Environ. Sci. Technol. 2004, 38, 6561-6570
Formation of Metal-Arsenate
Precipitates at the Goethite-Water
Interface
M A R K U S G R Ä F E , *
MAARTEN NACHTEGAAL, AND
DONALD L. SPARKS
Environmental Soil Chemistry Group,
Department of Plant and Soil Sciences, 152 Townsend Hall,
University of Delaware, Newark, Delaware 19717-1303
Little information is available concerning cosorbing
oxyanion and metal contaminants in the environment, yet
in most metal-contaminated areas, cocontamination by
arsenate [AsO4, As(V)] is common. This study investigated
the cosorption of As(V) and Zn on goethite at pH 4 and
7 as a function of final solution concentration. Complimentary
extended X-ray absorption fine structure (EXAFS)
spectroscopic data were collected at the As and Zn
K-edges in order to glean information about the coordination
environment of As and Zn at the goethite-water interface.
Macroscopic sorption studies revealed that As(V) and
Zn sorption on goethite increased in cosorption experiments
beyond that suggested by single sorption isotherms. At
pH 4 and 7, As(V) surface saturation was 3.2 and 2.2 µmol
m-2, respectively, and Zn surface saturation was absent
at pH 4 and ∼ 1.0 µmol m-2 at pH 7. Arsenate sorption on
goethite increased in the presence of Zn by 29% and by
more than 500% at pH 4 and 7, respectively. In the presence
of As(V), Zn sorption on goethite increased by 800 and
1300% at pH 4 and 7, respectively. More As(V) than Zn sorbed
on goethite below surface saturation at pH 7. Above
surface saturation, the Zn:As surface density ratio (SDR)
remained constant at 0.91 ( 0.03. At pH 4, the Zn:As SDR
was less than 1 throughout the concentration range.
Below As(V) surface saturation on goethite, As(V) formed
bidentate binuclear bridging complexes on Fe and/or Zn
octahedra, while Zn mainly formed edge-sharing complexes
with Fe at the goethite surface. Above surface saturation,
Zn was increasingly complexed by AsO4, gradually
forming an adamite-like [Zn2(AsO4)OH] surface precipitate
on goethite. Precipitated contaminants are more stable
due to the limited dissolution kinetics of their solid phase.
This study may therefore prove useful in remediation
strategies of sites knowingly contaminated with oxyanions
and metals.
Introduction
The affinity of many di- and trivalent metal cations (Co, Ni,
Cu, Zn, Cd, Pb) and oxyanions (AsO4, PO4, SeO4) for pHdependent surface sites of iron and aluminum (hydr)oxides
(e.g., goethite and gibbsite) could potentially result in
competitive reactions at the mineral-water interface (1-4).
Metal-contaminated sites are often burdened, in addition to
* Corresponding author phone: +61 08 8302 5062; fax: +61 08
8302 3057; e-mail: markus.grafe@unisa.edu.au.
10.1021/es035166p CCC: $27.50
Published on Web 10/26/2004
 2004 American Chemical Society
the metal contaminants, by minor yet sufficiently toxic
quantities of arsenic (As) due to the use of common parent
materials in smelting operations or applications of metalarsenate bearing materials (e.g. copper chromated arsenate
for wood treatment, zinc-arsenate pesiticides, etc.) (3-5).
Aqueous concentrations of As in mine drainage range
between 5 and 7000 µg L-1 and in isolated cases reaches
70 000 µg L-1 (6, 7). The little information available about the
sorption behavior of cosorbing metals [e.g. Zn(II)] and
oxyanions (e.g. AsO4) is contradictory. Langner and coworkers (8) found that Fe3+ and AsO4 coprecipitate in thermal
springs following the oxidation of arsenite [As(III)] to arsenate
[As(V) ) AsO4]. Tournassat and co-workers (9) observed the
formation of a manganese-arsenate precipitate at the
birnessite-water interface ca. 162 h following the oxidation
of initially 11 mM As(III) to As(V) by the birnessite surface.
Alternatively, Waychunas and co-workers (4) reported that
AsO4 inhibited the formation of ferrihydrite by binding to
growth sites on the iron hydroxide. Sadiq (10) suggested that
on the basis of thermodynamic calculations Cu, Ni, Zn, and
Pb arsenates were less soluble than Ca3(AsO4)2. At alkaline
pH, Ca3(AsO4)2 controlled As concentration in soil solutions,
while at acidic pH, As concentrations in solution are
controlled by Al- and Fe-arsenates.
Arsenic, and to a lesser extent Zn, are highly toxic to plants,
animals, and humans at elevated concentrations (11, 12).
Their stabilization in contaminated soils, sediments, or
drainage ponds requires an understanding of how the ions
behave in each other’s presence under differing environmental conditions (e.g., pH, type of sorbent, solid:solution
ratio, concentration, ionic strength).
Arsenate binds via a ligand-exchange mechanism to
surface hydroxyl functional groups on iron and aluminum
(hydr)oxides, forming inner-sphere sorption complexes, as
elucidated by both extended X-ray absorption fine structure
(EXAFS) spectroscopy and infrared spectroscopy (4, 13-15).
On iron oxides, AsO4 forms a bidentate, binuclear surface
complex (2C) on two equatorially bonded Fe-octahedra with
As-Fe distances ranging between 3.25 and 3.30 Å. An As-Fe
distance of 3.45 Å is characteristic of another 2C AsO4 surface
complex with two distinct, nonbonded Fe-octahedra (4).
Arsenate-Fe distances beyond 3.50 Å are characteristic of
a single corner sharing complex (1V) (15). The rate of the
As(V) sorption reaction is very fast, occurring on a time scale
of microseconds with 80 to 90% of the reaction complete
within the first 2 h (16-18). With increasing pH, the stability
of the As(V) sorption complex is lowered by the increasing
competition from OH- groups in solution or by other ligands
such as PO4 or dissolved organic acids (16-21).
In contrast, Zn sorption on metal oxides increases with
rising pH (1, 22, 23). A wide range of Zn sorption complexes
have been identified with EXAFS spectroscopy, with the
nature of the surface complex depending on the properties
of the sorbent and the coordination environment around
Zn. Waychunas and co-workers (3) observed tetrahedral Zn
in ferrihydrite suspensions and Zn-Fe distances from 3.10
to 3.50 Å for adsorbed and coprecipitated samples. Trivedi
and co-workers (24) also observed mostly tetrahedral Zn on
goethite, with a Zn-Fe radial distance ∼ 3.51 Å, indicating
a bidentate corner-sharing complex. Schlegel et al. (25)
observed edge-sharing complexes between Zn and Fe
octahedra with an average Zn-Fe distance of 3.00 Å and a
vertex-sharing complex at 3.20 Å.
The objectives of the study were to determine the solidphase partitioning of cosorbing AsO4 and Zn at two common
(soil) pHs (4 and 7) and to investigate the effects of changing
VOL. 38, NO. 24, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6561
surface loadings on the atomic arrangement of As(V) and Zn
at the goethite surface using EXAFS spectroscopy. Arsenate
and Zn were chosen for this study, because they commonly
occur together in mining areas (6, 7). Additionally, goethite
was chosen because of its ubiquity in nature and its
importance as a sorbent for As(V) and Zn (26). Finally, the
pH (4 and 7) used in this study is commonly found in the
environment and reflects pH environments of opposing
affinities for As(V) and Zn on the goethite surface.
Experimental Section
Materials. Goethite (R-FeOOH) was prepared in an open
atmosphere following standard procedures described elsewhere (16, 27). The formation of goethite was confirmed
using X-ray diffraction (XRD) and thermogravimetric analysis
(TGA, data not shown). Scanning electron microscopy
(current ) 1 keV; Hitachi SEM S4700; Delaware Biotechnology
Institute, Newark, DE) identified a characteristic needle- or
rod-shaped material with an average particle size of 30 × 200
nm. A five point BET N2(g) adsorption isotherm indicated a
specific surface area of 70 m2 g-1 and ca. 2.4% porosity. The
goethite was stored as a freeze-dried powder in a capped
centrifuge bottle inside a desiccator.
Sorption Isotherms. Sorption isotherms were conducted
by hydrating a 100 mg L-1 goethite suspension overnight at
pH 4 or 7 and in 0.01 M NaCl. All wet chemistry experiments
were conducted in a N2(g)-filled glovebox in duplicate;
averaged results are reported. All reagents used in these
experiments were reagent grade, and stock solutions, acids,
and bases were also prepared in the glovebox. All glassware
was acid washed prior to contact with the reagents. The pH
was monitored using a Metrohm 718 pH stat (Metrohm,
Herisau, Switzerland). Arsenate or Zn stock solutions were
added in increments to achieve a cumulative concentration
range from 1 to 100 µM. The equilibration period between
each increment was g2 h. Kinetic experiments were conducted to test the rate of sorption of 0.25 mM As(V) and Zn
in 1000 mg L-1 goethite suspensions. This experiment showed
that 80-90% of the reaction occurred within 2 h and assured
that As(V) and Zn concentrations were such that oversaturation did not occur [IAP/Kso < 1 for a Zn3(AsO4)2‚2.5H2O (Kso
) 1026.2)] (28). A 4 mL aliquot was removed for As, Zn, and
Fe ICP analysis, and the total volume was recalculated on
the basis of the amount of titrant added and the sampling
volume removed in order to determine the appropriate
amount of As(V) or Zn(II) stock solution needed for the next
increment. Data were plotted for final solution concentration
versus site density (Figure 1a,b). The Zn:As surface density
ratio (SDR) on goethite, calculated by taking the ratio of the
Zn site density vs As, was plotted as a function of the
cumulative concentration range (Figure 1c).
EXAFS Sample Preparation. Arsenic and Zn EXAFS
samples were prepared by hydrating 2.5 g L-1 goethite
suspensions overnight at pH 4 or 7 in 0.01 M NaCl. The surface
loadings of As(V) and Zn(II) on goethite were increased by
applying an increasing number of 0.25 mM As(V) and Zn
increments to the suspensions (Table 1). The equilibration
period between each increment was determined from the
time required for the pH to stabilize and was not less than
30 min. After the final increment was added and the pH had
stabilized, the samples were equilibrated for an additional
24 h on a reciprocating shaker at 150 rpm.
A homogeneous zinc-arsenate precipitate (HZAP) reference phase was prepared by reacting 10 mM ZnCl2 with 10
mM Na2HAsO4 at pH 7 in 0.01 M NaCl. The pH of the
precipitate suspension was maintained with 1 M NaOH and
monitored for at least 2 h (∆pH ) 0) before placing it on a
reciprocating shaker at 150 rpm for 24 h. After 24 h, 9.996
mM Zn and 6.389 mM As had been removed from solution,
resulting in a Zn:As solid-phase ratio of ∼1.57:1 (see Table
6562
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 24, 2004
FIGURE 1. Sorption and cosorption isotherms of (A) As(V) and (B)
Zn, and (C) the Zn:As site density ratio (IS ) individual sorption,
solid symbols 2 and 9; CS ) cosorption, open symbols 4 and 0).
1). After 24 h the pH was reconfirmed for all samples (7.00
( 0.03) and all samples were centrifuged at room temperature
at 13 000 rpm for 20 min. The solid samples were mounted
at the beamline into Teflon sample holders between two
sheets of Kapton tape (Type K-104, E. I. DuPont, Wilmington,
DE).
EXAFS Spectroscopy Data Collection. Arsenic and Zn
K-edge EXAFS spectroscopy data were collected at beamline
X-11A of the National Synchrotron Light Source (Brookhaven
National Laboratory, Upton, NY). The electron storage ring
was operated at 2.5 GeV with a current ranging from 300 to
130 mA. The monochromator consisted of two parallel Si(111) crystals with a vertical entrance slit separation of 0.5
mm. The ionization chamber was filled with 90% N2 and
TABLE 1. Sample Preparation for EXAFS Spectroscopya
a
SDR ) surface density ratio. No. incr ) number of increments of 0.25 mM Zn(II) and As(V).
10% Ar gas for As K-edge EXAFS experiments and 85 and
15%, N2 and Ar, respectively, for Zn K-edge EXAFS experiments. All sorption samples were oriented at 45° to the
incident beam and a Lytle Cell detector was used to collect
spectra in fluorescence mode. For As EXAFS experiments,
the Lytle Cell detector was purged every 6 h with Kr gas,
while for Zn EXAFS experiments, the Lytle Cell detector was
purged continuously with Ar gas. The monochromator
crystals were detuned by 30% in I0 to reject higher order
harmonics. For signal optimization and removal of elastically
scattered radiation, the fluorescence signal was filtered by
a Ge (As) or Cu (Zn) foil, one or two Al foils, and Soller slits.
The monochromator angle was calibrated to the As(V) K-edge
(11.874 keV) using a diluted Na2HAsO4 standard (10 wt %
BN). The Zn K-edge (9.659 keV) was calibrated using a metallic
zinc foil. These standards were monitored in transmission
mode simultaneous to sample collection to check for
potential shifts in their respective K-edges. Multiple scans
were collected at room temperature for each sample to
improve the signal-to-noise ratio for data analysis. Spectra
of the homogeneous zinc-arsenate precipitate (HZAP),
adamite [Zn2(AsO4)(OH)], scorodite (FeAsO4·2H2O), ojuelaite
[ZnFe3+2(AsO4)2(OH)2·4H2O] (all obtained from Excalibur) and
aqueous Na2HAsO4 and ZnCl2 reference standards were
collected in transmission mode.
EXAFS Data Analysis. All data reduction was performed
using WinXAS 2.1 (29) and the following procedure (30).
Individual spectra were background-corrected and normalized prior to averaging. The spectra were then converted
from energy to photoelectron wave vector (k) units (k ) is
the wave vector number with units of Å-1) by assigning the
origin, E0, to the first inflection point of the absorption edge
[As(V) ) 11.874 keV; Zn ) 9.659 keV]. EXAFS spectra were
extracted using a cubic spline function consisting of e7 knots
applied over an average range in k-space (As, 2.0-13.0 Å-1;
Zn, 1.5-11.0 Å-1). Fourier transformation (FT) of the raw
k3χ(k) function was performed over a consistent region in
k-space (As, 2.75-12.50 Å-1; Zn, 2.0-10.0 Å-1) to obtain a
radial structure function (RSF) using a Bessel window function
and a smoothing parameter (β) of 4 to minimize the effects
from truncation in the RSFs. The FEFF 7.02 code (31) was
used to calculate theoretical phase and amplitude functions
of As-O, As-Fe, As-Zn, and Zn-O, Zn-As, Zn-Fe, and
Zn-Zn scattering paths using input files based on the
structural refinement of scorodite (FeAsO4‚2H2O), ojuelaite
[ZnFe3+2(AsO4)(OH)2‚4H2O], mapimite [Zn2Fe3(AsO4)3(OH)4‚
10H2O], adamite [Zn2(AsO4)OH], and franklinite (ZnFe2O4).
Individual shell contributions in the RSFs of reference
standards and model sorption samples were Fourier backtransformed for nonlinear least-squares shell fitting. The
VOL. 38, NO. 24, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6563
TABLE 2. Structural Parameters Derived from Nonlinear Least-Square Fitting of the raw k3-weighted χ(k) As Spectraa
As-O
sample
atom
As-metal
CN ( 20%
R ( 0.02
σ2 ( 0.0004
ΓAs ) 1.41
O
4.2
1.68
0.0028
ΓAsZn ) 1.48
ΓAsZn ) 4.16
O
O
5.1
4.7
1.69
1.69
0.0030
0.0024
ΓAs ) 1.01
O
4.3
1.68
0.0024
ΓAsZn ) 1.08
O
5.0
1.70
0.0022
ΓAsZn ) 3.77
ΓAsZn ) 7.80
O
O
4.8
5.0
1.70
1.70
0.0025
0.0028
ΓAsZn ) 13.20
O
4.6
1.69
0.0023
Na2HAsO4
HZAP
O
O
4.6
3.8
1.68
1.70
0.0034
0.0026
adamite
O
4.4
1.71
0.0017
ojuelaite
scorodite
O
O
4.5
5.1
1.70
1.68
0.0031
0.0032
CN ( 30%
R ( 0.05
σ2 ( 0.0016
∆E0
res
pH 4
Fe
Fe
Fe/Zn
Fe/Zn
1.5
0.8
1.3
1.4
3.28
3.47
3.32
3.30
0.0032
0.0032
0.0054
0.0045
2.12
36.95
3.23
2.97
24.68
15.52
pH 7
Fe
Fe
Fe/Zn
Fe/Zn
Fe/Zn
Fe/Zn
Zn
Zn
1.2
0.9
1.6
0.7
1.7
2.5
2.2
5.9
3.27
3.43
3.32
3.54
3.33
3.29
3.44
3.34
0.0030
0.0030
0.0036
0.0036
0.0073
0.0061
0.0061
0.0087
2.07
40.68
3.29
22.41
3.47
3.53
18.01
13.87
3.10
15.07
9.4
1.8
2.3
7.9
3.3
1.8
3.7
3.10
3.28
3.43
3.36
3.53
3.34
3.34
0.0034c
0.0045
0.0045
0.0061
0.0061
0.0071
0.0050
3.43
3.34
12.88
13.04
4.58
17.22
3.21
3.16
15.31
24.04
atom
References
O-O-Asb
Zn
Zn
Zn
Zn
Fe/Zn
Fe
a CN ) coordination number. R ) bond distance (Å). ∆E ) phase shift (eV). σ2 ) Debye-Waller factor (Å-2); values of the second shell are
0
correlated to each other. b O-O-As represents multiple scattering (MS) in the As-tetrahedron with an ideal CN of 12. c σ2-value of the MS path
is correlated to the σ2-value of the As-O path.
obtained parameters were then used for the final nonlinear
least-squares multishell fit of the raw k3-weighted χ(k)
function. The FEFF 7.02 reference code was used to calculate
the theoretical phase and amplitude function of a noncollinear O-O-As multiple scattering path using an input file
based on the structural refinement of scorodite. Noncollinear
O-O-As multiple scattering contributions to the EXAFS was
fit to the aqueous Na2HAsO4 spectrum by correlating the
disorder of the As-O single scattering path to that of the MS
O-O-As path. The number of permissible free floating
parameters (Npts) in all samples ranged between 15 and 18,
depending on Zn versus As EXAFS data (32). For each fit, the
coordination number (CN), the radial distance (R), the
disorder of the radial distance (i.e., the Debye-Waller factor,
σ2 ∼ Å2), and a single, cross-correlated phase shift value (∆E0)
for all backscatter paths were allowed to vary. The amplitude
reduction factor (S02) was fixed to unity for As fits and to 0.85
for Zn fits. Results reported are for multi-shell fits of raw
k3-weighted χ(k) functions (Table 2).
Results and Discussions
Sorption Isotherms. Arsenate sorption on goethite was
greater at pH 4 than at pH 7 and indicated surface saturation
on goethite at 3.2 and 2.2 µmol of As m-2 goethite, respectively,
in the absence of Zn (Figure 1A). In the presence of Zn, As(V)
sorption increased by 29% at pH 4 and by more than 500%
at pH 7.
Zinc sorption on goethite was greater at pH 7 than at pH
4 and indicated surface saturation on goethite at ∼1 µmol
of Zn m-2 goethite, with little or no sorption (∼0.3 µmol of
Zn m-2) occurring in pH 4 suspensions (Figure 1B). In the
presence of As(V), Zn sorption on goethite increased by
∼800% at pH 4 and by ∼1300% at pH 7.
The Zn:As surface density ratio (SDR) suggested greater
As(V) than Zn sorption on goethite in all cosorption experiments (Figure 1C). At pH 7, the Zn:As SDR was 0.91 ( 0.03
and at pH 4 it was 0.66 ( 0.07. These results were in contrast
to the predicted Zn:As SDRs by the individual sorption
isotherms (IS, solid symbols in Figure 1c). The Zn:As SDRs
at pH 4 and 7 are consistent with the accepted theory that
6564
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 24, 2004
As(V) sorption complexes on goethite are more stable at lower
than at higher pH and vice versa for metal cations such as
Zn (1). However, the solid-phase partitioning of cosorbing
As(V) and Zn fractions on goethite were greater at pH 7 than
at pH 4 and exceeded sorption levels in all single-sorption
isotherms. This result is in contrast to the accepted theory
that As(V) stability on goethite is greater at lower pH (1). The
results suggested that the presence of Zn induced a more
stable As(V) surface complex at pH 7 than at pH 4. Therefore,
it is important to evaluate the possible presence and
concentration of counterions in solution with respect to
oxyanion and metal cation sorption on goethite. Sorption
above surface saturation for either ion may be due to
precipitation reactions, solid solution formation, or sorption
to low affinity sites (22, 33, 34).
As K-Edge EXAFS Spectroscopy. Complimentary extended X-ray absorption fine structure (EXAFS) spectroscopy
experiments were conducted at the As K-edge to glean
information about the coordination environment of As in
order to describe the nature of the surface complexes as a
function of surface loading. Details about each sample’s
preparation, surface loading, and Zn:As SDR may be found
in Table 1. Structural parameters derived from nonlinear
least-squares fitting of the raw k3-weighted χ(k) As spectra
are provided in Table 2.
Figure 2A,B shows the raw and fitted k3-weighted χ(k)
spectra of As reference compounds and the corresponding
radial structure functions (RSFs). The EXAFS was dominated
by first-shell oxygen contributions whose amplitude decreased in χ(k) spectra with increasing wavenumbers (Å-1),
as evident from the aqueous AsO4 χ(k) spectrum and a single
peak in the RSF at R + ∆R ) 1.4 Å (uncorrected for phase
shifts). The average first-shell CNAs-O for all samples and
reference standards was 4.5 ( 0.3 at an average R ) 1.69 (
0.01 Å and suggested that As(V) was coordinated by oxygen
atoms in tetrahedral configuration (4, 15, 17, 20, 35, 36) (Table
2). In Na2HAsO4(aq), O-O-As multiple scattering (MS) was
demonstrated by fitting a MS path to the aqueous AsO4
standard with CNO-O-As of 9.4 at R ) 3.10 Å. This result is
comparable to O-O-Cr MS observed in spectra of aqueous
FIGURE 2. (A) Raw (solid line) and fitted (dotted line) k3-weighted χ(k) spectra of As reference compounds, and (B) corresponding radial
structure functions (RSFs).
K2CrO4 solutions (37). In contrast to the aqueous As(V)
standard, all precipitated and mineral reference spectra had
contributions from neighboring second-shell Zn and/or Fe
atoms observed from a beat pattern in their χ-spectra and
a RSF peak between R + ∆R ) 2.5-3.0 Å. Splitting in the first
oscillation between 4 and 5.5 Å-1 in the χ(k) spectra of the
HZAP, adamite, and scorodite was due to multiple scattering
in the As tetrahedron (O-O-As) and Zn-O-As or Fe-OAs (scorodite) multiple scattering (Figure 2A).
For the HZAP spectrum, nonlinear least-squares fitting
suggested that As(V) was coordinated by 1.8 Zn atoms at 3.28
Å and 2.3 Zn atoms at 3.43 Å, consistent with the structure
of koettigite [Zn3(AsO4)2‚8H2O] (38) and corroborated by the
Zn:As ratio (1.57) of the homogeneous precipitate (Table 2).
In adamite, As(V) was coordinated by 7.9 Zn atoms at 3.35
Å and 3.2 Zn atoms at 3.53 Å. The EXAFS fit results for adamite
were in good agreement with the bonding environment of
As(V) in adamite observed elsewhere (39). In ojuelaite,
nonlinear least-squares fitting suggested that As(V) was
coordinated by 1.8 Fe/Zn atoms at 3.34 Å, consistent with
this mineral’s structure (40). One should observe that the
imaginary phase of ojuelaite is shifted to a lower R + ∆R (Å)
in comparison to the imaginary phases of the HZAP and
adamite reflecting a phase shift of ∼45° or 0.25π. A similar
method to differentiate between second-shell atomic neighbors was used by Manceau and co-workers (41) deciphering
Ni speciation in lithiphorite [(Al,Ni)MnO2(OH)2]. In our study,
it reflects a ∼14% difference in ionic radius between Zn2+
and Fe3+ (42) and the distance(s) at which the neighboring
metal atoms occur (41). In scorodite, the imaginary phase
occurs approximately between that of ojuelaite and adamite,
which stems from three Fe atoms at an average distance of
∼3.27 Å (3.14, 3.27, and 3.39 Å) and a fourth Fe atom at a
distance of 3.74 Å (all distances per FEFF 7.02 calculations
from atomic coordinates in ref 43). In our scorodite sample,
nonlinear least-squares fitting suggested that As(V) was
coordinated by 3.7 Fe atoms at R ) 3.34 Å, consistent with
the scorodite structure. The most notable structural difference
between the 1:1 metal-arsenate, scorodite, and the 1.5:1
and 2:1 zinc-arsenates, koettigite and adamite, respectively,
is the absence of edge- or corner-sharing M-M octahedra,
all of which are present in the other reference minerals and
precipitates we investigated. Therefore, the imaginary phase
in the As scorodite RSF does not reflect any contributions
from bidentate binuclear As(V) complexes. Instead, contributions to the EXAFS stem from four Fe atoms at four different
Fe-As distances. This is an important feature in the imaginary
phase of RSFs of single and cosorption spectra (Figure 3A,B).
Figure 3A,B shows the raw and fitted k3-weighted χ(k) As
spectra at pH 4 (bottom) and pH 7 (top) for single (ΓAs, with
units of µmol of As m-2 goethite) and cosorption (ΓAsZn, with
units of µmol of As m-2 goethite) EXAFS samples and their
corresponding RSFs. The EXAFS was dominated by firstshell oxygen contributions similar to the aqueous AsO4 χ(k)
spectrum and a single peak in the RSF at R + ∆R ∼ 1.4 Å.
The first RSF peak did not change its position as a function
of surface loading, suggesting similar first ligand shell
coordination in all sorption samples (Figure 3B). The average
first-shell CNAs-O in all single and cosorption EXAFS spectra
was 4.5 ( 0.3 at an average R ) 1.69 ( 0.01 Å and suggested
that As(V) was coordinated by oxygen atoms in tetrahedral
configuration (Table 2).
Contributions to the EXAFS from neighboring secondshell Fe and Zn atoms were identifiable from a characteristic
beat pattern in the k3-weighted χ(k) spectra. The magnitude
of these contributions varied with As and Zn surface loading
on goethite and produced a peak of increasing magnitude
in the RSFs at R + ∆R ∼ 2.5-3.0 Å (Figure 3B). In the pH 7
χ(k) spectra of cosorption samples (ΓAsZn ) 1.08-13.20),
shoulder features and oscillations emerged between 6 and
7, 8 and 9, and 10 and 11 Å-1 (Figure 3A). The appearance
of these beat patterns caused the imaginary phase of
cosorption samples to shift to higher R + ∆R (Å) (Figure 3B).
In the single sorption spectrum (pH 7, ΓAs ) 1.01), the trough
of the imaginary phase is at R + ∆R ∼ 3 Å (see dashed vertical
line at 3 Å in Figure 3B). With increasing surface loading and
the presence of Zn, the imaginary phase shifts increasingly
to greater R +∆R (Å) until the peak of the imaginary phase
is almost at R +∆R ) 3 Å at ΓAsZn ) 13.20. In comparison to
the χ(k) spectra and RSFs of ojuelaite, HZAP, and adamite,
VOL. 38, NO. 24, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6565
FIGURE 3. (A) Raw (solid line) and fitted (dotted line) k3-weighted χ(k) As spectra of sorption (ΓAs) and cosorption (ΓAsZn) samples at pH
4 (bottom) and pH 7 (top), and (B) their corresponding radial structure functions (RSFs).
one can now see that the greatest similarities exist between
ojuelaite and cosorption samples at pH 4 and 7 with ΓAsZn e
4.16, HZAP and cosorption sample with ΓAsZn ) 7.80, and
adamite and cosorption sample ΓAsZn ) 13.20. Therefore, with
increasing surface loading, the nature of the As second shell
conformed increasingly to the backscattering of Zn over that
of Fe, suggesting that Zn was increasingly controlling the
coordination of As(V). In the pH 4 cosorption experiments,
the shift of the imaginary phase in the RSFs remained very
small, which reflects the preference of As(V) for the goethite
surface and a Zn:As SDR of 0.66 (Table 1).
Nonlinear least-squares multishell fitting of the single As(V) sorption samples on goethite at pH 4.0 and 7.0 showed
that As(V) was coordinated to the surface by ∼1.3 Fe atoms
at R ) 3.28 Å and 0.8 Fe atoms at R ) 3.45 Å (Table 2). These
fit results suggested that As(V) formed bidentate binuclear
bridging (2C) complexes (at 3.28 Å) on the goethite surface
with edge-sharing Fe octahedra (4, 15, 17). A previous study
suggested that the contribution from Fe at 3.45 Å is due to
a bidentate binuclear bridging complex of As with isolated
Fe-octahedra (4). Alternatively, this contribution may develop
from a single corner-sharing (or monodentate mononuclear)
As-Fe complex (1V; see Figure 5) at edge-sharing Feoctahedra, when the As-O-Fe angle is ∼132°, as is observed
with metal(+2)-arsenate minerals such as koettigite or
6566
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 24, 2004
erythrite [Co3(AsO4)2‚8H2O] (38). We did not attempt to fit a
bidentate mononuclear complex at R ∼ 2.80 Å, as Sherman
and Randall (35) suggested that these sorption complexes
were energetically unfavorable to form.
In cosorption samples of low surface loading (ΓAsZn e 3.77
µmol of As m-2), the second shell CNAs-Fe/Zn ranged between
1.2 (ΓAs ) 1.01) and 1.7 (ΓAsZn ) 3.77) atoms at 3.27 to 3.32
Å, suggesting bidentate binuclear complexation of As(V) on
goethite. Additional contributions from 0.7 Fe/Zn atoms
could be fit at R ∼ 3.54 Å in the first cosorption sample (ΓAsZn
) 1.01). These contributions probably stem from an adjacent
Zn atom sorbed on goethite, as indicated by the Zn K-edge
EXAFS fit results (see sample ΓZnAs ) 1.25 and Table 3). At
pH 4 and in the presence of Zn, As(V) also formed bidentate
bridging (2C) complexes on the surface with Fe and or Zn.
Contrary to the results at pH 7, however, neither CNAs-Fe/Zn
nor RAs-Fe/Zn increased significantly, despite a cumulative
application of 2.5 mM As(V) and Zn (sample 4.2). The best
fit results, χ(k) spectra and RSFs for cosorption samples at
pH 4 and 7 with ΓAsZn e 4.16, were most similar to those of
ojuelaite, suggesting mixed bidentate binuclear bridging (2C)
complexes of As(V) with Zn and Fe-octahedra at the goethite
surface.
At ΓAsZn ) 7.80, As(V) was coordinated by 2.5 Zn atoms
at R ∼ 3.29 Å and ∼2.2 Zn atoms at R ) 3.44 Å. The RSFs,
FIGURE 4. (A) Raw (solid line) and fitted (dotted line) k3-weighted χ(k) Zn spectra of sorption (ΓZn) and cosorption (ΓZnAs) samples at pH
7 and Zn reference compounds, and (B) their corresponding radial structure functions (RSFs).
FIGURE 5. Graphical presentation of the possible arrangements of Zn and As(V) at the goethite surface.
χ(k) spectra, and best fit results are in good agreement with
those of the HZAP and suggested that a koettigite-like surface
precipitate had formed. At ΓAsZn ) 13.20, As(V) was coordinated by 5.9 Zn atoms at 3.34 Å. The RSFs, χ(k) spectra, and
VOL. 38, NO. 24, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6567
TABLE 3. Structural Parameters Derived from Nonlinear Least-Square Fitting of the Raw k3-Weighted χ(k) Zn Spectraa
Zn-O
sample
CN ( 20%
R ( 0.02
Zn-As, Fe, Zn
σ2 ( 0.0024
ΓZn ) 0.66
ΓZnAs ) 1.25
5.3
4.8
2.04
2.06
0.0118
0.0072
ΓZnAs ) 3.61
5.4
2.01
0.0111
ΓZnAs ) 7.95
ΓZnAs ) 13.76
3.1
4.6
3.7
2.6
1.99
2.14
2.00
2.10
0.0070d
0.0070
0.0064d
0.0064
Zn(Cl)2(aq)
HZAP
6.7
5.8
2.09
2.09
0.0091
0.0106
adamite
3.2
2.1
1.98
2.11
0.0045d
0.0045
CN ( 30%
atom
Fe
Fe/Zn
As
Zn/Fe
As
As
Zn
Zn
As
Zn
R ( 0.05
σ2 ( 0.0033
∆ E0
res
1.0
1.0
0.5
1.1
1.1
0.7
1.2
1.5
2.6
1.8
3.14
3.14
3.62
3.18
3.32
3.41
3.70
3.05
3.37
3.60
0.0118b
0.0045
0.0045
0.0044c
0.0044
0.0059c
0.0059
0.0088c
0.0088
0.0088
1.47
1.31
31.90
44.07
0.98
24.57
3.34
24.99
2.73
18.81
1.9
1.5
0.8
1.0
0.9
3.38
3.66
3.00
3.38
3.63
0.0093c
0.0093
0.0021c
0.0021
0.0021
1.15
3.37
20.73
22.10
2.64
15.04
References
As
Zn
Zn
As
Zn
a CN ) coordination number. R ) radial distance (Å). σ2 ) Debye-Waller factor (Å-2). ∆E ) phase shift (eV).
0
is correlated among second metal shells. d σ2 in the first shell are correlated.
best fit results are in good agreement with those of adamite
and suggested that an adamite-like surface precipitate had
formed. The formation of zinc-arsenate surface precipitates
is suggested by CNs in the second shell exceeding a
cumulative value of 4 and was observed for samples with
ΓAsZn g 7.80. A nearly 4-fold increase in surface saturation
cannot be supported by As(V) sorption to low affinity sites
only or formation of a goethite-As(V) solid solution. The
increasing shift of the imaginary phase in RSFs, the similarity
between χ(k) spectra of cosorption samples and HZAP and
adamite, and the similarity of the best fit results therefore
support the formation of zinc-arsenate surface precipitates
with increasing surface loading.
Zinc K-Edge EXAFS Spectroscopy. Zinc K-edge EXAFS
spectroscopy data were collected to elucidate and corroborate
the role of Zn in zinc-arsenate surface precipitates as
suggested by As K-edge EXAFS results. Details about each
sample’s preparation, surface loading, and Zn:As SDR may
be found in Table 1. Structural parameters derived from
nonlinear least-squares fitting of the raw k3-weighted χ(k)
Zn spectra are provided in Table 3.
Figure 4A,B shows raw and fitted k3-weighted χ(k) Zn
spectra for sorption (ΓZn, with units of µmol of Zn m-2
goethite) and cosorption (ΓZnAs, with units of µmol of Zn m-2
goethite) samples at pH 7 and Zn reference phases and their
corresponding RSFs, respectively. The Zn EXAFS was dominated by first-shell oxygen contributions whose amplitude
decreased in k3-weighted χ(k) spectra with increasing wavenumbers (Å-1), as evident from the aqueous ZnCl2 χ(k)
spectrum and a single peak in the RSF between R + ∆R )
1.5-1.6 Å (uncorrected for phase shifts). A small shift of the
spectra to higher wavenumbers (Å-1) was apparent for ΓZnAs
) 3.61, suggesting a different coordination sphere around
Zn in this sample. A small shift of the first major peak to
lower R + ∆R (Å) between 1.4 and 1.5 Å in the RSFs of ΓZnAs
) 3.61 suggested that the Zn-O bond distance had decreased.
In all other samples the average distance of Zn to the first
ligand was 2.05 ( 0.03 Å with CN averaging 5.6 ( 1.0,
suggesting octahedral coordination of Zn by oxygen atoms.
In well-oxygenated environments, the first atomic shell of
Zn is usually occupied by four or six oxygen atoms, but
coordination environments of five and seven atoms surrounding Zn are also known (3, 24, 25, 44-46). Due to
common distortions in Zn’s first atomic shell and the
possibility, though rare, of mixed Zn coordination (e.g.,
6568
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 24, 2004
b
Correlated to σ2 of Zn-O. c σ2
adamite, CN ) 5 and 6; hydrozincite, CN ) 4 and 6), the true
coordination environment is usually determined from a
characteristic bond distance: for 4-fold coordination by
oxygen atoms, the Zn-O bond distance is e1.98 Å and for
6-fold coordination by oxygen atoms, the Zn-O bond
distance is g2.03 Å (3). In ΓZnAs ) 3.61, the first shell radial
distance was 2.01 Å with a CN of 5.4. The RZn-O ) 2.01 Å
could suggest minor contributions from 5-fold-coordinated
Zn atoms along with dominantly 6-fold-coordinated Zn
atoms, which is a common mixture in zinc-arsenate minerals
[adamite, CN ) 5 and 6 (39); paradamite, CN ) 5]. Similarities
of the Zn-O coordination in samples with ΓZnAs g 7.95 and
adamite corroborated the possible formation of pentahedrally
coordinated Zn at ΓZnAs ) 3.61. Contributions to the EXAFS
from neighboring second-shell Fe, Zn, or As atoms were
noticeable from a beat pattern in the k3-weighted χ(k) spectra.
The magnitude of these contributions varied with As and Zn
surface loading on goethite and resulted in several peaks of
varying magnitude and position in the RSFs between R + ∆R
) 2.0-4.0 Å (Figure 4A,B).
Best fit results (Table 3) showed that in the HZAP, Zn was
coordinated to 1.9 As atoms at 3.38 Å and 1.5 Zn atoms at
3.66 Å. There were minor contributions from ∼0.7 Zn atoms
at 3.08 Å. The fit results are in good agreement with those
of a koettigite-like precipitate, as suggested earlier by As
K-edge EXAFS. In adamite, the first shell was fitted with two
Zn-O scattering paths showing Zn coordination by 3.2 O
atoms at 1.98 Å and 2.1 O atoms at 2.11 Å. This distribution
reflects that for one 5- and one 6-fold-coordinated Zn
polyhedron in adamite, nine out of 11 O atoms occur at an
average distance of 2.03 Å and the remaining two O atoms
occur (ideally) at a distance of 2.26 Å (39). In the second
shell, Zn was coordinated by 0.8 Zn atoms at 3.00 Å, 1.0 As
atoms at 3.38 Å, and 0.9 Zn atoms at 3.63 Å. These results
are consistent with the average bonding environment of Zn
in adamite (39) consisting of edge-sharing Zn-octahedra, 2C/
2*1V complexed with AsO4 ligands in the zinc hydroxide
matrix, and a Zn-Zn distance between a 5- and a 6-foldcoordinated Zn atom, respectively (Figure 5).
At ΓZn ) 0.66, Zn was coordinated by ∼1 Fe atom at 3.14
Å, suggesting a corner-sharing complex on goethite (25) or
an edge-sharing complex (2E) (Table 3) (3, 47). In the presence
of As(V), the initial sorption complexes formed at low surface
loadings (ΓZnAs e 3.61) showed a consistent fit with ∼1 Fe
atom at a distance of 3.16 Å, suggesting similar vertex- or
edge-sharing complexes with Fe similar to Zn coordination
in the absence of As(V) (25, 47). The CNZn-Fe and RZn-Fe did
not change between ΓZn ) 0.66 and ΓΖnAs ) 1.25, 3.61, due
to the presence of As(V) in the cosorption samples; however,
the second shell disorders of the cosorption spectra (ΓZnAs e
3.61) were significantly lower, suggesting that As(V) promoted
edge- or vertex-sharing complexes of Zn on goethite. With
ΓZnAs increasing from 1.25 to 3.61, the RZn-As decreased from
3.62 to 3.32 Å, while the CNZn-As increased from 0.5 to 1.5,
suggesting a change from adjacent isolated Zn and As atoms
on the goethite surface at ΓZnAs ) 1.25 (see Figure 5, mapimite)
to a 2C Zn-As complex at ΓZnAs ) 3.61. These results are in
good agreement with fit results for the same samples at the
As K-edge and support our earlier hypothesis that Zn
promoted As(V) complexation at the goethite surface. This
should occur when Zn octahedra form edge-sharing complexes, creating additional O apexes on the goethite surface
for bidentate binuclear As(V) bonding.
Similarities observed for the χ(k) spectra and RSFs of
spectra at ΓZnAs ) 7.95 and the HZAP and of spectra at ΓZnAs
) 13.76 and adamite suggested that similar solid phases may
have formed in the cosorption samples and was consistent
with spectral fingerprinting of the same sample pairs in As
K-edge EXAFS. At ΓZnAs ) 7.95, Zn was coordinated by 3.1 O
atoms at 1.99 Å and 4.5 O atoms at 2.14 Å, suggesting mixed
Zn coordination in the first ligand shell. In the second shell,
Zn was coordinated by 0.7 As atoms at 3.41 Å and 1.2 Zn
atoms at 3.70 Å. Second-shell fit results were not different
from those of the HZAP; however, the absence of two different
Zn-O contributions in the HZAP spectrum suggested that
the surface precipitate formed in sample 7.3 did not have
quite the same structure as the HZAP. At ΓZnAs ) 13.76, Zn
was coordinated by 3.7 O atoms at 2.00 Å and 2.6 O atoms
at 2.10 Å in the first ligand shell, which was in good agreement
with the first-shell coordination of Zn in adamite. In the
second shell, Zn was coordinated by 1.5 Zn atoms at 3.06 Å,
2.6 As atoms at 3.37 Å, and 1.8 Zn atoms at 3.60 Å. The
disorder (σ2) for second-shell distances was ∼4 times greater
than in adamite. Bond distance disorder (σ2) and CN are
directly related (30), explaining why the magnitude of the
second-shell CN in spectrum ΓZnAs ) 13.76 was greater than
in adamite. The fit results suggested therefore that an
adamite-like precipitate had formed on the goethite surface
at the highest surface loading. On the basis of As and Zn
K-edge EXAFS data, we suggest that recurring, undersaturated
(IAP/Kso < 1) Na2HAsO4 and ZnCl2 solutions may result in
the formation of adamite-like zinc-arsenate surface precipitates on goethite at pH 7.
Surface Precipitation and Environmental Implications.
Zinc and As(V) sorption at the goethite-water interface were
significantly enhanced in cosorption experiments. Zinc
sorption on goethite in the presence of As(V) increased by
800 and 1300% at pH 4 and 7, respectively, while As(V)
sorption on goethite in the presence of Zn increased by 30%
and more than 500%. Zinc sorption on goethite was enhanced
probably by As(V) decreasing the positive surface charge,
while edge-sharing surface complexes of Zn on goethite
probably increased the number of available O/OH/H2O
functional groups for As(V) bonding at the goethite surface.
Below As(V) surface saturation on goethite at pH 7, 1.21.8 times more As(V) sorbed on goethite than Zn in cosorption
experiments (Figure 1c). At pH 7, As(V) exists as the negatively
charged HAsO42- (60%) and H2AsO4- species [40%, pKa2 )
6.96, (48)], while Zn occurs as the positively charged Zn(H2O)62+ molecule and the surface charge on goethite is
usually still positive (no experimental point of zero charge
was determined) (1). Hence, As(V) (over Zn) sorption on the
goethite surface was favored. The Zn:As site density ratio
(SDR) in pH 4 cosorption experiments (0.66 ( 0.07) corroborated this trend as well as the smaller shifts of the
imaginary phase in As RSFs of cosorption samples (ΓAsZn e
3.77).
Above As(V) surface saturation on goethite, the marginal
Zn:As SDR stabilized at 0.91 ( 0.03, suggesting a nearly 1:1
uptake of Zn and As(V) and the possible formation of a similar
1:1 ZnHAsO4‚3H2O surface precipitate. Arsenic and Zn K-edge
EXAFS experiments, however, showed that an adamite-like
surface (Zn:Asunitcell ∼ 2:1) precipitate was forming on the
goethite surface between 7 and 14 µmol of As/Zn m-2. The
discrepancy between macroscopic Zn:As SDR and the Zn:As
SDR suggested by EXAFS spectroscopy results may be due
to additional As(V) sorption occurring on the surface
precipitate, i.e., the surface precipitate provided functional
groups for additional As(V) complexation. Theoretically, a
unit cell of koettitgite (Zn:As ratio ) 1.5) or adamite (Zn:As
ratio ) 2) provides sufficient functional groups through O,
OH, or H2O ligands on Zn-octahedra to lower the Zn:As ratio
to 0.5. In kinetic studies of the formation of the HZAP, the
Zn:As ratio decreased from ∼1.5 after 24 h to ∼1 after 21 days
(manuscript in preparation). The Zn:As SDR helped to
quantify the Zn:As stoichiometry at the goethite surface;
however, it could not indicate which possible structures were
forming. It is unlikely that a 1:1 ZnHAsO4 surface precipitate
formed, because (a) 1:1 metal-arsenates form from trivalent
rather than divalent cations, e.g., scorodite (Fe3+AsO4‚2H2O)
or mansfieldite (Al3+AsO4‚2H2O); (b) 1:1 metal-arsenates do
not have certain structural elements observed in cosorption
samples, e.g., edge-sharing among (Zn) octahedra, bidentate
binuclear As-Zn coordination, M-M distances < 4 Å.
Our research showed that chronic or frequent exposure
to aqueous zinc-arsenate solutions (and potentially other
metal-oxo acid solutions) may be dealt with in a mutually
beneficial way (Zn:As removal ∼ 0.9) if the solutions enter
well-buffered, pH 7 stabilized (soil) environments. Goethite
effectively enhanced Zn and As(V) solid-phase partitioning
by providing surface OH groups for adsorption and subsequent precipitation reactions. The occurrence of metalarsenate precipitates is not merely a laboratory phenomenon.
In a copper chromated arsenate contaminated site, microfocused EXAFS spectroscopy suggested that more than 80 %
of As occurred in coprecipitated phases with copper and
zinc (manuscript in preparation). Coprecipitated rather than
adsorbed contaminants have generally lower bioavailability
and have lower contaminant loss due to leaching. Metalarsenate precipitates could therefore potentially increase the
time frame over which contaminated sites need to be treated
and could potentially promote less intrusive and less costly
techniques, e.g. phytoremediation vs excavation/landfilling.
The results of this study may therefore be useful in devising
remediation strategies for sites cocontaminated with Zn and
As(V) and possibly other metal-oxo acid pairs, e.g. Cu and
PO4. Additional research is needed to describe metal-oxo
anion sorption on other ubiquitous metal-oxides (gibbsite,
kaolinite) and the factors controlling cosorption reactions.
Acknowledgments
We are thankful to Dr. Matthew J. Eick (VPI&SU) for BETSSA, XRD, and TGA measurements; to Dr. Kirk Czymmek
and Deborah Powell (DBI) for help with the SEM analysis of
goethite; and to Dr. K. Pandya (X11A, NSLS/BNL) for her
support during data collection. Markus Gräfe appreciates a
competitive graduate fellowship from the University of
Delaware.
Literature Cited
(1) Sparks, D. L. Environmental Soil Chemistry; 2nd ed.; Academic
Press: San Diego, CA, 2002.
(2) McBride, M. B. Environmental Chemistry of Soils; 1st ed.; Oxford
University Press: New York, 1994.
VOL. 38, NO. 24, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6569
(3) Waychunas, G. A.; Fuller, C. C.; Davis, J. A. Geochim. Cosmochim.
Acta 2002, 66, 1119-1137.
(4) Waychunas, G. A.; Rea, B. A.; Fuller, C. C.; Davis, J. A. Geochim.
Cosmochim. Acta 1993, 57, 2251-2269.
(5) Welch, A. H.; Westjohn, D. B.; Helsel, D. R.; Wanty, R. B. Ground
Water 2000, 38, 589-604.
(6) Carlson, L.; Bigham, J. M.; Schwertmann, U.; Kyek, A.; Wagner,
F. Environ. Sci. Technol. 2002, 36, 1712-1719.
(7) Williams, M. Environ. Geol. 2001, 40, 267-278.
(8) Langner, H. W.; Jackson, C. R.; McDermott, R.; Inskeep, W. P.
Environ. Sci. Technol. 2001, 35, 3302-3309.
(9) Tournassat, C.; Charlet, L.; Bosbach, D.; Manceau, A. Environ.
Sci. Technol. 2002, 36, 493-500.
(10) Sadiq, M. Water Air Soil Pollut. 1997, 93, 117-136.
(11) Hall, A. H. Toxicol. Lett. 2002, 128, 69-72.
(12) Roberts, D. R.; Scheinost, A. C.; Sparks, D. L. Environ. Sci. Technol.
2002, 36, 1742-1750.
(13) Goldberg, S. Soil Sci. Soc. Am. J. 2002, 66, 413-421.
(14) Goldberg, S.; Johnston, C. T. J. Colloid Interface Sci. 2001, 234,
204-216.
(15) Fendorf, S.; Eick, M. J.; Grossl, P.; Sparks, D. L. Environ. Sci.
Technol. 1997, 31, 315-320.
(16) Grafe, M.; Eick, M. J.; Grossl, P. R. Soil Sci. Soc. Am. J. 2001, 65,
1680-1687.
(17) O’Reilly, S. E.; Strawn, D. G.; Sparks, D. L. Soil Sci. Soc. Am. J.
2001, 65, 67-77.
(18) Grossl, P. R.; Eick, M.; Sparks, D. L.; Goldberg, S.; Ainsworth, C.
C. Environ. Sci. Technol. 1997, 31, 321-326.
(19) Grafe, M.; Eick, M. J.; Grossl, P. R.; Saunders: A. M. J. Environ.
Qual. 2002, 31, 1115-1123.
(20) Arai, Y.; Elzinga, E. J.; Sparks, D. L. J. Colloid Interface Sci. 2001,
235, 80-88.
(21) Pierce, M. L.; Moore, C. B. Water Res. 1982, 16, 1247-1253.
(22) Stumm, W. Chemistry of the Solid-Water Interface; Wiley: New
York, 1992.
(23) Stumm, W.; Hohl, H.; Dalang, F. Croat. Chem. Acta 1976, 48,
491-504.
(24) Trivedi, P.; Axe, L.; Tyson, T. A. J. Colloid Interface Sci. 2001,
244, 230-238.
(25) Schlegel, M. L.; Manceau, A.; Charlet, L. J. Phys. IV 1997, 7,
823-824.
(26) Schwertmann, U.; Cornell, R. M. Iron Oxides in the Laboratory;
VCH: Weinheim, 1991.
(27) Schwertmann, U.; Cornell, R. M. Iron Oxides in the Laboratory:
Preparation and Characterization; VCH: Weinheim, 1991.
(28) Schecher, W. D.; McAvoy, D. C. MINEQL+ Environmental
Research Software, 1998.
6570
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 24, 2004
(29) Ressler, T. J. Synchrotron Radiat. 1998, 5, 118-122.
(30) Bunker, G. Overview of the standard XAFS data analysis
procedure. http://gbxafs.itt.edu/training/tutorials.html (accessed August 8, 2003).
(31) Zabinsky, F.; Rehr, J. J.; Ankudinov, A.; Albers, R. C.; Eller, M.
J. Phys. Rev. B Condens. Mater. 1995, 52, 2995-3009.
(32) Stern, E. A. Phys. Rev. B 1993, 48, 9825-9827.
(33) Sposito, G. The Surface Chemistry of Soils; Oxford University
Press: New York, 1984.
(34) Farley, K. J.; Dzombak, D. A.; Morel, F. M. M. J. Colloid Interface
Sci. 1985, 106, 226-242.
(35) Sherman, D. M.; Randall, S. R. Geochim. Cosmochim. Acta 2003,
67, 4223-4230.
(36) Randall, S. R.; Sherman, D. M.; Ragnarsdottir, K. V. Geochim.
Cosmochim. Acta 2001, 65, 1015-1023.
(37) Pandya, K. I. Phys. Rev. B 1994, 50, 15509-15515.
(38) Hill, R. J. Am. Mineral. 1979, 64, 376-382.
(39) Hill, R. J. Am. Mineral. 1976, 61, 979-986.
(40) Hughes, J. M.; Bloodaxe, E. S.; Kobel, K. D.; Drexler, J. W. Mineral.
Magn. 1996, 60, 519-521.
(41) Manceau, A.; Tamura, N.; Marcus, M. A.; MacDowell, A. A.;
Celestre, R. S.; Sublett, R. E.; Sposito, G.; Padmore, H. A. Am.
Mineral. 1998, 87, 1494-1499.
(42) Manceau, A.; Schlegel, M. L.; Musso, M.; Sole, V. A.; Gauthier,
C.; Petit, P. E.; Trolard, F. Geochim. Cosmochim. Acta 2000, 64,
3643-3661.
(43) Hiriyana, R.; Sakurai, K. X-ray Stud. Polymorphism Mag. 1949,
85-88.
(44) Manceau, A.; Marcus, M. A.; Tamura, N. In Applications of
Synchrotron Radiation in Low-Temperature Geochemistry and
Environmental Science; Fenter, P., Sturchio, N. C., Eds.; Mineralogical Society of America: Washington DC, 2002; Vol. 49,
pp 341-428.
(45) Roberts, D. R.; Ford, R. G.; Sparks, D. L. J. Colloid Interface Sci.
2003, 263, 364-376.
(46) Trainor, T. P.; Brown, G. E., Jr.; Parks, G. A. J. Colloid Interface
Sci. 2000, 231, 359-372.
(47) Ginderow, D.; Cesborn, F. Acta Crystallogr. 1981, B 37, 10401043.
(48) O’Neill, P. In Heavy Metals in Soils; Alloway, B. J., Ed.; Blackie:
London, 1990.
Received for review October 20, 2003. Revised manuscript
received July 26, 2004. Accepted August 16, 2004.
ES035166P