spectroscopy:
Evidence for green rust (fougerite) in a
gleysol and its mineralogical transforma-
tions with time and depth. Geochimica et
Cosmochimica Acta 69: 4463-4483 (2005).
Fennelly, J.P and A.L. Roberts. Reaction of
1,1,1-trichloroethane with zero-valent metals
and bimetallic reductants. Environmental
Science & Technology 32: 1980-1988 (1998).
Ferrey, M.L., R.T. Wilkin, R.G. Ford and J.T.
Wilson. Non-biological removal of cis-
dichloroethylene and 1,1-dichloroethylene
in aquifer sediment containing magnetite.
Environmental Science & Technology
38: 1746-1752 (2004).
Furukawa, Y, J. Kim, J. Watkins and R.T.
Wilkin. Formation of ferrihydrite and asso-
ciated iron corrosion products in perme-
able reactive barriers of zero-valent iron.
Environmental Science & Technology 36:
5469-5475 (2002).
Gancedo, J.R., M.L. Martinez and J.M. Oton.
Mossbauer spectroscopy study of corrosion
products of iron with ammonium nitrate
in aqueous solutions. Journal of Physics
Colloquium C6:297-299 (1976).
Gander, J.W., G.F. Parkin and M.M. Scherer.
Kinetics of 1,1,1-trichloroethane transfor-
mation by iron sulfide and a methanogenic
consortium. Environmental Science &
Technology 36: 4540-4546 (2002).
Genin, J.-M. R., G. Bourrie, F. Trolard, M.
Abdelmoula, A. Jaffrezic, P. Refait, V.
Maitre, B. Humbert, and A. Herbillon.
Thermodynamic equilibria in aqueous sus-
pensions of synthetic and natural Fe(II)-
Fe(III) green rusts: occurrences of the min-
eral in hydromorphic soils. Environmental
Science & Technology 32: 1058-1068
(1998).
Genin J.-M.R., P. Refait, and M. Abdelmoula.
Green rusts and their relationship to iron cor-
rosion; a Key Role in Microbially Influenced
Corrosion. Hyperfine Interactions
139: 119-131 (2002).
Gillham, R.W., and S.F. O'Hannesin. Enhanced
degradation of halogenated aliphatics by
zero-valent iron. Ground Water 32: 958-967
(1994).
Gregory, K.B., P. Larese-Casanova, G.F. Parkin
and M.M. Scherer. Abiotic transformation
of hexahydro-l,3,5-trinitro-l,3,5-triazine
by Fell bound to magnetite. Environmental
Science & Technology 38: 1408-1414 (2004).
Hanoch, R.J., H. Shao and E.G. Butler.
Transformation of carbon tetrachloride by
bisulfide treated goethite, hematite, magne-
tite, and kaolinite. Chemosphere 63: 323-334
(2006).
Hansel, C.M., S.G. Benner and S. Fendorf.
Competing Fe(II)-induced mineralization
-------
pathways of ferrihydrite. Environmental
Science & Technology 39: 7147-7153 (2005).
Hansen, H.C.B. Composition, stabilization, and
light absorption of Fe(II)Fe(III) hydroxy-
carbonate (green rust). Clay Minerals 24:
663-669 (1989).
Hansen, H.C.B. Environmental chemistry of
iron(II)-iron(III) LDHs. In Layered Double
Hydroxides (ed. V. Rives), pp. 413-434.
Nova Science Publisher, Huntington, NY
(2001).
Hansen, H.C.B., S. Guldberg, M. Erbs and
C.B. Koch. Kinetics of nitrate reduction by
green rusts-effects of interlayer anion and
Fe(II):Fe(III) ratio. Applied Clay Science
18: 81-91 (2001).
Hansson, E.B., M.S. Odziemkowski and R.W.
Gillham. Influence of Na2S on the degrada-
tion kinetics of CC14 in the presence of very
pure iron. Journal of Contaminant Hydrology
98: 128-134 (2008).
Hashsham, S. A., R. Scholze and D.L. Freedman.
Cobalamin-enhanced anaerobic bio-
transformation of carbon tetrachloride.
Environmental Science & Technology
29: 2856-2863 (1995).
He, Y.T., J.T. Wilson, and R.T. Wilkin. Impact
of iron sulfide transformation on trichlo-
roethylene degradation. Geochimica et
Cosmochimica Acta 74: 2025-2039 (2010).
He, Y.T., J.T. Wilson and R.T. Wilkin.
Transformation of reactive iron minerals in
a permeable reactive barrier (biowall) used
to treat TCE in groundwater. Environmental
Science & Technology 42: 6690-6696 (2008).
Henderson, J.K., D.L. Freedman, R.W. Falta, T.
Kuder and J.T. Wilson. Anaerobic biodegra-
dation of ethylene dibromide and 1,2-dichlo-
roethane in the presence of fuel hydrocar-
bons. Environmental Science & Technology
42: 864-870 (2008).
Henneke, E., G.W. Luther, G.J. de Lange and
J. Hoefs. Sulphur speciation in anoxic
hypersaline sediments from the east-
ern Mediterranean Sea. Geochimica et
Cosmochimica Acta 61: 307-321 (1997).
Herbert, R.B., S.G. Benner and D.W. Blowes.
Solid phase iron-sulfur geochemistry of a
reactive barrier for treatment of mine drain-
age. Applied Geochemistry 15: 1331-1343
(2000).
Herbert, R.B., S.G. Benner, A.R. Pratt and D.W.
Blowes. Surface chemistry and morphology
of poorly crystalline iron sulfides precipitated
in media containing sulfate-reducing bacte-
ria. Chemical Geology 144: 87-97 (1998).
Heron, G., C. Crouzet, A.C. Bourg and T.H.
Christensen. Speciation of Fe(II) and
Fe(III) in contaminated aquifer sediments
using chemical extraction techniques.
Environmental Science & Technology
28:1698-1705 (1994).
Horneman, A., A. van Geen, D.V. Kent,
PE. Mathe, Y Zheng, R.K. Dhar, S.
O'Connell, M.A. Hoque, Z. Aziz, Z., M.
Shamsudduha, A.A. Seddique and K.M.
Ahmed. Decoupling of As and Fe release
to Bangladesh groundwater under reducing
conditions. Part I: Evidence from sediment
profiles. Geochimica et Cosmochimica Acta.
68: 3459-3473 (2004).
Hsieh, YP, S.-W. Chung, Y.-J. Tsau and C.-T.
Sue. Analysis of sulfides in the presence
of ferric minerals by diffusion methods.
Chemical Geology 182: 195-201 (2002).
Hunger, S. and L.G. Benning. Greigite: a true
intermediate on the polysulfide pathway to
pyrite. Geochemical Transactions 8: 1-20
(2007).
Hunkeler, D., R.U. Meckenstock, B. Sherwood
Lollar, T.C. Schmidt and J.T. Wilson. A
Guide for Assessing Biodegradation and
Source Identification of Organic Ground
-------
Water Contaminants using Compound
Specific Isotope Analysis (CSIA). EPA
600/R-08/148 (2008).
Hurtgen, M.T., T.W. Lyons, E.D. Ingall and
A.M. Cruse. Anomalous enrichments of
iron monosulfide in euxinic marine sedi-
ments and the role of H2S in iron sulfide
transformations:Examples from Effingfham
Inlet, Orca basin, and the Black Sea.
American Journal of Science 299: 556-588
(1999).
Jeong, H.Y. and K.F. Hayes. Impact of transi-
tion metals on reductive dechlorination
rate of hexachloroethane by mackinawite.
Environmental Science & Technology
37: 4650-4655 (2003).
Jeong, H.Y. and K.F. Hayes. Reductive dechlo-
rination of tetrachloroethylene and trichloro-
ethylene by mackinawite (FeS) in the pres-
ence of metals: reaction rates. Environmental
Science & Technology 41: 6390-6396
(2007).
Jeong, H.Y, H. Kim and K.F. Hayes. Reductive
dechlorination pathways of tetrachloroeth-
ylene and trichloroethylene and subsequent
transformation of their dechlorination prod-
ucts by mackinawite (FeS) in the presence of
metals. Environmental Science & Technology
41: 7736-7743 (2007).
Jeong, H.Y, J.H. Lee and K.F. Hayes.
Characterization of synthetic nanocrystal-
line mackinawite: Crystal structure, particle
size, and specific surface area. Geochimica et
Cosmochimica Acta 72: 493-505 (2008).
Karim, Z. Formation of ferrihydrite by inhibi-
tion of green rust structures in the presence
of silicon. Soil Science Society of America
Journal 50: 247-250 (1986).
Kennedy, L., J.W. Everett and J. Gonzales.
Aqueous and mineral intrinsic bioremedia-
tion assessment: natural attenuation. Journal
of Environmental Engineering 130: 942-950
(2004).
Kennedy, L.G., J.W. Everett and J. Gonzales.
Assessment of biogeochemical natural atten-
uation and treatment of chlorinated solvents,
Altus AFB, Altus, Oklahoma. Journal of
Contaminant Hydrology 83: 221-236 (2006).
Kenneke, J.F. and E.J. Webber. Reductive
dehalogenation of halomethanes in iron- and
sulfate-reducing sediments, 1. Reactivity
pattern analysis. Environmental Science &
Technology 37: 713-720 (2003).
Klausen, J., S.P. Troeber, S.B. Haderlein and R.P.
Schwarzenbach. Reduction of substituted
nitrobenzenes by Fe(II) in aqueous min-
eral suspensions. Environmental Science &
Technology 29: 2396-2404 (1995).
Kneedler, E.M., J. Rothe, K.W. Weissmahr,
K. Pecher, and B.P Tonner. Identification
of green rust in environmental compounds
using XANES of Fe-Ln, m Edges. (1997).
http://www.als.lbl.gov/als/compendium/
AbstractManager/uploads/Kneedlerl.pdf
(accessed September 22, 2009).
Ko, S.-O., S.-Y Jun, D.-H. Lee, J. Park and
W.-S. Shin. Effects of oxidative coupling
reaction of 4-chlorophenol with manganese
oxide on the phenanthrene sorption. Journal
of Environmental Science and Health Part A.
42:257-263 (2007).
Koch, C.B. and S. M0rup. Identification of
green rust in an ochre sludge. Clay Minerals
26: 577-582 (1991).
Komadel, P. and J. W. Stucki. Quantitative
assay of minerals for Fe2+ and Fe3+ using
1,10-phenanthroline: III. A rapid photochem-
ical method. Clays and Clay Minerals
36: 379-381 (1988).
Kostka, J.E. and G.W. Luther. Partitioning and
speciation of solid phase iron in saltmarsh
-------
sediments. Geochimica et Cosmochimica
Acta58: 1701-1710(1994).
Kriegman-King, M. and M. Reinhard.
Transformation of carbon tetrachloride by
pyrite in aqueous solution. Environmental
Science & Technology 28: 692-700 (1994).
Kriegman-King, M.R. and M. Reinhard.
Transformation of carbon tetrachloride in the
presence of sulfide, biotite, and vermiculite.
Environmental Science & Technology
26: 2198-2206 (1992).
Kuder, T., Y. He, J.T. Wilson and R.P. Philp.
Reactions of 1,2-dibromoethane in FeS-
hydrogen sulfide systems: kinetics and iso-
tope fractionation. Environmental Science &
Technology (In Prep.).
Lasorsa, B. and A. Casas. A comparison of
sample handling and analytical methods for
determination of acid volatile sulfides in
sediment. Marine Chemistry 52: 211-220
(1996).
Lee, W. and B. Batchelor. Abiotic reductive
dechlorination of chlorinated ethylenes by
iron-bearing soil minerals. 1. Pyrite and mag-
netite. Environmental Science & Technology
36: 5147-5154(2002a).
Lee, W. and B. Batchelor. Abiotic reductive
dechlorination of chlorinated ethylenes by
iron-bearing soil minerals. 2. Green rust.
Environmental Science & Technology
36: 5348-5354 (2002b).
Lee, W. and B. Batchelor. Abiotic reductive
dechlorination of chlorinated ethylenes by
iron-bearing phyllosilicates. Chemosphere
56: 999-1009 (2004).
Lee, W. and B. Batchelor. Reductive capacity of
natural reductants. Environmental Science &
Technology 37: 535-541 (2003).
Lennie, A.R., S.A.T. Redfern, PE. Champness,
C.P Stoddart, PF. Schofield and D.J.
Vaughan. Transformation of mackinawite to
greigite: An in situ X-ray powder diffraction
and transmission electron microscope study.
American Mineralogist 82: 302-309 (1997).
Lennie, A.R., S.A.T. Redfern, PF. Schofield and
D.J. Vaughan. Synthesis and Rietveld crystal
structure refinement of mackinawite, tetrago-
nal FeS. Mineralogical Magazine
59: 677-683 (1995).
Leonard, E.N., V.R. Mattson, D.A. Benoit, R.A.
Hoke and G.T. Ankley. Seasonal varia-
tion of acid volatile sulfide concentration
in sediment cores from three northeastern
Minnesota lakes. Hydrobiologia 271: 87-95
(1993).
Leonard, E.N., D.R. Mount and G.T. Ankley.
Modification of metal partitioning by supple-
menting acid volatile sulfide in freshwater
sediments. Environmental Toxicology and
Chemistry 18: 858-864 (1999).
Liang, X., Y. Dong, T. Kuder, L.R. Krumholz,
R.P. Philp and E.G. Butler. Distinguishing
abiotic and biotic transformation of tetrachlo-
roethylene and trichloroethylene by stable
carbon isotope fractionation. Environmental
Science & Technology 41: 7094-7100 (2007).
Liang, X., R.P. Philp, and E.G. Butler. Kinetic
and isotope analyses of tetrachloroethylene
and trichloroethylene degradation by model
Fe(II)-bearing minerals. Chemosphere
75: 63-69 (2009).
Lipczynska-Kochany, E., S. Harms, R. Milburn,
G. Sprah andN. Nadarajah. Degradation
of carbon tetrachloride in the presence of
iron and sulphur containing compounds.
Chemosphere 29: 1477-1489 (1994).
Liu, X.M., P. Peng, J.M. Fu and W.L. Huang.
Effects of FeS on the transformation kinetics
of y-hexachlorocyclohexane. Environmental
Science & Technology 37: 1822-1828
(2003).
-------
Liu, Y, S.A. Majetich, R.D. Tilton, D.S. Sholl
and G.V. Lowry. TCE dechlorination rates,
pathways, and efficiency of nanoscale
iron particles with different properties.
Environmental Science & Technology
39: 1338-1345 (2005).
Loyaux-Lawniczak, S., P. Refait, J.-J. Ehrhardt,
P. Lecomte, and J.-M. R. Genin. Trapping
of Cr by formation of ferrihydrite during the
reduction of chromate ions by Fe(II)-Fe(III)
hydroxysalt green rusts. Environmental
Science & Technology 34: 438- 443 (2000).
Maithreepala, R.A. and R.A. Doong. Enhanced
dechlorination of chlorinated methanes and
ethenes by chloride green rust in the pres-
ence of copper(II). Environmental Science &
Technology 39: 4082-4090 (2005).
McCormick, M.L., and P. Adriaens. Carbon
tetrachloride transformation on the surface
of nanoscale biogenic magnetite particles.
Environmental Science & Technology
38: 1045-1053 (2004).
McCormick, M.L., E. J. Bouwer and P. Adriaens.
Carbon tetrachloride transformation in
a model iron reducing culture: Relative
kinetics of biotic and abiotic reactions.
Environmental Science & Technology
36: 403-410 (2002).
MDH. Public Health Assessment Baytown
Township Groundwater Contamination
Site Washington County, Minnesota. EPA
Facility ID: MND982425209. Prepared
by: Minnesota Department of Health under
Cooperative Agreement with the Agency
for Toxic Substances and Disease Registry
(September 14, 2004). http://www.health.
state, mn. us/divs/eh/hazardous/sites/wash-
ington/baytown/baytownpha.pdf (accessed
06-07-2009).
Mihelcic, J.R. and R.G. Luthy. Degradation of
polycyclic aromatic hydrocarbon compounds
under various redox conditions in soil-water
systems. Applied and Environmental
Microbiology 54: 1182-1187 (1988).
Morse, J.W. and J.C. Cornwell. Analysis and
distribution of iron sulfide minerals in recent
anoxic marine sediments. Marine Chemistry
22: 55-69 (1987).
Morse, J.W. and D. Rickard. Chemical dynam-
ics of sedimentary acid volatile sulfide.
Environmental Science & Technology
38: 131A-136A(2004).
Murad, E. and R.M. Taylor. The Mossbauer
spectra of hydroxycarbonate green rusts.
Clay Minerals 19: 77-83 (1984).
Murad, E. and R.M. Taylor. The oxidation of
hydroxycarbonate green rusts, pp. 585-593.
In Industrial Applications of the Mossbauer
Effect, GJ. Long and J.G. Stevens (eds),
Plenum Press, New York, NY (1986).
Myneni, S.C.B., T.K. Tokunaga, and G.E.
Brown, Jr. Abiotic selenium redox transfor-
mations in the presence of Fe(II,III) oxides.
Science 278: 1106-1109(1997).
Neal, A.L., S. Techkarnjanaruk, A. Dohnalkova,
D. McCready, B.M. Peyton and G.G. Geesey.
Iron sulfides and sulfur species produced
at hematite surfaces in the presence of
sulfate-reducing bacteria. Geochimica et
Cosmochimica Acta 65: 223-235 (2001).
Neumann, A., T.B. Hofstetter, M. Skarpeli-Liati
and R.R. Schwarzenbach. Reduction of
polychlorinated ethanes and carbon tetra-
chloride by structural Fe(II) in smectites.
Environmental Science & Technology
43: 4082-4089(2009).
Neumann, A., T.B. Hofstetter, M. Liissi, O.A.
Cirpka, S. Petit and R.P Schwarzenbach.
Assessing the redox reactivity of structural
iron in smectites using nitroaromatic com-
pounds as kinetic probes. Environmental
Science & Technology 42: 8381-8387 (2008).
-------
Nooten, T.V., D. Springael and L. Bastiaens.
Positive impact of microorganisms on the
performance of laboratory-scale permeable
reactive iron barriers. Environmental Science
& Technology 42: 1680-1686 (2008).
Nzengung, V.A., R.M. Castillo, W.P. Gates and
G.L. Mills. Abiotic transformation of per-
chlorothylene in homogeneous dithionite
solution and in suspensions of dithionite-
treated clay minerals.Environmental Science
& Technology
35: 2244-2251 (2001).
O'Loughlin, EJ. and D.R. Burris. Reduction
of halogenated ethanes by green rust.
Environmental Toxicology and Chemistry
23: 41-48(2004).
O'Loughlin, E.J., K.M. Kemner and D.R. Burris.
Effects of Ag1, Au11, and Cum on the reduc-
tive dechlorination of carbon tetrachloride
by green rust. Environmental Science &
Technology 37: 2905-2912 (2003).
O'Loughlin, E., P. Larese-Casanova,
M. Scherer, and R. Cook. Green Rust
Formation from the Bioreduction of y
-FeOOH (Lepidocrocite): Comparison
of Several Shewanella Species.
Geomicrobiology Journal 24: 211-230
(2007).
Oita, I.J. Coulometric determination of sub-part-
per-million levels of sulfur in volatile liquids.
Analytical Chemistry 55: 2434-2436 (1983).
Ona-Nguema G., M. Abdelmoula, F. Jorand,
O. Benali, A. Gehin, J.-C. Block, and J.-M.
R. Genin. Iron(II,III) hydroxycarbonate
green rust formation and stabilization from
lepidocrocite bioreduction. Environmental
Science & Technology 36: 16-20 (2002b).
Ona-Nguema G., M. Abdelmoula, F. Jorand,
O. Benali, A. Gehin A., J.-C. Block, and
J.-M. R. Genin. Microbial reduction of lepi-
docrocite y-FeOOH by Shewanella putrefa-
ciens; the formation of green rust. Hyperfine
Interactions 139: 231-237 (2002a).
Ona-Nguema, G., C. Carteret, O. Benali, M.
Abdelmoula, J.-M.R. Genin, and F. Jorand.
Competitive formation of hydroxycarbonate
green rust 1 versus hydroxysulphate green
rust 2 in Shewanella putrefaciens cultures.
Geomicrobiology Journal 21: 79-90 (2004).
Ostwald, J. A note on occurrences of nick-
eliferous and cupriferous mackinawite.
MineralogicalMagazine 42: 516-517 (1978).
Park, S., S. Kim, J. Kim, S. Choi, H.I. Inyang
and S. Tokunaga. Particle surface hydropho-
bicity and the dechlorination of chloro-com-
pounds by iron sulfides. Water, Air, and Soil
Pollution 6: 97-110(2006).
Pecher, K., S.B. Haderlein and R.P
Schwarzenbach. Reduction of polyhalo-
genated methanes by surface-bound Fe(II)
in aqueous suspensions of iron oxides.
Environmental Science & Technology
36: 1734-1741 (2002).
Pesch, C.E., D.J. Hansen, W.S. Boothman, W.J.
Berry and J.D. Mahony. The role of acid
volatile sulfide and interstitial water metal
concentrations in determining bioavailability
of cadmium and nickel from contaminated
sediments to the marine polychaete Neanthes
arenaceodentata. Environmental Toxicology
and Chemistry 14: 129-141 (1995).
Petrie, R.A., J.E. McLean and R.C. Sims.
Treatment of pentachlorophenol with man-
ganese oxide addition to biotic and abiotic
sediments. Hazardous Waste & Hazardous
Materials 12: 271-281 (1995).
Pham, H.T., M. Kitsuneduka, J. Kara, K. Suto
and C. Inoue. Trichloroethylene transforma-
tion by natural mineral pyrite: The deciding
role of oxygen. Environmental Science &
Technology 42: 7470-7475 (2008).
-------
Phillips, D.H., B. Gu, D.B. Watson, Y. Roh, L.
Liang and S.Y. Lee. Performance evalu-
ation of a zerovalent iron reactive barrier:
Mineralogical characteristics. Environmental
Science & Technology 34: 4169-4176 (2000).
Popa, R. and B.K. Kinkle. Discrimination
among iron sulfide species formed in micro-
bial cultures. Journal of Microbiological
Methods 42: 167-174 (2000).
Posfai, M., PR. Buseck, D.A. Bazylinski and
R.B. Frankel. Iron sulfides from magneto-
tactic bacteria: Structure, composition, and
phase transitions. American Mineralogist
83: 1469-1481 (1998a).
Posfai, M., PR. Buseck, D.A. Bazylinski and
R.B. Frankel. Reaction sequence of iron
sulfide minerals in bacteria and their use as
biomarkers. Science 280: 880-883 (1998b).
Poulton, S.W. Sulfide oxidation and iron disso-
lution kinetics during the reaction of dis-
solved sulfide with ferrihydrite. Chemical
Geology 202: 79-94 (2003).
Poulton, S.W., M.D. Krom and R. Raiswell. A
revised scheme for the reactivity of iron
(oxyhydr)oxide minerals towards dissolved
sulfide. Geochimica et Cosmochimica Acta
68: 3703-3715 (2004).
Pruden, G. and C. Bloomfield. The determina-
tion of ferrous sulfide in soil in the presence
of ferric oxide. Analyst (London) 93: 532-
534(1968).
Pyzik, AJ. and S.E. Sommer. Sedimentary iron
monosulfides: Kinetics and mechanisms of
formation. Geochim ica et Cosmochimica
Acta 45: 687-698(1981).
Raiswell, R. and D.E. Canfield. Rates of reac-
tion between silicate iron and dissolved sul-
fide in Peru Margin sediments. Geochimica
et Cosmochimica Acta 60: 2777-2787 (1996).
Randall, S.R., D.M. Sherman, and K. Vala
Ragnarsdottir. Sorption of As(V) on green
rust (Fe4(II)Fe2(III)(OH)12SO4 3H2O) and
lepidocrocite (y-FeOOH): Surface complexes
from EXAFS spectroscopy. Geochimica et
Cosmochimica Acta 65: 1015-1023 (2001).
Refait, P., M. Abdelmoula, F. Trolard, J.-M.
R. Genin, JJ. Ehrhardt, and G. Bourrie.
Mossbauer and XAS study of a green rust
mineral; the partial substitution of Fe2+ by
Mg2+. American Mineralogist 86: 731-739
(2001).
Rickard, D. Kinetics of pyrite formation by
the H2S oxidation of iron(II) monosulfide in
aqueous solutions between 25 and 125°C: the
rate equation. Geochimica et Cosmochimica
Acta 61: 115-134(1997).
Rickard, D. The solubility of FeS. Geochimica
et Cosmochimica Acta 70: 5779-5789 (2006).
Rickard, D. and G.W. Luther. Chemistry of iron
sulfides. Chemical Review 107, 514-562
(2007).
Rickard, D. and G.W. Luther. Kinetics of pyrite
formation by the H2S oxidation of iron(II)
monosulfide in aqueous solutions between 25
and 125°C: the mechanism. Geochimica et
Cosmochimica Acta 61: 135-147 (1997).
Rickard, D. and J.W. Morse. Acid volatile
sulfide (AVS). Marine Chemistry 97 (2005)
141-197 (2005).
Rimstidt, J.D. and DJ. Vaughan. Pyrite oxi-
dation: a-state-of-the-art assessment of
the reaction mechanism. Geochimica et
Cosmochimica Acta 67: 873-880 (2003).
Roberts, A.L., L.A. Totten, W.A. Arnold, D.R.
Burris and TJ. Campbell. Reductive elimi-
nation of chlorinated ethylenes by zero-
valent metals. Environmental Science &
Technology 30: 2654-2659 (1996).
Scherer, M.M., E. O'Loughlin, G.F. Parkin, R.
Valentine, H. Al-Hosney, R. Handler, C.
Just, P. Larese-Casanova, T. Pasakarnis, S.
L. Smith. Sustainability of long-term abiotic
-------
attenuation of chlorinated ethenes. Final
Report, SERDPProjectER-1369.
(28 September 2007).
Schoonen, M.A.A. Mechanisms of sedimentary
pyrite formation. In Sulfur Biogeochemistry:
Past and Present. Geological Society of
America, Special Paper 379, Boulder, CO
(2004).
Schoonen, M.A.A. and H.L. Barnes. Reactions
forming pyrite and marcasite from solu-
tion: II. via FeS precursors below 100°C.
Geochimica et Cosmochimica Ada
55: 1505-1514(1991).
Schoonen, M.A. and D.R. Strongin. Catalysis
of Electron Transfer Reactions at Mineral
Surfaces. In: Environmental Catalysis,
by V.H. Grassian, Taylor and Francis,
pages 37-60, 2005.
Schwarzenbach, R.P, W. Giger, C. Schaffner
and O. Wanner. Groundwater contamina-
tion by volatile halogenated alkanes: abiotic
formation of volatile sulfur compounds under
anaerobic conditions. Environmental Science
& Technology 19: 322-327 (1985).
Shao, H. and E.G. Butler. Influence of soil
minerals on the rates and products of abiotic
transformation of carbon tetrachloride in
anaerobic soils and sediments. Environmental
Science & Technology 43: 1896-1901 (2009).
Shen, H. and J.T Wilson. Trichloroethylene
removal from groundwater in flow-through
columns simulating a permeable reac-
tive barrier constructed with plant mulch.
Environmental Science & Technology
41: 4077-4083 (2007).
Sivavec, T.M. and D.P Horney. Reduction of
chlorinated solvents by Fe (II) minerals. In:
Preprints of Papers Presented at the 213th
ACS National Meeting, vol. 37. American
Chemical Society: San Francisco, CA, pp.
115-117(1997).
Stucki, J. W. The quantitative assay of minerals
for Fe2+ and Fe3+ using 1,10-phenanthroline.
II. A photochemical method. Soil Science
Society of America Journal 45: 638-641
(1981).
Su, C. and R.W. Puls. Significance of iron(II,
III) hydroxycarbonate green rust in arsenic
remediation using zerovalent iron in labora-
tory column tests. Environmental Science &
Technology 38: 5224-5231 (2004).
Su, C. and R.T. Wilkin. Arsenate and arse-
nite sorption on and arsenite oxidation by
iron(II,III) hydroxycarbonate green rust.
American Chemical Society Symposium
Series 915: 25-40 (2005).
Sumoondur, A., S. Shaw, I. Ahmed, and L.G.
Benning. Green rust as a precursor for
magnetite: An in situ synchrotron based
study. Mineralogical Magazine 72: 201-204
(2008).
Suzuki, S., K. Shinoda, M. Sato, S. Fujimoto, M.
Yamashita, H. Konishi, T. Doi, T. Kamimura,
K. Inoue, and Y. Waseda. Changes in chemi-
cal state and local structure of green rust by
addition of copper sulphate ions. Corrosion
Science 50: 1761-1765 (2008).
Szecsody, I.E., J.S. Fruchter, M.D. Williams,
V.R. Vermeul and D. Sklarew. In situ
chemical reduction of aquifer sediments:
Enhancement of reactive iron phases and
TCE dechlorination. Environmental Science
& Technology 38: 4656-4663 (2004).
Tack, F.M., F. Lapauw and M.G. Verloo.
Determination and fractionation of sulphur
in a contaminated dredged sediment. Talanta
44: 2185-2192 (1997).
Tandoi, V, T.D. DiStefano, PA. Bowser, J.M.
Gossett and S.H. Zinder. Reductive deha-
logenation of chlorinated ethenes and halo-
genated ethanes by a high-rate anaerobic
enrichment culture Environmental Science &
Technology 28: 973-979 (1994).
-------
Taylor, L. A. and L. W. Finger. Structural
refinement and composition of macki-
nawite. Carnegie Institute of Washington
Geophysical Laboratory Annual Report
69: 318-322(1970).
Taylor, R.M. Crystal structures of some double
hydroxide minerals. MineralogicalMagazine
39: 377-389 (1973).
Taylor, R.M. Formation and properties of
Fe(II)/Fe(III) hydroxy-carbonate and its
possible significance in soils. Clay Minerals
15: 369-382 (1980).
Taylor, R.M. Stabilization of colour and struc-
ture in the pyroaurite-type compounds Fe(II)
Fe(III)-hydroxycarbonates. Clay Minerals
17: 369-372 (1982).
Tobler, N.B., T.B. Hofstetter, K.L. Straub, D.
Fontana and R.P Schwarzenbach. Iron-
mediated microbial oxidation and abiotic
reduction of organic contaminants under
anoxic conditions. Environmental Science &
Technology 41: 7765-7772 (2007).
Trolard, F., G. Bourrie, M. Abdelmoula, P.
Refait, and F. Feder. Fougerite, a new
mineral of the pyroaurite-iowaite group:
Description and crystal structure. Clays and
Clay Minerals 55: 323-334 (2007).
Ulrich, G.A., L.R. Krumholz and J.M. Suflita. A
rapid and simple method for estimating sul-
fate reduction activity and quantifying inor-
ganic sulfides. Applied and Environmental
Microbiology 63: 1627-1630 (1997).
Vaughan, DJ. and J.R. Craig. Mineral
Chemistry of Metal Sulfides. Cambridge,
MA, Cambridge University Press, 1978.
Vaughan, DJ. and M.S. Ridout. Mossbauer
studies of some sulphide minerals. Journal
of Inorganic Nuclear Chemistry 33: 741-746
(1971).
Vikesland, P.J., A.M. Heathcock, R.L. Rebodos
and K.E. Makus. Particle size and
aggregation effects on magnetite reactivity
toward carbon tetrachloride. Environmental
Science & Technology 41: 5277-5283 (2007).
Vogel, T.M., C.S. Criddle and PL. McCarty.
Transformations of halogenated aliphatic
compounds. Environmental Science &
Technology 21: 722-736 (1987).
Vogel, T.M. and PL. McCarty. Abiotic and
biotic transformations of 1,1,1-trichloro-
ethane under methanogenic conditions.
Environmental Science & Technology
21: 1208-1213 (1987).
Wang, Q. and J.W. Morse. Pyrite formation
under conditions approximating those in
anoxic sediments: I. Pathway and morphol-
ogy. Marine Chemistry 52: 99-121 (1996).
Weerasooriya, R. and B. Dharmasena. Pyrite-
as si sted degradation of trichloroethene
(TCE). Chemosphere 42: 389-396 (2001).
Wiedemeier, T.H., M.A. Swanson, D.E.
Moutoux, E.K. Gordon, J.T. Wilson, B.H.
Wilson, D.H. Kampbell, PE. Haas, R.N.
Miller, I.E. Hansen and F.H. Chapelle.
Technical Protocol for Evaluating Natural
Attenuation of Chlorinated Solvents in
Ground Water. EPA/600/R-98/128 (1998).
Wieder, R.K., G.E. Lang and V.A. Granus. An
evaluation of wet chemical methods for
quantifying sulfur fractions in freshwater
wetland peat. Limnology and Oceanography
30: 1109-1115 (1985).
Wilkin, R.T. Ground Water Issue: Mineralogical
Preservation of Solid Samples Collected
from Anoxic Subsurface Environments.
EPA/600/R-06/112 (2006).
Wilkin, R.T. and K. J. Bischoff. Coulometric
determination of total sulfur and reduced
inorganic sulfur fractions in environmental
samples. Talanta 70: 766-773 (2006).
Wilkin, R.T. and R.W. Puls. Capstone Report
on the Application, Monitoring, and
-------
Performance of Permeable Reactive Barriers
for Ground-Water Remediation, Volume 1:
Performance Evaluations at Two Sites, EPA
Report, EPA/600/R-03/045a, 156 pp. (2003).
Williams, A.G.B and M.M. Scherer. Kinetics
of Cr(VI) reduction by carbonate green
rust. Environmental Science & Technology
35: 3488-3494(2001).
Williams, A.G.B. and M.M. Scherer.
Spectroscopic evidence for Fe(II)-Fe(III)
electron transfer at the iron oxide-water
interface. Environmental Science &
Technology 38: 4782-4790 (2004).
Wilson, J.T., K. Banks, R.C. Earle, Y.T. He, T.
Kuder and C. Adair. Natural attenuation
of the lead scavengers 1,2-dibromoethane
(EDB) and 1,2-dichloroethane (1,2-DCA)
at motor fuel release sites and implica-
tions for risk management. USEPA, EPA
600/R-08/107, Ada, OK (2008).
Wilson, J.T., D.H. Kampbell, M. Ferrey and
P. Estuesta. Evaluation of the Protocol
for the Natural Attenuation of Chlorinated
Solvents: Case Study at the Twin Cities Army
Ammunition Plant. EPA/600/R-01/025, 2001.
Wolthers, M., SJ. Van der Gaast and D. Rickard.
The structure of disordered mackinawite.
American Mineralogist 88: 2007-2015
(2003).
Zegeye, A., G. Ona-Nguema, C. Carteret,
L. Huguet, M. Abdelmoula, F. Jorand.
Formation of hydroxysulphate green rust 2 as
a single iron(II-III) mineral in microbial cul-
ture. Geomicrobiology Journal 22: 389-399
(2005).
Zegeye, A, C. Ruby, and F. Jorand. Kinetic and
thermodynamic analysis during dissimilatory
y-FeOOH reduction: Formation of green rust
1 and magnetite. Geomicrobiology Journal
24: 51-64(2007).
Zwank, L., M. Eisner, A. Aeberhard, R.P
Schwarzenbach. and S.B. Haderlein.
Carbon isotope fractionation in the reduc-
tive dehalogenation of carbon tetrachloride
at iron (hydr)oxide and iron sulfide miner-
als. Environmental Science & Technology
39: 5634-5641 (2005).
-------
Appendix A
Iron Sulfides
A.I Experimental Details Related to
TCE Degradation by FeS
FeS was synthesized by mixing a solution of
Fe2+ (prepared with ferrous ammonium sulfate)
and a solution of NaHS inside an anaerobic
glovebox. The mixed solution contained equal
molars of iron and sulfide. NaHS was prepared
by reacting N2-sparged NaOH with 10% H2S
(g/g). The mixed FeS suspension was allowed
to settle overnight before removing supernatant.
Then, the FeS suspension was transferred into
a dialysis bag and dialyzed against filtered tap
water that had been sparged with oxygen-free
nitrogen gas. The dialysis water was amended
with Na2S to a S2~ concentration of approxi-
mately 100 mg/L. Based on prior experiments
using deionized water, a total of seven washes
over three to four days were sufficient to remove
all the ions except sodium and sulfide in the
synthesized FeS suspension. However, washing
with tap water would result in Na+, S2~ and ions
in the tap water being transferred to the solution.
After dialysis, the FeS suspension was stored in
a glass bottle inside an anaerobic glovebox until
use. The content of solid FeS in the suspension
was determined gravimetrically.
Filtered tap water was used for dialysis because
of its capacity to buffer pH. The tap water was
naturally buffered with a bicarbonate buffer
system (alkalinity 320 mg/L). Biological buffers
were avoided in this study because Danielsen
et al. (2005) and other studies have shown that
biological buffers have significant impact on the
rate of TCE degradation.
FeS directly from the FeS suspension after
dialysis was described as non-freeze dried FeS
(ND FeS). To prepare freeze dried FeS (FD
FeS), the dialyzed FeS suspension was freeze
dried with a LABCONCO® Freeze Dry System.
Filtered tap water sparged with oxygen-free
nitrogen was used to prepare a 13C-labeled TCE
stock solution. Experimental incubations were
carried out in glass serum vials with a measured
volume of 21 mL. To begin an incubation, 20
mL of well mixed suspension containing 20 g/L
FeS was transferred to each vial, then 1.05 mL
of 13C-labeled TCE stock solution was added,
and then the vial was sealed without headspace
with a Teflon-faced septum and a crimp cap.
Controls were prepared with filtered tap water
without FeS. Triplicates of control and FeS
treatment were prepared for each time interval.
To determine the effect of pH, the FeS suspen-
sions were adjusted to pH 5.4, 6.2, 7.2 and 8.4
using 1 M HC1 or 1 M NaOH before adding the
TCE stock solution. Samples were incubated for
different time periods on a Stovall® Low Profile
Roller in the dark at room temperature.
A suspension of freeze dried FeS was prepared
by sparging filtered tap water with oxygen-free
nitrogen, then adding freeze dried FeS to pro-
duce a final solid FeS content of 20 g/L. Then
the pH was adjusted to 7.2 with 1 M NaOH or
1 M HC1.
To generate a sample for analysis of TCE and
daughter products, 0.2-mL samples of the pore
water in the experimental systems were added to
20.0 mL filtered tap H2O (preserved with 2 drops
of concentrated HC1) in a glass vial with crimp
seal. The diluted sample was contained in a
21-mL glass serum vial that was sealed with a
Teflon faced septum and crimp cap. To generate
a sample for isotope analysis, 2 mL of original
sample were added to 18 mL of filtered tap water
in a sealed vial. One original replicate experi-
mental system was used directly to sample for
dissolved gases. A helium headspace was cre-
ated in the vial (2 mL). The vial was shaken to
distribute dissolved gasses to the headspace, and
-------
then the headspace was sampled for analysis of
methane, ethane, ethylene, and acetylene. The
appropriate Henry's Law Constant was used to
calculate the original concentration in the water
in the experimental system. In addition, pH,
Fe2+, and S2~ concentrations were measured after
filtering samples with 0.2-um Millipore® filter.
Concentrations of [1,2-13C] TCE, along with cis-
DCE, tmns-DCE, 1,1-DCE, and vinyl chloride
were analyzed using gas chromatography/mass
spectrometry (GC/ MS) according to a modi-
fication of the procedures established in EPA
Method 5021A, "Volatile Organic Compounds
in Various Sample Matrices using Equilibrium
Headspace Analysis," in conjunction with EPA
Method 8260C, "Volatile Organic Compounds
by Gas Chromatography/Mass Spectrometry
(GC/MS)". The method detection limit for each
compound was 0.001 |jM in the diluted sample,
or 0.1 |jM in the original experimental system.
Methane, ethylene, ethane, and acetylene were
analyzed by injecting headspace gases into a
HP Series P200H GC Chromatograph. The
detection limit for each hydrocarbon was below
0.076 |jM in the liquid phase. Dissolved organic
carbon (DOC) and dissolved inorganic carbon
(DIG) in water was analyzed using a Dohrman
DC-80 Carbon Analyzer. The method detec-
tion limit was 15 |jM. Stable isotope ratios
were analyzed by Isotech Laboratories, Inc.
(Champaign, IL). The precision of the analyses
was 0.1%o. Sulfide and Fe2+ in water samples
were analyzed with a spectrophotometer (Hach
DR/2010) using the methylene blue method and
the 1,10-phenanthroline method, respectively.
The pH was measured using a BNC pH TestrlO®
pH meter.
A.2 Experimental Details Related to
EDB Degradation by FeS
The FeS was synthesized the same way as
described in Section A. 1. An EDB stock solu-
tion was prepared in filtered tap water that had
been sparged with oxygen-free nitrogen gas.
The EDB microcosms in FeS suspension were
prepared in serum vials as described above with
the preparation of TCE microcosms. The pH of
the FeS suspensions was adjusted to pH 5.4, 6.2,
7.2, and 8.4 using 1M HC1 or 1 M NaOH before
adding EDB stock solution. The experimental
systems were incubated for different time peri-
ods on a Stovall® Low Profile Roller in the dark
at room temperature.
To analyze EDB and daughter products, the 0.2
mL original sample was added to 20.0 mL fil-
tered tap H2O (preserved with 2 drops of concen-
trated HC1) in a crimp sealed vial. For isotope
analysis, 2 mL of original sample were added to
18 mL of filtered tap water in crimp sealed vial.
For dissolved gases measurement, 2 mL of origi-
nal sample were added into 18 mL of filtered
tap water in crimp sealed vial. In addition, pH,
Fe2+, and S2~ concentration were measured after
filtering samples with 0.2-|j,m Millipore® filter.
EDB analytical methods followed procedures in
appendix D of Wilson et al. (2008).
Experiments described in Section A. 1 and
Section A.2 were conducted at the Kerr
Environmental Research Center.
-------
Appendix B
Magnetite
At the time of this writing (summer 2009),
Ferrey et al. (2004) provides the only published
information on the rates of degradation of
chlorinated organic compounds on naturally
occurring magnetite in an aquifer sediment. This
appendix describes additional microcosm studies
done at the Kerr Center. The microcosms were
constructed with sediment from aquifers con-
taminated with chlorinated solvents.
B.I Building 102 Site and Site A on
the TCAAP, North of St. Paul,
Minnesota
The Building 102 site is located in the shallow
water table aquifer on the west side of the Twin
Cities Army Ammunition Plant (TCAAP) (See
Ferrey et al., 2004 and Wilson et al., 2001 for
background on the TCAAP). The water table
occurs in sand approximately 1.2 to 2.4 m below
ground surface. The horizontal ground water
velocity was estimated to be 230 m yr"1. The
contamination flowed to the north, resulting in
a plume that was approximately 40 m wide and
140 m long (Compare Table B.I and Figure B.I).
The contaminants are essentially depleted in
72 m of travel from the most contaminated well.
The contaminated ground water is oxygenated
much of the time, with some traces of reduced
iron and reduced manganese.
Site A was also in the shallow water table
aquifer. It is located on the northwest side of
the TCAAP. Ground water at Site A was 3 to
6 m below ground surface, with the aquifer
ranging in thickness from 3 m near the source
area to 9 m to the west. The aquifer sedi-
ments are composed of lacustrine silt and fine
or medium sands. The ground water seepage
velocity is estimated to be 60 m yr1 at Site A.
The area near well 01U108 is considered the
primary source of contaminants to the aquifer
(Figure B.2). The wells depicted in Figure B.2
have been sampled annually for over 15 years.
Over time, the plume is shrinking back toward
the source. The ground water at Site A is
anoxic. PCE and TCE are reduced to cis-DCE,
but the concentrations of vinyl chloride are low
(Table B.2).
-------
O 01U582
01L581
O 01U583
01U584 O
100 feet
30 meter
N
01U581
O 01U580
O 01U579
O 01U578
Ground
Water
Flow
\
Figure B.1. Orientation of monitoring wells at the Building 102 site on the TCAAP.
Table B.1. Distribution of chlorinated organic compounds and geochemical parameters in ground water at the
Building 102 site on the TCAAP in 2005.
Building 102
Site
01U578
01U579
01U580
05U581
05L581
01U584
01U583
01U582
Distance
from source
m
Background
0.0
21.3
71.6
71.6
71.6
74.7
156.1
PCE
ug/L
<1
<1
1
0.4
<1
<1
1.2
<1
TCE
ug/L
<1
380
17000
1.3
9.1
<1
<1
<1
cis-DCE
ug/L
<1
520
4400
8.8
7.6
<1
<1
<1
Vinyl
Chloride
ug/L
<1
6.9
19
<1
<1
<1
<1
<1
02
mg/L
1.7
1.6
2.6
1.7
0.7
1.3
2.3
4.6
Fe2+
mg/L
1
ND
ND
ND
ND
ND
1.1
ND
pH
S.U.
6.5
6.4
6.6
6.2
6.7
6.6
6.1
6.6
-------
Ground
Water
Flow
n
n
Residential Wells
Q01U904
Q01U902
o 01U115
Q01U117
01U108
300 feet
100 meter
Figure B.2. Orientation of monitoring wells at Site A on the TCAAP.
Table B.2. Distribution of chlorinated organic compounds and geochemical parameters in ground water at Site
A on the TCAAP in 1998.
Site A
01U067
01U108
01U117
01U115
01U904
Distance
from source
m
Background
0
61
137
293
PCE
^g/L
<1
1100
78.2
1.2
1.7
TCE
^g/L
<1
750
47
13
71
cis-DCE
^g/L
<1
800
200
50
48
Vinyl
Chloride
^g/L
<1
<1.9
<1
<1
<1
02
mg/L
5.3
<1
<1
<1
<1
Fe2+
mg/L
<0.1
0.4
0.06
<0.1
<0.1
pH
S.U.
-
6.8
7
7
-
-------
Sediment was acquired from the Building 102
site and Site A in an acetate sleeve within a
coring device that was pushed into the ground
water sediment below the water table. The
sediments were used to construct a laboratory
study of degradation and isotopic fractionation
of PCE, TCE, and cis-DCE in sediment contain-
ing magnetite. To preclude biodegradation of
the organics by bacteria that use the chlorinated
organic compounds as an electron acceptor, or
by iron reducing bacteria that can degrade cis-
DCE., the sediments were dried to a powder, then
autoclaved over night, and then used to construct
microcosms. Aseptic technique was used to
construct the microcosms. Microcosms were
constructed in January 2006.
In previous microcosm experiments, there was
one to two mL of water standing above the
sediment in each microcosm. These conditions
confound experiments intended to monitor
isotopic fractionation, because the chlorinated
organic compounds in the standing water in
the microcosm are not exposed to the reac-
tive minerals, and as a result, the chlorinated
organic compounds in the standing water are not
fractionated (personal communication Tomasz
Kuder, U. Oklahoma, see discussion in Hunkeler
et al., 2008).
To construct the microcosms, approximately
35 g of sediment was added to a 25-mL serum
bottle, leaving 3 mL of headspace. Then 1.0
mL of water containing the dose of PCE, TCE,
or cis-DCE was added. The microcosm was
sealed and vortexed to mix the dose into the pore
water of the microcosm. Then the sediment in
the microcosm was allowed to settle, the micro-
cosm was opened, and the standing water was
removed. Then the microcosm was sealed using
a flat Teflon-coated, grey, butyl rubber septum
and crimp cap. To sample the microcosm, the
septum was removed and 2.5 mL of RO water
was added to the microcosm. The microcosm
was resealed, vortexed to mix the water, and
then opened for sampling. One mL of standing
water was transferred to 14 mL of RO water in
a 20 mL serum bottle, preserved with two drops
of HC1, sealed with a septum and crimp cap,
and analyzed for volatile organic compounds as
described in Ferrey et al. (2004). An additional
1 mL aliquot of standing water was transferred
to 14 mL of RO water in a 20-mL serum bottle,
preserved with two drops of HC1, sealed with a
septum and crimp cap, and analyzed for the ratio
of stable carbon isotopes by Tomasz Kuder in
Paul Philp's lab at the University of Oklahoma.
Container controls were prepared by adding
the dose solutions to sterile RO water in 25-mL
serum bottles, and sealing the bottles with the
same Teflon faced septa and crimp caps.
Figures B.3, B.4, and B.5 present the removal of
PCE, TCE and cis-DCE respectively in sediment
from the Building 102 site and Site A. Although
PCE and TCE were removed in microcosms
constructed with sediment from the Building 102
site, the removal of PCE and TCE in the sedi-
ment from Site A was significantly slower than
the removal in the container controls. The rates
of removal of PCE and TCE in the Site A sedi-
ment were 0.21 ±0.15 per year and 0.29 ±0.10
per year respectively, while that in the container
control was 0.80 ± 0.21 per year.
The microcosms were incubated on their sides
with wet sand against the septum. The sand
would have reduced the surface area available at
the septum for diffusion of the organic com-
pounds from water into the septum. Because
pore water occupied only 25% to 30% of the
total porosity in the sediment, the surface area
available for diffusion may have been reduced
by a factor of three. In addition, convection
currents in the water controls may have allowed
fresh water to be presented to the Teflon faced
septum that sealed the controls, while mass
transfer limitations imposed during diffusion
through pore water in the sand in the micro-
cosms may have further reduced loss of TCE
from the microcosms. The sediment from Site A
on the TCAAP and the sediment from the nearby
Baytown Superfund site were similar in texture.
-------
The rate of removal of PCE and TCE in micro-
cosms constructed with sediment from Site A
was not used in the calculations which appeared
in Section 3, Tables 3.2 and 3.4. However,
the rates of removal of PCE and TCE in the
microcosms constructed with sediment from
Site A were probably better controls on losses
from the container than the formal container
controls. The rates of removal of PCE and TCE
in microcosms constructed with sediment from
the Building 102 site were used in calculations
in Tables 3.2 and 3.4 without correction for
removal in the container controls.
In contrast to the behavior of PCE and TCE, cis-
DCE was degraded in microcosms constructed
with sediment from both sites (Figure B.5).
1000 n
§100
LU
O
D.
10
TCAAP Site 102
container
controls
killed
microcosms
0 200 400 600 800 1000
Time of Incubation (days)
1000
LU
O
-100 -
10
1000
-100 -
- TCAAP Site 102-
container
controls
killed
microcosms
0 200 400 600 800 1000
Time of Incubation (days)
LLJ
O
0 200 400 600 800 1000
Time of Incubation (days)
Figure B.4. Removal of TCE in autoclaved sedi-
ment from the Building 102 Site and
Site A on the former TCAAP, north of
St. Paul, MN.
1000 -i
LU
O
Q_
10
TCAAP Site A
container
controls
killed
microcosms
0 200 400 600 800 1000
Time of Incubation (days)
Figure B.3 Removal of PCE in autoclaved sedi-
ment from the Building 102 Site and
Site A on the former TCAAP, north of
St. Paul, MN.
-------
1000 -i
TCAAP Site 102
O)
UJ100 -
Q
CO
c
-in -
" container
controls
° killed
microcosms
8 ;^
killed
0 200 400 600 800
Time of Incubation (days)
1000
1000 -i
O)
LU
O
Q
CO
O
100 -
10
- TCAAP Site A
killed
0 200 400 600 800 1000
Time of Incubation (days)
Figure B.5. Removal of cis-DCE in autoclaved sedi-
ment from the Building 102 Site and
Site A on the former TCAAP, north of
St. Paul, MN.
B.2 Baytown Superfund Site,
Minnesota
TCE releases from a former metal manufacturing
shop have contaminated water supply aquifers
in portions of Baytown and West Lakeland
Townships, and portions of the cities of Baytown
and Lake Elmo, in Minnesota. The source of
the TCE plume was discovered in 2004. At the
source, the water table was in sands and grav-
els in glacial drift overlying the water supply
aquifers in the Prairie du Chien dolomite and
underlying Jordan Sandstone. There was up to
50 mg/L TCE in ground water at the source. A
hollow stem auger rig and a split spoon sampler
were used to acquire sandy sediment from the
glacial drift, at a location down gradient from the
source of contamination near monitoring well
10B (see Figure B.6) from a depth interval 4.5 to
6.0 meters below the water table. The micro-
cosm study was constructed in January 2004.
The microcosms were constructed following
procedures described in Ferrey et al. (2004).
Figure B.6 depicts the distribution of TCE in the
Prairie du Chien aquifer, and the locations of the
source area, and monitoring wells down gradient
of the release. Table B.3 compares the concen-
trations of TCE, cis-DCE, and dissolved oxygen
along the flow path. Concentrations of oxygen
in the ground water were high. There was very
little evidence of reductive dechlorination of
TCE in the monitoring wells.
Source MW10B MW3B MW2B MW1B 14715
Figure B.6. Relationship between monitoring wells
in Table B.3 and TCE contamination in
the underlying Prairie Du Chien Aqui-
fer at Baytown Township, MN. TCE
concentration contours redrawn from
Figure 5a of Minnesota Department of
Health (2004).
-------
Table B.3. Distribution of chlorinated organics and
dissolved oxygen in the ground water in
the Baytown TCE plume.
Location
MW10B
MW3B
MW1B
MW2B
14715
2003
TCE
ug/L
100 to 130
80
29 to 38
20
4
2005
TCE
ug/L
224
90.7
0.75
3.29
cis-
DCE
ug/L
0.29
<0.2
<0.2
<0.2
0.2
Dissolved
Oxygen
mg/L
6.0
5.8
7.7
pH
S.U.
7.8
Figure B.7 compares removal of TCE in the
microcosm study conducted with the sediment
from Baytown. There was more removal of
TCE in the living microcosms than in the auto-
claved killed microcosms. To be conservative,
the rate of removal of TCE in the autoclaved
sediment was taken as the best estimate of
removal from abiotic processes. The first order
rate constants and the 95% confidence interval
on the rate constants are presented in Tables 3.2
and 3.4. There was a small increase in removal
of TCE in the killed microcosms over the water
control. As discussed previously for sediment
from Site A and Site 102 at the TCAAP, the
loss of TCE from the water controls prob-
ably overestimated loss from the microcosms.
The removal of TCE in the killed Baytown
microcosms was 0.98 + 0.49 per year at 95%
confidence. The sediment from Baytown and
Site A on the TCAAP were similar in texture.
The removal of TCE in the Site A sediment was
0.29 + 0.10 per year at 95% confidence. The
removal in the killed microcosms was used in
calculations in Tables 3.2 and 3.4 without correc-
tion for removal in the water controls.
1000
D)
LU
O
100-
10
Baytown
» live
n killed
1000 -i
LU
o
100 -
10
0 200 400 600 800
Time of Incubation (days)
H Baytown
container control
killed
0 200 400 600
Time of Incubation (days)
800
Figure B.7. Removal of TCE in sediment from the
Baytown Site, north of St. Paul, Minne-
sota.
B.3 Thermo-Chem Site, East of
Muskegon, Michigan
At the Thermo-Chem site, wastes from a solvent
recycling facility had been disposed to a lagoon,
resulting in ground water contamination from
a mixed NAPL containing petroleum hydro-
carbons and chlorinated solvents. The water
table aquifer is in glacial outwash sand. The
plume of contamination extended to the south
toward Black Creek (Figure B.8). The plume
discharged from the glacial outwash sand, and
entered the sediments in the flood plain of Black
Creek, and eventually ground water discharged
to Black Creek. Sediment was collected from
the bed sediments on the north side of Black
Creek. A Geoprobe rod was driven to a depth of
6 meters below land surface. The rod contained
a knockout point. The point was dislodged, and
-------
then the rod was pulled back a few centimeters.
Sediment and ground water was pumped from
the bottom of the rod with plastic tubing and a
peristaltic pump. The sediment was retained in a
jar as water was allowed to overflow until the jar
was filled with sediment.
The sediment was used to construct a micro-
cosm study following the procedures in Ferrey
et al. (2004). The microcosm study began in
December 2003.
Figure B.8 depicts the distribution of the plume
when the sediment sample was collected in
1998. By the time the plume reached the bank
of Black Creek, most of the PCE and TCE had
been reduced to cis-DCE. There was relatively
little vinyl chloride detected at the site. The
ground water was anoxic and iron-reducing
(Table B.4).
Source
1000 feet
300 meter
Northern Margin
Flood Plain
Sediment Sample for
Microcosm Study
Figure B.8. Location of sediment used for micro-
cosms in the plume of contamination
at Thermo-Chem site near Muskegon,
Michigan.
Table B.4. Distribution of chlorinated organic compounds and geochemical parameters in ground water on
the bank of Black Creek at the Thermo-Chem site in Michigan.
Date
April 1998
April 1998
April 1998
Oct 1998
Oct 1998
Oct 1998
Oct 1998
Depth below
land surface
m
3.5
4.4
6.2
1.7
2.6
3.5
4.4
PCE
^g/L
150
6.6
6
17.6
43.7
9.4
3.7
TCE
Hg/L
30.4
9
4
0
9.4
11.9
5.6
c/s-DCE
Hg/L
1100
6920
8580
1.3
31.4
1340
7500
Vinyl
Chloride
Hg/L
350
94.9
17
3.1
65.8
235
135
02
mg/L
0.2
0.2
0
0.5
0.32
0.3
0.17
Fe2+
mg/L
1.6
1.9
1.65
1.0
0.8
2.7
6.75
pH
S.U.
6.8
7
7.4
7.1
-------
Removal of cis-DCE in sediment from Black
Creek in Michigan was similar to removal in
sediment from Minnesota. The rate of removal
was the same in both living and autoclaved sedi-
ments. No vinyl chloride was detected in the
microcosms (data not shown).
1200-1
1000-
800-
o
400 -
200-
0
Thermo Chem site
killed
* container controls
° killed microcosms
<- living microcosms
0 100 200 300 400 500
Time of Incubation (days)
Figure B.9. Removal of cis-DCE in sediment from
the Thermo-Chem site, east of Muske-
gon, Michigan.
B.4 Products of Degradation of cis-
DCE in Sediment from TCAAP
Minnesota Site
In an attempt to identify the products of abi-
otic degradation of cis-DCE in the autoclaved
sediment described in Ferrey et al. (2004), the
experiment was repeated using freshly collected
sediment from near monitoring well 03U094
on the TCAAP. The sediment was collected
from an interval approximately 1 m below the
water table. The sediment corresponded to the
"shallow" sediment in Ferrey et al. (2004). One
treatment repeated the procedure in Ferrey et al.
(2004). The other treatment added uC-cis-DCE
to autoclaved sediment and sterile container
controls. Figure B. 10 presents the time course
of removal of cis-DCE in the microcosm experi-
ment with unlabelled cis-DCE. The microcosm
experiment began in December 2003.
After one day and after 375 days of incuba-
tion, the contents of the microcosms containing
14C- cis-DCE were stirred with a vortex mixer,
and three microcosms and three container
controls were opened and the pore water was
sampled and distributed for analysis. Two ml
of the standing water was preserved with HC1
and shipped to Paul Bradley with the U.S.G.S.
District Office at Columbia, South Carolina, for
analysis of radioactivity in volatile components
by gas chromatography and a proportional coun-
ter. One ml of the standing water was analyzed
for label remaining using a liquid scintillation
counter.
In the parallel study using microcosms that were
dosed with cis-DCE that was not radioactive
(see Figure B.10, Table 3.1), the first order rate
of disappearance that would be expected for 14C-
cis-DCE in the microcosms is 0.82 + 0.39 per
year at 95% confidence. At this rate of transfor-
mation, after 375 days of incubation, only about
43% of the original 14C label should remain as
14C- cis-DCE and about 57% would be trans-
formed to something else. At 95% confidence,
from 37% to 81% of the original label should be
transformed to something else.
Results are presented in Table B.5. In three
microcosms, from 14% to 34% of the total
volatile label was recovered as carbon dioxide.
The balance of the volatile label was 14C- cis-
DCE. No other labeled products were detected.
In two of three container controls, no carbon
dioxide was detected, but in one control, 24% of
the total label was recovered as carbon dioxide.
The production of carbon dioxide in this one
container control has not been explained.
-------
Table B.5. Distribution of 14C from 14C-c/s-DCE in
microcosms and container controls.
Microcosm
2-T5
7-T5
11-T5
Container
Control
(sterile water, no
sediment)
90W
100W
108W
Carbon
Dioxide
cis-DCE
Percent of total 14C
recovered as a vapor
14
34
24
0
24
0
86
66
76
100
76
100
The only labeled product identified in this study
was carbon dioxide, but the yield of 14C label in
carbon dioxide was less than would be expected
from the transformation of cis-DCE. It is
likely that other materials were produced in the
microcosms, similar to the materials described as
non-strippable residue in the microcosm experi-
ments of Darlington et al. (2008).
1200 -
1000-
^800 -
1
— 600 -
LU
Q 400-
co
o 200 -
n -
TCAAP deep T5 A container controls
o n ° killed microcosms
y
^4 n o living microcosms
killed o
*-*----,....
a
0 100 200 300 400 500
Time of Incubation (days)
Figure BAG. Removal of cis-DCE in autodaved sedi-
ment from the former TCAAP, north of
St. Paul, MN.
B.5 Removal of Vinyl Chloride in
Sediment from TCAAP Minnesota
Site
Figure B.ll presents data on the removal of
vinyl chloride in sediment from the TCAAP site.
The data was collected as part of the data set
for Ferrey et al. (2004), but was not included in
that manuscript because a container control was
not done for vinyl chloride. There was a trend
toward lower rates of removal in the container
controls with fewer chlorine atoms on the parent
molecule. The rate of removal of PCE in the
container controls was 0.80 + 0.21 per year, the
rate of removal of TCE was as much as 0.49 +
0.12 per year, and rate of removal of cis-DCE
was as much as 0.26 + 0.08 per year (Table 3.1
and 3.2). Based on this trend in the rate of
removal of PCE, TCE and cis-DCE, the rate of
removal of vinyl chloride should be less than
the rate of removal of cis-DCE. In any case, the
removal of vinyl chloride in container controls
should not be any faster than the removal of
cis-DCE.
Work presented in this appendix shows that the
removal of PCE from sediment microcosms can
be as much as fourfold lower in microcosms
filled with sediment compared to the container
controls, and the removal of TCE could be
twofold lower than removal in container controls
(Table 3.2). The true removal of vinyl chloride
through diffusion out of the microcosm is prob-
ably on the order of 0.13 per year or less. The
removal of vinyl chloride in sediment micro-
cosms was 0.311 + 0.12 per year. The values for
vinyl chloride in Tables 3.2 and 3.4 (Section 3)
were not corrected for removal in the container
controls.
Removal of Vinyl Chloride in deep
oxidized sediment from the TCAAP site
Killed
120 -i
„ 100 -
§ 80 -
g 60 -
o
0 40-
20 -
0
0 100 200 300 400 500 600 700
Time of Incubation (days)
Figure B.11 .Removal of vinyl chloride in sediment
from the former TCAAP, north of St.
Paul, MN.
o live
° killed
o
-------
Appendix C
Materials and Methods for Laboratory Studies of Abiotic
Transformation of TCE and Carbon Tetrachloride by
Green Rusts
Stoichiometric carbonate green rust and sulfate
green rust were synthesized by precipitation of
ferrous ions in sulfate-containing solution using
ferrous sulfate heptahydrate, sodium hydrox-
ide, sodium carbonate, and sodium bicarbonate
followed by oxidation using air bubbles (Su
and Wilkin, 2005). After synthesis, green rust
suspensions were stored inside an anaerobic
glovebox (3-6% H2 in N2) before use. The solid
concentrations of green rusts in suspensions
were determined gravimetrically after wash-
ing off dissolved salts with deionized water and
drying the solids inside the glovebox. Green
rust mineralogy was confirmed by X-ray dif-
fraction. Samples were mixed with glycerol to
minimize oxidation during handling and XRD
analysis. It was found that carbonate green
rust was not stable as it changed to magnetite
and Fe(OH)2 within 48 hours. Therefore, fresh
samples were used in the dechlorination tests.
The surface area of the preparation of green rust
was measured by the BET N2 gas adsorption
method. Approximately 7 mL of the suspen-
sion of sulfate green rust was filtered through
a 0.22 |j,m membrane and then rinsed with 10
mL of deionized and N2-purged water in the
anaerobic glovebox (humidity = 35 to 60%).
The sample was then ground to desired fineness
by rolling the sample with a Teflon roller for
7 minutes. Samples were then transferred to a
Coulter 3100 surface area analyzer in a portable
anaerobic chamber to maintain anaerobic condi-
tions. The samples were then out-gassed at 80°C
for 15 min before N2 adsorption took place.
Exposure time of the sample to the atmosphere
was less than 45 seconds during the procedure.
The measured surface area was 28.4 m2 g"1.
In the microcosm study involving TCE, appro-
priate amounts of suspensions of synthetic
carbonate green rust or sulfate green rust (0.25 g
equivalent dry mass) were added to 60-mL
glass serum bottles, then approximately 50 mL
of degassed and deionized water was added.
Appropriate amounts of a 1000 mg L"1 TCE
stock solution were added to create desired
initial TCE concentrations (2 to 10 mg L"1)
followed by addition of degassed and deion-
ized water to fill the bottles leaving no head-
space. The electrolyte concentrations in the
60-mL serum bottles were 0.037 M Na2SO4 +
0.030 M Na2CO3 (or NaHCO3) in the carbonate
green rust experiments and 0.074 M Na2SO4 in
the sulfate green rust experiments. The bottles
were sealed with Teflon-lined septa and alumi-
num caps, and then they were removed from
the glovebox and shaken on a shaker at 100
rpm in the dark at 23 + 1°C for up to 30 days.
To evaluate the effect of Fe(II) sorbed to the
green rust on TCE degradation, ferrous sul-
fate was added to some of the suspensions of
green rust two hours before TCE addition to
bring about a total dissolved Fe(II) concen-
tration of 35 mM. To evaluate the catalytic
effect of copper ions, appropriate amounts
of a 0.1 M CuSO4 stock solution were added
to make final Cu(II) concentrations at 0.5 to
10 mM to some of the suspensions of green
rust, with or without a Trizma buffer (pH
8) two hours before the addition of TCE.
Batch tests involving carbon tetrachloride (ini-
tial concentrations near 2 mg L"1) were per-
formed in 60-mL serum bottles with 0.25 g
of sulfate green rust at different pH levels
-------
(adjusted with H2SO4 and NaOH) in the pres-
ence or absence of added CuSO4 solution over
a wide range of time periods up to 15 days.
At pre-set time intervals, the bottles were centri-
fuged at 3600 rpm for 30 minutes and then they
were moved inside the anaerobic glovebox. The
clear supernatant solution was transferred using
a glass syringe into a 43-mL glass bottle and
capped without headspace for analysis of TCE
and carbon tetrachloride and their degradation
products (dichloroethylene isomers, vinyl chlo-
ride, chloroform, chloride, acetylene, ethylene,
and ethane). The pH and redox potential were
measured on the remaining solution. Dissolved
chlorinated hydrocarbons were determined
using gas chromatography-mass spectroscopy
and dissolved non-chlorinated hydrocarbons
by gas chromatography. Finally, 15 mL of
solution was filtered through a 0.22-|j,m mem-
brane and the filtrates were acidified with
HC1 for total dissolved elements (Fe, B, Si)
analyses using Inductively-Coupled Plasma-
Optical Emission Spectrometry (ICP-OES).
Data on ratioed concentrations were fitted by
linear regression to a pseudo first order rate
equation (Ln C/C0 = -kt), where C is the con-
centration at reaction time t (h), C0 is the initial
concentration, and k is the rate constant (h"1).
-------
Appendix D
Quality Assurance Documentation
D.I Analysis of Halogenated Organic
Compounds in Water Samples
Concentrations of Carbon Tetrachloride, TCE,
13C- labeled TCE, cw-DCE, 13C-labeled cis-
DCE, Vinyl Chloride and EDB were determined
by headspace gas chromatography/mass spec-
trometry (GC/MS) using a modification of EPA
Method 5021A, "Volatile Organic Compounds
in Various Sample Matrices using Equilibrium
Headspace Analysis," June 2003. Samples were
collected for analysis with an automated static
headspace sampler. Analytes were determined
by gas chromatography/mass spectrometry using
an Ion Trap Detector. The quantitation limit
for Carbon Tetrachloride, TCE, 13C- labeled
TCE, c/5-DCE, 13C-labeled cis-DCE and Vinyl
Chloride was 1.0 ng/L; the quantitation limit for
EDB was 5.0 [ig/L.
These are the quantitation limits in the samples
provided for analysis to the in-house analytical
contractor. In most cases, the samples were
acquired from small microcosms, and the
samples were diluted before they were delivered
to the analyst. Samples for Section 2 were
diluted one hundredfold, samples from Section
3 were diluted fifteenfold, and samples from
Section 4 were not diluted.
The quality assurance parameters for measure-
ments of data presented in Section 2 are com-
piled in Tables D.I and D.2. The symbols under
the column labeled "parameter" in the table are
as follows. The Continuing Calibration Check
(CCC) standards are determined at beginning of
each set of samples, after every 20 samples and
end of each sequence of samples. The purpose
of the CCC standards is to check for changes in
the sensitivity of the instrument over time during
the analytical run.
A method blank (MB) is determined at the
beginning and end of each set of samples. The
blank is acquired from the same source of water
used to prepare standards and dilute samples
to bring them into the range of the calibration
curve. The purpose is to determine if there is
any background contribution of the analyte of
interest from the analytical procedures.
The Matrix Spike (MS) is an analysis of a
specially prepared replicate of a sample that
has already been analyzed. It is spiked with a
known amount of the analyte, and then the total
concentration of analyte is determined. The
difference between the concentration determined
in the sample and the concentration determined
in the spiked sample is compared to the con-
centration that was spiked into the sample. The
purpose is to determine if there is a change in
the recovery of the analyte from the sample in
preparation or analysis of sample. The agree-
ment between the measured increase in concen-
tration with the spike and the expected increase
in concentration is an estimate of the accuracy of
the determination.
The Secondary Standards (SS) are determined
at the beginning of the sample set. They are
freshly prepared from some other source of
material than the source that was used to prepare
the standards used to calibrate the instrument,
and CCC standards. Their purpose is to provide
an independent check on the standards used to
calibrate the instrument.
Duplicates are replicate analyses of the same
sample. They are usually prepared by diluting
a sample, and then splitting and preparing the
sample for duplicate determination of concen-
trations. Results are expressed as the relative
percent difference (RPD), which is the differ-
ence between the two separate determinations,
-------
divided by the average of the two determina-
tions, expressed as a percentage. Agreement
between duplicates is an expression of analytical
precision of the determination.
The column labeled "No." in the tables is the
number of determinations performed for that
particular quality parameter. For duplicates, the
table reports the number of determinations with
results above the method detection limit out of
the total number of determinations performed.
Table D.1. Quality of Data on Concentrations of Halogenated Organic Compounds as Determined in Experi-
ments Described in Section 2.
Analyte
Parameter
Acceptable Value, Concentration
or Range of Values
Relative
Percent
Difference
(%)
Hg/L
%
Expected
Value
Max
Min
No.
Section 2, Figures 2.1 and 2.2.
13C TCE
13C TCE
13C TCE
13C TCE
13C TCE
TCE
TCE
TCE
TCE
TCE
13C c-DCE
13C c-DCE
13C c-DCE
13C c-DCE
13C c-DCE
c-DCE
c-DCE
c-DCE
c-DCE
c-DCE
ccc
MB
MS
ss
Duplicates
CCC
MB
MS
SS
Duplicates
CCC
MB
MS
SS
Duplicates
CCC
MB
MS
SS
Duplicates
<25
<25
<25
<25
<0.10
<0.10
<0.10
<0.10
120 to 80
130 to 70
120 to 80
120 to 80
130 to 70
120 to 80
120 to 80
130 to 70
120 to 80
120 to 80
130 to 70
120 to 80
104
<0.10
109
96
19.8
117
<0.10
112
113
25
Not
<0.10
Not
Not
-------
Table D.2. Quality of Data on Concentrations of Acetylene as Determined in Experiments Described in Sec-
tion 2.
Analyte
Parameter
Acceptable Value, Concentration
or Range of Values
Relative
Percent
Difference
(%)
ppm
(v/v)
%
Nominal Value
Max
Min
No.
Section 2 Figure 2.1
C2H2
C2H2
C2H2
C2H2
ccc
MB
ss
Duplicates
<20
<1.63
115 to 85
115 to 85
103
<1.63
101
13.8
91
94
0
11
10
5
8 of 15
All the quality parameters for concentrations of
halogenated organic compounds and acetylene
were within the acceptance limits, and data from
all of the samples in the experiments were used
in the figures and in calculations.
The quality assurance parameters for mea-
surements of data presented in Section 3 are
compiled in Table D.3. As indicated by the *
in the table, the Duplicates were out of range
for determination of PCE, and both the MS and
Duplicates were out of range for determination
of TCE in three sample sets. Data from these
sample sets were not used in the figures and in
calculations. The quality parameters for con-
centrations of halogenated organic compounds
in all of the other sample sets were within the
acceptance limits, and data from all of the other
sample sets in the experiments were used in the
figures and in calculations.
-------
Table D.3. Quality of Data on Concentrations of Halogenated Organic Compounds as Determined in Experi-
ments Described in Section 3 and Appendix B.
Analyte
Parameter
Acceptable Value, Concentration
or Range of Values
RPD
(%)
Hg/L
%
Nominal Value
Max
Min
No.
Appendix B, Site A and Site 102, Figures B.3, B.4, and B.5.
PCE
PCE
PCE
PCE
PCE
TCE
TCE
TCE
TCE
TCE
c-DCE
c-DCE
c-DCE
c-DCE
c-DCE
ccc
MB
MS
ss
Duplicates
CCC
MB
MS
SS
Duplicates
CCC
MB
MS
SS
Duplicates
<25
<25
<25
<0.10
<0.10
<0.10
120 to 80
130 to 70
120 to 80
120 to 80
130 to 70
120 to 80
120 to 80
130 to 70
120 to 80
119
<0.10
112
110
63*
107
<0.10
106
105
64*
110
<0.10
114
17
9192
82
91
5
93
57*
91
o
5
86
82
2.4
45
37
15
20
12 of 19
45
37
15
20
11 of 10
45
37
11
15
11 of 19
Appendix B, Baytown Site, Figures B.7 and B.8.
TCE
TCE
TCE
TCE
TCE
CCC
MB
MS
SS
Duplicates
<25
<0.10
120 to 80
130 to 70
120 to 80
117
<0.10
117
118
15
92
96
96
5.7
12
16
6
13
6 of 6
Appendix B, Thermo Chem and TCAAP5 Site, Figures B.9 and B.10.
c-DCE
c-DCE
c-DCE
c-DCE
c-DCE
CCC
MB
MS
SS
Duplicates
<25
<0.10
120 to 80
130 to 70
120 to 80
104
<0.10
109
108
21
97
96
91
0.4
7
11
5
12
5 of5
Appendix B, TCAAP deep, Figure B.ll
VC
vc
VC
vc
CCC
MB
MS
Duplicates
<25
<0.10
120 to 80
130 to 70
115
<0.10
113
-------
The quality assurance parameters for mea-
surements of data presented in Section 4
are compiled in Table D.4. All the quality
parameters for concentrations of halogenated
organic compounds were within the acceptance
limits, and data from all of the samples in the
experiments were used in the figures and in
calculations.
Table D.4. Quality of Data on Concentrations of Halogenated Organic Compounds as Determined in Experi-
ments Described in Section 4.
Analyte
Section 4
CC14
CC14
CC14
CC14
CC14
TCE
TCE
TCE
TCE
TCE
Parameter
ccc
MB
MS
ss
Duplicates
CCC
MB
MS
SS
Duplicates
Acceptable Value, Concentration
or Range of Values
RPD
(%)
<25
<25
Hg/L
<0.10
<0.10
%
Expected Value
120 to 80
130 to 70
120 to 80
120 to 80
130 to 70
120 to 80
Max
112
<0.10
111
113
-------
containing Ni-Pt kept at 940°C. An auxiliary O2
stream is mingled with carrier gas upstream of
the reactor to facilitate more efficient combus-
tion of the analytes. Combustion to CO2does not
affect chromatographic resolution.
The CO2 peaks are introduced into the IRMS
for determination of carbon isotope ratios. Raw
output of the IRMS consists of three simultane-
ously acquired signal channels, corresponding to
CO2 with variable C and O isotope substitution.
Integration of the individual channel outputs
over the peak's retention time window provides
13C/12C ratios of each GC peak.
Rather than measuring the absolute ratios of
isotope species, IRMS technique relies on data
normalization relative to internal standard of
known isotopic composition. A number of
pulses of standard gas (CO2) are introduced into
the IRMS's source during the run to provide
a reference 13C/12C signal to compare to the
signal derived from the analytical sample. This
configuration permits high-precision determina-
tion of 513C (the total analytical error of 513C
is + 0.5%o or better) in chlorinated ethenes at
concentrations as low 2 to 3 |j,g/l in a 25 ml
sample.
The analysis was conducted using the following
parameters.
The analytes were purged from a 25-ml water
sample. The purge gas was helium. The water
sample was purged at a flow rate of 40 ml/min
for 12 min. The sample temperature was 50 °C.
The trap was a Vocarb 3000. The temperature
program that was used is the manufacturer's
default for this type of sorbent. After the sample
was purged, dry helium was delivered to the trap
for 3 min. The PT trap was then desorbed for
5 min using helium at a flow rate of 6 to 8 ml
min.
The pre-column was a 0.5 |j,m film
DB-Carbowax column, 25m x 0.25mm. The
flow rate of helium carrier gas was 6 to
8 ml/min., isothermal at 40° C for 6 minutes
with LN2 cryofocusing. The final separation on
analytes was on a DB-MTBE, 60m x 0.32mm
column. The flow rate of helium carrier gas
was 1.8 ml/min. The GC column was held
isothermal at 40 °C for 9 min, and then ramped
at 6 °C/min. After elution of the last peak of
interest, the columns were cleaned by baking
them at 240 °C for 15 min.
The MAT 252 IRMS was operated at 8kV,
focused for "high linearity." Due to sample size
limitations, some of the data were obtained with
the IRMS focused for "high sensitivity." Tuning
for "high sensitivity" permitted analysis of
lower concentrations in samples but required a
more stringent QAQC procedure in the samples
to assure that their respective peak sizes were
matched. The concentrations of the spiked
compounds in the external standard runs were
very close to the target analyte concentrations.
Method bias associated with sample preparation
and analytical reproducibility were determined
daily by analysis of an external standard (water
spiked with DCE, TCE, and PCE of known
isotope composition) that was run through the
PT-GCIRMS cycle, using identical method
parameters as the analyzed samples. The exter-
nal standard is the equivalent of a CCC standard
as discussed above. At regular intervals, the
external samples are checked against a known
reference sample (Vienna PDB, provided by the
International Atomic Energy Agency in Vienna,
Austria). The V-PDB sample serves as the SS
standard.
The quality assurance parameters for measure-
ments of data presented in Section 3 are com-
piled in Table D.5. The acceptance value for
replicate analyses of the external standards was
a sample standard deviation of + 0.5%o or less.
All three sample sets were within the acceptance
level (Table D.5). The determined values in
the analytical samples were corrected for the
bias associated with sample preparation in their
particular sample set, before they were plotted in
figures or used in calculations.
-------
Duplicate analyses for 513C of cis-DCE were
performed on 4 out of 14 samples. Duplicates
differed from each other in a range of 0.2%o to
0.6%o. This was within the expected agreement
between duplicates of l%o. Duplicate analyses
for 513C of TCE were performed on 3 out of 14
samples. Two of the sets of duplicates (analyzed
in "high linearity" mode) differed from each
other at 0.2%o and 0.8%o. One of the set of
duplicates was analyzed in "high sensitivity"
mode. The agreement was 0.1%o. Two samples
depicted in Figure 3.7 were analyzed in "high
sensitivity" mode. The two samples with greater
uncertainty were identified in the figure.
Table D.5. Quality parameters for analysis of 813C in TCE and cis-DCE.
TCE
Set 1
10/10/2008 to 10/14/2008
Run#
4164
4165
4171
4172
4178
4179
4181
4182
average
stdev
reference 813C
bias correction
average
stdev
513C
%0
-30.0
-30.2
-30.6
-30.2
-30.6
-30.2
-30.1
-30.1
-30.3
0.2
-30.7
-0.4
-30.3
0.2
Set 2
10/14/2008 to 10/16/2008
Run#
4184
4190
4191
4197
4198
4202
4206
average
stdev
reference 813C
bias correction
average
stdev
513C
%0
-30.4
-31.3
-30.3
-31.2
-30.6
-30.8
-30.6
-30.7
0.4
-30.7
0.0
-30.7
0.4
cis-DCE
Set 1
10/8/2008 to 10/17/2008
Run#
4143
4144
4150
4153
4159
4164
4165
4171
4172
average
stdev
reference 813C
bias correction
average
stdev
513C
%0
-24.8
-24.7
-24.9
-24.8
-25.3
-25.2
-24.8
-24.8
-24.6
-25.0
0.3
-26.1
-1.1
-25.0
0.3
-------
D.3 Analysis of Total Iron in Sediment
Samples of the sediment used to construct the
microcosms were extracted in nitric acid by
microwave digestion using a CEM Corporation
Microwave Accelerated Reaction System, Model
MARS5. Samples were digested at approxi-
mately 175°C in 10% nitric acid. The extracts
were analyzed for the concentration of total
iron by Inductively Coupled Plasma-Atomic
Emission Spectroscopy using a Perkin Elmer
Optima 3000DV emission spectroscope. All the
quality parameters for concentrations of total
iron in nitric acid digests were within the accep-
tance limits, and data from all of the samples in
the experiments were used in the calculations.
Table D.6. Quality of Data on Concentrations of Total Iron in Sediment as Determined in Experiments De-
scribed in Section 3.
Analyte
Total Iron
Total Iron
Total Iron
Total Iron
Total Iron
Parameter
ccc
MB
MS
ss
Duplicates
Acceptable Value, Concentration
or Range of Values
RPD
(%)
<20
mg/L
<0.10
%
Expected Value
110 to 90
110 to 90
110 to 90
Max
102
<0.10
108
106
8.4
Min
99.5
90
101
0.1
No.
8
9
4
4
7
DA Analysis of Sodium, Potassium,
Calcium, Magnesium, and Iron in
Water Samples
Water samples were analyzed for the concentra-
tion of total iron by Inductively Coupled Plasma-
Atomic Emission Spectroscopy using a Perkin
Elmer Optima 3000DV emission spectroscope.
If concentrations of iron in samples collected
in November 2008 were below the detection
limit for Inductively Coupled Plasma-Atomic
Emission Spectroscopy, they were analyzed by
Inductively Coupled Plasma Mass Spectrometry
on a Thermo Elemental PQExcell Plasma Mass
Spectrometer.
The quality assurance parameters for measure-
ments of data presented in Section 1 are com-
piled in Table D.7. One of the method blanks
for sodium was slightly over twice the accept-
able value according to the relevant SOP. The
value in the blank was 0.2 mg/L and the values
returned from the field were all in excess of 300
mg/L. The presence of sodium in the blank at
0.2 mg/L has no effect on the data quality for
sodium.
The MS samples were routinely spiked to
increase the concentration of analyte by 2 mg/L.
The concentration of calcium and magnesium in
the matrix spikes were less than the concentra-
tion determined in the sample. The concentra-
tions in the samples were greater than 300
mg/L for calcium and greater than 100 mg/L
for magnesium. The spiked concentration was
too low to allow adequate resolution over the
ambient concentrations. All the other quality
parameters for concentrations of sodium, potas-
sium, calcium, magnesium, and iron in water
were within the acceptance limits, and data from
all of the samples in the experiments were used
in the calculations.
-------
Table D.7. Quality of Data on Concentrations of Sodium, Potassium, Calcium, Magnesium and Iron used in the
Geochemical Modeling in Section 1.
Analyte
Sodium
Sodium
Sodium
Sodium
Sodium
Potassium
Potassium
Potassium
Potassium
Potassium
Calcium
Calcium
Calcium
Calcium
Calcium
Magnesium
Magnesium
Magnesium
Magnesium
Magnesium
Parameter
ccc
MB
MS
ss
Duplicates
CCC
MB
MS
SS
Duplicates
CCC
MB
MS
SS
Duplicates
CCC
MB
MS
SS
Duplicates
Acceptable Value, Concentration
or Range of Values
RPD
(%)
<20
<20
<20
<20
mg/L
<0.091
<0.092
<0.001
<0.023
Expected
Value (%)
110 to 90
110 to 90
110 to 90
110 to 90
110 to 90
110 to 90
110 to 90
110 to 90
110 to 90
110 to 90
110 to 90
110 to 90
Inductively Coupled Plasma-Atomic Emission Spectroscopy
Iron
Iron
Iron
Iron
Iron
CCC
MB
MS
SS
Duplicates
<20
<0.008
110 to 90
110 to 90
110 to 90
Inductively Coupled Plasma Mass Spectrometry
Iron
Iron
Iron
Iron
Iron
CCC
MB
MS
SS
Duplicates
<20
<0.0003
110 to 90
110 to 90
110 to 90
Max
102
0.217*
Not
102
1.47
105
<0.092
Not
101
8.3
104
<0.001
NA
100
1.92
106
<0.023
NA
99.8
1.42
100
<0.008
106
103
1.39
108
<0.0003
100
100
5.17
Min
98.6
Done
98.8
0.95
100
Done
98.2
1.15
99
98.6
0.45
97.5
97.4
0
97.5
96.5
100
1.03
99.1
No.
4
10
5
5 of 5
4
10
5
5 of 5
8
10
1
5
5 of 5
10
1
5
3 of 5
10
10
4
5
2 of 5
3
3
1
1
1
One sample out of acceptable range.
-------
D.5 Analysis of Sulfate and Chloride in
Water Samples
Concentrations of sulfate and chloride were
determined by capillary ion electrophoresis
with indirect UV detection using a Quanta 4000
Capillary Ion Analyzer.
One of the SS samples was slightly out of range.
All the other quality parameters for concentra-
tions of chloride and sulfate in water were within
the acceptance limits, and data from all of the
samples in the experiments were used in the
calculations.
Table D.8. Quality of Data on Concentrations of Sulfate and Chloride used in the Geochemical Modeling in
Section 1.
Analyte
Sulfate
Sulfate
Sulfate
Sulfate
Sulfate
Chloride
Chloride
Chloride
Chloride
Chloride
Parameter
ccc
MB
MS
SS
Duplicates
CCC
MB
MS
SS
Duplicates
Acceptable Value, Concentration
or Range of Values
RPD
(%)
<20
<20
mg/L
<0.137
<0.113
%
Expected Value
110 to 90
120 to 80
110 to 90
110 to 90
120 to 80
110 to 90
Max
105
<0.137
101
98.4
2.08
105
<0.113
100
95.1
2.1
Min
90.8
84.7
92
0
92.6
87.8
88.6*
0
No.
18
10
7
6
13
18
10
7
6
13
'One sample out of acceptable range.
D.6 Analysis of S LI I fide in Water
Samples
Sulfide concentrations in water samples as
described in Table 1.2 and Figure 1.3 were mea-
sured with a spectrophotometer (Hach DR/2010)
at wavelength 665 nm using the methylene blue
method (Standard Method 4500 D, Clesceri et
al., 1999). Many of the samples exceeded 100
mg/L sulfide. If a sample was out of range, it
was diluted in oxygen free distilled water to
bring it into range.
D.7 Determination of Magnetic
Susceptibility
In each sample set, the nominal value of the
Magnetic Susceptibility of the calibration stan-
dard (2.66 x 10"6 m3 kg"1) was compared to the
actual reading. For five sets of samples, the rela-
tive percent difference (RPD) between the mea-
sured value of the calibration standard and the
nominal value of the calibration standard ranged
from 0.02% to 0.53%. Each sediment sample
was determined in duplicate, and sometimes in
triplicate. In the case of triplicate analyses, the
-------
RPD was calculated between the first and second
replicate, and the first and third replicate. For 24
duplicate determinations on sediment samples,
the RPD varied between 1.2% and 21.5%. The
median value of RPD was 6.4%.
The sediment cores were blended by hand
before they were used to construct microcosms,
and were sampled for determination of mass
magnetic susceptibility. The RPD included
error in the measurement of the mass magnetic
susceptibility as well as variation in the content
of magnetic materials from one 10 cm3 sample to
the next. It is likely that most of the variability
from sediment sample to sediment sample in the
material used to construct each of the treatments
of the microcosm studies reflected variation in
the concentration of magnetite in the sediment.
D.8 Scintillation Counting
The following applies to the experiments in
Section B.5. Liquid Scintillation Counting was
used to estimate the amount of 14C radio-label
that was associated with pore water in the micro-
cosms or container controls, and with sediment
solids in the microcosms.
The solution of 14C cis-DCE that was used to
spike the microcosms and container controls
and to prepare standards was prepared from neat
14C cis-DCE that was supplied by American
Radiolabeled Chemicals, Inc. The specific
activity was 2.0 mCi/mMole. The 14C cis-DCE
was provided without a carrier such as ethanol,
and without a stabilizer. The liquid contents of
the glass ampoule (approximately 6.5 |jL) were
taken up into a 10 uL syringe, and transferred to
70 ml of RO water in a glass serum bottle. The
bottle was sealed with a Teflon faced septum,
and stirred over night with a magnetic stir bar
to dissolve the 14C cis-DCE. This became the
stock solution used to dose the microcosms and
container controls, and to estimate the count-
ing efficiency of the scintillation counter. Each
microcosm or container control received 1.0 ml
of stock solution. The 14C in the stock solution
produced 3.49 x 106 dpm/ml.
The counting efficiency of the water samples
was determined by counting a 14C Standard
Spike of the stock solution. To prepare the 14C
Standard Spike, 100 |il of the cis-DCE dis-
solved in 900 |jl of autoclaved RO water was
transferred to 6 ml of Beckman Ready Safe
Scintillation Cocktail. Samples were calculated
against the efficiency of this known standard
by counting 1 ml of water from the microcosms
in 6 ml Beckman Ready Safe Scintillation
Cocktail. Program A2 was used on the LKB
Liquid Scintillation Counter with a count time
of 60 minutes. The Error Range was 0.744% to
1.33%. The sample counts used in calculations
were corrected for counting efficiency.
Soil samples were counted as follows: 0.2 g soil
was suspended in 8 ml Beckman Ready Value
Scintillation Cocktail and 7 ml autoclaved RO
water in a 20 ml scintillation vial (the vial was
vigorously shaken to suspend the soil particles
for scintillation counting). Program 3 on the
Beckman LS5000TD Liquid Scintillation
Counter was used for counting. The count time
was 60 minutes. The counting efficiency for soil
gel samples was 70% to 81%. Sample counts
used in calculations were corrected for counting
efficiency.
-------
-------
United States
Environmental Protection
Agency
PRESORTED STANDARD
POSTAGES FEES PAID
EPA
PERMIT NO. G-35
Office of Research and Development (8101R)
Washington, DC 20460
Official Business
Penalty for Private Use
$300
-------
Q.
CD
O
tt
I-H
o'
Q.
O
o
I-H
CD
I'
rt-
o'
(D
l-H
3^
O
Q.
V>
(D
&)
O
(D
(D
CA
T3
O
C£
o;
(D
C
CD
C
o'
a
o
3^
O
Q.
O
-------
------- |