ISSUE PAPER ON THE
ENVIRONMENTAL CHEMISTRY
OF METALS
Donald Langmuir,1 Paul Chrostowski,2
Bernard Vigneault,3 and Rufus Chaney4
Submitted to:
U.S. Environmental Protection Agency
Risk Assessment Forum
1200 Pennsylvania Avenue, NW
Washington, DC 20460
Contract #68-C-02-060
Submitted by:
ERG
110 Hartwell Avenue
Lexington, MA 02421
August 19, 2004
(corrected reference on January 25, 2005)
^ydrochem Systems Corp./Colorado School of Mines, Golden, CO
2CPF Associates, Inc., Takoma Park, MD
3CANMET Mining & Mineral Sciences Laboratories/
Natural Resources Canada, Ottawa, ON
4U.S. Department of Agriculture, Beltsville, MD
-------
TABLE OF CONTENTS
1. INTRODUCTION 1
2. METALS OF CONCERN 4
3. NATURAL OCCURRENCE OF METALS IN THE ENVIRONMENT 4
3.1 Baseline and Background Metal Concentrations 4
3.2 Mean and Median Metal Concentrations 5
3.3 General Sources of Metals Data and Maps 12
3.4 National and Regional Metals Problem Areas 14
3.5 Local Sources of Metals Data 15
4. METAL SPECIATION: METAL COMPLEXES 15
4.1 Metal Complexes: Concepts and Importance 17
4.2 Hard and Soft Acids and Bases: The Stability of Complexes and Metal Toxicity .. 18
4.3 Predominant Inorganic Species in Fresh Waters 21
4.4 Effect of pH on Metal Complexing 22
5. ADSORPTION CONTROLS ON METAL CONCENTRATIONS AND MOBILITIES ... 26
5.1 Introduction 26
5.2 Metal Adsorption and Desorption 27
5.3 Adsorption Behavior of the Metals of Concern 28
5.4 Surface Complexation Adsorption Models 31
5.5 Applications of the Diffuse Layer Model to Natural Systems 32
5.6 WHAM and Related Models for Predicting Metal Activities in Soil Pore Waters .. 35
6. SOLUBILITY CONTROLS ON METAL CONCENTRATIONS AND MOBILITIES .... 35
6.1 The Importance of pH 35
6.2 Oxidation Potential and pH 35
6.2.1 Iron and Sulfur 41
6.2.2 Aluminum, Beryllium, Strontium, and Barium 46
6.2.3 Cadmium, Zinc, Cobalt, Nickel, and Lead 49
6.2.4 Manganese 49
6.2.5 Silver, Copper, and Mercury 49
6.2.6 Thallium 60
6.2.7 Arsenic, Antimony, Chromium, Molybdenum, Selenium, and Vanadium
60
7. SOIL TRACE METAL TRANSFER TO PLANTS AND THE BIOAVAILABILITY OF
SOIL-METALS 70
8. LINKING METAL SPECIATION AND METAL TOXICITY: THE BIOTIC LIGAND
MODEL APPROACH 74
ii
-------
8.1 Overview of Biotic Ligand Model Development 74
8.2 Metal Speciation Estimation 74
8.3 Validation Studies 75
8.4 Current Limitations and Future Research 77
9. ATMOSPHERIC CHEMISTRY OF THE METALS 77
9.1 Discussion and Recommendations About Atmospheric Metal Chemistry 79
10. DETERMINATION OF METAL SPECIATION IN WATERS AND SOILS 80
10.1 Unfiltered Versus Filtered Water Samples 80
10.2 Unstable Parameters or Species Including Metal Redox Pairs 81
10.3 Determination of Individual Species Concentrations in Water 82
11. SOILS 82
11.1 Sampling of Soil Pore Waters 82
11.2 Sequential Extractions To Determine the Nature of Metal/Soil Associations .... 83
12. APPLICATION IN A REGULATORY CONTEXT 83
13. APPLICATIONS TO SPECIATION 84
13.1 Regulatory Defaults 85
13.2 Literature/Limited Measurement 85
13.3 Analytical Chemistry 86
13.4 Geochemical Modeling 87
13.5 Chemical-Specific Modeling 87
14. APPLICATIONS TO MOBILITY 88
15. RESEARCH RECOMMENDATIONS 92
16. LITERATURE CITED 92
iii
-------
LIST OF TABLES
Table 1. Median or Mean Concentrations of Some Metals and Related Elements in Natural
Waters, Soils, and the Earth's Crust 7
Table 2. Mean Baseline Metal Concentrations in Soils (|ig/g-dry) by State (from U.S. EPA,
2002d) 8
Table 3. Sample Population Distribution Parameter Estimates for Trace Elements Dissolved in
Ground Water 9
Table 4. Fe, Al, and Major Constituents (>1 mg/L) or Parameters in Some Surface and Ground
Waters 10
Table 5. Metals in an Acid Cambisol (10% Clay) Formed on a Mixed Shale-Gneiss Moraine in
Southern Sweden: Metal Concentrations in Dried Soil and in Soil Solution (pH = 5.2) 11
Table 6. U.S. Environmental Protection Agency (2002b) Drinking Water Standards for Select
Inorganic Species in Community Water Supplies 12
Table 7. Dominant Chemical Species of Metals in Soils and Natural Waters, Not Considering
Most (Especially Weak) Metal Complexing 19
Table 8. Hard and Soft Acids (Cations) and Bases (Ligands) 20
Table 9. Major Inorganic Species in a Hypothetical Natural Water 25
Table 10. Surface Areas, Surface Site Densities, and Cation Exchange Capacities (CECs) of
Some Important Sorbent Phases and the pH Dependence of Metal Sorption 30
Table 11. Comparison of Model Predictions and Measured Values of Percent Metals Associated
with the Suspended Particulate Fraction of Mine-Drainage Waters from Selected Sites 34
Table 12. Oxidation States of Trace Metals as They Occur in Natural Waters and Mineral
Systems, Their Redox Sensitivity, and Their Tendency to Form Sulfides at Low Eh 40
Table 13. Maximum Tolerable Levels of Dietary Minerals for Domestic Livestock in
Comparison with Levels in Forages 71
Table 14. Atmospheric Metals 78
Table 15. Partition Coefficients as a Function of pH for Several Important Elements of Potential
Concern (U.S. EPA, 1998) 90
iv
-------
LIST OF FIGURES
Figure 1. Cumulative percentages showing the frequency distribution of various constituents in
potable (chiefly surface) waters 6
Figure 2. Cumulative percent plot of cadmium concentrations in Colorado Front Range ground
waters 17
Figure 3. Speciation of Pb(II) (10~9 M) and Zn(II)(10"8 M) under freshwater conditions 23
Figure 4. Titrations of fulvic acid (FA) and copper 24
Figure 5. Adsorption of Pb from DuPage landfill leachate by kaolinite at 25°C, as a function of
pH 27
Figure 6. Adsorption of various metal cations and oxyanions, each at 5 x 10"7 M, by ferrihydrite
(£Fe[III] = 10"3 M) as a function of pH at an ionic strength of 0.1 mol/kg 29
Figure 7. Percent metal adsorbed: (a) by Al(OH)3(am) at total metal concentrations of 1.25 x 10"4
M in 1 M NaN03 solutions, and (b) by humic acid at total metal concentrations of 5 x 10"4 M
31
Figure 8. Comparison of experimental data (symbols) for Pb, Cu, Zn, Ni, and Cd sorption onto
streambed sediment from St. Kevin Gulch, Colorado, with computer-model simulations (curves)
for sorption onto hydrous ferric oxide, at a streambed concentration of 2.9 g/L 33
Figure 9. Approximate positions of some natural environments in terms of Eh and pH 37
Figure 10. Eh-pH diagram for the system C-02-H2 at a total carbonate concentration of 10"3 M
38
Figure 11. Sequence of microbially mediated oxidation-reduction reactions 39
Figure 12. Solubility of amorphous Fe(OH)3, pKsp = 37.1 (top curve), and goethite [a-FeOOH],
pKsp = 44.2 (bottom curve) as a function of pH at 25° C 42
Figure 13. Eh-pH diagram for the system Fe-02-C02-H20, assuming that total dissolved
carbonate equals 10"3 mol/kg and total dissolved iron is 10"3 mol/kg at aqueous solid boundaries
43
Figure 14. Eh-pH diagram for the system S-02-H20 at 25° C, showing the fields of predominance
of the aqueous species and of elemental sulfur (S°) for SS(aq) = 10"3 mol/kg at aqueous/S°
boundaries 44
v
-------
Figure 15. Eh-pH diagram for the system Fe-02-S-H20 at 25° C, showing stability fields of
goethite (a-FeOOH), pyrite (FeS2), and monoclinic pyrrhotite (Fe7S8) for SS(aq) = 10"2 mol/kg,
and total carbonate 10"4 mol/kg 45
Figure 16. Solubility of (a) amorphous Al(OH)3, Ksp = 10"312 and (b) gibbsite (Al[OH]3), Ksp =
10"33 9, as a function of pH at 25° C 47
Figure 17. Solubility of three forms of Be(OH)2 as a function of pH 48
Figure 18. Eh-pH diagram for the system Cd-02-C02-S-H20 for SCd = 10"7 mol/kg, ES = 10"5
mol/kg, and total carbonate 10"2 and 10"3 mol/kg 50
Figure 19. Eh-pH diagram for the system Zn-02-C02-S-H20, assuming that SZn = 10"6 and 10"4
mol/kg, EC = 10"3 mol/kg, and 2S = 10"3 mol/kg 51
Figure 20. Eh-pH diagram for the system Co-02-C02-S-H20, assuming that SCo = 10"6 mol/kg,
EC = 10"3 mol/kg, and 2S = 10"3 mol/kg 52
Figure 21. Eh-pH diagram for the system Ni-02-C02-S-H20, assuming that SNi = 10"6 mol/kg
and 2S = 10"5 mol/kg 53
Figure 22. Eh-pH diagram for the system Pb-02-C02-S-H20, assuming that SPb = 10"6 and 10"6 5
mol/kg, SC = 10"4 mol/kg, and SS = 10"5 mol/kg 54
Figure 23. Eh-pH diagram for the system Pb-02-P04-S-H20, assuming that SPb = 10~8,
10"7, and 10"6 mol/kg at solid/liquid boundaries; £S = 10"5 mol/kg; and SP04 = 10"6 mol/kg ... 55
Figure 24. Eh-pH diagram for the system Mn-02-C02-S-H20, assuming that SMn = 10"6 mol/kg,
SC = 10"3 mol/kg, and SS = 10"3 mol/kg 56
Figure 25. Eh-pH diagram for the system Ag-02-Cl-S-H20, assuming that SAg = 10"8 mol/kg,
SCI = 10 3 5 mol/kg, and 2S = 10"3 mol/kg 57
Figure 26. Eh-pH diagram for the system Cu-02-S-H20, assuming that SCu = 10"6 mol/kg and
SS = 10"2 mol/kg 58
Figure 27. Eh-pH diagram for the system Hg-02-Cl-S-H20, assuming that SHg = 10"6 9 mol/kg,
SCI = 10"3 mol/kg, and SS = 10"3 mol/kg 59
Figure 28. Eh-pH diagram for the system Tl-02-S-H20, assuming that ST1 = 10"8 and 10"6 mol/kg
and 2S = 10"3 mol/kg 61
Figure 29. Eh-pH diagram for the system As-02-H20 showing the relative stabilities of arsenate
vi
-------
(arsenic acid, As[V]) and arsenite (arsenious acid, As [III]) species under oxidizing and reducing
conditions 62
Figure 30. Eh-pH diagram for the system As-02-S-H20, assuming that 2 As = 10"6 mol/kg and £S
= 10"3 mol/kg 63
Figure 31. Eh-pH diagram for the system Sb-02-S-H20, assuming that SSb = 10" mol/kg and £S
= 10"2 and 10"5 mol/kg 64
Figure 32. Eh-pH diagram for the system Cr-02-H20, assuming that the concentration of SCr =
10"6 mol/kg at solid/liquid boundaries 66
Figure 33. Eh-pH diagram for the system Mo-02-S-H20, assuming that SMo = 10"6 mol/kg and
SS = 10"2 and 10"5 mol/kg 67
Figure 34. Eh-pH diagram for the system Se-02-H20 68
Figure 35. Eh-pH diagram for the system V-02-H20, for dissolved EV<10 mol/kg (11.5 mg/L
as V04) 69
Figure 36. Effect of pH on the BLM-predicted (HydroQual, 2002) Cu2+ and Ag+ accumulation in
the gills of Pimephales promelas 76
vu
-------
ENVIRONMENTAL CHEMISTRY
1. INTRODUCTION
Metals in the environment are of concern to almost all EPA programs. The Safe Drinking
Water Act mandates maximum contaminant levels and maximum contaminant level goals for a
variety of metals; the Clean Water Act mandates the development of ambient water quality
criteria for the protection of human health and aquatic life against the potential toxic effects of
metals; CERCLA requires that hazardous metals released into the environment be remediated to
levels which do not pose an unacceptable risk to human health or the environment; and RCRA
regulates the management of waste containing metals and metal compounds. To varying degrees,
each of these programs relies on scientific information regarding the metals. For example,
toxicity, bioavailability, bioaccumulation, and environmental fate and transport are all
significant characteristics of metals that should be considered by EPA risk managers.
Over time, EPA has developed a paradigm for the assessment of the behavior and effects
of organic compounds in the environment. Key elements of this framework are the fundamental
principle of lipophilicity as a driving force for many environmental and exposure events, the
significance of Henry's law for intermedia transfers, the degradability and potential for chemical
transformation of a hazardous material, and the fact that most organics in the environment
typically exist in one form.1 With few exceptions (mainly for strongly covalent inorganic
compounds or organo-metallics), metals do not exhibit these behaviors. In particular, although
metals can undergo a variety of intermedia transfers and chemical reactions, they are indefinitely
persistent and conservative in the environment. Possibly most important, the speciation, or
chemical form, of a metal has a profound impact on its environmental behavior and effects.
The term "speciation" as used by many environmental chemists is poorly defined.
Forstner (1995) refers to speciation as those aspects of metal chemistry including precipitation-
dissolution, adsorption-desorption, and complex formation in relation to pH; redox conditions;
and the content of soluble chelating agents. Bodek et al. (1988) identify a group of properties or
processes that are important for the environmental behavior of metals, including volatilization,
photolysis, sorption, atmospheric deposition, acid/base equilibria, polymerization, complexation,
electron-transfer reaction, solubility and precipitation equilibria, microbial transformation, and
diffusivity. For regulatory purposes, the recommended IUPAC definition might be more useful
(Templeton et al., 2000). In that definition, "speciation" refers to the distribution of an element
among defined chemical species, which are the specific forms of an element defined as to
isotopic composition, electronic or oxidation state, and/or complex or molecular structure.
The oxidation state of a metal is often a significant feature of its speciation due to its
impact on other processes. The well-known example of the high mobility of Cr(VI) compared to
1 There are a few exceptions to this rule. Organic acids and bases can exist in ionized or neutral forms, and some
organics exist as structural isomers (e.g., the hexachlorocyclohexanes).
1
-------
Cr(III) demonstrates the significance of oxidation state. Toxicity depends both on the oxidation
state and form of a metal (as cation or anion) and its tendency to form complexes with ligands.
For example, the toxicity of As(III) to aquatic life is significantly different from the toxicity of
As(V). Cr(VI) is considered a known human inhalation carcinogen, whereas Cr(III) is generally
considered to have low human toxicity. Cupric ion is more toxic to fish than is the cupric
carbonate complex. Dimethyl mercury is generally thought to be more toxic and to have a
greater bioaccumulation potential than mercuric chloride.
Mobility is affected by a variety of factors. The ability of a metal to sorb to a substrate is
usually the determining factor in its mobility. Physical adsorption—which is important for
molecular organic compounds—is largely inapplicable to the sorption of the trace metals, which
are usually adsorbed as ionic species. Ion exchange, too, rarely applies, in part because of the
relatively low concentrations of trace metals compared to those of major ions. In fact toxic metal
adsorption is often relatively independent of the concentrations of the major ions (Langmuir,
1997a). Metal sorption is usually strongly pH-dependent and a function of metal complex
formation and ionic strength. The most accurate and mechanistic approach to modeling and
predicting metal adsorption is surface complexation modeling, which ideally can take into
account all of these variables (cf. Stumm and Morgan, 1996; Langmuir, 1997a). Factors that are
important to metal sorption in soils, aquifers, and sediments include:
Soil Solids
Soil mineral composition
Specific surface areas
of metal sorbing solids
Surface site density or
cation exchange capacity
of metal sorbing solids
Aeration status
Microbial type, activity, and
population
Organic matter content
and character
Temperature
Soil Solution
pH
Eh
Dissolved oxygen
Solute composition
and concentrations
(activities)
Dissolved organic carbon
Ionic strength
Temperature
Solutes
Chemical identity
Complexation chemistiy
Solubility
Precipitation chemistiy
Redox behavior
Vapor pressure
The degree of mobility of organic contaminants is often expressed by means of a single
partition coefficient that describes the extent of equilibrium between sorbed and dissolved forms
of a compound (U.S. EPA, 1995). Mobility is then calculated from the partition coefficient. Such
an approach is only applicable to metal adsorption when the conditions listed above are
practically constant, which is rarely the case. In fact when metal adsorption is described using
partition coefficients, the value of such coefficients typically needs to be varied by two or more
orders of magnitude to reproduce metal adsorption behavior.
The bioavailable fraction of metals includes metal species that are environmentally
available and are absorbed or adsorbed by an organism (see the issue paper on bioavailability
2
-------
and bioaccumulation of metals for a detailed definition of metal bioavailability).
"Bioavailability" refers to a specific route of exposure—oral, inhalation, or dermal in the case of
a human; oral, gill, or dermal in the case of a fish. Oral bioavailability varies with oxidation
state, chemical form, and mineralogy. Inhalation bioavailability is strongly dependent on
solubility and particle size. Uptake of metals by biota or bioaccumulation is important for
determining concentrations and species of metals to which organisms can be exposed. EPA is
typically interested in the biouptake of metals by aquatic life (including aquatic vegetation),
terrestrial vegetation, terrestrial and avian wildlife, and farm animals. The concentration of a
chemical in an organism is usually calculated by application of a bioconcentration factor or a
transfer coefficient (U.S. EPA, 1998) that relates the concentration of a chemical in an organism
to the concentration in its food. Sorption, solubility, oxidation state, chemical form,
complexation, and competing chemical species are all important to the biouptake of metals.
In a few cases, metabolism may impact the speciation and subsequent fate and effects of
a chemical at a given location in a food chain. For example, fish are capable of metabolizing
carcinogenic inorganic arsenic to non-carcinogenic forms (Nriagu, 1994). Various
microorganisms are capable of creating organic forms of mercury (e.g., methyl-, ethyl-) from
inorganic mercury. The organic mercury forms are more toxic and better absorbed by animals
than the inorganic forms (U.S. EPA, 1997).
Some regulatory programs rely heavily on chemical characterization. In
SARA/CERCLA, the characterization of a material as a hazardous substance and the reportable
quantity depend on metal speciation. For example, reportable quantities for nickel range from
0.454 kg for nickel carbonyl, to 454 kg for nickel hydroxide, to 227 kg for nickel chloride. The
RCRA characteristic of toxicity is based on performance in a leaching test that is used to predict
the mobility of a metal under specified laboratory conditions.
Thus, at least in theory, EPA needs to have methods, either analytical or mathematical,
for evaluating and/or predicting dissolved vs. sorbed metals, oxidation states, solubility,
complexation, and chemical form for a group of elements with widely diverse physicochemical
and chemical properties.
EPA has identified three types of regulatory risk assessments where information
regarding speciation of metals is useful and desirable: national hazard/risk ranking and
characterization, site-specific assessments, and National Regulatory Assessments. National
hazard or risk ranking assessments are typically used by EPA for broad priority setting. An
example of this type of assessment is the Hazard Ranking System (HRS) used to determine the
eligibility of a contaminated site for inclusion on the National Priorities List. Site-specific
assessments are performed to determine if a site requires remediation and, if so, what the type
and extent of remediation should be. A Superfund risk assessment is an example of this type of
assessment. The third type, the National Regulatory Assessment, is typically used for standard
setting. Development of ambient water quality criteria is an example of this type of assessment.
These three examples will be used to illustrate the concepts of metal chemistry developed in this
paper.
3
-------
2. METALS OF CONCERN
Because of their abundance at contaminated sites and potential toxicity to plants or
animals, the elements of concern are aluminum, antimony, arsenic, barium, beryllium, cadmium,
chromium, cobalt, copper, lead, manganese, mercury, molybdenum, nickel, selenium, silver,
strontium, thallium, vanadium, and zinc. Phillips and Williams (1965) consider all of these
elements to be metals, with the exception of arsenic and selenium, which are nonmetals.
Webster's New World Dictionary (1970) defines a metal as an element that acts as a cation in
chemical reactions, forms a base with the hydroxyl radical, and can replace the hydrogen of an
acid to form a salt. Unlike the other elements listed, antimony, arsenic, molybdenum, selenium,
and vanadium generally occur as oxyanions in waters and soils, and not as cations. These
elements are sometimes described as metalloids. For simplicity, though, the following discussion
describes all the elements listed above as metals.
3. NATURAL OCCURRENCE OF METALS IN THE ENVIRONMENT
For reasons that may be unrelated to human activity, many surface and ground waters
contain natural concentrations of metals that exceed the EPA drinking water standards (cf.
Runnells et al., 1992). Soils can also contain naturally elevated levels of metals (cf. McBride,
1994; Salminen and Tarvainen, 1997). When a regulatory decision is made to restore affected
waters or soils to a presumed earlier state, it is obviously unrealistic to assign clean-up goals that
are below preexistent metal levels. It is critically important therefore, to attempt to distinguish
between metal amounts that were naturally present and amounts added as a result of human
activities.
3.1 Baseline and Background Metal Concentrations
It is useful to introduce two concepts at this point. The first is that of geochemical
baseline (cf. Salminen and Tarvainen, 1997; Salminen and Gregorauskiene, 2000), which may
also be described as ambient concentration. This is the regional metal concentration in a
medium, a concentration that has not been further increased by a local source of contamination.
Baseline values may have been elevated by regional contamination. For example, arsenic
concentrations in soils exceed 9 parts per million (ppm) in a 100-kilometer-wide, 1,000-
kilometer-long belt that extends southwest from New York state across Pennsylvania, Ohio, and
Kentucky (Gustavsson et al., 2001). There is evidence that these high arsenic values are chiefly
derived from the burning of coal (Smith, 2003). An assessment of arsenic contamination of soils
by a local source in this area would need to account for the high regional baseline values.
The second concept is that of background, which is the concentration of a metal in a
medium as it existed before being affected by human activity. Background concentrations are a
function of regional geology and local soil and sediment conditions. The metal background in
soils also depends on the depth of the sampling (i.e., location within the soil profile) and whether
the analysis is of the total soil or of a specific size fraction (cf. Salminen and Tarvainen, 1997).
Metal concentrations are generally higher in the fine-grained fraction of soils than in the coarse
4
-------
fraction (sand fractions and medium/coarse silt fraction). It is important to note that baseline and
background metal concentrations are not single values, but a distribution of values that can range
over orders of magnitude over distances of a few centimeters in porous media and meters in
surface waters (cf. Gustavsson et al., 2001). It is also important to recognize that background
metal concentrations can exceed environmental quality criteria at some sites.
3.2 Mean and Median Metal Concentrations
Though metal concentrations in the environment can range widely, it is still useful to
consider their mean and median values in waters and soils and the Earth's crust (Table 1).
Values for iron and sulfur are included in Table 1 for purposes of comparison and later
reference. Data on typical baseline levels of several metals in every state have recently been
compiled (Table 2). It is often assumed that species present in water at less than 1 milligram per
liter (mg/L) but more than 1 microgram per liter (|ig/L) are minor species, while those present at
below 1 |ig/L are termed trace species (Langmuir, 1997a). Based on these definitions, most of
the metals of concern are minor species and some are trace species, particularly in surface waters
which are in general more dilute than soil or ground waters.
Median groundwater concentrations of metals and major species that exceed 1 mg/L, as
per Table 3, do not differ much from the median metal values for ground waters in Table 1, or
the averages for major species in Table 4. However, given the wide variability of metal
concentrations in waters, soils, and rocks, average and median values can be deceiving. Figure 1
shows a plot of the cumulative percentages of some major, minor, and trace constituents in
mostly surface and some ground waters. The data for trace metals, which are most complete for
ground waters (see Table 3), show that their concentrations range over 2 to 6-7 orders of
magnitude in the case of As, Cd, Cu, and Zn.
100
80
I 60
I 40
8
20
0
0.01 0.1 1 10 100 1000
mg/L (ppm)
Figure 1. Cumulative percentages showing the frequency distribution of various
constituents in potable (chiefly surface) waters. From Davis and DeWiest, 1966.
5
-------
Although the data are limited for metals, their concentrations in soils are generally higher
than in surface or ground waters, presumably because of concentration in soils by
evapotranspiration and because of the relatively high solid to water ratio of soils. In recent years,
methods of chemical analysis such as ICP-ES and ICP-MS have been adapted for the
determination of metals in soils and soil moisture. These methods have vastly lowered metal
detection limits and improved the accuracy of analysis of trace metals at low concentrations.
Shown in Table 5 is a chemical analysis of the metals of concern and some major elements in a
Swedish Cambisol (Tyler and Olsson, 2001a). The soil has 8% organic matter, 10% clay, and a
pH of 5.2. Metal concentrations were determined either by ICP-ES or ICP-MS.
6
-------
Table 1. Median or Mean Concentrations of Some Metals and Related Elements in Natural Waters, Soils, and the Earth's
Crust
Ground Waters
(mean detects)
Ground Waters
(mean)
Soils
(mean)
Earth's Crust
(mean)
Metal
Symbol
(U.S. EPA, 2002a)
(Newcomb and Rimstidt,
2002)
(Shacklette
and Boerngen,
1984)
(Fortescue, 1992)
(Hg/L)
(Hg/L)
(mg/kg)
(mg/kg)
Aluminum
A1
72,000
83,600
Antimony
Sb
5
0.66
0.2
Arsenic
As
8
13.9
7.2
1.8
Barium
Ba
140
77
580
390
Beryllium
Be
2
0.92
2.0
Cadmium
Cd
1
0.35
0.16
Chromium
Cr
7
4.4
54
122
Cobalt
Co
4.3
9.1
29
Copper
Cu
70.6
25
68
Iron
Fe
26,000
62,200
Lead
Pb
2.6
19
13
Manganese
Mn
550
1,060
Mercury
Hg
2
0.09
0.086
Molybdenum
Mo
0.97
1.2
Nickel
Ni
11.5
19
99
Selenium
Se
5
2.5
0.39
0.050
Silver
Ag
0.5
0.05
0.08
Strontium
Sr
240
384
Sulfur
S
340
Thallium
T1
6
0.2b
0.72
Vanadium
V
12.5
80
136
Zinc
Zn
265
60
76
aU.S. EPA (2002a), occurring as sulfate.
bBowen, 1979.
7
-------
able 2. Mean Baseline Metal Concentrations in
Soils (|ig/g-dry) by State (from U.S. EPA
, 2003a)
State
Al
As
Ba
Be
Cd
Co
Cr
Cu
Mn
Ni
Pb
Se
V
Zn
Alabama
23100.0
4.7
200.0
0.6
4.4
30.6
9.6
420.0
11.0
9.3
0.3
38.0
26.0
Arkansas
33429.0
9.7
336.0
0.9
12.0
53.1
17.0
731.0
18.0
21.0
0.7
52.0
39.0
Arizona
32933.0
9.6
364.0
1.0
0.4
9.9
37.3
23.0
447.0
23.0
16.0
0.4
42.0
51.0
California
75633.0
5.1
598.0
1.1
0.4
14.0
119.9
39.0
640.0
48.0
26.0
0.2
118.0
113.0
Colorado
61557.0
6.7
662.0
1.4
6.8
41.7
21.0
343.0
13.0
31.0
0.4
74.0
87.0
Connecticut
85000.0
4.1
400.0
0.5
7.5
40.0
15.0
450.0
13.0
5.0
0.8
60.0
40.0
Deleware
22500.0
1.4
400.0
0.5
3.3
30.0
5.0
85.0
6.0
15.0
0.3
20.0
23.0
Florida
9944.0
3.0
48.0
0.6
0.1
1.6
15.4
5.6
86.0
8.5
12.0
0.3
11.0
12.0
Georgia
38250.0
5.0
232.0
0.6
6.9
32.4
21.0
252.0
17.0
19.0
0.4
43.0
47.0
Iowa
64667.0
7.3
617.0
1.3
11.0
64.7
31.0
603.0
26.0
19.0
0.4
97.0
57.0
Idaho
58500.0
6.4
757.0
1.1
12.0
52.1
28.0
580.0
22.0
22.0
0.3
90.0
83.0
Illinois
48714.0
7.1
551.0
0.7
9.8
48.4
24.0
646.0
19.0
39.0
0.5
62.0
67.0
Indiana
50000.0
7.5
500.0
0.7
10.0
46.8
27.0
518.0
18.0
18.0
0.4
74.0
56.0
Kansas
61818.0
6.8
694.0
1.0
8.9
49.0
25.0
452.0
17.0
32.0
0.4
77.0
67.0
Kentucky
54123.0
7.8
349.0
1.1
11.0
79.8
17.0
483.0
23.0
16.0
0.5
66.0
35.0
Louisiana
42188.0
7.6
441.0
0.6
8.6
60.8
33.0
470.0
33.0
16.0
0.7
76.0
55.0
Massachusetts
34083.0
8.6
203.0
1.3
0.2
7.8
39.5
16.0
439.0
13.0
13.0
1.9
87.0
54.0
Maryland
39167.0
3.8
393.0
1.3
7.5
47.9
20.0
291.0
13.0
22.0
0.2
63.0
39.0
Maine
65385.0
9.4
319.0
1.6
10.0
71.2
28.0
581.0
30.0
19.0
0.7
98.0
80.0
Michigan
10964.0
4.2
127.0
0.7
0.9
4.6
13.8
12.0
230.0
12.0
9.2
0.3
44.0
33.0
Minnesota
49457.0
5.5
571.0
0.7
0.3
7.2
25.4
20.0
583.0
14.0
9.9
0.3
72.0
38.0
Missouri
42094.0
10.0
499.0
1.0
12.0
50.0
19.0
940.0
20.0
23.0
0.5
72.0
53.0
Mississippi
45368.0
8.8
390.0
0.9
12.0
53.2
20.0
471.0
21.0
18.0
0.5
68.0
45.0
Montana
70938.0
8.8
739.0
1.1
7.5
63.3
29.0
366.0
20.0
14.0
0.4
101.0
69.0
Nebraska
59474.0
5.5
711.0
1.1
5.9
32.5
15.0
306.0
15.0
16.0
0.4
62.0
54.0
North Carolina
60105.0
4.8
356.0
0.6
15.0
64.8
34.0
563.0
24.0
17.0
0.4
107.0
56.0
North Dakota
62857.0
7.0
682.0
0.9
6.9
53.2
23.0
530.0
20.0
13.0
0.4
83.0
64.0
New Hampshire
66667.0
4.4
500.0
2.3
0.6
5.3
18.4
12.0
633.0
10.0
28.0
0.3
57.0
23.0
New Jersey
10075.0
7.0
54.0
0.3
0.3
1.7
13.9
14.0
221.0
3.8
35.0
0.9
30.0
22.0
New Mexico
54423.0
5.9
727.0
1.0
8.8
55.5
21.0
367.0
28.0
18.0
0.3
72.0
44.0
Nevada
66078.0
9.0
822.0
1.3
8.4
36.8
25.0
481.0
15.0
25.0
0.3
78.0
69.0
New York
58800.0
6.4
666.0
1.4
0.2
9.1
66.9
36.0
418.0
21.0
20.0
0.3
132.0
82.0
Ohio
54615.0
12.0
469.0
1.0
13.0
55.0
28.0
550.0
25.0
23.0
0.6
88.0
69.0
Oklahoma
39200.0
7.0
430.0
1.1
7.1
46.0
16.0
465.0
15.0
18.0
0.3
50.0
50.0
Oregon
94412.0
5.1
682.0
0.9
16.0
121.6
53.0
725.0
23.0
15.0
0.3
168.0
70.0
Pennyslvania
63438.0
13.0
366.0
1.4
15.0
52.8
37.0
609.0
24.0
23.0
0.5
80.0
81.0
Rhode Island
100000
3.5
500.0
0.5
10.0
50.0
15.0
500.0
15.0
15.0
0.9
70.0
30.0
South Carolina
39143.0
3.9
151.0
1.4
3.5
21.4
16.0
87.1
7.8
5.0
0.3
45.0
25.0
South Dakota
74333.0
8.5
1043.0
1.4
7.7
58.7
29.0
1013.0
28.0
16.0
0.5
108.0
75.0
Tennessee
31894.0
16.0
193.0
0.8
0.2
14.0
40.3
17.0
1112.0
18.0
23.0
0.6
49.0
57.0
Texas
41958.0
6.4
404.0
0.9
5.3
39.6
15.0
303.0
12.0
14.0
0.3
52.0
39.0
Utah
45638.0
8.0
493.0
0.9
6.6
45.6
26.0
371.0
13.0
35.0
0.3
70.0
96.0
Virginia
60438.0
5.1
436.0
0.9
9.7
54.3
33.0
441.0
17.0
36.0
0.4
77.0
233.0
Vermont
56667.0
3.6
333.0
1.7
12.0
66.7
18.0
800.0
25.0
20.0
0.4
70.0
43.0
Washington
66834.0
4.5
606.0
0.9
0.8
18.0
49.9
31.0
760.0
23.0
14.0
0.3
160.0
78.0
Wisconsin
48000.0
4.4
543.0
2.0
7.7
40.3
12.0
365.0
14.0
12.0
0.3
48.0
44.0
West Virginia
67000.0
8.6
360.0
1.0
14.0
46.0
22.0
770.0
23.0
17.0
0.5
65.0
60.0
Wyoming
56125.0
6.5
756.0
0.7
8.3
47.9
21.0
416.0
16.0
17.0
0.5
84.0
57.0
8
-------
Table 3. Sample Population Distribution Parameter Estimates for Trace Elements
Dissolved in Ground Water. Trace element sample populations are censored at multiple reporting
limits.
Element
No. of
Records
% Censored
Min.
Max.
Mean
Median
Antimony
1893
94
0.2
44
-
-
Arsenic
7199
53
0.03
42,600
13.9
1.4
Barium
9957
6
1.0
5,000
77
35
Beryllium
5550
94
0.03
10
-
-
Cadmium
7088
90
0.02
19,200
-
-
Chromium
9097
76
0.5
3,800
4.4
1.0
Copper
9582
70
0.1
74,000
70.6
1.1
Lead
9061
76
0.1
2,020
2.6
0.3
Mercury
2031
91
0.02
7.3
-
-
Nickel
8460
70
1.0
10,300
11.5
2.0
Selenium
4674
81
1.0
186
2.5
0.2
Silver
9528
87
0.04
139
0.5
0.2
Thallium
670
90
0.02
60
-
-
Vanadium
4498
80
0.4
19,000
12.5
1.1
Zinc
9464
40
0.09
633,000
265
5.7
This table is based on 104,280 dissolved-fraction ground water concentration records for samples obtained in
1996-1998 from the STORET database (U.S. EPA, 2000). Concentrations are in ng/L.
From Newcomb and Rimstidt, 2002. Robust data analysis techniques (Helsel, 1990; Helsel and Hirsch, 1992) were
used to determine central value estimates and variables in sample populations with <90% censored (below detection)
data. See discussion in Newcomb and Rimstidt, 2002.
9
-------
Table 4. Fe, Al, and Major Constituents (>1 mg/L) or Parameters in Some Surface and
Ground Waters
Constituent or
Parameter
Average River,
World"
Average River, N.
America"
Average Ground
Water, Worldb
Na+
6.3
9
30
K+
2.3
1.4
3
Mg2+
4.1
5
7
Ca2+
15
21
50
Fe
0.67
0.16
Al
0.07
cr
7.8
8
20
N03"
1
1
hco3
58.4
68
200
S042-
11.2
20
30
Si02(aq)
13.1
9
16
pH
7.4
TDS
120
142
350
aLivingstone, 1963.
"Turekian, 1977.
10
-------
Table 5. Metals in an Acid Cambisol (10% Clay) Formed on a Mixed Shale-Gneiss
Moraine in Southern Sweden: Metal Concentrations in Dried Soil and in Soil Solution (pH
= 5.2)
Dried Soil
(jig/kg)
Soil Solution
(jig/L)
Ag
0.88
0.1
A1
75,500
297
As
65.9
3.2
Ba
646
132
Be
3
0.025
Cd
0.72
0.36
Co
7.72
0.4
Cr
66
1.3
Cu
142
59
Hg
0.16
0.13
Mn
231
66
Mo
93.1
5.7
Ni
11.4
11
Pb
43.9
2.4
Sb
0.4
0.86
Dried Soil
(jig/kg)
Soil Solution
jig/L
mg/L
Se
1.74
1.7
Sr
73.6
77
T1
3.9
0.15
V
815
7.2
Zn
63
90
Ca
1,620
27,300
27
Fe
50,800
280
0.28
K
25,000
6,800
6.8
Mg
2,750
2,170
2.2
Na
170
4,550
4.6
P
527
119
0.12
S
4,938
23,900
23.9
c(org)
80,000
45
From Tyler and Olsson, 2001b.
11
-------
Table 6. U.S. Environmental Protection Agency (2002b) Drinking Water Standards for
Select Inorganic Species in Community Water Supplies
Contaminant
MCL(|ig/L)
Note
Contaminant
SMCL (jxg/L)
Note
Antimony
6
a
Aluminum
50 to 200
a
Arsenic
10
a
Iron
300
a
Barium
2,000
a
Manganese
50
a
Beryllium
4
a
Silver
100
a
Cadmium
5
a
Copper
1,000
a
Copper
1,300
b
Uranium
30
a
Chromium
100
a
Zinc
5,000
a
Lead
15
b
Mercury
2
a
Nickel
100
a
Selenium
50
a
Thallium
2
a
Maximum contaminant levels (MCLs) are enforceable standards for substances that may constitute a health hazard
at higher concentrations. Secondary maximum contaminant levels (SMCLs) are not enforceable, but are set for
aesthetic reasons, to avoid tastes, odors, and staining of plumbing fixtures.
aCurrent standard.
•"Treatment triggered at 1.3 mg/L Cu and 0.015 mg/L Pb.
It is instructive to compare EPA drinking water standards for inorganic species in water
supplies (Table 6) to the average concentrations of the same species in ground waters (Table 3).
The comparison shows that the median concentration of As is 14 percent of the standard. This
suggests that a large percentage of background waters will have natural concentrations of As that
exceed drinking water standards.
3.3 General Sources of Metals Data and Maps
The U.S. Environmental Protection Agency, the U.S. Geological Survey of the
Department of Interior, and the Department of Energy have all compiled extensive databases of
chemical analyses of various environmental media that are readily available on CD-ROM or for
downloading from the World Wide Web. EPA's STORET (STOrage and RETrieval) data
12
-------
system, which can accessed at http://www.epa.gov/storet/, contains 200 million water sample
observations from about 700,000 sampling sites for both surface and ground waters. Reported
data include stream flow information and measured concentrations for most of the metals of
concern. There is no requirement that the data in STORET be evaluated using quality assurance
procedures, however. For the data from STORET, or from the other sources listed below, it is
important that prospective users be aware that such data may or may not have been screened for
accuracy. Because of the potential for sample contamination and sampling/analysis artifact, trace
metal concentration data could be unreliable—in particular those data collected before the late
1980s, since the necessity for using "trace metal clean techniques" had not been recognized by
that time.
The U.S. Geological Survey Water Web site (http://waterdata.usgs.gov/nwis) provides
access to chemical and physical records for 1.5 million sampling sites in all 50 states, the District
of Columbia, and Puerto Rico as part of the National Water Quality Assessment Program
(NAWQA). A stated goal of NAWQA is "to describe the status and trends in the quality of a
large representative part of the Nation's surface and ground-water resources and to identify the
natural and human factors that affect their quality." Available through the Web site are chemical
analyses and physical data for streams, stream sediments, lakes, springs, and wells that include
their metals concentrations. A map of arsenic in 31,350 samples of ground water in the
contiguous United States based on Welch et al. (2000) is featured on the Web site. As part of the
NAWQA program, Rice (1999) reported on the concentrations of As, Cd, Cr, Cu, Pb, Hg, Ni, Se,
and Zn in 541 streambed samples obtained from across the conterminous United States. In the
analysis she takes into account background and baseline metal concentrations.
The USGS also maintains the National Geochemical Data Base (NGDB), which contains
more than 2 million data records for samples of stream sediment, soil, rocks, water, and
vegetation. This database is also accessible via the Web. The NGDB includes approximately
260,000 sediment and soil samples collected by the U.S. DOE in the 1970s and 1980s as part of
the National Uranium Resource Evaluation (NURE) program. Most of the soil, sediment, and
rock samples from which the NGDB was generated are held in archival storage and are available
for further study. Data in the NGDB are available on CD-ROM, and can be used to generate
metal concentration maps, so as to help determine metal baseline and background concentrations
and contamination levels. The Geological Survey also maintains the PLUTO.RASS database,
which lists geochemical data for about 500,000 samples—chiefly rocks, but also stream
sediments, soils, heavy-mineral concentrates, waters, and vegetation.
A number of national surveys of metals concentrations in soils have been published. The
most famous and still useful was that of Boerngen and Shacklette (1981) and Shacklette and
Boerngen (1984), which involved collecting soil and other regolith samples from 1,323 sites in
the conterminous United States. Samples were obtained at a depth of 20 centimeters from
untilled, naturally vegetated soils away from roads in the period 1961-1975. The results of
Shacklette and Boerngen have been reviewed and statistically assessed by Gustavsson et al.
(2001), who present colored national maps for Al, As, Ba, Cr, Cu, Fe, Hg, Mn, Ni, Pb, Se, Sr, V,
and Zn among other metals. Metals in soils results from the U.S. study of Shacklette and
13
-------
Boerngen (1984) are included in an assessment of metals and other elements in soils worldwide
in a report by Darnley et al. (1995).
Detailed soil survey maps are available from the offices of the USDA/ARS Natural
Resources Conservation Service (formerly the U.S. Soil Conservation Service). These generally
include areal maps that show soil types and classifications by county, along with the pH and
organic matter contents of typical soil types. The soil surveys also describe soil associations and
soil-forming processes for the soils within specific areas.
3.4 National and Regional Metals Problem Areas
The previous section cites a number of published and Web-accessible sources of metals
data for streams, ground waters, and soils. Of particular value to anyone identifying and
prioritizing metal problems regionally or nationally are maps on which metals data for soils,
sediments, and waters have been plotted and the metals concentrations contoured. Such maps
have been published for soils, although they are based on limited data (see Gustasvsson et al.,
2001), but are less available for surface and ground waters. Welch et al. (2000) have published a
map of arsenic in U.S. ground waters, and discuss geologic and climatic (e.g.,
evapotranspiration) controls on elevated arsenic values. David B. Smith of the U.S. Geological
Survey maintains a national database of metals data for stream sediments and has colored
regional maps of such data for Al, As, Ba, Be, Co, Cr, Cu, Mn, Hg, Ni, Pb, Se, Sr, V, and Zn for
a large part of the U.S. (See http://water.usgs.gov/osw/techniques/workshop/smith.html.)
Dissolved metals in lower order streams, which are often fed by ground water, can be
expected to correlate with local geology. The chemical compositions of river waters are the
results of rainfall, rock-weathering reactions, and evaporation-crystallization (cf. Langmuir,
1997a). The relative contribution of climate-related factors will be more important for major
rivers.
A comparison of maps for the metals with a geologic map of the United States often
shows a correlation between local or regional geologic formations and metal levels in soils and
ground waters. The correlations are most obvious in areas away from major urban and industrial
centers, which have contributed important metal amounts to the environment through their waste
disposal and land use practices. Colored national and state geological maps are available in
digital form on the Web at http://pubs.usgs.gov/dds/ddsll/ and
http://nationalatlas.gov/geologym.html.
Broad-scale soil survey data for the contiguous United States are available in the State
Soil Geographic (STATSGO) database (cf. Schwarz and Alexander, 1995). Shirazi et al. (2001)
developed mathematical models relating the soil characteristics of soil map units from the
STATSGO to predict water quality parameters. They found that with statistical information on
soil particle size distribution, they could estimate soil, lake, and stream water quality parameters
including acid neutralizing capacity, pH, dissolved organic carbon, nitrate, and turbidity.
14
-------
Because such water quality parameters importantly influence the concentrations and mobility of
the metals of concern, such soil information can be used to predict metal behavior.
3.5 Local Sources of Metals Data
In studies of localized environmental contamination by metals (e.g., as associated with a
Superfund Site cleanup), it is especially important to define background and baseline metals
concentrations, which will in some cases exceed health standards for aquatic life and drinking
water. If data and maps describing metal concentrations at a site prior to its contamination are
not available, the preferred approach is to take water, soil, sediment, rock, or plant samples from
a nearby area with similar characteristics, but that is unaffected by contamination as a measure
of background (cf. Banks et al., 1995; Lahermo et al., 1995; Miller and McHugh, 1999). In
populated and urbanized or extensively mined areas, however, it may be difficult to locate a
sample that has not been contaminated. All that may be available are baseline values. A second
approach to determine background for waters is to assume that background water quality is the
same as that measured in similar streams or well waters in similar topographic, climatic, and
geological settings. Lacking this information for ground waters, White et al. (1963) have
summarized typical ground water compositions from different rock types that may be useful.
Perhaps the most defensible way to distinguish background or baseline concentrations
and more elevated metal values in the waters of a given area is to use statistical analysis.
Cumulative probability plots can be drawn onto which all of the concentration data for a metal
are plotted (Levinson et al., 1987; Fleischhauer and Korte, 1990). Such plots may allow the
classification of samples into a background or baseline group and one or more contaminated
groups, with an estimate of the median and standard deviation for each group. Such a plot for Cd
in ground waters of Front Range of Colorado is shown in Figure 2 (Langmuir and Klusman,
1997).
4. METAL SPECIATION: METAL COMPLEXES
Chemical analytical laboratories generally determine and report total metal
concentrations. However, in recent years it has become increasingly apparent that metal toxicity
is a function of the concentrations of specific metal species, not of the total metal (cf. Stumm and
Morgan, 1996; Langmuir, 1997a). In fact, toxicity was often found to be best predicted by the
free metal ion activity (Morel and Herring, 1993). In natural waters, its concentration, or those of
other metal species, can sometimes be measured directly, although often with difficulty.
Alternatively, given the appropriate solution analysis and thermodynamic database, the
concentrations of individual species can be computed using a geochemical equilibrium modeling
program such as MINTEQA2 (U.S. EPA, 1991) or PHREEQC (Parkhurst and Appelo, 1999).
Errors in the thermodynamic database of the earlier versions of MINTEQA2 (Serkiz et al., 1996)
have been corrected in version 4.0 (1999). Unfortunately, MINTEQA2 is a DOS-based program
and cannot be used on computers with Windows operating systems more recent than Windows
98. However, a Windows version of MINTEQA2 version 4.0 (1999), a program called Visual
15
-------
MINTEQ (VMINTEQ, version 2.30; Gustafson, 2004), is available and may be downloaded free
of charge from the Web.
However, PHREEQ and MINTEQA2 are not suitable for predicting metal complexation by
natural organic matter. The presence of several potential complexation sites with different
affinities and electrostatic interactions will affect metal complexation by humic substances,
which are the main components of natural organic matter. The PHREEQC model approximates
complexation by natural organic matter by using two generalized complexation constants for
fulvate and humate. MINTEQA2, version 4.0, allows the use of a competitive Gaussian model to
estimate the complexation of metals by natural organic matter, but the database provided is not
adequate (Allison and Perdue, 1994).
More elaborate databases and models have also been developed to account for multiple binding
sites on humic substances and electrostatic interactions. For example, the Windermere Humic
Aqueous model (WHAM) is based on an extensive literature review and applies the humic ion-
binding model VI (Tipping, 1994, 1998), in which discrete sites with a range of affinities as well
as bidentate complexation sites are considered. Another example of such models is the non-ideal
competitive adsorption (NICA)-Donnan model (Kinniburgh et al., 1996), which is included in
the chemical speciation and transport program ECOSAT (Keizer and van Riemsdjik, 1994). The
model uses the NICA to model specific binding to humic substances and the Donnan
electrostatic model to consider non-specific binding.
In all cases, the capacity of geochemical equilibrium models to predict metal speciation in waters
containing natural organic matter is limited and has not yet been validated in the field. The
development and calibration of both models are based on laboratory titrations of isolated natural
organic matter; very few studies are available to validate their use for natural waters. Even with
more elaborated models, only the humic substances fraction of natural organic matter is
considered in terms of complexation. In fact, input parameters for these models are the humic
and fulvic acid concentrations, while the dissolved organic carbon pool is generally not
characterized. For some natural waters, however, other organic ligands with higher affinities
might be more important ligands than humic substances (Xue and Sigg, 1999).
16
-------
99.9
99
95
K~
i 80
cc
HI
CL 50
Ui
>
< 20
-i
D
3 5
O
1
0.1
Figure 2. Cumulative percent plot of cadmium concentrations in Colorado Front Range
ground waters. Cd concentrations in the 101—102 |ig/L range are anomalously elevated.
Concentrations below 10 |ig/L are presumably background values. The linear trend of higher
values can be derived from contamination. From Langmuir and Klusman, 1997.
4.1 Metal Complexes: Concepts and Importance
Dissolved species in water can be described as free ions or aquo-complexes, or simply
complexes. A complex is a dissolved species that exists because of the association of a cation
with an anion or neutral molecule (Langmuir, 1997a). A ligand is an anion or neutral molecule
that can combine with a cation to form a complex. The total analytical concentration of a given
metal in water is the sum of the concentrations of its free ion and complexes and any metal
associated with suspended solids, whether organic or mineral. For example, the total molal
concentration of lead, SPb, in a natural water might equal:
SPb = mPb2+ + mPbOH+ + mPbC03° + mPbHC03+ + mPbS04° + mPb(suspended solids) (1)
In most natural waters the concentration of free lead ion, mPb2+, is less than the sum of
the concentrations of its complexes, which in this case are lead complexes with hydroxyl,
;
\
\
\
\
\
•
r
*
*
1 1 1 . 1 _L J J
_l 1 1 1 1 1 1 1 1 1 1 L_
0 50 100 150
Cd (ng/L)
17
-------
carbonate, bicarbonate, and sulfate ions. Other metals that are found in natural waters most often
as complexes and not as free ions include Al3+, Ag+, Cu2+, Fe3+, and Hg2+. As, Se, Cr, Mo, Sb,
and V occur most often in aerobic waters and soils not as cations but covalently bonded to
oxygen in oxyanions that under oxidizing conditions include arsenate, selenate, chromate,
molybdate, and vanadate, which themselves are complexes. Important chemical species
(including metal complexes common in soils and waters at intermediate pH values) are listed in
Table 7.
Complexes that incorporate metals play a major role in controlling the availability and
fate of metals in the environment. Increasing the fraction of a metal that is complexed increases
the solubility of minerals of that metal (Langmuir, 1997a). For example, the solubility of lead
sulfate is related to the molal concentrations of free lead and sulfate ions, through the expression:
Ksp = [(yPb.mPb2+)(YS04.mS042-)] (2)
where the terms yPb and yS04 are the activity coefficients of the ions.2 The product of the ion
activity coefficient and molal concentration of each species equals the activity of the ion.
Equation 2 shows that the activity of free lead ion controls the solubility of lead sulfate. For a
given total lead concentration (Equation 1), the more of the lead that is complexed, the lower
will be the concentration of free lead ion. This means that as the extent of lead complexing
increases, the total lead concentration must also increase in order to reach saturation with lead
sulfate. In other words, metal complexing increases total metal solubility.
Metal complexing also has a direct influence on metal adsorption to organic matter or
mineral surfaces. For example, metal carbonate, sulfate, and fluoride complexes are usually
poorly adsorbed, whereas metal hydroxide complexes are strongly adsorbed (Langmuir, 1997a).
In summary, metal complexing generally increases the solubility and mobility of metals in
surface and ground waters.
4.2 Hard and Soft Acids and Bases: The Stability of Complexes and Metal Toxicity
Complexes are formed between metals (acids) and ligands (bases) both in solution and at
the surfaces of minerals and of organisms. Toxic reaction of organisms to metals can be directly
related to the nature of the metal complexes formed in solution and at the surface of the
organism.
2
By definition, the product of the activity coefficient and the molal (or molar) concentration for an ion, i, is equal to
the activity (a,) of that ion. That is: a, = Yrmr Activity coefficients of ions are generally less than one in fresh waters,
and decrease with increasing salinity or ionic strength (cf. Stumm and Morgan, 1996; Langmuir, 1997a). The activity
of an ion can be considered its effective concentration in water.
18
-------
Table 7. Dominant Chemical Species of Metals in Soils and Natural Waters, Not
Considering Most (Especially Weak) Metal Complexing
Metal
Soils
Waters
Best Predictor of
Metal Toxicity
Ag
Ag+
Ag+, AgCl
Ag+
As
As043"
As043", As033~
As033"
Ba
Ba2+
Ba2+
Ba2+
Be
Be2+, BexOy2x 2y
Be2+
Be2+
Cd
Cd2+
Cd2+
Cd2+
Co
Co2+
Co2+
Co2+
Cr
Cr3+
Cr042", Cr3+
Cr042"
Cua
Cu2+, Cu-OMb
Cu2+, -fulvate
Cu2+
Hg
Hg2+, CH3Hg
Hg(OH)2°, HgCl2°
CH3Hg
Mn
Mn4+, Mn2+
Mn2+
Mn2+
Mo
Mo042"
Mo042"
Mo042"
Ni
Ni2+
Ni2+
Ni2+
Pb
Pb2+
Pb(OH)+
Pb2+
Sb
SbmOx?
Sb(OH)6"
?
Se
Se042", HSe03
Se042"
Se042"
V
VwOx?
HV042"
?
Zn
Zn2+
Zn2+
Zn2+
Modified after Logan and Traina, 1993.
aTypically, much or most of Cu in soils is complexed with organic matter.
bCu-OM denotes copper complexed with organic matter.
19
-------
A useful definition that helps to explain the strength of metal complexing and metal
toxicity is that of hard and soft acids and bases (Pearson, 1973). In this definition, cations are
Lewis acids and ligands Lewis bases, with the metal cation and ligand in a complex acting as
electron acceptor and donor, respectively. Soft implies that the species' electron cloud is
deformable or polarizable with the electrons mobile and easily moved. Such species prefer to
participate in covalent bonding. Hard species are comparatively rigid and nondeformable, have
low polarizability, hold their electrons firmly, and prefer to participate in ionic bonds in complex
formation (Langmuir, 1997a). Hard acids form strong, chiefly ionic bonds with hard bases,
whereas soft acids and soft bases form strong, chiefly covalent bonds when they form
complexes. In contrast, the bonds formed between hard-soft or soft-hard acids and bases are
weak, such that their complexes tend to be rare. Table 8 summarizes hard and soft acid and base
relationships for the metals of concern. Its footnotes summarize the applicability of hard and soft
concepts to the formation of metal complexes.
Table 8. Hard and Soft Acids (Cations) and Bases (Ligands)
Hard acids
Al3+, Ba2+, Be2+, Co3+\ Cr3+, Fe3+, Mn2+, Sr2+,
u4+, uo22+, vo2+
Borderline acids (between hard and soft)
Co2+, Cu2+, Fe2+,Ni2+, Pb2+, Zn2+
Soft acids
Ag+, Cd2+, Cu+, Hg2+, Hg+, CH3Hg+, Tl3+, Tl+
Hard bases
F, H20, oxyanions: OH", S042", C032, HC03
, C2042", Cr042", Mo042" HnP04n 3, HnAs04n"3,
Se042, H2V04, NH3, RNH2, N2H4, ROH,
RO , R20, CH3COO, etc.
Borderline bases (between hard and soft)
CI, Br, N02, S032, HnAs03n 3, C6H5NH2,
c5h5n, n3-, n2
Soft bases
I , HS, S2, CN, SCN , Se2, S2032, -SH, -SCH3,
-NH2, R-, C2H4, C6H6, RNC, CO, R3P, (RO)3P,
R3As, r2s, RSH, RS
Modified after Huheey et al., 1993, and Langmuir, 1997a. "R" refers to an organic molecule.
Hard acids and hard bases. Complexes formed between divalent hard acid cations and monovalent or divalent hard
bases are ionic and relatively weak, and are often termed "ion pairs." Complexes formed between Be2+ or trivalent
hard acids, and hard bases tend to be ionic and relatively strong.
Soft acids and soft bases. Strong, relatively covalent bonds are formed in complexes between soft and borderline
soft acid cations and soft bases. Ligand binding sites on the external or internal surfaces of organisms are often of
soft base character, and so bond strongly with soft and borderline soft acid cations.
20
-------
The Lewis acids in natural waters include H+ and metabolically essential metal cations
such as Na+, K+, Mg2+, Ca2+, Mn2+, Fe3+, Co2+, Ni2+, Cu2+, and Zn2+, and the toxic metal species
Hg2+, CH3Hg+, Pb2+, and Cd2+, which are soft acids. These species, along with the soft acid
thallium (Tl) and the essential protein and enzyme metals Fe2+, Cu2+, Ni2+ and Zn2+, which are
borderline soft, form strong bonds with soft base sulfur species. The ligands at the surfaces of
biota are chiefly sulfur, nitrogen and oxygen electron donor groups, and also include the solute
bases HC03, HP042, and OH" (Stumm and Morgan, 1996).
Although the hard acid metals Ca, Mg, and Na bond weakly with soft base sites on the
surfaces of biota, their concentrations in most waters are typically 103 times greater than those of
the trace metals, so that they often effectively compete with trace metals such as Zn2+ and Cd2+
for surface bonding sites. However, when metal adsorption is by organic matter, metals such as
Cu2+ in particular may be strongly adsorbed, practically independent of concentrations of the
major metals (cf. Lu and Allen, 2002).
The proton is the most effective competitor of all with adsorbed metals. The proton can
displace essential and trace metals from surface sites at pH values as high as 6. The role of pH as
a fundamental control on metal concentrations and metal transport is discussed extensively
below.
Mechanisms by which toxic metals poison plants and animals relate to their tendency to
form strong complexes with the generally soft functional groups on biomolecules (cf. Sposito,
1989; Morel and Herring, 1993). Sposito (1989) proposes several processes by which soft-metal
cations cause phytotoxicity. First, a soft metal such as Cd can displace an essential metal such as
Ca bound to a bioligand. Also, complexation of a bioligand by a soft-metal cation can block that
ligand from reacting normally or modify it structurally and thus interfere with its intended
activity. Enzymes have active or catalytic sites with which they bind to biological substrates and
that facilitate enzyme function. These sites are especially vulnerable to damage by soft-metal
cations. The amino acids cysteine and methionine present at the active sites in some enzymes
contain -SH and SCH3 groups (Manahan, 1994). These sulfur-containing groups are soft ligands
and form strong covalent bonds with soft-metal cations such as Hg, Ag, Cd, Cu, and Pb. Such
bond formation can result in the breakdown of normal enzyme function and a toxic reaction by
the affected organism.
4.3 Predominant Inorganic Species in Fresh Waters
It is instructive to compute the distribution of dissolved metal species including their
complexes in a typical natural water. Such computations are readily accomplished using a
geochemical computer code such as MINTEQA2 (U.S. EPA, 1991). Stumm and Morgan (1996)
discuss such a calculation for an oxygenated fresh water with a pH of 8.0 that has the following
total concentrations: Na = 5.7 mg/L, Mg = 73 mg/L, Ca = 40 mg/L, CI = 8.9 mg/L, S04 = 29
mg/L, and HC03" =122 mg/L. Aqueous speciation for the metals of concern is given in Table 9.
Also shown is the percentage of the metal present as the free ion, which in most cases is the
21
-------
percentage of the total metal present in toxic form. Note that at pH 8, Al, Be, Cu, Fe(III), Hg, Ni,
Pb, Tl, and Zn are chiefly present as complexes. In contrast, it was recently suggested that metal-
sulfide complexes could dominate the speciation of some soft acid metals such as Ag and Cu,
given that such inorganic complexes are stable in oxic waters (e.g., Rozan et al., 2000).
The anionic species of As, Cr, Mo, Se, and V generally form weak complexes with
monovalent or divalent cations, but may be precipitated as minerals by high concentrations of
dissolved Fe(III) and Al(III), which are generally present only in acid systems.
4.4 Effect of pH on Metal Complexing
In natural waters, sulfate, phosphate, and fluoride tend to be important in acid waters (c.f.
Langmuir, 1997a), whereas carbonate and hydroxide complexes become increasingly important
at pH values above 6-8. This is evident from Figure 3, which shows the concentrations of
species of Pb and Zn as a function of pH, as computed with a geochemical model for the same
concentrations used to derive Table 9. Total lead and zinc concentrations assumed for these
figures are 10"9 M (0.2 pg/Lj for lead and 10"8 M (0.65 |ig/L).for zinc. As the salinity and thus
chloride content of water increases, the borderline soft chloride ion forms important complexes
with the soft and borderline soft metal cations Ag+, Cd2+, Hg2+, Ni2+ and Zn2+.
22
-------
Figure 3. Speciation of Pb(II) (109 M) and Zn(II)(108 M) under freshwater conditions.
Total carbonate equals 2 x 10"3 M. Figure is computed. From Stumm and Morgan, 1996.
Several milligrams of dissolved organic carbon (DOC) per liter of soil solution is
common in humid climate soils. The DOC is composed of humic substances, which include
humic and fulvic acids. These acids form metal complexes under the same pH conditions that
favor metal carbonate and hydroxide complexes (Stumm and Morgan, 1996). Metal bonding is
with carboxylic and phenolic hydroxyl groups of the acid anions. Recent work has also
highlighted the importance of sulfide-thiolate groups as high-affinity binding sites for soft acid
metals (e.g., Smith et al. 2002). Shown in Figure 4 is the effect of increasing pH and total copper
on the extent of copper-fulvic complexing. Free Cu2+ ion concentrations can be seen to decrease
with increasing DOC and increasing pH in the presence of Suwannee River fulvic acid.
Increasing DOC should thus reduce the toxicity of a given concentration of dissolved Cu2+ or
other toxic metal that forms fulvic acid complexes.
23
-------
5
6
7
pCu
8
9
10
7 6 5 4
pCur
(a)
11
10
9
pCu
8
7
6
5
4 5 6 7 8 9
PH
(b)
Figure 4. Titrations of fulvic acid (FA) and copper, (a) shows dissolved Cu2+ ion
concentration as pCu (-log Cu [mol/L]), plotted as a function of total copper ([Cu]x) for different
fulvic concentrations at pH 7. (b) shows pCu as a function of pH for two total copper
concentrations. From Cabaniss and Shuman, 1988.
Suwannee FA
pH 7.00
« / DOC lmg liter1
. +
* o ~
* * "* DOC 2 mg liter1
t ** DOC 5 mg I iter1
* * DOC 10 mg liter1
24
-------
Table 9. Major Inorganic Species in a Hypothetical Natural Water (see text)
Condition
Metal/Element
Major Species
Percent as Free Metal Cation
Oxyanions
As(V)
HAs042"
Cr(VI)
Cr042"
Mo(VI)
Mo042"
Se(VI)
Se042"
V(V)
hvo42-, h2vo4-
Predominantly free
Na+
Na+
100
aquo-ions
K+
K+
100
Mg2+
Mg2+
94
Ca2+
Ca2+
94
Sr2+
Sr2+
94
Ba2+
Ba2+
95
Complexed with
Ag(I)
Ag+, AgCl0
60
OH\ C032\ HC03\
cr
Al(III)
A1(OH)3(s), Al(OH)2+, Al(OH)4-
1 x 10"7
Be(II)
BeOH+, Be(OH)2°
0.15
Cd(II)
Cd2+, CdC03°
50
Co(II)
Co2+, CoC03°
50
Cu(II)
CuC03°, Cu(OH)2°
2
Fe(III)
Fe(OH)3(s), Fe(OH)2+, Fe(OH)4"
2 x 10"9
Hg(II)
Hg(OH)2°
1 x 10"8
Mn(IV)
Mn02(s)
Ni(II)
Ni2+, NiC03°
40
Pb(II)
PbC03°
5
T1(I),
Tl(III)a
Tl+, Tl(OH)3, Tl(OH)4-
2 x 10"19
Zn(II)
Zn2+, ZnC03°
40
The right-hand column is the percent of the total metal concentration present as the free metal cation. Based on
Stumm and Morgan (1996).
aRedox state of T1(I) under natural conditions is uncertain; ratio is for Tl(III).
25
-------
5. ADSORPTION CONTROLS ON METAL CONCENTRATIONS AND MOBILITIES
5.1 Introduction
Except for A1 and sometimes Mn, concentrations of trace metals in the environment are
generally too low for those metals to exceed the solubility products of their pure metal solids and
thus to precipitate. Instead, toxic metal concentrations are generally limited by sorption onto the
surfaces of minerals, and onto organic matter including microbial cell wall surfaces. For
aluminosilicate-rich soils, however, surface precipitation of Zn, Ni, and Co as layered double
hydroxides was reported (e.g., Ford and Sparks, 2000). As metal concentrations further increase
and fill available sorption sites, most metals tend to be incorporated in the structures of major
mineral precipitates as "coprecipititates" in which they substitute for major metal cations,
forming so-called solid solutions (Langmuir, 1997a). At higher metal concentrations, the metals
may be precipitated in pure metal phases, limiting further increases in metal concentration. For a
hypothetical metal, M, the general trend of such reactions with increasing metal concentrations
in soil pore waters or surface or ground water might be:
• Adsorption control of M
(M<100 ng/L)
• Coprecipitation of M in a major metal hydroxide, carbonate, sulfate or silicate mineral (etc.)
(M>100 |ig/L and <10 mg/L)
• Precipitation of M in a "pure" metal phase
(M = 10 mg/L)
Practically, it is difficult to distinguish adsorption and coprecipitation reactions. These
concepts are illustrated in Figure 5, which shows adsorption of lead from a landfill leachate by
kaolinite clay as a function of pH. In the figure the adsorbed lead in mg/g of clay is plotted
against the total lead in the leachate. The plot indicates that lead adsorption increases with
increasing pH. Geochemical modeling of the leachate solution at pH 5 and 6 shows that at about
240 and 40 mg/L lead, respectively, (as indicated by the vertical dotted lines) further increases in
dissolved lead are limited by precipitation of solid lead hydroxy-carbonate (Pb3[0H]2[C03]2).
26
-------
8.0
7.0
6.0
5.0
B 4.0
I
3.0
2.0
1.0
40 80 120 160 200 240 280 320 360 400
XPb(aq) (mg/L)
Figure 5. Adsorption of Pb from DuPage landfill leachate by kaolinite at 25°C, as a
function of pH. Dashed vertical lines show the Pb concentration at saturation with Pb-hydroxy-
carbonate solid. From Griffin and Shimp, 1976.
5.2 Metal Adsorption and Desorption
Distribution Coefficients, Isotherm Equations and Ion Exchange:
They Don't Generally Apply to Adsorption of the Metals of Concern
The distribution coefficient (Kd) for lead adsorption by kaolinite is defined as
Kd = (x/m)/EPb(aq), where x/m is the weight of lead adsorbed, x, divided by the weight of
sorbent, m, at a specific concentration of total dissolved lead, SPb(aq). Based on this definition,
every tangent to a curve in Figure 5 has a different value of Kd, The solid curves in Figure 5 are
termed sorption isotherms, and can be model-fit with Langmuir or Freundlich isotherm equations
(cf. Stumm and Morgan, 1996; Langmuir, 1997a). Clearly, a single distribution coefficient or
isotherm equation cannot predict lead adsorption for all of the conditions described in Figure 5.
This is because the extent of lead adsorption depends on the total lead concentration and the pH.
Further analysis shows that adsorption also varies with the amount of lead complexed by
carbonate, sulfate, chloride, and organic ligands, and by the nature and amounts of one or more
sorbent phases present. Given all of these variables, Kd values for metal adsorption can vary by
two to three orders of magnitude for a single soil or sediment. In other words, adsorption of
27
-------
metals is far more complicated than adsorption of organics, which can often be adequately
modeled using a distribution coefficient or isotherm approach.
Ion exchange models are also generally inappropriate to explain and model adsorption of
trace metals. Simple ion exchange models best apply when concentrations of competing ions are
at comparable values, which is not the case for most trace metals most of the time. Except in
highly metal-contaminated waters and sediments, concentrations of toxic metal cations are
typically 102 to 103 times lower than concentrations of the major cations Na+ and Ca2+ which are
competing for sorption sites on clays, for example . Concentrations of the anionic trace metals
are also in most cases less than 1% of the concentrations of major anions such as sulfate. (See
Tables 1 and 3.) At trace concentrations, metals are preferentially adsorbed relative to major
metal ions, but not according to the principles of simple ion exchange (Langmuir, 1997a).
5.3 Adsorption Behavior of the Metals of Concern
In porous media the most important sorbent solids for metals are clay minerals, organic
matter, and oxyhydroxides of Fe and Mn. Their important surface properties are given in Table
10. For a given weight of sorbent, metal sorption capacity is proportional to surface area and
surface site density. The greatest surface site densities (positively or negatively charged sites)
and cation exchange capacities (negatively charged sites only) are those of organic material and
the oxyhydroxides. These phases are the strongest and most important sorbents of trace metals.
The clays, except for kaolinite, have a surface charge that is largely independent of pH, whereas
the surface charge of organic matter and the oxyhydroxides is strongly pH dependent. The pH at
which a solid surface changes sign is referred to as the zero point of charge (ZPC). The ZPCs of
organic matter and kaolinite are below the range of natural pHs and, while their surface charge
will decrease with decreasing pH, their surfaces will remain negatively charged at low natural
pH. In contrast, some oxyhydroxides (such as goethite, a-FeOOH) have ZPCs within the range
of natural pHs and will have positive surface charges at low pH and negative surface charge at
high pH. Thus, the negative surface charge of the oxyhydroxides increases with increasing pH,
which means their sorptive capacity for metals increases with increasing pH. Conversely, the
positive surface charge of the oxyhydroxides increases as the pH drops, making these phases
more effective sorbents for anions under low pH conditions.
These effects are shown in Figure 6, which is a plot of percent sorbed versus pH for
metal adsorption by ferrihydrite, or hydrous ferric oxide (HFO). The curves are called sorption
edges for each metal. The diagram shows that the oxyanions are strongly adsorbed by HFO
under acid conditions, but are desorbed (become mobile) with increasing pH. Based on Figure 6,
combined with adsorption data assembled by Dzombak and Morel (1990), the order of
desorption from HFO with increasing pH is, selenate, antimonate, molybdate, chromate,
vanadate, arsenate, and phosphate. Selenate desorbs between pH 3 and 8, whereas arsenate is
strongly held at lower pH values and desorbed between pH 9 and 11. Also based on Figure 6 and
Dzombak and Morel (1990), with increasing pH, HFO preferentially adsorbs metals in the order
Hg2+>Be2+>Ba2+>Cr3+>Pb2+>Cu2+>Cd2+ - Zn2+>Ni2+.
28
-------
2
4 6
8
10 12
(a)
pH
100
*
(3
3
O
£>
80
60
-
* Jr
* /
* /
t /
I /
Cr3+ : / Pb2+
! I
/Cu2+ /
/ / / /
/ /Ni2+
/ / <
/ / #
/ / *
3
40
—
t / ,
* I /
* / /
/ /
ti
20
-
l/ l
Cd2+/ /¦
-/¦Zn2+
/ ; Ca2+~
1
0
3 4 5
6
7 8
(b)
pH
Figure 6. Adsorption of various metal cations and oxyanions, each at 5 x 107 M, by
ferrihydrite (2Fe[III] = 103 M) as a function of pH at an ionic strength of 0.1 mol/kg. There
are 2 x 10~4 M of reactive sites on the oxyhydroxide. The dashed curves are calculated. After
Stumm, 1992.
29
-------
Table 10. Surface Areas, Surface Site Densities, and Cation Exchange Capacities (CECs) of
Some Important Sorbent Phases and the pH Dependence of Metal Sorption
Sorbent
Mineral/Phase
Surface
Area (m2/g)
Site Density
(sites/nm2)
Site Density
(mmol sites/g)
CEC
(meq/lOOg)
pH
Dependence
Kaolinite
10 to 38
1.3 to 3.4
0.022 to 0.21
3 to 15
Strong
Illite and chlorite
65 to 100
0.4 to 5.6
0.043 to 0.93
10 to 40
Slight
Smectite-
montmorillonite
600 to 800
0.4 to 1.6
0.4 to 2.1
80 to 150
Absent or
negligible
Organics in soils,
humic materials3
260 to 1300
2.31
1.0 to 5.0
110 to 500
Strong
Mn
oxyhydroxides
143 to 290
2 to 18
0.48 to 8.7
100 to 740
Strong
Fe(III)
oxyhydroxides
(ferrihydrite)
250 to 600
20
8.3 to 20
100 to 740
Strong
Modified after Langmuir (1997a).
aPaulson and Balistrieri (1999) suggest 1 mmol of sites per gram of organic carbon.
Metal adsorption onto soils and sediments is probably more dependent on changes in pH
than on any other solution variable. For a divalent metal cation, M2+, the general sorption
reaction can be written:
SOH + M2+ = SOM+ + H+ (3)
where SOH and SOM+ are surface sites with an adsorbed proton and a metal ion. For adsorption
of a divalent metal anion, L2, we can similarly write:
SOH2+ + l2- = SOHj; (4)
Sorption edges for metal adsorption, by amorphous Al(OH)3 and by soil humus material,
are given in Figure 7. The plots show a strong pH dependence of metal adsorption on both
substrates. Metal selectivity with increasing pH is similar to what was observed for adsorption
by HFO.
30
-------
Figure 7. Percent metal adsorbed: (a) by Al(OH)3(am) at total metal concentrations of 1.25
x 104 M in 1 M NaN03 solutions, and (b) by humic acid at total metal concentrations of 5 x
10"4 M. From Sposito, 1984.
5.4 Surface Complexation Adsorption Models
The most useful and mechanistic sorption models for predicting and modeling toxic
metal adsorption in surface and ground waters are probably the electrostatic or surface
complexation (SC) adsorption models (cf. Westall and Hohl, 1980; Davis and Kent, 1990;
Langmuir, 1997a). These include the constant capacitance (CC), diffuse layer (DL), and triple
layer (TL) models, all of which are available in the EPA geochemical program MINTEQA2
(U.S. EPA, 1991). The DL model is also included in the U.S. Geological Survey program
PHREEQC (Parkhurst and Appelo, 1999).
The SC models employ the electrical double layer (EDL) theory, according to which the
sorbent surface charge results in an electrostatic potential gradient between the solid surface and
the bulk solution. The layer close to the surface has a fixed charge equal to the surface charge.
This layer is associated with specific adsorption, which is not influenced by ions in the bulk
solution. In all SC models, H+ and OH" are specifically adsorbed within the first layer. In the
second layer, however, the accumulation of counter ions results in a charge with a sign opposite
to the surface charge and in an absolute charge that decreases toward the bulk solution.
According to EDL, it is assumed that the concentration of a sorbed ion (Xsz), which is not
measurable, is related to the ion's concentration in the bulk solution (Xz) by an exponential
Boltzmann expression:
(Xsz) = (Xz)[e"fF/RT]z (5)
in which z is the charge of the ion, e"fF/RT is the Boltzmann factor (unitless), i|/ is the potential in
volts at the plane of adsorption, and F, R, and T are the Faraday constant (96,480 C/mol), the
ideal gas constant (8.314 J/mol), and the absolute temperature, respectively.
31
-------
The SC models consider and compute activity coefficients of ions and ion activities, as
well as the concentrations of free and complexed metals in solution. The models treat metal
adsorption onto surface sorption sites as a complexation reaction. Required model inputs include
the specific surface area of sorbent phases (m2/g), the site density (moles of sites/moles of
sorbent), and intrinsic surface complexation constants for adsorption of each metal by each
sorbent phase. These constants are analogous to the equilibrium constants that describe the
formation of metal complexes in solution. Because the SC models are relatively atomistic and
mechanistic in the way they treat adsorption, they are far better at predicting metal adsorption
behavior for conditions beyond those used to determine model parameters than are models based
on adsorption isotherms or ion exchange. In fact, even if the SC models have been developed in
the laboratory, they have been used with success to predict adsorption on natural particles (e.g.,
Tessier et al., 1996).
The diffuse layer (DL) model is the simplest of the SC models (U.S. EPA, 1991;
Langmuir, 1997a). In the DL model, adsorption is considered to occur specifically at the surface,
and the effect of adsorbed electrolyte ions on the potential away from the surface is neglected.
Both MINTEQA2 and PHREEQC contain a database of sorption parameters (intrinsic constants)
for the adsorption of metals by HFO from Dzombak and Morel (1990). Sorption parameters for
all of the toxic metal cations and anions of concern, except for Al and Tl, are included in the
database. Diffuse layer model parameters for adsorption of H+, Cd2+, Cu2+, Ni2+, Pb2+, and Zn2+
by 5-Mn02 have recently been published by Pretorius and Linder (2001), but are not yet
included in MINTEQA2 or PHREEQC.
5.5 Applications of the Diffuse Layer Model to Natural Systems
It has been observed that in many soils and sediments low in organic matter, HFO is the
most important metal sorbent, and the only sorbent that needs to be considered in predictions of
toxic metal sorption behavior (cf. Benjamin and Leckie, 1981). The diffuse layer model (also
called the generalized two-layer model, or GTLM) in MINTEQA2 has been extensively applied
in aquatic environmental studies of metal transport and attenuation. Loux et al. (1990) used the
DL model and MINTEQA2 to predict the adsorption and precipitation behavior of eight metals
in an oxidized, sandy aquifer as a function of pH. Assuming that HFO was the only sorbent, DL
model adsorption adequately described changes in the concentrations of Ni, Pb, and Zn.
Cadmium behavior was better understood, assuming its precipitation in CdC03 (otavite).
Changes in Cu, Ba, Be, and Tl were not simply explained. Copper may have been adsorbed by
organic matter, which was not considered in the modeling.
More recently, adsorption of metals by organic matter and aluminum oxyhydroxides, as
well as HFO, has been included in DL modeling with MINTEQA2 (Paulson and Balistrieri,
1999). These authors studied neutralization of acidic ground waters by ambient surface and
ground waters using a mixing model approach. Particulate organic matter (POC) and HFO were
the chief metal sorbents. In pristine systems, Cu is usually the chief metal associated with POC;
32
-------
however, in their study, Zn and Cd were mostly adsorbed by POC and Cu was mostly absorbed
by HFO.
In another study of metals in acid mine waters, Smith et al. (1998) measured and
modeled the adsorption of Cd, Cu, Ni, Pb, and Zn by streambed sediments as a function of pH
using MINTEQA2 and the DL (GLTM) model. Figure 8 shows their measured adsorption
results, obtained in batch experiments, assuming 2.9 g/L of HFO as the only sorbent. The plot
shows fair to excellent agreement between measured and predicted metal adsorption values.
These authors also measured and predicted metal concentrations in three mine drainage waters,
assuming that metal adsorption was by suspended HFO. The results, listed in Table 11, show
good general agreement between measured and predicted metal concentrations.
Tailings Site, October
Downstream Site, October
Figure 8. Comparison of experimental data (symbols) for Pb, Cu, Zn, Ni, and Cd sorption
onto streambed sediment from St. Kevin Gulch, Colorado, with computer-model
simulations (curves) for sorption onto hydrous ferric oxide, at a streambed concentration
of 2.9 g/L. From Smith et al., 1998.
33
-------
Table 11. Comparison of Model Predictions and Measured Values of Percent Metals
Associated with the Suspended Particulate Fraction of Mine-Drainage Waters from
Selected Sites
Argo-3
(pH 5.6, HFO = 0.007 g/L)
Rawley-3
(pH 6.2, HFO = 0.11 g/L)
Leadville Drain
(pH 7.2, HFO = 0.001 g/L)
Predicted
(%)
Measured
(%)
Predicted
(%)
Measured
(%)
Predicted
(%)
Measured
(%)
As
—
—
98
<78a
—
—
Pb
82
<71a
80
<93a
86
<71a
Cu
18
27
60
63
-
—
Zn
<1
0 to 8
<1
0 to 9
2
3
Ni
<1
<1
<1
1
-
—
Cd
<1
<1
<1
6
<1
<1
Model predictions made with the DL model and MINTEQA2. From Smith et al., 1998.
aDissolved concentration was below the detection limit. Value computed using the limit of detection for the
dissolved concentration.
Cederberg et al. (1985) and Yeh and Tripathi (1991) considered surface complexation
modeling of metal adsorption and metal transport in ground water. Parkhurst (2002) has
developed a computer model called PHAST,3 which is a 3D reactive transport model that
combines PHREEQC, which has the DL metal adsorption model, with HST3D, a ground-water
flow and transport model.
Several recent studies have measured and modeled trace metal adsorption and metal
transport in streams using a surface complexation approach to adsorption. U.S. Geological
Survey researchers of the Toxic Substances Hydrology Program have published a number of
papers using the OTEQ and OTIS models. OTEQ is a one-dimensional model for studying the
fate and transport of metals in streams and rivers. The model couples the OTIS transient storage
model with MINTEQ, which includes DL model adsorption of metals by HFO (cf. Ball et al.,
1999; Runkel et al., 1999). In their study, Runkel et al. (1999) considered in-stream metal
transport, metal oxide precipitation-dissolution, and pH-dependent sorption of copper and zinc.
3
See http://wwwbrr.cr.usgs.gov/projects/GWC_coupled.
34
-------
5.6 WHAM and Related Models for Predicting Metal Activities in Soil Pore Waters
MINTEQA2 (U.S. EPA, 1991) and VMINTEQ (Gustafson, 2003) both contain
subroutines that allow estimates of the importance of metal-organic complexing if the
concentration of dissolved organic carbon (DOC) is known. Perhaps more useful in studies of
metals in soil pore waters are programs such as WHAM (Tipping, 1994, 1998), and NICA
(Gooddy et al., 1995) (see Section 4 for more information). Application of the chemical
speciation model WHAM has been discussed by Tye et al. (2003), who successfully predicted
Zn2+ and Cd2+ activities in soil pore water assuming the metals were adsorbed by soil humus
according to a pH-dependent Freudlich isotherm model. Competitive adsorption between Ca2+
and Zn2+ and Cd2+ could be ignored because it did not improve model fits.
6. SOLUBILITY CONTROLS ON METAL CONCENTRATIONS AND MOBILITIES
6.1 The Importance of pH
The solubility of most metals that occur as cations is strongly pH dependent. Their
greatest solubilities are usually measured in acid systems, with solubilities that decrease as the
pH rises. For a few metals (e.g., Be[II], Zn[II], A1[III], and Fe[III]), metal solubility increases
again at alkaline pH values, a property which is termed amphoteric behavior (Figures 12 and 16).
In the following discussions of solubility, we will focus on the pH range between 4 and 9, which
includes that of most natural waters and soils.
Tyler and Olsson (2001a, 2001b) mixed calcium carbonate with an acid Swedish
Cambisol to vary the soil pH from 5.2 to 7.8. They then studied the effect of the pH change in
the oxidized soil on concentrations of 60 elements in soil pore waters. With increasing pH,
concentrations of As, Mo, S, Sb, and to a lesser degree Co, Cr, Hg, and Sr increased, whereas Al,
Ba, Fe, Mn, and T1 concentrations decreased. The pH effect on Be and Cu concentrations was
poorly defined. Metal concentration changes with increasing pH may have been caused by
increasing desorption of anionic elements (As, Mo, S, Sb, and Cr), increasing adsorption of
cationic species, the precipitation of oxyhydroxide solids (Al, Fe, Mn) and, at least for Ba, the
precipitation of sulfate or carbonate. The ill-defined behavior of Cu with rising pH may reflect
its participation in competing reactions: (1) complexation by increasing amounts of dissolved
humic substances with increasing pH, which tend to solubilize Cu2+; and (2) increased Cu2+
adsorption by solid organic matter and metal oxyhydroxides.
6.2 Oxidation Potential and pH
Shown in Table 124 are the possible oxidation states and speciation of the trace metals in
natural systems. Also indicated are their hard or soft acid or base characters, which depend on
the oxidation state, their forms in oxidized and reduced systems, and whether they can
4
Information for Table 12 was obtained from the following sources: Baes and Mesmer, 1976; Bodek et al., 1988;
Brookins, 1988; Langmuir, 1978, 1997a; Pourbaix, 1966; and Rai and Zachara, 1984.
35
-------
precipitate in sulfide minerals. This distinction is important because the metal sulfides are
extremely insoluble, and if precipitated tend to reduce metal concentrations to below the
microgram per liter range.
The table shows that all metals forming sulfide phases (sulfide is a soft ligand) are either soft or
borderline soft.
The Eh-pH diagrams allow a graphical representation of the simultaneous influence of
pH and redox conditions on metal speciation. In the following discussion, the figures describe
the oxidation state in terms of Eh and/or pe or pE. At 25° C and 1 bar pressure, the two are
related through the expression:
pe or pE = Eh(volts)/(0.0592) (6)
Figure 9 describes the locus of measured pH and Eh values in natural waters, and the
types of waters in which the measurements have been made.5 Oxygenated surface and near
surface environments have Eh values that plot near the area titled "Environments in contact with
the atmosphere." Ground waters, because they are out of direct contact with the atmosphere, tend
to be more reducing. Waterlogged soils and sediments are among the most reduced aqueous
systems. Differences in the oxidation potential or Eh of these environments are usually related to
the abundance of organic matter and their isolation from air. Oxygen is relatively insoluble in
water, with a solubility of 8.25 mg/L at 25° C. In principle, it takes only 2-3 mg/L of labile DOC
in water to deplete this oxygen content (Langmuir, 1997a). Leenheer et al. (1974) reported a
median DOC of 0.7 mg/L for U.S. ground waters. In temperate and tropic regions, soils have a
mean organic matter content of about 2-4% (Bohn et al., 1985). Langmuir (1997a) has observed
that 4 mg/L of DOC in ground water is sufficient to make the water anaerobic.
An Eh-pH diagram for carbon is shown in Figure 10. The stability field of organic
matter, generally, if it could be plotted, would overlay the stability fields of methane and native
C (graphite) in the low Eh part of the diagram. When organic matter is in excess, and oxygen not
replenished fast enough, microbial activity can generate reducing conditions. The sequence of
reductions that result is shown in Figure 11. In waters and sediments where oxygen or other
oxidants are in excess, the sequence of oxidations shown in Figure 11 may proceed.
Oxidation of the organic matter in an isolated aquatic environment can deplete the
oxygen and provide conditions suitable for sulfate reduction. Precipitation of metal sulfides is
preceded by reduction of more abundant soil Fe(III) oxyhydroxides to dissolved Fe(II) (Figure
11). This is then followed by precipitation of Fe(II) sulfides in association with the less abundant
sulfides of Cd, Zn, Co, Ni, Pb, Ag, Cu, Hg, and Mo.
5 Problems with measuring Eh and the difference between measured and theoretical Eh values are discussed at
length elsewhere (cf. Langmuir, 1997a).
36
-------
JS
W
+ 1.0
+0.8
+0.6
+0.4
+0.2
0.0
-0.2
-0.4
-0.6
G/^,- re .
" rO/}^. Sj&/,
\ Vv<^
^>$>4
. % ,
VSn. <*
er%i
0/v, ^
«u
er
'*S>
-0.8
•1.0
2 4 6 8 30 12 14
pH
Figure 9. Approximate positions of some natural environments in terms of Eh and pH. The
dashed line represents the limits of measurements in natural environments, as reported by Baas-
Becking et al. (1960).
37
-------
6 8 10 12 14
pH
Figure 10. Eh-pH diagram for the system C-02-H2 at a total carbonate concentration of 103
M. Native C is graphite. If shown, the upper boundary of the stability field for carbohydrates
would be at slightly higher Eh values than the methane/carbonate boundary. After Brookins,
1988.
38
-------
-0.5
+0.5
__l
-10
+1.0
I
-5
+5
+10
+15
Reductions
\ 02 - Reduction
y 1 ... 1 -
Sulfide S042
Oxidations
Oxidat. of Fe(ll)^>
NH4+^ NOa'
Oxidat. of Mn(ll)
N2 -"N03-
, i
O2 - Formation)
-10 -5 0 +5 +10 +15 +20 p8
Figure 11. Sequence of microbially mediated oxidation-reduction reactions. This sequence is
often observed in natural systems and represents the thermodynamic order of the reactions.
Modified from Stumm and Morgan, 1996.
39
-------
Table 12. Oxidation States of Trace Metals as They Occur in Natural Waters and Mineral
Systems, Their Redox Sensitivity, and Their Tendency to Form Sulfides at Low Eh
Metal
Symbol
Hard
or Soft
Oxidation
States
Oxidized
Forms
Reduced
Forms3
Can Form
Sulfides at
Low Eh
Aluminum
A1
H
3+
Al3+
Al3+
no
Beryllium
Be
H
2+
Be2+
Be2+
no
Barium
Ba
H
2+
Ba2+
Ba2+
no
Strontium
Sr
H
2+
Sr2+
Sr2+
no
Cadmium
Cd
S
2+
Cd2+
Cd2+
yes
Zinc
Zn
B
2+
Zn2+
Zn2+
yes
Cobalt
Co
B
(3+), 2+
(Co3+),Co2+
Co2+
yes
Nickel
Ni
B
(3+), 2+
(Ni3+), Ni2+
Ni2+
yes
Manganese
Mn
H
(4+), (3+), 2+
(4+), 2+
(3+), 2+
nob
Lead
Pb
B
(4+), 2+
Pb2+
Pb2+
yes
Silver
Ag
S
1+, (0)
Ag+
Ag+/Ag(s)
yes
Copper
aCu
bB/S
2+, 1+, 0
Cu2+
Cu+/Cu(s)
yes
Mercury
Hg
S
2+, 1+, 0
Hg2+
Hg+/Hg(l)
yes
Thallium
T1
s
(4+), (3+), 1+
(Tl4+), (Tl3+)
T120(s)/T1+
yes
Arsenic
As
bH/S
5+, 3+, 0
HAs042"
H3As037As(s)
yes
Antimony
Sb
H/S
5+, 3+, 0
Sb(OH)s"
Sb(OH)3°/Sb(s)
yes
Chromium
Cr
H
6+, 3+
Cr042"
Cr3+, Cr(OH)3(s)
no
Molybdenum
Mo
H/S
6+, (5.33+), 5+,
(4+)
HMo04"
Mo02+/Mo308(s)/
MoS2(s)
yes
Selenium
Se
H/S
6+, 4+,(0), 2-
Se042"
Se0327Se(s)/ HSe
no
Vanadium
V
H
5+, 4+, 3+
h2vo4-
V02+, V(OH)3°
no
Oxidation states in parentheses are found in mineral systems only. Hard (H), soft (S), and borderline soft (B) metals
are indicated.
aCu2+ is borderline soft, and Cu+ is a soft cation.
bB/S or H/S denotes whether the oxidized species is borderline hard or hard, and the reduced species is soft.
40
-------
Other ligands that can limit maximum metal concentrations include oxygen and
hydroxide, which react with the hard acid cations Al(III), Mn(III, IV), and Cr(III) to produce
insoluble oxyhydroxides. Phosphate, sulfate, and carbonate also form relatively insoluble
mineral precipitates when they react with divalent and trivalent metal cations. These are noted on
a case by case basis below.
6.2.1 Iron and Sulfur
Although iron and sulfur are not elements of concern, their behavior needs to be
discussed because it has a pivotal impact on the fate and transport of the trace metals. Shown in
Figure 12 is a solubility diagram for the Fe(III) oxyhydroxides as a function of pH. Plotted are
curves showing the solubility of Fe(OH)3(am), which is also termed ferrihydrite or HFO, and a
solubility curve for the crystalline mineral goethite. Initial precipitates of Fe(III) oxyhydroxide
tend to be relatively amorphous, with a strong capacity to adsorb or coprecipitate trace metals.
As they crystallize with time towards goethite, they lose surface area and adsorption capacity
and tend to desorb trace metals. Figure 12 also shows that as the pH is reduced below 7, HFO
tends to dissolve, becoming quite soluble below pH 3-4. Thus, soils and sediments at low pH
typically have little capacity to retain the metals of concern, whereas soils at higher pH values
that contain ferric oxyhydroxides may limit the release of the trace metals to the environment.
An Eh-pH diagram for iron in the presence of carbonate is shown in Figure 13. The ferric
oxyhydroxides occupy most of the diagram for oxidizing conditions, which reflects their great
stability and insolubility. The figure shows that the stability field of amorphous Fe(OH)3 (pKsp =
37.1) is much smaller than that of the crystalline phase, goethite (pKsp = 44.2.), and that under
strongly reducing conditions the ferric oxyhydroxides are unstable relative to dissolved ferrous
iron and the mineral siderite (FeC03).
A sulfur Eh-pH diagram is given in Figure 14. The large size of the sulfate field is
consistent with the fact that sulfate is the dominant form of sulfur in most environments. The
position of the stability fields for native sulfur (S°), hydrogen sulfide, and bisulfide ion indicate
that these forms are stable only under highly reducing conditions. When we add sulfur to the iron
Eh-pH diagram as shown in Figure 15, a large stability field for pyrite (FeS2) appears at low Eh
values. Acid mine drainage is produced when pyrite is exposed to atmospheric oxygen.
Oxidation of the iron and sulfur in pyrite generates strong acidities and pH values as low as 1-2.
This tends to dissolve any nearby HFO as well as trace metals that were present as sulfides or
adsorbed to the HFO.
41
-------
Figure 12. Solubility of amorphous Fe(OH)3, pKsp = 37.1 (top curve), and goethite [a-
FeOOH], pKsp = 44.2 (bottom curve) as a function of pH at 25° C. Also shown are fields of
dominance of Fe3+ ion and Fe3+-OH complexes. After Langmuir, 1997a.
42
-------
pH
Figure 13. Eh-pH diagram for the system Fe-02-C02-H20, assuming that total dissolved
carbonate equals 103 mol/kg and total dissolved iron is 103 mol/kg at aqueous solid
boundaries. Also shown is the position of the aqueous/solid boundaries for amorphous Fe(OH)3
with pKsp = 37.1 and goethite with pKsp = 44.2. The figure shows that siderite (FeC03) is
metastable in the presence of goethite. After Whittemore and Langmuir, 1975.
43
-------
PH
Figure 14. Eh-pH diagram for the system S-02-H20 at 25° C, showing the fields of
predominance of the aqueous species and of elemental sulfur (S°) for ES(aq) = 103 mol/kg
at aqueous/S° boundaries. After Langmuir, 1997a.
44
-------
1.00 -
_c
W
Figure 15. Eh-pH diagram for the system Fe-02-S-H20 at 25° C, showing stability fields of
goethite (a-FeOOH), pyrite (FeS2), and monoclinic pyrrhotite (Fe7S8) for 2S(aq) = 102
mol/kg, and total carbonate 10"4 mol/kg. SFe(aq) = 10~6 and 10"4 mol/kg at aqueous/solid
boundaries. The diagram shows that dissolved iron occurs chiefly in sulfate complexes. From
Barnes and Langmuir, 1979.
45
-------
6.2.2 Aluminum, Beryllium, Strontium, and Barium
As shown in Table 12, the hard acid cations, Al, Be, Sr, and Ba occur in only one
oxidation state. Their least soluble solids (strongest bonds) are formed with hard bases such as
OH", S042", C032, and P043. Maximum Al(III) concentrations are generally limited by the
solubility of aluminum oxyhydroxide solids. Depending on the degree of crystallinity of Al(OH)3, the
maximum dissolved Al concentrations will range between the solubility of gibbsite, the least soluble form
, and the solubility of amorphous Al(OH)3, the most soluble form (Figure 16). The solubility of
amorphous aluminum hydroxide is about 0.17 mg/L at pH 6.5, and 6 mg/L at pH 5. If the
aluminum in soils or water is derived from the leaching of the crystalline gibbsite, its
equilibrium concentration is 0.34 |ig/L at pH 6.5 and 12 |ig/L at pH 5. These calculations suggest
that the Al concentration of 0.07 mg/L in average world rivers (Table 4) that have a pH near 7
must be largely in suspension, probably in the colloidal size range (particles less than ~10"5 m).
Figure 16 also suggests that high Al concentrations in soils—concentrations toxic to many
plants—will generally not be found except in acid soils.
It is unclear what solid or solids control maximum beryllium concentrations, but the least
soluble Be phases for which there are such data are P-Be(OH)2 and beryllium silicate (Be2Si04)
(Bodek et al., 1988). The computed solubility of P-Be(OH)2 decreases from about 7.8 mg/L Be at
PH 5 to 1.0 |ig/L at pH 9 (Figure 17). The silicate is considerably less soluble, with a solubility
of about 7.2 |ig/L at pH 5, decreasing to 0.006 |ig/L at pH 8-9. These values may be compared to
the median Be concentration in surface and ground waters, which is 5 |ig/L (Table 1).
The least soluble minerals of Sr2+ and Ba2+ are strontianite (SrC03), celestite (SrS04), and
barite (BaS04). The solubility product of strontianite is 10"9 27, that of celestite is 10"6 63, and for
barite Ksp = 10"996 (Nordstrom et al., 1990). At a sulfate concentration of 96 mg/L, PHREEQC
modeling calculations give a barite solubility of 32 |ig/L Ba. This value is not far from the
median Ba concentration of 20 |ig/L in surface and ground waters (Table 1), suggesting that
barite solubility must often limit maximum Ba concentrations in soils and natural waters.
Modeling calculations indicate that celestite and strontianite are too soluble to limit Sr
concentrations in general, so that strontium is most often limited by coprecipitation with the Ca
carbonates or adsorption by clays.
46
-------
Ml
3
a
<
(a)
"o
a
<
00
o
(b)
Figure 16. Solubility of (a) amorphous Al(OH)3, Ksp = 10312 and (b) gibbsite (Al[OH]3), Ksp
_ jq-33.9^ as a function of pH at 25° C. Also shown are lines indicating the solubility
contributions of Al3+ and individual Al-hydroxy complexes. From Langmuir, 1997a.
47
-------
PH
Be(0H)2(a) -B- alpha-Be(OH)2(c) beta-Be(0H)2(c)
Figure 17. Solubility of three forms of Be(OH)2 as a function of pH. Diagram computed using
the geochemical model PHREEQC (Parkhurst and Appelo, 1999).
48
-------
6.2.3 Cadmium, Zinc, Cobalt, Nickel, and Lead
These metals are classified as either soft or borderline soft. Their metal sulfides are quite
insoluble at low Eh, even at low metal and sulfide concentrations (Figures 18 through 23). For
example, CdS (greenockite) precipitates and has a large stability field even for Cd = 11 |ig/L
(10~7 mol/kg) and total sulfur at 0.3 mg/L. The insolubility of the sulfides of Zn, Co, Ni, and Pb
is similar. Other important Cd and Zn minerals include their carbonates, although the carbonates
are relatively soluble at pH values below 8. Pure metal-containing mineral phases generally do
not control the dissolved concentrations of Cd, Ni, or Zn in aerobic soils. These metals are more
often controlled through adsorption or coprecipitation by oxyhydroxides of iron, manganese, and
aluminum.
Lead is relatively immobile in soils, sediments, and ground waters. This reflects its
strong tendency to be adsorbed by Fe and Mn oxides, but also the insolubility of a number of
lead minerals including lead hydroxycarbonate (Figure 22), which limits lead concentrations in
some public water systems, and pyromorphite (Figure 23), which controls lead concentrations in
some soils adjacent to highways affected by road salt and leaded gasoline exhaust.
6.2.4 Manganese
Unlike the other metals of concern, in most cases except aluminum, manganese is often
precipitated in soils and sediments as manganese minerals. Environmentally, the most important
of these minerals are the Mn(III/IV) oxides. Figure 24 shows that these phases are stable and
relatively insoluble in oxidized systems above pH 6-7. The Mn oxides are often stronger
sorbents of trace metals than is HFO. Suarez and Langmuir (1975) found that most of the Co, Ni,
Cu, Zn, Pb, Ag, and Cd in a Pennsylvania soil were associated with Mn and Fe oxides. The Mn
oxides held the highest metal amounts relative to their abundance. Rhodochrosite is an important
Mn mineral, but only under reducing conditions. Manganese does form a sulfide, but it has a
small stability field and is a rare mineral.
6.2.5 Silver, Copper, and Mercury
Eh-pH diagrams for these metals are shown in Figures 25 through 27. As soft metals,
silver and mercury form strong complexes with borderline soft CI ion, which may dominate the
solution chemistry of these metals. Silver and mercury form even stronger complexes with the
soft halogens bromide and iodide. Silver, copper, and mercury are highly insoluble in reduced
environments, where they precipitate as metals or as sulfides. In the absence of ligands other
than OH" and depending on the copper concentration, the solubility of copper above pH 7-8 can
be very low, due to the precipitation of tenorite (CuO). Moreover, in the presence of abundant
carbonate, relatively insoluble Cu2+ carbonate minerals can precipitate.
Bodek et al. (1988) offer a useful summary of the behavior of mercury. Hg(II) is usually
complexed—in pure water as Hg(OH)2°, and at chloride concentrations typical of fresh waters
49
-------
(<10~2 mol/kg) as HgCl2°. Both Hg(II) and Cu(II) form strong humate complexes, so that in soils
>99.9% of the metals may be complexed.
pH
Figure 18. Eh-pH diagram for the system Cd-02-C02-S-H20 for SCd = 107 mol/kg, SS =
105 mol/kg, and total carbonate 102 and 103 mol/kg. From Barnes and Langmuir, 1978.
50
-------
Figure 19. Eh-pH diagram for the system Zn-02-C02-S-H20, assuming that EZn = 106 and
10"4 mol/kg, EC = 10"3 mol/kg, and ES = 10"3 mol/kg. After Brookins, 1988.
51
-------
6 8 10 12 14
pH
Figure 20. Eh-pH diagram for the system Co-02-C02-S-H20, assuming that ECo
mol/kg, EC = 10"3 mol/kg, and ES = 10"3 mol/kg. After Brookins, 1988.
52
-------
-c
Uj
- - 1
.75 -
N(*2
25 -
- .25
.5 -
^ "T
, Y-.C-0
J \~!x - ' i~ i L . •
\ ** J\ ' ' v* < v*" ' ^ ' N
•-V¦ ^,.v'V.I~
;'mW/ (OH)5-
0 .«» >4 * v */ / «" \ I . 1
''2 *-"Mv 1 ' s- 'TS
OrV^v,-
x l ' > * -
N15S4
N i S f>*
¦« •- " , '
N i 3S2
J L_
L . , ~ I K
¦-j-X-i -r
2.
3.
4 .
6.
7.
e.
R.
IB.
11 .
12
pH
Figure 21. Eh-pH diagram for the system Ni-02-C02-S-H20, assuming that ENi = 10'
mol/kg and SS = 10"5 mol/kg. After Barnes and Langmuir, 1978.
53
-------
pH
Figure 22. Eh-pH diagram for the system Pb-02-C02-S-H20, assuming that SPb = 106 and
10"6'5 mol/kg, EC = 10"4 mol/kg, and 2S = 10"5 mol/kg. Diagram shows the stability fields of
lead hydroxycarbonate and galena (PbS). After Barnes and Langmuir, 1978.
54
-------
pH
Figure 23. Eh-pH diagram for the system Pb-02-P04-S-H20, assuming that 2Pb = 108,
107, and 106 mol/kg at solid/liquid boundaries; SS = 105 mol/kg; and SP04 = 106 mol/kg.
Diagram shows the stability fields of the lead phosphate, pyromorphite, and galena (PbS). After
Barnes and Langmuir, 1978.
55
-------
6 8 10 12 14
pH
Figure 24. Eh-pH diagram for the system Mn-02-C02-S-H20, assuming that EMn = 106
mol/kg, 2C = 10"3 mol/kg, and ES = 10"3 mol/kg. MnS is the very rare mineral alabandite. After
Brookins, 1988.
56
-------
6 8 10 12 14
PH
Figure 25. Eh-pH diagram for the system Ag-02-Cl-S-H20, assuming that EAg = 108
mol/kg, 2C1 = 103 5 mol/kg, and ES = 103 mol/kg. Diagram shows the importance of Ag
chloride complexing and the large stability field for metallic silver. After Brookins, 1988.
57
-------
Figure 26. Eh-pH diagram for the system Cu-02-S-H20, assuming that ECu = 10"6 mol/kg
and SS = 102 mol/kg. After Drever, 1997.
58
-------
pH
Figure 27. Eh-pH diagram for the system Hg-02-Cl-S-H20, assuming that EHg = 106 9
mol/kg, 2C1 = 10"3 mol/kg, and 2S = 10"3 mol/kg. Diagram shows the importance of Hg-Cl
complexing. After Bodek et al., 1988.
59
-------
Mercury is biologically methylated only in environments low in sulfide. The observed
decreases in biological methylation at high sulfide concentrations are generally attributed to the
precipitation of HgS(s). Alternatively, Benoit et al. (1999, 2001) have suggested that the
reduction of biological methylation at high sulfide concentrations is associated with a decrease in
Hg uptake by sulfide-reducing bacteria, due to decreases in the concentrations of the neutral
complexe HgS°. The stable methyl mercury species in fresh waters is CH3HgOH. The methyl
mercury cation, CH3Hg+, complexes with ligands containing C, P, O, N, and the halogens, and
forms very stable complexes with sulfur-containing ligands. In oxidized fresh waters, the Hg(II)
methyl hydroxo and methyl chloro complexes predominate.
Mercury and copper are strongly adsorbed by organic matter. All three metals are also
strongly adsorbed by Fe(III) and Mn oxides, and secondarily by clays. Sorption of mercury is
very fast and practically irreversible (Bodek et al., 1988). By inhibiting mercury sorption, Hg-Cl
complexing helps to mobilize the metal. Dimethyl mercury is very insoluble in water and tends
to be volatilized from soils.
6.2.6 Thallium
In oxidizing environments, dissolved thallium is limited to concentrations of less than 2
|-ig/L (10~8 mol/L) by the low solubilities of the thallium 4+, 3+ and 2+ oxides (Fig. 29). The
solubility product of Tl(OH)3, which should be more soluble than T1203 in Figure 28, is 10"45 2,
which makes this phase highly insoluble between pH 4 and 10. Only in highly reducing systems
as Tl+ is the metal soluble. Tl+ is a weak complex former. The sulfide of Tl+ is stable only above
pH 12. Limited data indicate that T1 is strongly adsorbed by montmorillonite clays and
manganese oxides (Bodek et al., 1988).
6.2.7 Arsenic, Antimony, Chromium, Molybdenum, Selenium, and Vanadium
These six elements occur chiefly as oxyanions in oxidizing environments. As such they
are relatively mobile, although they are adsorbed by ferrihydrite under acid to neutral conditions.
Consistent with Figure 6, on a number of different sorbent phases, the decreasing order of
adsorption is usually As > Cr > Mo > Se « S (Bodek et al., 1988). Arsenic and phosphorus
chemistry are similar under oxidizing but not reducing conditions. Figure 29 shows the stability
fields of dissolved As(V) and As(III) species in pure water. With the addition of sulfur, insoluble
As sulfide minerals are precipitated at low Eh (Figure 30) in the field otherwise occupied by
arsenious acid species.
As suggested by the absence of a solid phase field in Figure 31, antimony may be too
soluble for its concentration to be limited by mineral precipitation. The least soluble Sb phase is
probably Sb(OH)3, which does not precipitate until SSb concentrations exceed about 10"6 67
mol/kg, or 26 |ig/L (Barnes and Langmuir, 1978). Antimony is a weak complex former, except
for its reaction with sulfur at low Eh to form sulfide complexes.
60
-------
6 8 10 12 14
PH
Figure 28. Eh-pH diagram for the system Tl-02-S-H20, assuming that 2T1 = 108 and 106
mol/kg and SS = 10"3 mol/kg. The diagram shows the insolubility of the higher-valent T1 oxides
under oxidizing conditions. After Brookins, 1988.
61
-------
pH
Figure 29. Eh-pH diagram for the system As-02-H20 showing the relative stabilities of
arsenate (arsenic acid, As[V]) and arsenite (arsenious acid, As[III]) species under oxidizing
and reducing conditions.
62
-------
pH
Figure 30. Eh-pH diagram for the system As-02-S-H20, assuming that SAs = 106 mol/kg
and SS = 10"3 mol/kg. The diagram shows that in the presence of reduced S, the As sulfides
control As solubility and take over the stability field of dissolved As(III) species in Figure 25.
After Brookins, 1988.
63
-------
pH
Figure 31. Eh-pH diagram for the system Sb-02-S-H20, assuming that SSb = 108 mol/kg
and SS = 10"2 and 10"5 mol/kg. After Barnes and Langmuir, 1978.
64
-------
Chromate (Cr[VI]) species (Figure 32) predominate and are highly mobile in oxidized
systems, except for their tendency to be adsorbed, particularly by Fe(III) and Mn oxides below
pH 8. Chromate is in general weakly complexed. In contrast, Cr(III), which dominates in
reducing environments, is a strong, hard-acid substance that readily forms complexes. Cr(III)
complexes with hydroxyl, sulfate, organic ligands, and other species, which increases its stability
and thus raises the Cr(III)/Cr203 boundary to higher pH values. This increases the solubility of
Cr203. Organic matter, Fe2+, and H2S can reduce Cr(VI) to Cr(III). Manganese oxides in soil can
adsorb Cr(III) and oxidize it to Cr(VI). As shown in Figure 6, adsorption of chromate by HFO
decreases with increasing pH, whereas Cr(III) adsorption increases as the pH rises.
Molybdate is highly mobile in oxidized environments (Figure 33), except for its tendency
to be adsorbed by Fe(III) oxides below pH 7. Also under acid conditions, in soils and waters
high in Fe(III), molybdate may precipitate as ferrimolybdite (Fe2[Mo04]3[s]), which is stable up
to pH 5 for SFe(III) = SMo = 10"6 mol/kg (Barnes and Langmuir, 1978). Under reducing
conditions, Mo is immobilized by the low solubility of the Mo oxides and MoS2.
The redox behavior of selenium (Figure 34) is similar to that of sulfur (Figure 14) in that
(1) Se(VI) oxyanions predominate under oxidizing conditions, (2) the element has a stability
field under reducing conditions, and (3) metal cations react with Se(2-) to form insoluble
selenides. Selenate is highly stable and not readily reduced by H2S or Fe2+. The very low
solubility of native Se indicates that it is an important sink for dissolved Se in reducing
environments. Se(2-) forms very insoluble metal selenides with the following -log Ksp values:
26.0 (FeSe), 60.8 (Cu2Se), 48.1 (CuSe), 29.4 (ZnSe), 35.2 (CdSe), 64.5 (HgSe), and 42.1 (PbSe)
(Bodek et al„ 1988).
Selenite salts are less soluble than selenate salts. Selenite and selenate are both strongly
adsorbed by the Fe and A1 oxyhydroxides. Phosphate and sulfate effectively compete with
selenite and selenate for sorption sites on Fe oxides.
Like chromium, vanadium occurs as an oxyanion at high Eh values and in cationic form
under reducing conditions (Figure 35). Vanadate probably forms an insoluble precipitate with
Fe(III) below pH 7, although the evidence is limited. The V(III) and Fe(III) oxides are
isostructural, suggesting that V(III) may substitute for Fe(III) in the iron oxides (Rai and
Zachara, 1984). A problem with this idea is the low Eh stability of V(III) versus the higher Eh
stability of the Fe(III) oxides. In soils, the distribution of V closely follows that of secondary
Fe(III) oxides, probably because of V adsorption by the oxides. Vanadium is readily reduced and
mobilized by soil organic matter, even under oxidizing conditions.
65
-------
2 4 6 8 10 12 14
PH
Figure 32. Eh-pH diagram for the system Cr-02-H20, assuming that the concentration of
SCr = 106 mol/kg at solid/liquid boundaries. After Brookins, 1988.
66
-------
pH
Figure 33. Eh-pH diagram for the system Mo-02-S-H20, assuming that EMo = 106 mol/kg
and SS = 10"2 and 10"5 mol/kg. After Barnes and Langmuir, 1978.
67
-------
pH
Figure 34. Eh-pH diagram for the system Se-02-H20. Solid/liquid boundaries drawn for SSe
= 10"7 and 10"9 mol/kg. After Barnes and Langmuir, 1978.
68
-------
PH
Figure 35. Eh-pH diagram for the system V-02-H20, for dissolved EV<104 mol/kg (11.5
mg/L as V04). After Langmuir, 1978.
69
-------
7. SOIL TRACE METAL TRANSFER TO PLANTS AND THE BIOAVAILABILITY OF
SOIL-METALS
The potential risk that trace elements in soils pose to the feed- and food-chain has been
intensively examined during the last 35 years. One purpose of that investigation has been to
understand the risk from application of biosolids (municipal sewage sludge) and other metal
contamination sources to soil.
During this period, the "soil-plant barrier" concept was introduced to communicate how
metal addition, soil chemistry, and plant chemistry affected the risk to animals from metals
mixed in soil (Chaney, 1980). Reactions and processes that take place at the soil-plant barrier are
influenced by the following: (1) solid metal sources (e.g., Fe, Al, and Mn oxyhydroxides and
organic matter) may have adsorptive surfaces that influence soil chemistry; (2) adsorption or
precipitation of metals in soils or in roots limits uptake-translocation of most elements to shoots;
(3) the phytotoxicity of Zn, Cu, Ni, Mn, As, B, Al, F, etc., limits residues of these elements in
plant shoots to levels chronically tolerated by livestock and humans; and (4) food-chain transfer
of an element may not constitute a risk, but the direct ingestion of the contaminated soil may.
A summary of metal tolerances by plants and livestock is presented in Table 13. It should
be noted that the NRC (1980) committee which identified the maximum levels of metals
tolerated by domestic livestock based its conclusions on data from toxicological-type feeding
studies in which soluble metal salts had been mixed with practical or purified diets to examine
animal response to the dietary metals. If soil is incorporated into diets, metal solubility and
bioavailability will probably be much more limited than in the tests relied on by NRC (1980).
For example, it has been noted that until soil exceeds about 300 mg Pb kg"1, animals show no
increased body burden from ingesting the soil (Chaney and Ryan, 1993). Other metals in
equilibrium with poorly soluble minerals or strongly adsorbed in soils are often much less
bioavailable than they would be if they occurred in more soluble salts.
The chemistry of metals in soils is affected by the presence of ions which can cause
precipitation of the metal, organic matter and sesquioxides which can adsorb metals, redox
changes which affect the chemical species of the metal present, and similar factors discussed in
the section about water chemistry. Soils are usually in a relatively restricted pH range of 5.5 to 8
for high-producing soils, and as wide as 4 to 9 in nearly all soils in the general environment.
Industrial contamination with acids or bases can cause lower and higher pH than this practical
natural pH range: as low as pH 2 and as high as pH 11. Usually, such contaminated soils are
barren due to infertility or phytotoxicity of soil elements affected by extreme pH.
Many elements (e.g., Si, Ti, Fe, Pb, Hg, Al, Cr3+, Ag, Au, Sn, Zr, and other elements that
serve as a label for soil contamination of plants and diets) are so insoluble in oxidized soils
between pH 5.5 and 8 that they are not a risk even when soils with relatively high concentrations
are ingested by livestock. When present at very high concentrations, elements that may pose a
70
-------
risk because of absorption in the intestine when contaminated soils are ingested include F, Pb,
As, and Zn. This process is also important for animals that consume soil biota such as
earthworms. The earthworm is consumed without depuration on internal soil, giving exposure to
high levels of dietary soil—perhaps 35% of dry weight.
Table 13. Maximum Tolerable Levels of Dietary Minerals for Domestic Livestock in
Comparison with Levels in Forages
Element
"Soil-
Plant
Barrier"
Level in Plant Foliage8
Maximum Levels Chronically Toleratedb
Normal
Phytotoxic
Cattle
Sheep
Swine
Chicken
mg/kg"1 dry foliage
mg/kg"1 dry diet
As, inorg.
Yes
0.01-1.0
3-10
50
50
50
50
B
Yes
7-75
75
150
(150)
(150)
(150)
Cdc
Fails
0.1-1
5-700
0.5
0.5
0.5
0.5
Cr3+
Yes
0.1-1
20
(3,000)
(3,000)
(3,000)
3,000
Co
Fail?
0.01-0.3
25-100
10
10
10
10
Cu
Yes
3-20
25^10
100
25
250
300
F
Yes?
1-5
-
40
60
150
200
Fe
Yes
30-300
-
1,000
500
3,000
1,000
Mn
?
15-150
400-2,000
1,000
1,000
400
2,000
Mo
Fails
0.1-3.0
100
10
10
20
100
Ni
Yes
0.1-5
50-100
50
(50)
(100)
(300)
Pbc
Yes
2-5
-
30
30
30
30
Se
Fails
0.1-2
100
(2)
(2)
2
2
V
Yes?
0.1-1
10
50
50
(10)
10
Zn
Yes
15-150
500-1,500
500
300
1,000
10,00
aBased on literature summarized in Chaney et al., 1983.
bBased on NRC, 1980. Continuous long-term feeding of minerals at the maximum tolerable levels may cause
adverse effects. NRC estimated the levels in parentheses by extrapolating between animal species when data were
not available for an animal.
CNRC based the maximum levels tolerated of Cd or Pb in liver, kidney, and bone in foods for humans rather than
simple tolerance by the animals. Because of the simultaneous presence of Zn, Cd in animal tissues is less
bioavailable than Cd salts added to diets.
71
-------
Freshly applied metal salts are not in equilibrium with soil and have a greater
phytoavailability than they would exhibit upon equilibrating with the soil over time. The
phytoavailability and bioavailability of metals may also be reduced if the metals are adsorbed,
chelated, or precipitated before ingestion by children or grazing livestock.
An example of the interaction phenomenon is the toxicity of biosolids or manure-applied
Cu or Zn to animals. Cu-deficiency-stressed animals are more sensitive to dietary Zn than
animals fed Cu-adequate diets, but biosolids-fertilized crops are not low in Cu, so ordinarily Zn
phytotoxicity protects all livestock, including the most sensitive ruminants. Similarly, Cu
toxicity to sensitive ruminant animals is substantially reduced by increased dietary levels of Zn,
Cd, Fe, Mo, and S04 or sorbents such as soil organic matter. In contrast with the predicted
toxicity from Cu in ingested swine manure or biosolids, reduced Cu absorption has been
observed unless ingested biosolids exceed about 1,000 milligrams Cu kg"1 or manure is fed at
50% of diet.
An important interaction which reduces risk from Cd is the normal 1 Cd:100 Zn ratio of
geogenic Cd. Although culture of crops in strongly acidic soils allows uptake of increased levels
of Cd and Zn, the presence of Zn in the crops inhibits uptake of Cd. Because Cd and Zn are
taken up from acidic soils at about the ratio of the total metals in the soil, Zn phytotoxicity (at
500 mg Zn kg"1) serves as a natural maximal limit on crop Cd, and plant (intrinsic) Zn inhibits
absorption of plant Cd in animals. These factors very significantly reduce soil Cd risk compared
to risks observed in toxicological studies that tested risks from Cd salts.
The potential for plant uptake to allow metals to be transferred to feed- and food-chains
has been extensively studied. Uptake slopes measured in pot studies are much higher than those
found in the field, so greenhouse or growth chamber studies cannot be used to estimate
environmentally relevant uptake slopes. Unfortunately, the majority of the valid field studies
were conducted under poor management conditions rather than recommended "Best
Management Practices" for managing mobile nutrients and minimizing threats to the
environment. It is very clear that strongly acidic soils increase plant uptake of Zn, Cd, Ni, Mn,
and Co, and increase the potential for phytotoxicity from Cu, Zn, and Ni. Alkaline soil pH
increases uptake of Mo and Se, while Pb and Cr are not absorbed to any significant extent at any
pH (Chaney and Ryan, 1993).
For strongly adsorbed metal cations and in the absence of artifacts due to the research
methodology used, plant metal uptake does not increase linearly with biosolids-applied metal
concentrations, but instead saturates at elevated metal concentrations. First, comparison of
application rates is only valid after the system has equilibrated for a given period, for example,
because of the rapid biodegradation of biosolids-applied organic matter. When biosolids are
applied at high rates, uptake can be increased for several years due to the formation of
biodegradation by-products which increase metal diffusion and convection to the roots. The
effect is more significant for Poaceae than other species, perhaps due to the role of
phytosiderophores in metal uptake. Second, soil pH levels should be equal across rates studied;
72
-------
co-variance of soil pH should be used to correct for unequal soil pH. Studies by Bell et al. (1988)
found a strong plateau response when pH co-variance was incorporated in the data assessment.
Third, the metal concentration in the source applied affects the slope of metal uptake (or the
increase above the background plant metals when the plateau is reached). Higher metal
concentration in the source means higher phytoavailability at equal metal applications (Jing and
Logan, 1992). The presence of metal-adsorbing sesquioxides in many biosolids decreases the
slope or increment at the plateau. The natural limitations on metal uptake and bioavailability,
including the plateau response, prevent toxic levels of most metals being reached in plants used
as food or feed.
Protection of wildlife is similar to that of livestock; wildlife's diet can consist entirely of
plants grown on a contaminated site. In cases involving wildlife in unmanaged ecosystems,
maximal plant residues may exceed those allowed on managed farmland—wildlife may eat sick
plants which would not be harvested by a commercial grower. Evaluation of a rich literature on
wildlife exposure to metal-contaminated soils indicates that animals which consume earthworms
are the highly exposed individuals (Brown et al., 2002): 35% of an earthworm is soil, on a dry
matter basis. But the soil in the earthworm can adsorb metals, reducing their bioavailability. Of
the metals, only mercury in the methyl-mercury form is actually biomagnified, and very little of
the mercury in soils is in that form. Se can be incorporated into proteins and also be
biomagnified. Other metals are instead "biominified," according to many studies (Beyer and
Stafford, 1993). Nearly all of the metals ingested in forage materials or earthworms—in some
cases, >99.9%—are excreted. Thus, the increase in risk with increasing trophic level seen with
chlorinated hydrocarbons has not been seen with metals in soils .
Each element must be considered separately because of its unique chemistry. For
example, arsenate is more strongly adsorbed than arsenite; when a soil is flooded to grow rice,
soil microbes can reduce arsenate to arsenite and the higher concentration of dissolved arsenite
can be phytotoxic to rice in more highly contaminated soils. Generally, for As, the most sensitive
crop is rice because soils are reduced for rice production. Most elements have little potential for
redox change with change in the redox status of soils. Besides forming arsenite, reduced soils
also form less soluble forms of U. Reduced soils can form sulfide, and sulfide forms low-
solubility compounds with most of the metals of concern in soils, including Pb, Zn, Cd, Cu, and
Ni (Table 12). These low-solubility species can become deficient for rice. Upon oxidation of the
soil, sulfide is quickly oxidized and the metals are returned to more normal equilibrium reactions
of aerobic soils.
Much research has been conducted using multi-element analysis of plants and animals
exposed to soil metals. One group of studies used neutron activation analysis of biosolids, soils,
crops, and animals (e.g., Chaney et al., 1978). More recently, ICP-MS and other very sensitive
analytical methods have been used to examine soil solution and soil-plant transfer of 60 elements
as a function of soil pH (Tyler and Olsson, 2001a, 2001b). These studies provide evidence which
further supports the concept of the soil-plant barrier. A few elements may require further
evaluation in connection with industrial sources that can—despite the fact that they are normally
73
-------
poorly absorbed from soils—introduce them into soils at very high concentrations. When a high
metal concentration source reaches soil, the soil processes that limit risk may be overwhelmed,
at least for elements that are weakly accumulated by plants (e.g., Tl, Be).
8. LINKING METAL SPECIATION AND METAL TOXICITY: THE BIOTIC LIGAND
MODEL APPROACH
8.1 Overview of Biotic Ligand Model Development
The biotic ligand model (BLM) approach is used to predict metal toxicity by linking
chemical equilibrium models (which estimate metal speciation in solution) to metal
complexation at the biological surface. The framework of the BLM approach synthesizes over 30
years of research on the geochemistry and toxicity of trace metals (Paquin et al., 2002). The first
general descriptions of the approach were provided by Pagenkopf (1983), with his gill surface
interaction model, and by Morel (1983), who formulated a free ion activity model. About 10
years later, Playle et al. (1992, 1993a, 1993b) provided additional information on the effect of
anionic complexation and competition with other cations on trace metals binding to fish gills.
These data were later integrated with toxicity data by Di Toro et al. (2001), who developed a
version of the BLM model to predict the acute toxicity of copper and silver to several freshwater
species.
8.2 Metal Speciation Estimation
A key component of the BLM approach is the estimation of trace metal complexation by
inorganic and organic ligands in solution. While the use of chemical equilibrium models for
complexation by inorganic and synthetic ligands is relatively straightforward, predicting
complexation by dissolved natural organic matter is more challenging, given such matter's
heterogeneous character. Basically, two different estimation approaches have been used in the
BLM. The first (e.g., Di Toro et al., 2001), uses a modified CHESS speciation code (Santore and
Driscoll, 1995) that includes the WHAM V (Tipping, 1994) approach to model complexation by
organic matter. WHAM V is based on an extensive dataset for natural organic matter and
considers mono- and bidentate complexation and the influence of electrostatic interactions on
binding. Proton binding and competitive metal binding are described for two types of acid
groups. Each group is assigned a range of intrinsic pKs that are given median and range values.
The second approach uses the MINEQL+ (Schecher and McAvoy, 1994) speciation code, and a
single binding constant measured for a natural organic matter sample (e.g., McGeer et al., 2000).
The other key component of metal speciation by BLM models is the complexation by the
biotic ligand, for example by the gills of fish in the original models. Two approaches are
currently used to estimate the amount of metal associated with a biological surface. The first,
which is more mechanistically based, relies on measured conditional binding constants of the
metals to the biological surface (e.g., MacRae et al., 1999). The second BLM approach is more
empirical and uses complexing by a hypothetical biologically active surface ligand as a fitting
74
-------
parameter to relate metal speciation to the observed metal toxicity (e.g., De Schamphelaere and
Janssen, 2002). In the later approach, in addition to the free metal ion, binding of metal
complexes such as CuOH+ (De Schamphelaere and Janssen, 2002) and AgCl0 (Paquin et al.,
1999) may also be considered.
In toxicity testing at relatively high metal concentrations, the measured or estimated
binding constants are often roughly constant for different biological ligands. This is presumably
because surface complexation is associated with similar low-affinity chemical moieties in each
case. For example, Heijerick et al. (2002a) noted the similarity of published binding constants
for zinc: experimentally derived log KZnBL values of 5.3 to 5.6 (Alsop and Wood, 2000) and 5.1
(Galvez et al., 1998) for rainbow trout, and estimated values of 5.4 for steelhead trout (Cusimano
et al., 1986) and 5.3 for Daphnia magna (Heijerick et al., 2002b). Conditional stability constants
for copper are also almost constant for fish gills, with experimentally derived log KCuBL values of
7.5 for rainbow trout and 7.3 for brook trout (MacRae et al., 1999) and 7.4 for fathead minnows
(Playle et al., 1993a). But the binding constants are different for other organisms, with an
estimated log KCuBL of 8.0 for Daphnia magna (De Schamphelaere et al., 2002) and expected
differences for non-chitinous invertebrates such as Lumbriculus variegatus (Meyer et al., 2002).
Some of these differences may also be due to differences in experimental and model design, as
there are no standardized methodologies and procedures for deriving data for conditional
stability constants.
The effects of pH on metal toxicity can be illustrated with the BLM model, which can
consider changes in aqueous metal speciation and competition of H+ with metals for binding
sites at the biological surface. For example, using the BLM approach (Di Toro et al., 2001), we
estimated the competitive effect of H+ from pH 5 to pH 9 on copper and silver binding to the
gills of fathead minnows in Lake Superior water (Figure 36). The effect shown in the figure only
relates to competition for binding to the gill surface, not to changes in copper or silver speciation
in solution. This is because total dissolved metal concentrations were increased with pH so that
free metal concentrations remained constant. In a toxicologically relevant metal concentration
range and pH, the direct competition of protons for binding to the gill surface is weak due to the
higher affinity of Cu2+ and Ag+ ions for binding sites. Neither copper nor silver is completely
displaced from the surface even at pH 5. Copper associated with the gills almost doubles from
PH 5 to pH 7, and competition between Cu2+ and CuOH+ becomes significant at pH 8 and above.
In contrast to copper, there is almost no predicted competition of H+ with silver. In fact, no
experimental data support competition of H+ with silver; the small reduction of Ag+ bound to the
gill at pH below 6 is presumably an artifact related to the use of an empirically derived lower
proton binding constant for the silver BLM (Paquin et al., 1999).
8.3 Validation Studies
While the toxicity of trace metals can vary by several orders of magnitude, studies that
have validated the BLM approach with independent datasets indicate that BLM-predicted
toxicity is generally within 2 times the observed toxicity. De Schamphelaere et al. (2002)
75
-------
validated their BLM model for predicting acute copper toxicity to Daphnia magna using
independent toxicity test data conducted with 25 reconstituted media representative of European
natural waters and with 19 spiked European natural waters. Heijerick et al. (2002b) conducted a
similar validation using 17 reconstituted media to predict zinc acute toxicity to Daphnia magna.
Santore et al (2001) have also tested their BLM model for acute copper toxicity to fathead
minnows by comparing their prediction to the toxicity test results for two effluent-impacted
streams in the United States. Finally, the model developed by McGeer et al. (2000) to predict
acute silver toxicity to rainbow trout also accurately predicted the toxicity measured in 31
toxicity test results obtained from 10 different independent studies. The BLM approach has thus
been validated and shown to be mechanistically based for the prediction of trace metal acute
toxicity to certain indicator species.
O)
o
E
V)
5
c
ro
"S
¦o
ro
d
E
d
o
o
<
14
12 -
10 -
8 -
6 -
4 -
2 -
Silver : [Ag+] = 30 nM
Copper : [Cu2+] = 12 nM
10
PH
Figure 36. Effect of pH on the BLM-predicted (HydroQual, 2002) Cu2+ and Ag+
accumulation in the gills of Pimephales promelas. Lake Superior average composition
(Erickson et al., 1996) and freshwater criteria maximum total dissolved concentrations (U.S.
EPA, 2002b) of 205 nM for copper and 32 nM for silver assumed at pH 5. Total dissolved
concentrations were increased with increasing pH so that free copper and free silver
concentrations would remain constant, [Cu2+] = 12 nM and [Ag+] = 30 nM. Silver BLM: Log
K,
H+-gills
: 4.3 and Log K
Ag+-gills
7.3. Copper BLM: Log K,
H+-gills
5.4 and Log K,
Cu2+-gills
7.4.
76
-------
8.4 Current Limitations and Future Research
The BLM approach's limits are mostly related to its inherent assumption that (1) trace
metal speciation and metal uptake are in chemical equilibrium and (2) the biotic ligand remains
unmodified through exposure. The model also assumes that metal uptake or expression of the
biological response is the rate-limiting step. Trace metals in surface waters and exposure media
are not always in chemical equilibria, especially with natural organic matter. Campbell (1995)
reviewed other exceptions to the BLM's assumptions, such as direct uptake through passive
diffusion of lipophilic complexes or kinetically controlled uptake of some trace metals because
of their fast uptake. The assumption that the affinity of the biotic ligand is constant has also been
challenged. There are physiological responses to metal exposures that might affect metal uptake.
In addition, prior sublethal exposures to metal or the quality of the diet have been shown to
affect the binding of metal to fish gills (e.g., Alsop and Wood, 2000; Szebedinszky et al., 2001).
Variation in the affinity of the biotic ligand for metals will be especially important for the
prediction of chronic toxicity.
Extensive current research is focused on predicting chronic toxicity and metal mixture
toxicity. Applying the BLM model to chronic toxicity is not straightforward, since the
physiological mechanisms involved are likely to be much different. Similarly, predicting the
toxicity of metal mixtures that have different modes of action is difficult. For example, Cu and
Ag affect ionoregulation, whereas metals such as Ni have respiratory effects. Accounting for
such differences will require further refinement of the BLM model. However, it can be expected
that the toxicity of metal mixtures in which the metals have the same mode of action can be
predicted using the current BLM framework.
9. ATMOSPHERIC CHEMISTRY OF THE METALS
Metals are ubiquitous components of particulate matter in ambient air and are emitted by
numerous natural and anthropogenic sources. Richardson (2002) included volcanic eruptions and
emissions, entrainment of soil and dust, entrainment of sea salt spray, and natural forest fires as
significant metals emission sources. Most metals are removed from the atmosphere by
deposition and have relatively short atmospheric half-lives. Table 14, abstracted from Seinfeld
and Pandis (1998), presents average atmospheric residence time and airborne concentrations for
some metals in California.
77
-------
Table 14. Atmospheric Metals
Metal
Residence Time (days)
Average Concentration
Arsenic
Unknown
2.4 ng/m3
Cadmium
7
1-2.5 ng/m3
Nickel
Unknown
7.3 ng/m3
Beryllium
10
0.11-0.22 ng/m3
Lead
7-30
270-820 ng/m3
Mercury
0.3-2 years
0.37-0.49 ppb
From Seinfeld and Pandis, 1998.
A substantial part of the atmospheric chemistry of metals takes place in the aqueous
phase, where it is not significantly different from aqueous chemistry in other media. Constructs
described elsewhere in this report (e.g., pH-Eh diagrams and geochemical computer models)
may be used to determine metal speciation in atmospheric aquatic aerosols. However, the
application of such diagrams and models, which assume that metal species are in chemical
equilibrium, must be limited to describing metal speciation reactions that can equilibrate in
seconds or less. These include many acid/base and metal complex formation reactions, but not
most reactions involving adsorption, oxidation/ reduction, or mineral precipitation. Atmospheric
metal reactions that take longer to equilibrate must be studied in terms of their reaction kinetics
rather than equilibrium chemistry.
Seinfeld and Pandis (1998) present a detailed discussion of the acid-base chemistry of
atmospheric water that may be useful in modeling metal speciation. In addition, sorption to
particles, deposition, and gas phase redox chemistry may be important for some metals (Bodek
et al., 1988). Deposition is largely a function of the properties of particulates, not the sorbed or
incorporated metals. Redox chemistry takes place through oxidation with free radicals such as
OH or through direct photolysis of metal complexes (Finlayson-Pitts and Pitts, 1986). Due to the
oxidizing nature of the atmosphere, metals are often converted to their most oxidized stable
forms.
Metals have not been a major component of EPA air programs. Lead is listed as a criteria
pollutant under the Clean Air Act. A group of metals including antimony, arsenic, beryllium,
cadmium, chromium, cobalt, lead, manganese, mercury, nickel, and selenium are listed as
hazardous air pollutants under the Clean Air Act. Of these only a few have concerns related to
speciation. Arsenic is listed for "inorganic" and arsine forms; the listing for chromium makes a
distinction between Cr(III) and Cr(VI); mercury is listed for elemental, inorganic (apparently
divalent), and methyl forms; and nickel is listed as salts, refinery dust or subsulfide, and
carbonyl (U.S. EPA, 1994). Deposition of metals from the atmosphere is under consideration by
the Office of Water (U.S. EPA, 2003b). This latter program is primarily concerned with
deposition, speciation, and transformation after deposition of mercury, although lead and
78
-------
cadmium are involved to a lesser extent. The State of California has a comprehensive program
for monitoring airborne metals through its air toxics monitoring network.
With very few exceptions (e.g., hexavalent chromium, nickel refinery dust, mercury),
EPA does not consider bioavailability or speciation in risk assessment of atmospheric metals;
rather, risks are assessed on the basis of total or respirable airborne metal concentrations (U.S.
EPA, 1998). EPA assesses risks for lead with a set of pharmacokinetic models known as the
integrated uptake biokinetic (IEUBK) and Adult Lead models. Although these models explicitly
consider oral bioavailability of lead, they do not consider inhalation bioavailability.
Most sampling and analytical techniques published by EPA for metals in air are oriented
toward evaluation of particulate-phase total metals rather than metal species (U.S. EPA, 1999a).
These methods involve collection of either total or respirable particulate fractions with
subsequent analysis by X-ray fluorescence, atomic absorption, inductively coupled plasma,
proton-induced x-ray emission, or neutron activation analysis gamma spectroscopy techniques.
The one notable exception is a method for mercury (Method 10-5) that speciates vapor and
particulate forms. To the extent that metals are sorbed to particulate phases, analysis of
individual metal species can, at least theoretically, be accomplished by the same techniques used
to analyze those species in other solid media.
Of all the metals of interest, mercury has the most complex and best understood
atmospheric chemistry (Lindqvist, 1994; Munthe, 1994). In the gas phase, mercury is oxidized
by 03 and N02; aqueous-phase chemistry includes oxidation of elemental mercury by free
radicals such as OH and H02. Both vapor phase and aqueous atmospheric chemistry may involve
heterogeneous-phase components. EPA has made a substantial effort to evaluate the atmospheric
fate of mercury due to the requirements of the Clean Air Act. The Mercury Study Report to
Congress (RTC) (U.S. EPA, 1997) contains a complete qualitative and quantitative evaluation of
mercury's atmospheric fate based on the state of knowledge when the report was written. EPA is
also working with the University of Wisconsin to quantify mercury transformations in the
atmosphere, including depositional behavior.6
9.1 Discussion and Recommendations About Atmospheric Metal Chemistry
Metal speciation and chemistry have not been significant components of EPA's
evaluation and assessment of environmental metals. The impetus toward a more comprehensive
evaluation of metals must start with qualitative and quantitative toxicology. If there is a
significant difference among the various atmospheric metal species with respect to toxicology
(as is the case with divalent and elemental mercury), then expenditure of resources on
environmental fate models, development of analytical techniques, and monitoring may be
warranted. To some extent, the Mercury RTC (U.S. EPA, 1997) is a paradigm for a complete
multimedia exposure and risk assessment for metal species. The resources and management of
6 EPA Grant R829798. Project officer: Bill Stelz, NCER.
79
-------
the development of the RTC may be used as a benchmark by EPA for the development of similar
efforts for other metals.
10. DETERMINATION OF METAL SPECIATION IN WATERS AND SOILS
10.1 Unfiltered Versus Filtered Water Samples
The metals transported by surface and ground waters are present in true solution as ionic
or molecular species, and in suspended particles (cf. Stumm and Morgan, 1996; Langmuir,
1997a). The suspended particles can be colloidal in size and as small as 10 to 100 Angstroms
(10~9 to 10"8m). According to the widely used operational definition, metals that pass through a
0.45 |im (4.5 x 10"7 m) filter are dissolved. Because colloidal particles range in size from less
than 10"8 m to about 10"5 m, a significant fraction of particulate metal concentrations can pass
through a 0.45 |im filter and be listed as dissolved. For example, Kennedy and Zellweger (1974)
found that the concentration of colloidal-sized particles of Al, Fe, Mn, and Ti passing through a
0.45 |im filter could exceed the dissolved concentration of these metals by an order of magnitude
or more. (See also Bergseth, 1983; Puis and Barcelona, 1989; Pohlmann et al., 1994). Clark et al.
(1992) reported that the type of pump used to sample ground waters had a major effect on total
concentrations of Mn, Pb, and Zn, with higher turbidity and particulate metal concentrations
collected from centrifugal pumps than from slower-flow-rate bladder pumps. Furthermore, the
determination of dissolved metal concentrations is complicated by several potential sampling
and filtration artifacts. For example, Horowitz et al. (1992, 1996) have demonstrated that several
factors in addition to filter pore size can affect the dissolved metal measurements: the type of
filters, the volume of water filtered, the amount of suspended matter, etc.
Usually the metals of concern are associated with the smallest particle size fractions
(especially colloidal-sized) in soils and sediments which have the largest surface areas per
weight, rather than with larger particle sizes on a weight basis. A variety of methods are
available for separating dissolved and suspended metals. These include centrifugation, filtration,
ultrafiltration, ion-exchange chromatography, and organic extraction. These methods are
discussed in some detail by Salbu and Steinnes (1994).
The distinction between dissolved and particulate metal concentrations is important,
because reactions such as adsorption, precipitation, oxidation/reduction, and complexation
control metal amounts in true solution, but not the amounts in suspension. The latter are limited,
if at all, by chemical coagulation and/or filtration in porous media, or, if the particles are large
enough (larger than about 10"5 m), by gravitational settling. Metals in suspension are usually
adsorbed or coprecipitated with suspended Fe, Al, or Mn oxyhydroxides or adsorbed by
suspended organic matter or clays. Only rarely, as in streams affected by acid rock drainage or at
sources of metal contamination, do the metals of concern occur as pure metal oxyhydroxides or
metal salts.
80
-------
In public water supplies that are subject to pH control by lime addition, the addition of
flocculants, and/or filtration, suspended metal concentrations may be largely eliminated. But the
distribution systems that connect the utilities to their customers may contribute measurable
amounts of dissolved and suspended metals to drinking water supplies (e.g., Singh and Mavinic,
1991).
10.2 Unstable Parameters or Species Including Metal Redox Pairs
The reactions that control the mobility of the metals of concern in porous media are
usually strong functions of pH, and—except for Al, Ba, Be and Sr—also depend on the oxygen
content or Eh of the media (See Baedecker and Cozzarelli, 1992). An accurate assessment of
metals mobility at a specific site requires that the values of these parameters used in any
modeling calculation be the same as in the medium involved. Values of pH and Eh, temperature,
and gas pressures (e.g., the partial pressures of 02, C02, H2, CH4, and H2S) must be measured
immediately upon collection, or when possible fixed in the field for later analysis. Otherwise
their values will usually change rapidly due to temperature change, loss or gain of gases to or
from the atmosphere, and the growth of organisms such as bacteria and algae. For such reasons,
the pH of a water measured in the laboratory after a few weeks of sample storage at ambient
temperature will usually differ from its value in the field by ±0.2 to ±1 pH units (Langmuir,
1971, 1997a).
Methods of field analysis of pH and Eh are discussed in detail by Langmuir (1971) and
Wood (1976). (See also USGS, 1998.) Modern, battery-operated solid-state pH/millivolt meters
for field use have greatly improved the ease of pH and Eh measurement. If pH meters and
electrodes are properly calibrated using a double pH buffer check, field-measured values can be
accurate to ±0.02 to ±0.05 pH units. Accurate and meaningful Eh measurements are usually far
more difficult to obtain. In fact the U.S. Geological Survey (1998) states that Eh measurement
"is not recommended in general because of the difficulties inherent in its theoretical concept and
its practical measurement." An accurate Eh measurement using a platinum (or other noble metal)
or glassy carbon electrode depends upon the presence of significant concentrations (usually
greater than 10"5 M) of a pair of electroactive redox species that can attain rapid chemical
equilibrium and greatly exceed the concentrations of the non-electroactive species also present.
Most often the species that control an Eh measurement are species of iron, manganese, or sulfide
and native sulfur (Langmuir, 1997a). This tends to limit thermodynamically meaningful Eh
measurements to acid mine waters and iron-, manganese-, or sulfide-rich stagnant surface waters
or ground waters. In other waters, which include most surface waters in contact with the
atmosphere, Eh measurements are of qualitative value only. In such waters, a measurement of
dissolved oxygen (DO) is the most reliable indication of the oxidation state. In fact field DO
measurements should be performed any time redox sensitive metals of concern are of interest.
DO is measurable in the field at concentrations as low as 0.03 |iM using modern analytical
methods (cf. Kent et al., 1994).
81
-------
Lovely et al. (1994) and others have suggested that instead of Eh, field measurements of
dissolved H2 be performed to define the redox state of surface-water bottom sediments and
ground waters. This reflects the fact that bacteria employ H2 in the reduction of nitrate Mn(IV),
Fe(III), sulfate, and carbon dioxide.
Geochemical models such as MINTEQA2 (U.S. EPA, 1991) and PHREEQC (Parkhurst
and Appelo, 1999) can compute the relative amounts of redox sensitive species of As, Se, Mo,
Cr, N, C, and other metals of concern from the total metal concentration and measured Eh (at a
given pH). This assumes of course that this Eh represents the overall oxidation state of the
system (which is usually questionable), and that the redox pairs of these elements are
electroactive and in equilibrium with the measured Eh, which they usually are not.
The only way to determine accurate relative amounts of redox pairs of As, Cr, N, C, or
Se, for example, is to fix them upon collection and measure them directly (cf. Baedecker and
Cozzarelli, 1992; Sule and Ingle, 1996). For example, Tretner et al. (2001) have shown that for
non-sulfidic waters, the As(V)/As(III) ratio and As concentrations could be maintained for two
weeks by first filtering samples, then acidifying them with HC1 to pH 1.5-1.8 and refrigerating
them to below 4°C. In general, the accurate measurement of concentrations of individual redox
pairs is difficult and costly, and not recommended for routine studies.
10.3 Determination of Individual Species Concentrations in Water
There are various methods for directly measuring the aqueous activity or concentration of
individual metals, including specific ion electrodes and voltammetric techniques (cf. Salbu and
Steinnes, 1994). Most of these methods, though, have interference problems and require the
addition of reagents to the natural water, which may itself change the metal speciation. The
simplest approach is to compute metal ion activities from a geochemical model calculation,
based on a chemical analysis of concentrations of dissolved metal and related species in the
water. However, prediction of metal speciation in systems containing natural organic matter is
challenging (see Section 4).
11. SOILS
11.1 Sampling of Soil Pore Waters
A major challenge when sampling soil pore waters for chemical analysis is to avoid
changing metal speciation in the sampling process. This means that methods involving the
application of gas pressure should be avoided, in that they will change ambient gas pressures and
so change the pH, for example. The most appropriate moisture sampling methods use either
negative pressure as in suction lysimeters (cf. Knight et al., 1998; Davis et al., 1999; Tye et al.,
2003), or positive pressure without contact between air or other foreign gases and the soil. The
latter devices include gas pressure membrane extractors and pressure plate extractors (cf.
SoilMoisture Equipment Corp., 1997).
82
-------
11.2 Sequential Extractions To Determine the Nature of Metal/Soil Associations
The purpose of sequential extractions is to define the speciation of individual metals in
the soil through a series of soil leaching steps. Using sequential extractions, Tessier et al. (1979)
have categorized metals as exchangeable, bound to carbonates, bound to iron and manganese
oxides, bound to organic matter, or residual. The extractions often involve (1) addition of a
MgCl2 solution at pH 7 to define the exchangeable metals fraction; (2) addition of a pH 5 sodium
acetate/acetic acid solution to selectively dissolve carbonates (the acid extractable fraction) and
release their metals; (3) addition of an acidic hydroxylamine hydrochloride (NH2OH.HCl)
solution, which is a reductant, to dissolve the amorphous Fe-Mn oxyhydroxides and release their
metals; (4) addition of a hydrogen peroxide extraction at pH 2 and 85°C to determine organic
matter content and the metals associated with it; and (5) aqua regia destruction of remaining
minerals to obtain the residual metal concentration. Typical extraction results show minor to
below-detection concentrations of trace metals associated with the exchangeable fraction, and
major metal concentrations released by both the acid extractable and reductant leaching steps (cf.
Tack and Verloo, 1999). In another sequential extraction study, Tye et al. (2003) found that the
behavior of Cd2+ and Zn2+ in a variety of contaminated soils was best explained by assuming they
were adsorbed by soil humus.
Tack and Verloo (1999) list some of the problems associated with sequential extractions,
which include problems due to sample handling and preparation, the non-selectivity of the
extractants, and the redistribution of trace metals among remaining solids during extraction. It
has long been known that the five extraction steps do not cleanly distinguish the forms of metal
association and speciation identified above (cf. Suarez and Langmuir, 1975). Tye et al. (2003)
point out further that "single or sequential chemical soil extraction schemes provide a
comparative classification system for soil metals but do not predict (metal) the bioavailability."
The free metal ion activity in the soil solution phase has been shown to be a better
indicator of bioavailability and toxic response than is the total soil metal content (cf. Tye et al.,
2003). Most of the soil equilibrium models assume that the labile (bioavailable) metal content of
a soil can be predicted from the total or extractable soil metal content (e.g., NICA, Gooddy et al.,
1995; SCAMP, Lofts and Tipping, 1998; Tipping et al., 2000). Such an approach will generally
overestimate the labile metal fraction. More directly, the soil pore waters may be sampled and
chemically analyzed, and the solution speciation estimated from a model such as WHAM 6
(Tipping, 1998) to determine metal ion activities in the soil solution.
12. APPLICATION IN A REGULATORY CONTEXT
EPA needs to use the tools of environmental chemistry to determine metal speciation for
use in risk assessments. In addition, EPA needs to be able to assess the mobility of metals in the
environment, especially the subsurface. Each of these applications will be discussed in turn.
83
-------
13. APPLICATIONS TO SPECIATION
Metal speciation may be required for risk assessment. If so, it is desirable to know the
identity and form of a metal at the exposure point. If information regarding bioavailability,
pharmacokinetics, and toxicology is also known, speciation information may be used in risk
assessment. Note that to do so, risk assessors need to understand the dynamic nature of metal
speciation, because of the chemical equilibrium among the different metal species. For example,
even if metal toxicity is often best predicted from the free metal ion concentration, the free metal
ion cannot be labeled "the toxic species," since it is in equilibrium with all the metal species
present. Even if there is biological uptake of the free metal ion, its concentration will be
maintained through the dissociation of the other metal species present. EPA has developed
bioavailability and toxicological data for very few metal species. A review of the Integrated Risk
Information System (IRIS) shows that, disregarding organometallics and cyano complexes, EPA
has developed toxicological information for the following metals:
• Arsenic—inorganic, arsine
• Chromium—Cr(III), Cr(VI)
• Mercury—elemental, mercuric chloride
• Thallium—oxide, acetate, carbonate, chloride, nitrate, selenite, sulfate
• Vanadium—pentoxide
• Uranium—soluble, natural
• Zinc—phosphide
EPA has not formalized toxicity reference values (TRVs) for ecological risk assessment,
but there is little consideration currently given to metal speciation (with the possible exceptions
of chromium oxidation state and organometallic forms of mercury) in contemporary ecological
risk assessment. A limited number of Ambient Water Quality Criteria (Table 6) also exist for
metal species.
It is useful to hypothesize a future condition, however, in which there are no limits on
toxicological information and a real need exists to determine metal species. The exact
information required and questions to be asked of the environmental chemist will depend on the
nature of the assessment in the context of the framework:
• Site-Specific Risk Assessment—What are the existing metal species at the exposure points?
Is there likely to be a change in speciation over the exposure duration? Can normal human
behavior alter the speciation? If there is a strong likelihood of more than one species, should
84
-------
different species be selected for risk assessment (i.e., a common, but possibly less toxic
species for a central tendency risk and a rarer but possibly more toxic species for a high-end
risk)?
• National Regulatory Assessment—What generalized information is available to determine
the chemical species to which a receptor is exposed at any point in the United States? Is it
possible to generalize speciation across the U.S. or is it necessary to develop regional (e.g.,
physiographic province) or water-body type (stream, lake, estuary) speciation? How do
metals species change when traveling from a potentially regulated source to an exposure
point?
• National Hazard or Risk Ranking—Is there a property of metals analogous to organic
persistence that can be used to generalize across metals? How can chemical information be
developed to perform the generalization? What degree of stability is required to determine if
the property will persist?
Many of these questions need to be addressed by risk managers within their specific
context. From a chemistry standpoint, there are several levels of information that can be used to
answer these questions: regulatory defaults, literature information based on limited
measurement, analytical chemistry, geochemical modeling, and chemical-specific modeling.
13.1 Regulatory Defaults
Current regulatory practice is to use total metal concentrations in risk assessments, with a
few exceptions. The bioavailability of a metal is assumed to be the same as in the toxicological
or epidemiological studies in which dose-response curves have been derived. This practice is
conservative from an environmental health standpoint and is likely to continue as a default into
the future. In a tiered risk assessment approach, the production of an unacceptable risk or
exceedence of a standard or criterion by a total metal concentration could trigger a more refined
approach to evaluate the speciation of the metal. A tiered risk assessment approach will also
allow setting the level of details and the scope of the metal speciation information required based
on a periodic review of the results of the completed tiers. An alternative approach to the use of
total metal concentrations is to use the species that is the best predictor of toxicity from Table 7,
for example. This is predicated on the availability of dose-response information for the particular
species.
13.2 Literature/Limited Measurement
Obtaining speciation data at this level involves a combination of limited environmental
measurements along with literature information. This technique is already in use for some metals
to determine water quality criteria as a function of hardness. A more sophisticated use of
combined analytical and literature information is the determination of metal speciation using
look-up diagrams such as the pH-Eh diagrams presented in this paper. This requires reliable
85
-------
measurements of pH and Eh for the system under evaluation. For example, Figure 18 presents an
Eh-pH diagram for cadmium. At neutral pH in a mildly reducing environment, this diagram tells
us, cadmium will exist as the slightly soluble cadmium sulfide, while under oxidizing conditions
it will exist as the soluble divalent cadmium ion. This conclusion would, of course, depend on
the availability of sufficient sulfide in the system. There are limitations to this approach which
include the need to be aware of assumptions that were used in producing the diagrams and
interactions from other chemicals. The analyst also needs to be aware of the fact that, especially
in natural waters, pH and Eh are not static concepts. The values selected should be representative
of both the current status of the system and reasonable future geochemical changes.
13.3 Analytical Chemistry
Most site-specific risk assessments, some ranking assessments (e.g., HRS), and many
waste characterizations currently depend on analytical chemistry. With rare exceptions, the
metals are treated as total metals or some subset of total metals such as total recoverable metals.
EPA compendia of analytical methods such as SW-846 have traditionally focused on methods
for total metal analyses such as atomic absorption or inductively coupled plasma. The primary
exception is chromium. Several EPA methods currently exist to quantify Cr(III) and Cr(VI),
including ion chromatography (0218.6), differential pulse polarography (7198) and speciated
isotope dilution mass spectrometry (6800). EPA has also published methods for differentiating
As(III) from As(V) by anodic stripping voltametry.
Analytical methods for identifying and quantifying chemical species have been discussed
earlier in this paper. Most of these methods are used by academic or private research laboratories
and are not currently amenable to routine use in a regulatory context. Before being used for site-
specific risk assessment or waste characterization, for example, these methods require robust
evaluations to determine if they are capable of meeting data quality objectives. This would
include (but not be limited to) evaluations of accuracy, precision, and sensitivity when used on a
routine basis by a contract laboratory.
Another class of analytical methods could be termed "adjunct" methods. These methods
do not measure the metal of concern directly, but measure other aspects of environmental
chemistry that are useful in determining metal speciation. DO, Eh, pH, and hardness are
examples of adjunct measurements. If a metal species is to be determined by means of a look-up
table, for example, system measurements of pH, Eh (or a surrogate), and DO are required. The
quantification of major anions (chloride, bromide, nitrate, sulfate, carbonate) and cations (Na, K,
Ca, and Mg) is necessary to compute metal speciation using a chemical equilibrium modeling
approach. EPA has published methods for routine determination of these adjunct parameters.
86
-------
13.4 Geochemical Modeling
Geochemical modeling probably has the most utility for regulatory application for metal
chemistry of all the techniques discussed here. There are numerous geochemical models
available in the literature. EPA (1999b) has reviewed many of these models and should be
consulted for a more detailed discussion (see also Langmuir, 1997a). In general, geochemical
models are classified as "speciation-solubility" or "reaction path" models. Speciation-solubility
models may be used to calculate aqueous speciation/complexation and the degree of saturation
with respect to the solids in the model's database. Some specialization-solubility models also
include limited mass transfer capabilities. Examples of refined speciation-solubility models
include WATEQ, REDEQL, GEOCHEM, MINEQL, MINTEQ (see U.S. EPA, 1999b, for
detailed references) and their subsequent versions. Reaction path models have more advanced
mass transfer capabilities. At each step of a process (e.g., travel of an ion through an aquifer)
reaction path models calculate masses of minerals precipitated or dissolved and the resulting
composition of the solution. Reaction path models include PHREEQC, PATHCALC, and
EQ3/EQ6.
MINTEQ2 (U.S. EPA, 1991; HydroGeologic, 1998; HydroGeologic and Allison, 1999a)
has been widely used in a national regulatory context by EPA. Typically MINTEQ is used to
calculate partition coefficients that are used in a subsequent national regulatory analysis (U.S.
EPA, 1996; HydroGeologic and Allison, 1999b) or a site-specific risk assessment (U.S. EPA,
1998). This usage is discussed in more detail in previous sections of this paper. For systems with
low concentrations of natural organic matter, for example deep aquifers, there is little technical
reason why MINTEQ2 could not be used on a routine basis to determine chemical speciation for
site-specific risk assessments. However, MINTEQ2 is inadequate for systems where metal
speciation can be dominated by natural organic matter, such as shallow aquifers, soil pore
waters, and surface waters.
13.5 Chemical-Specific Modeling
Chemical-specific modeling involves the construction of a mathematical model that is
capable of predicting all of the significant fate and transport processes of a given chemical. It
should be capable of tracing a chemical from a source to an exposure point and predicting both
concentration and speciation at the exposure point with a sufficient degree of accuracy to reflect
the objectives of the assessment.
The only model of this type for a metal is the modeling reflected in the Mercury Study
Report to Congress (RTC) (U.S. EPA, 1997). In general, the RTC evaluates the behavior of three
mercury species (divalent, elemental, and methyl). In the atmosphere, fate and transport in both
particulate and vapor phases are assessed for long range and local behavior using established
atmospheric transport models. Terrestrial and aquatic modeling is accomplished by application
87
-------
of the IEM-2M model, which was specifically developed for the RTC.7 This model simulates the
behavior of mercury in watershed soils and water bodies using a mass balance approach. The
RTC was a National Regulatory Assessment in the context of the framework. It generalized
environments to the eastern and western United States and evaluated mercury behavior in these
environments. The models may be applied on a site-specific basis (Foster et al., 2003) if
sufficient environmental information is available. The IEM-2M model is data intensive. Each
mercury species evaluated with the model requires input values for molecular weight, Henry's
law constant, soil, bed sediment and suspended sediment partition coefficients, air diffusivity,
chemical reaction rate constants for six processes in three media and biotransfer factors f or all
biota of concern. Inputs of concentrations or emission rates of the various mercury species are
also required, but in practice, default assumptions of speciation were used by EPA. The
development and application of this type of model for other metals is limited by the amount of
resources and scientific information required.
14. APPLICATIONS TO MOBILITY
Quantification of the mobility of metals in the unsaturated zone or ground water is a
significant regulatory application of speciation information. For example, metals criteria for the
toxicity characteristic in RCRA depend on the development of a generic soil-to-groundwater
dilution attenuation factor (DAF), as do the Soil Screening Levels (SSLs) for metals. Site-
specific risk assessments for all programs need to be able to predict the rate of movement of
metals through soils and subsequent movement and concentrations in groundwater. The primary
processes governing the environmental fate and transport of metals in the subsurface are
advection, dispersion, matrix diffusion, and retardation (U.S. EPA, 1994). Advection and
dispersion are functions of the system rather than the contaminant. Matrix diffusion, which is a
function of the contaminant, is relatively unimportant and omitted in most model transport
algorithms. Retardation depends on a number of factors (U.S. EPA, 1994; Langmuir, 1997a) and
may involve or be affected by the following:
• Sorption—the attachment of chemical species to mineral or other surfaces.
• Ion exchange—competitive adsorption of ionic species, including ionic contaminants and
major ions, onto oppositely charged surfaces of geologic materials.
• Speciation—the distribution of a given constituent among its possible chemical forms,
including metal complexes, which have differing tendencies to be adsorbed or desorbed.
• Precipitation—the process by which dissolved species exceed the solubility limits of their
solids, so that some of the species precipitate from solution. When a metal species reaches
mineral saturation, addition of further amounts of the species to solution are precipitated, not
adsorbed.
7 The IEM-2M was also used by EPA to support development of mercury emissions limitations for hazardous waste
combustors.
88
-------
• Colloid formation—the process of forming colloids and the association of metal species with
them. The metals may be sorbed or coprecipitated with colloidal-sized particles.
• Biofixation—the binding of metals to solid materials due to the interactions of
microorganisms or plants.
• Natural organic matter interactions—sorption to natural organic matter particles and
complexation of metals by natural organic matter in the dissolved/colloidal fraction.
• Anion exclusion—negatively charged mineral or other surfaces repelling anions and so
preventing their sorption by those surfaces.
• Other importance processes—changes in pH, oxidation potential, salinity, concentrations of
competing ions, the nature of sorbent phases and their surface areas, and surface site
densities
Due to the complexity and multiplicity of the processes involved, recourse is often made
to the use of a single partition or distribution coefficient that describes the degree to which the
contaminant's transport is retarded relative to water. This approach starts with defining the
retardation factor:
R , -
J V
C
where Rf is the retardation factor, vp the velocity of water through a control volume, and vc the
velocity of contaminant through a control volume. The retardation factor is related to the
distribution coefficient through the expression:
where pb is the porous media bulk density and ne is the effective porosity at saturation given as a
volume fraction (cf. Langmuir, 1997a). This model only applies if fluid flow in the porous media
is isotropic and adsorption is fast, reversible, and linear (cf. Freeze and Cherry, 1979). These
assumptions are often not valid, particularly for metal adsorption. As discussed above, the
distribution coefficient approach can provide accurate modeling results for organic
contaminants, but is likely to be in serious error when applied to the transport of metals through
89
-------
porous media at specific sites. This is because, as emphasized previously, the reactions and
processes that control metal sorption are in general far greater in number than is the case for the
adsorption of organic substances. Some transport models assume a constant partition coefficient,
or assume linearity of the partition coefficient over all concentration ranges. To the extent that
sorption is not constant and follows a non-linear isotherm—which is the usual case for
metals—these models will be inaccurate and should be avoided. The best that can be hoped for
when single partition coefficients are used to describe metal adsorption is that they represent
bounding values in a given application.
Laboratory adsorption studies often find that, in simple systems, the value of log Kd for
metal adsorption increases linearly with pH. For example, for Zn2+ adsorption by HFO,
Langmuir (1997a) noted that adsorption followed the equation log Kd = -5.48 + 1.77 pH. For
natural assemblages of particles, such as lake sediments, reported log Kd values for a given metal
are highly variable, but log Kd values also clearly tend to increase with pH (Tessier, 1992).
EPA (1999b) presents a comprehensive review of the properties and applications of
partition or distribution coefficients for metals. These coefficients may be obtained from the
literature, estimated using mathematical models, or measured. Partition coefficients tabulated as
a function of pH by EPA (1998) are listed in Table 15 for several important elements of potential
concern. EPA (1999b) has also presented non-pH-dependent values for lead (900), mercuric
chloride (58,000) and elemental mercury (1,000), which for reasons just discussed should be
used with considerable caution.
Table 15. Partition Coefficients as a Function of pH for Several Important Elements of
Potential Concern (U.S. EPA, 1998)
Element
pH 4.9
pH 6.8
pH 8.0
Arsenic (V)
25
29
31
Barium
11
41
52
Beryllium
23
790
1E+5
Cadmium
15
73
4.3E+3
Chromium (III)
1.2E+3
1.8E+6
4.3E+6
Chromium (VI)
31
19
14
Nickel
16
65
1,900
Selenium (VI)
18
5.0
2.2
Silver
0.1
8.3
110
Thallium (I)
44
71
96
Zinc
62
EPA (1999b) discusses the advantages and disadvantages of several methods for
measuring partition coefficients, including laboratory batch testing, in situ field batch testing,
flow-through testing, and field modeling. In many national assessments, EPA has used MINTEQ
90
-------
and its subsequent versions to generate generic partition coefficients that may be applied to
regional or national mobility evaluations.
The application of single partition coefficient values for individual metals should be
limited to regional and national scale studies where accuracy is not required, and bounding or
representative values are adequate. As noted previously in the discussion of adsorption, metal
partition coefficients can vary by several orders of magnitude over short distances (meters or
less) in porous media because of changes not only in pH, but also in metal complex formation,
metal adsorption competition, the solid/solution ratio, the relative abundances, the surface areas,
and the surface charge densities of the different metal-sorbing phases.
A number of papers have measured adsorption of radionuclide elements by a variety of
minerals (including oxyhydroxides, clays, and other silicates), and have developed diffuse layer
adsorption parameters which allow the prediction of radionuclide adsorption by these minerals
(cf. Langmuir, 1997b). A similar effort has not been made for most of the metals of concern, for
which diffuse layer modeling of adsorption is largely limited to their adsorption by Fe(III), Mn,
and A1 oxyhydroxides.
It may be possible to estimate metal adsorption with some accuracy without having to
measure it, depending on the information available on a specific soil, surface water, or ground
water system. What is needed minimally is the amounts of potentially sorbing materials (e.g.,
metal oxides, clays and organic matter) in a soil or sediment or in suspension in a stream.
Literature information like that given in Table 10 can then be used to estimate the sorption
properties of these materials for use in a sorption model. For example, as noted above in
discussion of the diffuse layer adsorption model, where ferric oxyhydroxide (HFO) is the
dominant sorbent, and the amount of it suspended in a stream is known, estimation of metal
adsorption can be accurate to within 10-20%, as shown in Table 11 (Smith et al., 1998). As a
general observation, other conditions being equal, it has been found that the surface charge
density and thus the metal adsorption capacity of most minerals is largely a function of their
surface areas exposed to water (cf. Pabalan et al., 1998). Thus, the adsorption of a given metal by
Al and Fe(III) oxyhydroxides in a system at a given pH may be assumed to be the same if they
have the same surface areas.
If greater accuracy or site-specificity is required, it may be necessary to measure metal
adsorption in laboratory experiments. Such measurements can be performed on pure minerals or
on whole (usually sieved) soils. The sorption results may be used to develop diffuse layer model
parameters for metal adsorption (cf. Dzombak and Morel, 1990; Stumm, 1992). Model results
may then be used to compute partition or distribution coefficients as a function of pH, for
example.
Several chemical mass transport codes are available that can model metal transport
through porous media using the more accurate diffuse layer adsorption model for metals. These
models include PHREEQC (Parkhurst and Appelo, 1999) and CHMTRNS (Noorishad et al.,
91
-------
1987), and—for transport by streams—OTIS/OTEQ (Runkel et al., 1999). Even if model
parameters are estimated based on literature values, such models will generally predict metal
adsorption and retardation more accurately than is possible when using single or linearly varying
distribution coefficient values.
15. RESEARCH RECOMMENDATIONS
In general, environmental chemistry of metals research could benefit from the
development of more routine chemical-species-specific analytical methods, the development of
extraction techniques that have general utility in assessing bioavailability and/or mobility, and
the validation of geochemical and chemical-specific environmental fate and transport models.
The quality of EPA's use of environmental chemistry could benefit from the training of risk
managers and other decision-makers in inorganic environmental chemical science.
The environmental chemistry of metals occupies a key position in the regulatory
understanding of mobility, exposure, toxicity, and waste characterization. However interesting in
its own right, EPA-sponsored research into the environmental chemistry of metals should
structured to answer specific environmental problems. Typically, these problems are associated
with particular instances in which mobility, bioavailability, bioaccumulation, or toxicity are
important. Rather than environmental chemistry driving the research agenda, it is probably more
important for risk assessment to drive the research agenda.
16. LITERATURE CITED
Allison, J.D., and E.M. Perdue. 1994. Modeling metal-humic interactions with MINEQA2. In:
Senesi, N., and T.M. Miano, eds. Humic substances in the global environment and implications
on human health. Elsevier Science, pp. 927-942.
Alsop, D.H. and C.M. Wood. 2000. Kinetic analysis of zinc accumulation in the gills of juvenile
rainbow trout: Effects of zinc acclimation and implications for biotic ligand modeling. Environ.
Toxicol. Chem. 19:1911-1918.
Baas-Becking, L.G.M., I.R. Kaplan, and D. Moore. 1960. Limits of the natural environment in
terms of pH and oxidation-reduction potentials. J. Geol. 68:243-284.
Baedecker, M.J. and I.M. Cozzarelli. 1992. The determination and fate of unstable constituents
in contaminated groundwater. In: Lesage, S. and R. Jackson, eds. Ground water contamination
and analysis at hazardous waste sites. New York: Marcel Dekker, Inc., pp. 425-461.
Baes, C.F., Jr., and R.E. Mesmer. 1976. The hydrolysis of cations. New York: Wiley
InterScience.
92
-------
Ball, J.W., R.L. Runkel, and D.K. Nordstrom. 1999. Transport modeling of reactive and non-
reactive constituents from Summittville, Colorado: Preliminary results from the application of
the OTIS/OTEQ model to the Whitman Fork/Alamosa River System. In: Proc. Technical
Meeting, Charleston, SC, Mar. 8-12, 1999. Vol. 1 of 3. Contamination from hard-rock mining.
U.S. Geological Survey Water Resources Investigations Report 99-4018A.
Banks, D., C. Reimann, O. Royset, H. Skarphagen, and O. M. Saether. 1995. Natural
concentrations of major and trace elements in some Norwegian bedrock ground-waters. Appl.
Geochem. 10:1-16.
Barnes, H L. and D. Langmuir. 1978. Geochemical prospecting handbook for metals and
associated elements. National Science Foundation Grant No. AER77-06511 A02. Annual
Report.
Bell, P.F., C.A. Adamu, C.L. Mulchi, M. Mcintosh, and R.L. Chaney. 1988. Residual effects of
land applied municipal sludge on tobacco. I: Effects on heavy metals concentrations in soils and
plants. Tob. Sci. 32:33-38.
Benjamin, M.M. and J.O. Leckie. 1981. Competitive adsorption of Cd, Cu, Zn and Pb on
amorphous iron hydroxides. J. Colloid Interface Sci. 83:410-419.
Benoit, J.M., C.C. Gilmour, and R.P. Mason. 2001. The influence of sulfide on solid-phase
mercury bioavailability for methylation by pure cultures of Desulfobulbus propionicus (lpr3).
Environ. Sci. Technol. 35:127-132.
Benoit, J.M., C.C. Gilmour, R.P. Mason, and A. Heyes. 1999. Sulfide controls on mercury
speciation and bioavailability to methylation bacteria in sediment pore waters. Environ. Sci.
Technol. 33:951-957.
Bergseth, H. 1983. Effect of filter type and filtrate treatment on the measured content of A1 and
other ions in groundwater. Acta Agric. Scand. 33:353-359.
Beyer, W.N. and C. Stafford. 1993. Survey and evaluation of contaminants in earthworms and in
soils derived from dredged material at confined disposal facilities in the Great Lakes region.
Environ. Monitor. Assess. 24:151-165.
Bodek, I.B., W.J. Lyman, W.F. Reehl, and D.H. Rosenblatt, eds. 1988. Environmental inorganic
chemistry: Properties, processes, and estimation methods. SET AC Spec. Publ. Ser. New York:
Pergamon Press.
Boerngen, J.G. and H.T. Shacklette. 1981. Chemical analyses of soils and other surficial
materials of the conterminous United States. U.S. Geol. Survey Open-File Report 81-197. 143
pp.
93
-------
Bohn, H.L., B.L. McNeal, and G.A. O'Connor. 1985. Soil chemistry, 2nd ed. New York: John
Wiley & Sons, Inc.
Bowen, H.J.M. 1979. Environmental chemistry of the elements. London: Academic Press.
Brookins, D.G. 1988. Eh-pH diagrams for geochemistry. New York: Springer-Verlag. 176 pp.
Brown, S.L., R.L. Chaney, M. Sprenger, and H. Compton. 2002. Assessing impact to wildlife at
biosolids remediated sites: Soil-animal pathway. BioCycle 43(8):50-58.
Cabaniss, S.E. and M.S. Shuman. 1988. Copper binding by dissolved organic matter. I: Suwanee
River fulvic acid equilibria. Geochim. Cosmochim. Acta 52:185.
Campbell, P.G.C. 1995. Interactions between trace metals and aquatic organisms: A critique of
the free-ion activity model. In: Tessier, A. and D.R. Turner, eds. Metal speciation and
bioavailability in aquatic systems. John Wiley & Sons, Inc. pp. 45-102.
Cederberg, G.A., R.T.L. Street, and J.O. Leckie. 1985. A groundwater mass transport and
equilibrium chemistry model for multicomponent systems. Water Resour. Res. 21:1095-1104.
Chaney, R.L. 1980. Health risks associated with toxic metals in municipal sludge. In: Bitton, G.,
B.L. Damron, G.T. Edds, and J.M. Davidson, eds. Sludge—Health risks of land application. Ann
Arbor, MI: Ann Arbor Sci. Publ. Inc., pp. 59-83.
Chaney, R.L. and J.A. Ryan. 1993. Heavy metals and toxic organic pollutants in MSW-
composts: Research results on phytoavailability, bioavailability, etc. In: Hoitink, A.J. and H.M.
Keener, eds. Science and engineering of composting: Design, environmental, microbiological
and utilization aspects. Columbus, OH: Ohio State University, pp. 451-506.
Chaney, R.L. et al. 1983. Potential effects of waste constituents on the food chain. In: Parr, J.F.,
P.B. Marsh, and J.M. Kla, eds. Land treatment of hazardous wastes. Park Ridge, NJ: Noyes Data
Corp. pp. 152-240.
Chaney, R.L., G.S. Stoewsand, A.K. Furr, C.A. Bache, and D.J. Lisk. 1978. Elemental content of
tissues of Guinea pigs fed Swiss chard grown on municipal sewage sludge-amended soil. J. Agr.
Food Chem. 26:944-997.
Clark, S.B., N.M. Park, and R.C. Tuckfield. 1992. Effects of sample collection device and filter
pore size on concentrations of metals in groundwater samples. National Groundwater Sampling
Symposium. 1-7.
94
-------
Cusimano, R.F., D.F. Brakke, and G.A. Chapman, 1986. Effects of pH on the toxicities of
cadmium, copper and zinc to steelhead trout (Salmo gardneii). Can. J. Fish. Aquat. Sci. 43:1497-1503.
Darnley, A.G., A. Bjorklund, B. Bolviken, et al. 1995. A global geochemical database for
environmental and resource management. Earth Sciences 19. UNESCO Publishing.
Davis, S.N., and R.J.M. DeWiest. 1966. Hydrogeology. New York: John Wiley & Sons, Inc.
Davis, J.A., and D.B. Kent. 1990. Surface complexation models in aqueous geochemistry. In:
Hochella, M.F., Jr., and A.F. White, eds. Mineral-water interface geochemistry. Min. Soc.
Amer., Rev. Mineral. 23:177-260.
Davis, G.B., C. Barber, T.R. Power, B. Patterson, J.L. Rayner, and Q. Wu. 1999. J. Contam.
Hydrology 36(3-4): 265-290.
De Schamphelaere, K.A.C., and C.R. Janssen. 2002. A biotic ligand model predicting acute
copper toxicity for Daphnia magna: The effects of calcium, magnesium, sodium, potassium and
pH. Environ. Sci. Technol. 36:48-54.
De Schamphelaere, K.A.C., D.G. Heijerick, and C.R. Janssen. 2002. Refinement and field
validation of a biotic ligand model predicting acute copper toxicity to Daphnia magna. Comp.
Biochem. Physiol. Pt. C 133:243-258.
Di Toro, D.M., H.E. Allen, H.L. Bergman, J.S. Meyer, P.R. Paquin, and R.C. Santore. 2001.
Biotic ligand model of the acute toxicity of metals. 1: Technical basis. Environ. Toxicol. Chem.
20:2383-2396.
Drever, J.I. 1997. The geochemistry of natural waters, 3rd ed. Upper Saddle River, NJ: Prentice-
Hall.
Dzombak, D.A. and F.M.M. Morel. 1990. Surface complexation modeling: Hydrous ferric oxide.
New York: Wiley InterScience.
Erickson, R.J., D.A. Benoit, V.R. Mattson, H.P. Nelson, Jr., and E.N. Leonard. 1996. The effects
of water chemistry on the toxicity of copper to fathead minnows. Environ. Toxicol. Chem.
15:181-193.
Finlayson-Pitts, B.J. and J.N. Pitts. 1986. Atmospheric chemistry. New York: Wiley
InterScience.
Fleischhauer, H.L. and N. Korte. 1990. Formulation of cleanup standards for trace elements with
probability plots. Environ. Manage. 14(1):95-105
95
-------
Ford, R.G., and D.L. Sparks. 2000. The nature of Zn precipitates formed in the presence of
pyrophyllite. Env. Sci. Technol. 34:2479-2483.
Forstner, U. 1995. Land contamination by metals: Global scope and magnitude of the problem.
In: Allen, H. et al., eds. Metal speciation and contamination of soil. Boca Raton, FL: Lewis
Publishers.
Fortescue, J.A.C. 1992. Landscape geochemistry: Retrospective and prospect, 1990. Appl.
Geochem. 7:1-53.
Foster, S.A., P.C. Chrostowski, and D.V. Preziosi. 2003. A comparison of two mercury
environmental fate and transport models in evaluating incinerator emissions. International
Conference on Incineration and Thermal Treatment Technologies. Orlando, FL, May 12-16,
2003.
Freeze, R.A. and J.A. Cherry. 1979. Groundwater. Englewood Cliffs, NJ: Prentice Hall, Inc.
Galvez, F., N. Webb, C. Hogstrand, and C.M. Wood. 1998. Zinc binding to the gills of rainbow
trout: The effect of long-term exposure to sublethal zinc. J. Fish. Biol. 52:1089-1104.
Gooddy, D.C., P. Shand, D.G. Kinniburgh, and W.H. Van Riemsdijk. 1995. Field-based partition
coefficients for trace elements in soil solutions. Eur. J. Soil Sci. 46(2):265-286.
Griffin, R.A. and N.F. Shimp. 1976. Effect of pH in exchange-absorption or precipitation of Pb
from landfill leachates by clay minerals. Environ. Sci. Technol. 10(3): 1256-1261.
Gustafson, J.P. 2003. Visual MINTEQ, version 2.14; a Windows version of MINTEQA2,
version 4.0. http://www.lwr.kth.se/english/OurSoftware/Vminteq.
Gustafson, J.P. 2004. Visual MINTEQ, version 2.30; a Windows version of MINTEQA2,
version 4.0. http://www.lwr.kth.se/english/OurSoftware/Vminteq/.
Gustavsson, N., B. Bolviken, D.B. Smith, and R.C. Severson. 2001. Geochemical landscapes of
the conterminous United States—New map presentations for 22 elements. U.S. Geological
Survey Professional Paper 1648. http://geology.cr.usgs.gov/pub/ppapers/pl648/.
Heijerick, D.G., K.A.C. De Schamphelaere, and C.R. Janssen. 2002a. Biotic ligand model
development predicting Zn toxicity to the alga Pseudokirchneriella subcapitata: Possibilities and
limitations. Comp. Biochem. Physiol. Pt. C 133:207-218.
Heijerick, D.G., K.A.C. De Schamphelaere, and C.R. Janssen. 2002b. Predicting acute zinc
toxicity for Daphnia magna as a function of key water chemistry characteristics: Development
and validation of a biotic ligand model. Environ. Toxicol. Chem. 21:1309-1315.
96
-------
Helsel, D.R. 1990. Less than obvious: Statistical treatment of data below the detection limit.
Environ. Sci. Technol. 24:1766-1774.
Helsel, D.R. and R.M. Hirsch. 1992. Statistical methods in water resources. Studies in
environmental science 49. New York: Elsevier.
Horowitz, A.J., K.R. Lum, J.R. Garbarino, G.E.M. Hall, and C.R. Demas. 1996. Problems
associated with using filtration to define dissolved trace element concentrations in natural water
samples. Environ. Sci. Technol. 30: 954-963.
Horowitz, A.J., K.A. Elrick, and M.R. Colberg. 1992. The effect of membrane filtration artifacts
on dissolved trace element concentrations. Wat. Res. 26: 753-763.
Huheey, J.E., E.A. Keiter, and R.I. Keiter. 1993. Inorganic chemistry—principles of structure
and reactivity, 4th ed. New York: Harper Collins College Publishers, 964 pp.
HydroGeologic. 1998. Diffuse-layer sorption reactions for use in MINTEQA2 for HWIR metals
and metalloids. Prepared for EPA National Exposure Research Lab.
HydroGeologic and Allison. 1999a. MINTEQA2/PRODEFA@, A geochemical assessment
model for environmental systems: User manual supplement for version 4.0.
HydroGeologic and Allison. 1999b. Partition coefficients for metals in surface water, soil, and
waste. Prepared for EPA Office of Solid Waste under Contract 68-C6-0020. (June 22).
HydroQual. 2002. Biotic ligand model: Windows interface, version 1.0.0. Hydroqual, Inc.,
Mahwaj, NJ.
Jing, J. and T. Logan. 1992. Effects of sewage sludge cadmium concentration on chemical
extractability and plant uptake. J. Environ. Qual. 21:73-81.
Keizer, M. and W. H. van Riemsdijk 1994. ECOSAT: A Computer Program for the Calculation
of Speciation and Transport in Soil-Water Systems. Department of Soil Science and Plant
Nutrition, Wageningen Agricultural University, Wageningen, The Netherlands.
Kennedy, V.C. and G.W. Zellweger. 1974. Filter pore-size effects on the analysis of Al, Fe, Mn,
and Ti in water. Water Resour. Res. 10(4):785-790.
Kent, D.B., J.A. Davis, C.D. Anderson, B. Rea, and T.D. Waite. 1994. Transport of chromium
and selenium in the suboxic zone of a shallow aquifer: Influence of redox and adsorption
reactions. Water Resour. Res. 30(4):1099-114.
97
-------
Kinniburgh, D.G., C J. Milne, M.F. Benedetti, J.P. Pinheiro, J. Filius, L.K. Koopal, and W.H.
van Riemsdijk. 1996. Metal ion binding by humic acid: Application of the NICA-Donnan model.
Environ. Sci. Technol. 30(5): 1687-1698.
Knight, B.P., A.M. Chaudri, S.P. McGrath, and K.E. Giller. 1998. Determination of chemical
availability of cadmium and zinc in soils using inert soil moisture samplers. Envir. Pollut.
99:293-298.
Lahermo, P.R., R. Salminen, T. Tarvainen, and P. Vaananen. 1995. Geochemical mapping of
stream waters and sediments in Finland: selected results. Geol. Surv. Finland. Spec. Paper
20:155-166.
Langmuir, D. 1997a. Aqueous environmental chemistry. Upper Saddle River, NJ: Prentice-Hall.
600 pp.
Langmuir, D. 1997b. The use of laboratory adsorption data and adsorption models to predict
radionuclide releases from a geological repository: A brief history. In: Gray, W.J. and I.R. Triay,
eds. Scientific basis for nuclear waste management XX. Materials Research Soc. Symp. Proc.
465, pp. 769-780.
Langmuir, D., 1978. Uranium solution-mineral equilibria at low temperatures with applications
to sedimentary ore deposits. Geochim. et Cosmochim. Acta 42(6): 547-569.
Langmuir, D. 1971. Eh-pH determination. In: Carver, R.E., ed. Procedures in sedimentary
petrology. New York: John Wiley & Sons, Inc., pp. 597-634.
Langmuir, D., and R.L. Klusman. 1997. Natural background concentrations of metals that
exceed drinking water standards: Defining background and the effect of sample filtration.
Abstract, Geol. Soc. Am. Annual Meetings. Oct. 20-23. Salt Lake City UT, p. A-435.
Leenheer J.A., R.L. Malcolm, P.W. McKinley, and L.A. Eccles. 1974. Occurrence of dissolved
organic carbon in selected groundwater samples in the United States. U.S. Geol. Survey J. Res.
2:361-369.
Levinson, A.A., P.M.D. Bradshaw, and I. Thomson. 1987. Practical exploration geochemistry.
Wilmette, IL: Applied Publishing Ltd.
Lindqvist, O. 1994. Atmospheric cycling of mercury: An overview. In: Watras, C J. and J.W.
Huckabee, eds. Mercury pollution: Integration and synthesis. Boca Raton, FL: Lewis Publishers.
Livingston, D.A. 1963. Chemical composition of rivers and lakes, 6th ed. U.S. Geol. Survey
Professional Paper 440-G.
98
-------
Lofts, S. and E. Tipping. 1998. An assemblage model for cation binding by natural particulate
matter. Geochim. Cosmochim. Acta 62:2609-2625.
Logan, T.J. and S.J. Traina. 1993. Trace metals in agricultural soils. Chapter 10 in: Metals in
Groundwater. Eds. Allen, H.E., E.M. Perdue, and D.S. Brown. New York: Lewis Publishers, pp.
309-347.
Loux, N.T., D.S. Brown, C.R. Chafin, J.D. Allison, and S.M. Hassan. 1990. Modeling
geochemical processes attenuating inorganic contaminant transport in the subsurface region:
Adsorption on amorphous iron oxide. In: Murarka, I. and S. Cordle, eds. Proc. Envir. Res. Conf.
on Groundwater Qual. and Waste Disposal. Report EPRI EN-6749. Palo Alto, CA: Electrical
Power Research Inst.
Lovely, D.R., F.H. Chapelle, and J.C. Woodward. 1994. Use of dissolved H2 concentrations to
determine distribution of microbially catalyzed redox reactions in anoxic groundwater. Environ.
Sci. Technol. 28(7): 1205-1210.
Lu, Y. and H.E. Allen. 2002. Characterization of copper complexation with natural dissolved
organic matter (DOM)—Link to acidic moieties of DOM and competition by Ca and Mg. Water
Res. 36:5083-5101.
MacRae, R.K., D.E. Smith, N. Sxoboda-Colberg, J.S. Meyer, and H.L. Bergman. 1999. Copper
binding affinity of rainbow trout (Oncorhynchus mykiss) and brook trout (Salveil mis fonti nulls)
gills: Implication for assessing bioavailable metal. Environ. Toxicol. Chem. 118:1180-1189.
Manahan, S.E. 1994. Environmental chemistry, 6th ed. Chelsea, MI: Lewis Publ. Co.
McBride, M.B. 1994. Environmental chemistry of soils. New York: Oxford University Press.
McGeer, J.C., R.C. Playle, C.M. Wood, and F. Galvez. 2000. A physiologically based biotic
ligand model for predicting the acute toxicity of waterborne silver to rainbow trout in freshwaters.
Environ. Sci. Technol. 34:4199-4207.
Meyer, J.S., C J. Boese, and S.A. Collyard. 2002. Whole-body accumulation of copper predicts
acute toxicity to an aquatic oligochaete (Lumbriculus variegatus) as pH and calcium are varied.
Comp. Biochem. Physiol. Pt. C 133:99-110.
Miller, W.R., and J.B. McHugh. 1999. Calculations of geochemical baselines of stream waters in
the vicinity of Summitville, Colorado, before historic underground mining and prior to recent
open-pit mining. In: Filipek, L.H. and G.S. Plumice, eds. The environmental geochemistry of
mineral deposits. Part B: Case studies and research topics. Rev. Econ. Geol. 6B:505-514.
Morel, F.M.M. 1983. Principles of aquatic chemistry. New York: Wiley-InterScience.
99
-------
Morel, F.M.M., and J.G. Herring. 1993. Principles and applications of aquatic chemistry. New
York: John Wiley & Sons, Inc.
Munthe, J. 1994. The atmospheric chemistry of mercury: Kinetic studies of redox reactions. In:
Mercury Pollution: Integration and Synthesis. Watras, C.J. and J.W. Huckabee, eds. Boca Raton,
FL: Lewis Publishers.
Newcomb, W.D. and J.D. Rimstidt. 2002. Trace element distribution in US groundwaters: a
probabilistic assessment using public domain data. Appl. Geochem. 17:49-57.
Noorishad, J.C., L. Carnahan, and L.V. Benson. 1987. CHMTRNS: A temperature-dependent
non-equilibrium reactive transport code. Earth Sciences Div., Lawrence Berkeley Laboratory.
Nordstrom, D.K., N. Plummer, D. Langmuir, E. Busenberg, and H.M. May. 1990. Revised
chemical equilibrium data for major water-mineral reactions and their limitations. In: Chemical
modeling of aqueous systems II. Melchior, D.C. and R.L. Bassett, eds. Am. Chem. Soc. Ser.
416:398-413. Washington, D.C.: Am. Chem. Soc.
NRC (National Research Council). 1980. Mineral tolerance of domestic animals. National
Academy of Sciences, Washington, DC. 577 pp.
Nriagu, J.O. 1994. Arsenic in the environment, Part 2: Human health and ecosystem effects. New
York: John Wiley & Sons, Inc.
Pabalan, R.T., D.R. Turner, F.P. Bertetti and J.D. Prikryl. 1998. Uranium VI sorption onto
selected mineral surfaces: Key geochemical parameters. Chap. 3. In Adsorption of Metals by
Geomedia. ed. E.A. Jenne. Academic Press, New York: pp. 99-130.
Pagenkopf, G.K. 1983. Gill surface interaction model for trace-metal toxicity to fishes: Role of
complexation, pH, and water hardness. Environ. Sci. Technol. 17:342-347.
Paquin, P.R., J.W. Gorsuch, S. Apte, G.E. Batley, K.C. Bowles, P.G.C. Campbell, C.G. Delos,
D.M. Di Toro, R.L. Dwyer, F. Galvez, R.W. Gensemer, G.G. Goss, C. Hogstrand, C.R. Janssen,
J.C. McGeer, R.B. Naddy, R.C. Playle, R.C. Santore, U. Schneider, W.A. Stubblefield, C.M.
Wood, and K.B. Wu. 2002. The biotic ligand model: A historical overview. Comp. Biochem.
Physiol. Pt. C 133:3-35.
Paquin, P.R., D.M. Di Toro, R.C. Santore, D. Trivedi, and K.B. Wu. 1999. A biotic ligand model
of the acute toxicity of metals. Ill: Application to fish and Daphnia exposure to silver. Office of
Water, Office of Research and Development.
Parkhurst, D.L. 2002. PHAST: 3D reactive transport model. U.S. Geological Survey.
http://wwwbrr.cr.usgs.gov/projects/GWC_coupled/phast.
100
-------
Parkhurst, D.L., and C.A.J. Appelo. 1999. User's guide to PHREEQC (version 2)-A computer
program for speciation, batch-reaction, one-dimensional transport, and inverse geochemical
calculations. U.S. Geol. Survey Water-Resources Inv. Report 99-4259.
Paulson, A.J., and L.S. Balistrieri. 1999. Modeling removal of Cd, Cu, Pb, and Zn in acidic
groundwater during neutralization by ambient surface waters and groundwaters. Environ. Sci.
Technol. 33(21):3850-3856.
Pearson, R.G. 1973. Hard and soft acids and bases. Stroudsburg, PA: Dowden, Hutchinson &
Ross.
Phillips, C.S.H., and R.J.P. Williams. 1965. Inorganic chemistry. Vol. 2: Metals. New York:
Oxford University Press.
Playle, R.C., D.G. Dixon, and K. Burnison. 1993a. Copper and cadmium binding to fish gills:
Estimates of metal-gill stability constants and modeling of metal accumulation. Can. J. Fish.
Aquat. Sci. 50:2678-2687.
Playle, R.C., D.G. Dixon, and K. Burnison. 1993b. Copper and cadmium binding to fish gills:
Modification by dissolved organic carbon and synthetic ligands. Can. J. Fish. Aquat. Sci.
50:2667-2677.
Playle, R.C., R.W. Gensemer, and D.G. Dixon. 1992. Copper accumulation on gills of fathead
minnows: Influence of water hardness, complexation and pH of the gill micro-environment.
Environ. Toxicol. Chem. 11:381-391.
Pohlmann, K.F., G.A. Icopini, and R.D. McArthur. 1994. Evaluation of sampling and field-
filtration methods for the analysis of trace metals in ground water. EPA/600R/94/119.
Pourbaix, M. 1966. Atlas of electrochemical equilibria in aqueous solutions. New York:
Pergamon Press.
Pretorius, P.J. and P.W. Linder. 2001. The adsorption characteristics of 5-manganese dioxide: A
collection of diffuse double layer constants for the adsorption of H+, Cu2+, Ni2+, Zn2+, Cd2+ and
Pb2+. Appl. Geochem. 16:1067-1182.
Puis, R.W. and M.J. Barcelona. 1989. Ground water sampling for metals analysis. EPA
Superfund Ground Water Issue. EPA/540/4-89/001.
Rai, D. and J.M. Zachara. 1984. Chemical attenuation rates, coefficients, and constants in
leachate migration. Vol. 1: A critical review. EA-3356. Research Project 2198-1. Final Report.
Palo Alto, CA: Electrical Power Research Inst.
101
-------
Rice, K.C. 1999. Trace-element concentrations in streambed sediment across the conterminous
United States. Envir. Sci. Technol. 33(15):2499-2504.
Richardson, G.M. 2002. Natural sources and emissions of elements to the atmosphere.
International Council on Mining and Metals. (February).
Rozan, T. F., M.E. Lassman, D.P. Ridge, and G.W. Luther III. 2000. Evidence for iron, copper
and zinc complexation as multinuclear sulphide clusters in oxic rivers. Nature 406:879-82.
Runkel, R.L., B.A. Kimball, D.M. McKnight, and K.E. Bencala. 1999. Reactive solute transport
in streams: A surface complexation approach for trace metal sorption. Water Resour. Res.
35(12):3829-3840.
Runnells, D.D., T.A. Shepherd, and E.E. Angino. 1992. Metals in water: Determining natural
background concentrations in mineralized areas. Envir. Sci. Technol. 26(12):2316-2322.
Salbu, B., and E. Steinnes, eds. 1994. Trace elements in natural waters. Boca Raton, FL: CRC
Press, 314 pp.
Salminen, R., and V. Gregorauskiene. 2000. Considerations regarding the definition of a
geochemical baseline of elements in the surficial materials in areas differing in basic geology.
Appl. Geochem. 15:647-653.
Salminen, R. and T. Tarvainen. 1997. The problem of defining geochemical baselines: case study
of selected elements and geological materials in Finland. J. Geochem. Explor. 60:91-98.
Santore, R.C., and C.T. Driscoll. 1995. The CHESS model for calculating chemical equilibria in
soils and solutions. In: Loeppert, R., A.P. Schwab, and S. Goldberg, eds. Chemical equilibrium
and reaction models. Madison, WI: The Soil Society of America, American Society of
Agronomy, pp. 357-375.
Santore, R.C., D.M. Di Toro, P.R. Paquin, H.E. Allen, and J.S. Meyer. 2001. Biotic ligand model
of the acute toxicity of metals. 2: Application to acute copper toxicity in freshwater fish and
Daphnia. Environ. Toxicol. Chem. 20:2397-2402.
Schecher, W.D., and D.D. McAvoy. 1994. MINEQL+: A chemical equilibrium program for
personal computers (version 3.01). Environmental Research Software, Hallowell, ME.
Schwarz, G.E., and R.B. Alexander. 1995. State soil geographic (STATSGO) data base for the
conterminous United States. Edition 1.1. U.S. Geol. Survey Open-File Report 95-449. U.S. Geol.
Survey, Reston, VA. http://water.usgs.gov/lookup/getspatial7ussoils.
102
-------
Seinfeld, J.H., and S.N. Pandis. 1998. Atmospheric chemistry and physics. New York: Wiley
InterScience.
Serkiz, S.M., J.D. Allison, E.M. Perdue, H.E. Allen, and D.S. Brown. 1996. Correcting errors in
the thermodynamic database for the equilibrium speciation model MINTEQA2. Wat. Res. 30:
1930-1933.
Shacklette, H.T., and J.G. Boerngen. 1984. Element concentrations in soils and other surficial
materials of the conterminous United States. U.S. Geol. Survey Professional Paper 1270.
Shirazi, M.A., L. Boersma, C.B. Johnson, and P.K. Haggerty. 2001. Predicting physical and
chemical water properties from relationships with watershed characteristics. J. Envir. Qual.
30:112-120.
Singh, I., and D.S. Mavinic. 1991. Significance of building and plumbing specifics on trace-metal
concentrations in drinking water. Can. J. Civil Eng. 18:893-903.
Smith, D.B. 2003. Verbal communication. U.S. Dept. of Interior, U.S. Geological Survey, Denver
CO. dsmith@usgs.gov.
Smith, D.C., R.A. Bell, and J.R. Kramer. 2002. Metal speciation in natural waters with emphasis
on reduced sulfur groups as strong metal binding sites. Comp. Biochem. & Physiol. C, Toxicol.
Pharmacol. 133:65-74.
Smith, K.S., J.F. Ranville, G.S. Plumlee, and D.L. Macalady. 1998. Predictive double-layer
modeling of metal sorption in mine-drainage systems. Chapter 24, in: Jenne, E.A., ed. Adsorption
of metals by geomedia. New York: Academic Press, pp. 521-547.
SoilMoisture Equipment Corp. 1998. Using the gas pressure extractor. 0899-103. Dec. 1997.
http://www.soilmoisture.com.
Sposito, G. 1989. The chemistry of soils. New York: Oxford University Press.
Sposito, G. 1984. The Surface Chemistry of Soils. Oxford Univ. Press, New York.
Stumm, W. 1992. Chemistry of the solid-water interface: Processes at the mineral-water and
particle-water interface. New York: Wiley InterScience.
Stumm, W., and J.J. Morgan. 1996. Aquatic chemistry, 3rd ed. New York: John Wiley & Sons,
Inc.
Suarez, D.L. and D. Langmuir. 1975. Heavy metal relationships in a Pennsylvania soil. Geochim.
Cosmochim. Acta 40:589-598.
103
-------
Sule, P. A. and J.D. Ingle, Jr. 1996. Determination of the speciation of chromium with an
automated two-column ion-exchange system. Anal. Chim. Acta 326:85-93.
Szebedinszky, C., J.C. McGeer, G.D. McDonald, and C.M. Wood. 2001. Effects of chronic Cd
exposure via the diet or water on internal organ-specific distribution and subsequent gill Cd
uptake kinetics in juvenile rainbow trout (Oncorhynchus mykiss). Environ. Toxicol. Chem.
20:597-607.
Tack, F.M.G. and M.G. Verloo. 1999. Single extractions versus sequential extraction for the
estimation of heavy metal fractions in reduced and oxidized dredged sediments. Chem. Spec.
Bioavail. ll(2):43-50.
Templeton, D.M., F. Ariese, R. Cornelis, L.-G. Danielson, H. Muntau, H.P. Van Leeuwen, and R.
Lobinski. 2000. Gudelines for terms related to chemical speciation and fractionation of elements.
Definitions, structural aspects, and methodological approaches. Pure Appl. Chem. 72(8): 1453-70.
Tessier, A. 1992. Sorption of trace elements on natural particles in oxic environments . In
Environmental Particles. Edited by J. Buffle and H.P. Van Leeuwen. Lewis Publishers, Boca
Raton, FL, USA. pp. 425-453.
Tessier, A., D. Fortin, N. Belzile, R. DeVitre, and G.G. Leppard. 1996. Metal sorption to
diagenetic iron and manganese oxyhydroxides and associated organic matter: Narrowing the gap
between field and laboratory experiments. Geochim. Cosmochim. Acta. 60(3):387-404.
Tessier, A., P.G.C. Campbell, and M. Bisson. 1979. Sequential extraction procedure for the
speciation of particulate trace metals. Anal. Chem. 51:844-851.
Tipping E. 1998. Humic ion binding model VI: An improved description of the interactions of
protons and metal ions with humic substances. Aqua. Geochem. 4(l):3-48.
Tipping, E. 1994. WHAM—A chemical equilibrium model and computer code for waters,
sediments, and soils incorporating a discrete site electrostatic model of ion-binding by humic
substances. Comp. Geosci. 20:973-1023.
Tipping E., A.J. Lawlor, S. Lofts, and C.D. Vincent. 2000. Development of critical level
methodologies for toxic metals in soils and surface waters—Dynamic modeling. Environmental
Diagnostics GST/04/1709 final report. Centre for Ecology & Hydrology, NERC. UK.
Tretner, A., M. Kofod, C. Scholz, and M. Isenbeck-Schroter. 2001. Influence of the geochemical
conditions and the input species on the species distribution of As in groundwater. 10th Intl. Symp.
on Water Rock Interaction. June 10-15, 2001. Villasimius, Italy.
Turekian, K.K. 1977. The fate of metals in the oceans. Geochim. et Cosmochim. Acta 41(8): 1139.
104
-------
Tye, A.M., S.D. Young, N.M J. Crout, H. Zhang, S. Preston, V.L. Barbosa-Jefferson, W. Davison,
S.P. McGrath, G.I. Paton, K. Kilham and L. Resende. 2003. Predicting the activity of Cd2+ and
Zn2+ in soil pore water from the radio-labile metal fraction. Geochim. et Cosmochim. Acta
67(3):375-85.
Tyler, G., and T. Olsson. 2001a. Concentrations of 60 elements in the soil solution as related to
the soil acidity. Eur. J. Soil Sci. 52:151-165.
Tyler, G., and T. Olsson. 2001b. Plant uptake of major and minor mineral elements as influenced
by soil acidity and liming. Plant Soil 230:307-321.
U.S. EPA, 2003a. Guidance for Developing Ecological Soil Screening Levels. Office of
Emergency and Remedial Response, Washington DC. OSWER Directive 9285.77-55.
U.S. EPA. 2003b. Air pollution and water quality: Atmospheric Deposition Initiative,
http ://www.epa.gov/ owow/ oceans/ airdep/index.html.
U.S. EPA. 2002a. List of drinking water contaminants & MCLs.
http://www.epa.gov/safewater/mcl.html.
U.S. EPA. 2002b. National recommended water quality criteria. EPA-822-R-02-047. Office of
Water, Office of Science and Technology.
U.S. EPA. 2000. Welcome to STORET, EPA's largest computerized environmental data system.
http://www.epa.gov/storet.
U.S. EPA. 1999a. Compendium of methods for the determination of inorganic compounds in
ambient air. EPA/625/R-96/010.
U.S. EPA. 1999b. Understanding variation in partition coefficient, Kd, values. EPA 402-R-99-
004A&B.
U.S. EPA. 1998. Human health risk assessment protocol for hazardous waste combustion
facilities. EPA 530-D-98:001.
U.S. EPA. 1997. Mercury study report to Congress. Volume III: Fate and transport of mercury in
the environment. EPA 452/R-97-005.
U.S. EPA. 1996. Background document for EPACMTP: Metals transport in the subsurface.
Prepared for U.S. EPA Office of Solid Waste by HydroGeologic under Contract 68-W7-0035.
(August).
U.S. EPA. 1995. Water quality assessment: A screening procedure for toxic and conventional
pollutants in surface and ground water. EPA/600/6-85/002.
105
-------
U.S. EPA. 1994. EPA health effects notebook for hazardous air pollutants—Draft. EPA 452/D-
95-003.
U.S. EPA. 1991. MINTEQA2, A geochemical assessment data base and test cases for
environmental systems: Ver. 3.0 user's manual. EPA/600/3-91/-21. Athens, GA. Allison, J.,D.,
D.S. Brown, and K.J. Novo-Gradac.
USGS (U.S. Geological Survey). 1998. Reduction-oxidation potential (electrode method). Chapt
6.5. In: National field manual for the collection of water-quality data. U.S. Geological Survey,
Water Resources, Office of Water Quality.
http://water.usgs.gOv/owq/FieldManual/Chapter6/6.5.html.
Webster's New World Dictionary. 1970. 2nd college edition. New York: The World Publ. Co.
Welch, A.H., D.B. Westjohn, D.R. Helsel, and R.B. Wanty. 2000. Arsenic in ground water of the
United States: Occurrence and geochemistry. Ground Wat. 38(4):589-604.
Westall, J.C. and H. Hohl. 1980. A comparison of the electrostatic models for the oxide/solution
interface. Adv. Colloid Interface Sci. 12:265-294.
White, D.E., J.D. Hem, and G.A. Waring. 1963. Chemical composition of subsurface waters. In:
Data of geochemistry, 6th ed. U. S. Geol. Survey Professional Paper 440-F.
Whittemore, D.O., and D. Langmuir. 1975. The solubility of ferric oxyhydroxides in natural
waters. Ground Wat. 13(4):360-365.
Wood, W.W. 1976. Guidelines for collection and field analysis of ground-water samples for
selected unstable constituents. Chapt. D2, Book 1, Techniques of water-resource investigations of
the U. S. Geol. Survey.
Xue, H.B., and Sigg, L. 1999. Comparison of the complexation of Cu and Cd by humic or fulvic
acids and by ligands observed in lake waters. Aquat. Geochem. 5(4):313-335.
Yeh, G.T., and V.S. Tripathi. 1991. A model for simulating transport of reactive multispecies
components: Model development and demonstration. Water Resour. Res. 27(12):3075-3094.
106
-------
^ ^ _ y 'O
|®j
What Did You Think?
We strive to constantly provide the highest level of value for you. Please take a few minutes
to tell us about your experience using this product.
To be taken to a short consumer satisfaction survey, please click here or copy and paste the
following URL into your browser:
https://www.surveymonkev.eom/r/OSAconsumerfdbck?
product=Issue Paper On Environmental Chemistry of Metals
Thank you for your feedback.
Sincerely,
Office of the Science Advisor
United States Environmental Protection Agency
www.epa.gov/OSA@epa.gov
------- |