Share paper
OECD Field of science



Environmental Toxicology and Chemistry, Vol. 25, No. 12, pp. 3118-3124, 2006

© 2006 SETAC Printed in the USA 0730-7268/06 $12.00 + .00



Lei Luo,t Shuzhen Zhang,*! Xiao-Quan Shan,! Wei Jiang,! Yong-Guan Zhu,! Tao Liu,$ Ya-Ning Xie,$

and Ronald G. McLaren§

tState Key Laboratory of Environmental Chemistry and Ecotoxicology, Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, P.O. Box 2871, Beijing 100085, China ¿State Key Laboratory of Synchrotron Radiation, Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100039, China §Agriculture and Life Sciences Division, P.O. Box 84, Lincoln University, Canterbury, New Zealand

(Received 11 April 2006; Accepted 21 June 2006)

Abstract—Arsenic sorption is the primary factor that affects the bioavailability and mobility of arsenic in soils. To elucidate the characteristics and mechanisms of arsenate, As(V), sorption on soils, a combination of sorption isotherms, zeta potential measurements, and extended X-ray absorption fine-structure (EXAFS) spectroscopy was used to investigate As(V) sorption on two Chinese red soils. Arsenate sorption increased with increasing As(V) concentration and was insensitive to ionic strength changes at pH 6.0. Arsenate, mainly as H2AsOj in soil solution at pH 6.0, was strongly sorbed mainly through ligand exchange by the two soils. The sorption capacity was affected by the iron and aluminum mineral contents in the soils. The zeta potential measurements showed that As(V) sorption lowered the zeta potential and the points of zero charge of the soils. The EXAFS data indicate that adsorbed As(V) forms inner-sphere complexes with bidentate-binuclear configurations, as evidenced by an As-Fe bond distance of 3.28 ± 0.04 A and an As-Al bond distance of 3.17 ± 0.03 A. The two As(V) complexes were stable at different As(V) loadings, whereas the proportions were related to the aluminum and iron mineral contents in the soils. This study illuminated the importance of inclusion of microscopic and macroscopic experiments to elucidate sorption behavior and mechanisms.

Keywords—Arsenate Sorption Extended X-ray absorption fine structure Soil


Arsenic, both naturally or anthropogenically derived, is a highly toxic trace element in natural and agricultural environments and has received considerable attention in the scientific literature [1,2]. Arsenic is persistent in soil because of its strong affinity for soil clay minerals and metal oxides [3,4]. Understanding arsenic sorption on soil, especially soil sorption reactions and surface structures that control arsenic retention in soil, is therefore critical for assessing the mobility, toxicity, and bioavailability of arsenic.

Sorption of arsenate on individual minerals has been investigated extensively, and relevant sorption mechanisms have been proposed (e.g., [5-9]). Previous studies employing chemical and physical methods have provided direct evidence suggesting the formation of inner-sphere complexes of arsenate adsorbed on Fe- and Al-containing minerals [10-14]. However, these proposed sorption processes are based mainly on the reactions of arsenate with pure minerals, and microscopic studies of the mechanisms in whole soils have received much less attention. Because of the complex soil matrix, arsenate sorption on soil cannot be described completely by any of the mechanisms proposed for individual soil components; therefore, the use of whole soils to examine the complex sorption mechanisms is necessary. Although a great deal of research on arsenate sorption on soils has been conducted and related sorption mechanisms have been attributed to inner-sphere com-

* To whom correspondence may be addressed (

plex formation via ligand exchange [3,15], spectroscopic or direct structural evidence is scarce.

Both macroscopic and microscopic experimental methods can provide insight into arsenate sorption mechanisms. Shifts in points of zero charge (PZCs) corresponding with ion sorption have been used as evidence for strong specific ion adsorption and inner-sphere surface complex formation [8,15]. Evaluation of the effect of changes in ionic strength on sorption behavior is also considered another macroscopic means to infer sorption mechanisms [8]. However, these mechanisms, determined from macroscopic data, are often without microscopic data support. Extended X-ray absorption fine-structure (EX-AFS) spectroscopy is of particular importance for the characterization of sorbed species on sorbents at relatively low surface loadings. Furthermore, EXAFS spectroscopy can also be applied to poorly crystallized materials to determine the precise vicinity of sorbates.

Because of the structural and chemical complexity of soil matrices, EXAFS spectroscopy has rarely been used to evaluate sorption mechanisms in soil samples, as in the case of synthetic samples. A typical problem in EXAFS analysis for a whole soil is the "deconvolution" of backscattering contributions from two or more atoms of different atomic number at overlapping distances because, usually, more than one soil constituent is responsible for arsenate, As(V), sorption on soils [16]. However, knowledge of the structure of As(V) complexes sorbed on minerals gained from prior research has made it possible to apply the technique to real situations. Thus, a primary objective of this study was to employ EXAFS to fingerprint the structural environment of As(V) adsorbed in soils, which has not been reported to our knowledge.

Table 1. Physicochemical properties of the soils studied from southern China

Fe (mg/kg) Al (mg/kg)

Sampling pH SOC Clay SSA

site (1:1) (%)a CDb Oxalate CD Oxalate (%) (m/g)c Order

Yingtan 5.3 0.66 24,995 3,069 4,402 1,339 42.6 50.4 Ultisol

Guangzhou 4.9 0.81 71,059 3,438 12,807 2,646 74.8 62.6 Oxisol

a SOC = soil organic carbon.

b CD = citrate-dithionite extractable; oxalate = ammonium oxalate extractable. c SSA = specific surface area.

Two Chinese red soils of typical high aluminum and iron oxide contents were used in this study because of their high As(V) sorption capacity [17] and therefore suitable for EXAFS analysis. A combination of macroscopic and microscopic techniques was applied to elucidate the sorption mechanisms of As(V) on the soils. Sorption isotherms, the effects of ionic strength on As(V) sorption, and the effects of As(V) sorption on shifts in PZC and on K-edge EXAFS spectra were determined. These results allowed us to identify the dominant surface complex of As(V) on whole soils and to explain the As(V) sorption mechanism.


Two red soils from the B horizons (15-30 cm depth) of a Yingtan clay loam (Ultisol) and a Guangzhou clay (Oxisol) from southern China were used in this study. Soils were air dried and passed through a 2-mm sieve. Aluminum and Fe were extracted by citrate-dithionite and ammonium oxalate for the complete dissolution of aluminum and iron oxides, and the amorphous aluminum and iron oxides, respectively, as outlined by Manning and Goldberg [18]. Particle size distribution was obtained by laser particle analyzer (Malvern Mastersizer 2000, Malvern Instruments, Malvern, Worcs, UK) [17], and specific surface areas were determined with a Micromeritics Flowsorb II surface area analyzer (ASAP 2000, Micromeritics, Norcross, GA, USA) and by applying the Brunauer Emmett Teller equation to the sorption of N2 at a relative partial pressure of 0.3 at 77 K. Soil organic carbon and pH were determined by standard methods [19]. These properties are listed in Table 1. Crystalline clay minerals in the soils were analyzed by X-ray diffraction with CuK-a1 radiation.

Sorption isotherms

Sorption isotherms were conducted at pH 6.0 and a constant room temperature of 25°C. Triplicate 1.00-g air-dried soil samples were equilibrated with 20 ml of background electrolyte solution of 0.01 or 0.1 mol/L NaNO3 containing As(V) concentrations ranging from 0 to 2.0 mmol/L (from a stock solution of Na3AsO4 in 0.01 or 0.1 mol/L NaNO3 at pH 6.0) in 50-ml polypropylene centrifuge tubes. A wide range of As(V) concentrations was selected to compare the sorption capacity of the two soils. The mixtures were shaken at 15 rpm on a reciprocal shaker for 24 h and subsequently centrifuged for 10 min at 7,200 g at 25°C before filtration through a 0.45-^m Millipore membrane (Millipore, Bedford, MA, USA).

Arsenic concentrations were analyzed by hydride-generation atomic fluorescence spectrometry (AFS-610A, Beijing Rayleigh Analytical Instrument, Beijing, China). No arsenite was detected in the equilibrium solutions. Arsenate sorption was calculated from the difference between concentrations in the equilibrium supernatant solutions and those in the initial solutions. In the systems with a background electrolyte solu-

tion of 0.01 mol/L NaNO3, dissolved Fe, Al, Si, and P were determined by inductively coupled plasma atomic emission spectrometry (Optima 2000 DV, PerkinElmer, Wellesley, MA, USA), and dissolved organic carbon (DOC) was determined with a Phoenix 8000 total organic carbon analyzer (Tekmar-Dohrman, Cincinnati, OH, USA). The amount of OH~ released into solution during As(V) sorption was also calculated according to the methods of Duquette and Hendershot [20].

The PZCs for the soils were determined by zeta potential with a zeta potential analyzer (Zetasizer 2000, Malvern Instruments). The zeta potential of soil suspensions containing 0.1% solid in 0.01 mol/L NaNO3 in the presence and absence of As(V) (0.2 mmol/L) was determined at various pH values. Before determination, the suspension was exposed to supersonic dispersion for 20 min. After stabilizing, the pH value was registered and the suspension was submitted to zeta potential measurements. Plotting the zeta potential versus pH, the PZCs correspond to the pH value where zeta potential is zero.

Arsenate K-edge EXAFS

The EXAFS data at the As K-edge were collected at the 4W1B beamline of the Beijing Synchrotron Radiation Facility (Beijing, China) with a Si (111) double crystal monochromator. The storage ring operated at 2.2 GeV with a beam current of approximately 80 mA. Spectra of all the sorption samples were collected in fluorescence mode with a Lytle detector (The EX-AFS Company, Pioche, NV, USA) filled with krypton gas. The sorption samples, at initial As(V) concentrations of 0.2 and 2.0 mmol/L for the two soils, were prepared according to the same experimental procedures described in the Sorption isotherms section. After decanting the supernatant, the pastes were loaded into the Teflon® sample holders and sealed with Kapton tape (CHR Industries, New Haven, CT, USA). Energy calibration was monitored with an As(V) standard (i.e., sodium arsenate salt). A Ge filter was used to remove elastically scattered radiation from the fluorescence emissions. A reference of scorodite (FeAsO4-2H2O) powder was also collected. Other procedures were the same as the methods of Arai et al. [14]. All spectra were measured at room temperature. Three scans were averaged for each compound and standard.

Data analysis was performed with WinXAS2.1 (Faraday-weg 4-6, D-14195 Berlin, Germany) [21]. The pre-edge adsorption background was fitted and subtracted by a polynomial formula. Above the edge, a cubic spline fit was used to remove the background. Raw spectra \(k) were Fourier transformed in R space over a k range from 2.0 to 13.0 A_1. The Fourier-filtered EXAFS spectrum was least squares fitted with a theoretical function to get the structural parameters (e.g., atomic distance [R] and coordination number [CN]). Phase shifts and amplitude functions were corrected from theoretical calculation with the multiple-scattering computer code FEFF 8.20

Fig. 1. Sorption isotherms for arsenate, As(V), on the two Chinese red soils (pH 6.0, suspension density = 50 g/L) as a function of ionic strength (0.01 and 0.1 mol/L NaNO3).

^ 0.9 £

-o 0.6

o c/> TO 0)

■ Yingtan Al

□ Guangzhou Al

• Yingtan Fe 1

o Guangzhou Fe S ■

: i'u S s i f i

- ms o $ ° i.i.i, o 1 0 1

0.0 0.5 1.0 1.5 Arsenate added (mmol/L)

Fig. 2. Al and Fe released from the two Chinese red soils simultaneously with the arsenate sorption shown in Figure 1 (pH 6.0, ion strength = 0.01 mol/L NaNO3, suspension density = 50 g/L).

[22] with the use of scorodite (FeAsO4-2H2O) [23] and mans-fieldite (AlAsO4-2H2O) [24] mineral structures as the bases. The two FEFF codes in turn are simultaneously applied to generate the path files or the single scattering theoretical spectra and phase shifts for the As-O and higher shells of back-scattering atoms (As-Al and As-Fe) because chemical analysis revealed that both Al and Fe oxides contributed to As(V) sorption in these two soils.

When more than one backscattering atom or backscatterer at different distances were assumed in the fit model, fixing some of the parameters was necessary for the higher shells [16]. Although the amplitude reduction factor (S0) and the Debye-Waller factor (a2, which indicates disorder) possibly do not have the same values for all the backscattering atoms, it was implicitly assumed that the parameters were the same for the two backscatterers [16]. During fitting, a fixed value of 0.85 was used for S0 [16,25]. Reasonable values were first entered for the possible backscatterer (Al or Fe), respectively, and the other parameters were allowed to vary, one at a time, until the minimum fit error was obtained. The parameters that generated the best fit curves from the single-element back-scatterer were then used as the initial values for curve fitting the full range of unfiltered EXAFS spectra. The difference between the user-defined threshold energy and the experimentally determined threshold energy (A£0) can be determined by analysis of the first shell and fixed in subsequent fits of higher order shells in the same reference-unknown system [16]. When fitting, the total number of adjusted parameters did not exceed

Table 2. Langmuir parameters for arsenate sorption in the two soils from southern Chinaa

Soil sample (mmol/kg) (L/mmol) r2

Yingtan + 0.01 mol/L NaNO3 10.54 11.83 0.953

Yingtan + 0.1 mol/L NaNO3 12.22 11.34 0.951

Guangzhou + 0.01 mol/L NaNO3 78.05 33.15 0.940

Guangzhou + 0.1 mol/L NaNO3 69.76 33.61 0.973

the number of independent data points [16,25,26]. This procedure resulted in first shell accuracy of R = ±0.02 A and CN = ±25%; the accuracy decreased for higher shells.


Arsenate sorption isotherms

Arsenate sorption isotherms on the two soils as a function of ionic strength are shown in Figure 1. The lines represent the fits of the data to the Langmuir model. The fitted parameters

' Sm = the arsenate sorption maximum; K = a constant related to the sorption energy.

Fig. 3. Silicon, P, OH, and dissolved organic carbon (DOC) released from the two Chinese red soils simultaneously with the arsenate sorption shown in Figure 1 (pH 6.0, ion strength = 0.01 mol/L NaNO3, suspension density = 50 g/L).

Fig. 4. Zeta potential of the two soils as a function of pH in the presence and absence of 0.2 mmol/L arsenate, As(V) (ion strength = 0.01 mol/L NaNO3, suspension density = 1.0 g/L).

are listed in Table 2. The model showed significant correlations with sorption data with the r2 values ranging from 0.940 to 0.973 (p < 0.01). The As(V) sorption generally increased with increasing initial As(V) concentrations from 0 to 2.0 mmol/ L. Both soils showed a strong affinity for As(V) sorption compared with previously reported data [3,17], and a large difference in the As(V) sorption capacities was observed between the two soils. The maximum sorption capacity Sm of the Guangzhou soil is nearly five times more than that of the Yingtan soil (Table 2). The main constituents of the soils were quartz, aluminum silicates (predominantly kaolinite), and iron oxides (mainly poorly crystalline ones), as identified by X-ray diffraction and chemical analysis. The X-ray diffraction analysis indicated only minor amounts of iron oxides in the soils; however, chemical analysis confirmed the existence of large amounts of hydrous iron and aluminum oxides of low crys-tallinity in the soils on the basis of the ratios of oxalate-ex-tractable to dithionate-extractable Al and Fe [27], which accounted for the high sorption capacity of the soils [3,17,28]. Compared with the Yingtan soil, the Guangzhou soil, a strongly weathered oxisol, has a significantly higher specific surface area and much higher contents of clay and aluminum and iron minerals (Table 1), which could be responsible for the significant difference in As(V) sorption capacity between the two soils.

In the As(V) sorption isotherm experiments, total Al and Fe, DOC, and the main inorganic anions such as Si and P were subsequently monitored. Only slight increases in the dissolved Al and Fe contents occurred with the Guangzhou soil, but the release of Al and Fe, as well as DOC, Si, P, and OH, for the Yingtan soil was significantly enhanced with increasing As(V) sorption (Figs. 2 and 3). Significantly higher concentrations of the anions and DOC in the presence of As(V) than in the respective systems without As(V) addition indirectly suggests anion release as a result of As(V) sorption. It is possible that weakly complexed anions and soil organic carbon (mainly aliphatic carboxyl) on the soil surface are exchanged via competitive sorption processes because of the occupation of common sorption sites by As(V) (mainly as H2AsOj at pH 6.0; pKal « 2 and pKa2 ~ 7) and other anions [3,5,14]. The slight increase in the release of Al and Fe during As(V) sorption could be attributed to the complexation by anions released,

Fig. 5. (a) Normalized k3-weighted arsenate, As(V); K-edge extended X-ray absorption fine-structure (EXAFS) spectra; and (b) radial structure functions (uncorrected for phase shift) obtained by Fourier transformation of the EXAFS spectra for As(V) sorbed onto the two Chinese red soils at different As(V) loadings and the scorodite standard. r indicates the molar ratio of As(V) sorbed to the iron and aluminum contents in the sorption samples; the subscript 1 indicates Yingtan soil and 2 indicates Guangzhou soil; k (A_1) is the photoelectron wave vector unit; x(k)k3 is the normalized k-weighed EXAFS signal; R (A) is the interatomic distance; and FEFF fit means the data were fitted with the multiple-scattering computer code FEFF 8.20 [22].

especially DOC, which was validated by the significant correlation between Al/Fe and DOC released at various As(V) concentrations from the soils (p < 0.01; Figs. 2 and 3). Surface precipitates might also occur in the soils because the bulk

Table 3. Structural parameters summarizing the local coordination environment of arsenate sorbed on the two Chinese red soils and scorodite from the least squares analyses of As K-edge extended X-ray absorption fine-structure spectra3


Sample densityb Shell R CN CT2

Scorodite As/Fe = 1 As- -O 1.68 (1) 4.9 (2) 0.0028 (2)

As- Fe 3.33 (3) 2.3 (4) 0.0024 (4)

Yingtan + 0.2 mmol/L As r ~ 0.006 As- O 1.68 (2) 4.9 (3) 0.0031 (3)

As- Al 3.18 (2) 2.4 (2) 0.007 (1)

As- Fe 3.27 (4) 2.0 (3) 0.0027 (6)

Yingtan + 2.0 mmol/L As r ~ 0.02 As- O 1.68 (1) 4.9 (2) 0.0038 (3)

As- Al 3.18 (3) 2.2 (4) 0.003 (1)

As- Fe 3.28 (3) 1.7 (3) 0.006 (1)

Guangzhou + 0.2 mmol/L As r ~ 0.002 As- O 1.68 (1) 4.8 (3) 0.0019 (3)

As- Al 3.16 (3) 3.0 (3) 0.008 (1)

As- Fe 3.28 (2) 1.7 (4) 0.0040 (5)

Guangzhou + 2.0 mmol/L As r ~ 0.02 As- O 1.69 (2) 4.8 (3) 0.0026 (2)

As- Al 3.17 (3) 2.4 (3) 0.007 (1)

As- Fe 3.29 (3) 2.2 (4) 0.007 (1)

a R = interatomic distance; CN = coordination number; ct2 = Debye-Waller factor (A2). Estimated standard deviations are given in parentheses. b r indicates the approximate molar ratio of arsenate sorbed to the total contents of iron and aluminum in the sorption subsamples.

solution was far beyond saturation with respect to the solubility products of As(V) compounds of Al and Fe [3,10]. However, if precipitation predominated in the soils, we would not expect the obvious enhancement in the release of dissolved Al and Fe because these cations would otherwise precipitate with As(V) in the system. During As(V) sorption on Guangzhou soil, little enhancement in the release of dissolved Al, Fe, DOC, and Si was observed, and P and OH~ were not detected at all. This soil, as an oxisol, is strongly weathered and intensely eluviated, and few anions are available for complexing with metal oxides (mainly sesquioxides) [29]; thus, the relatively enriched hydroxyl sesquioxides (i.e., Al and Fe oxides) could directly sorb As(V) in the system. As a result, we did not observe as much increase in dissolved Al, Fe, and DOC with the Guangzhou soil as with the Yingtan soil during As(V) sorption (Fig. 3). Arsenate was therefore probably sorbed on the two soils mainly via ligand exchange.

Although the As(V) sorption isotherms for the two soils were significantly different, a similar surface complex formation occurred, as revealed by the experimental results of ionic strength effects on As(V) sorption and the shifts in PZC resulting from As(V) sorption. Ionic strength-independent anion sorption is often linked to distinguishing inner-sphere sorption from outer-sphere sorption mechanisms [8]. Figure 1 shows that As(V) sorption is not affected by the ionic strength; therefore, it is reasonable to conclude an inner-sphere sorption mechanism for As(V) occurred on the two soils. However, it is often difficult to elucidate surface complexation without any spectroscopic evidence. Figure 4 presents the zeta potential versus pH curves obtained with and without As(V) sorption on the two soils. The control PZCs occurred at pH 5.7 for Guangzhou soil and at pH 4.4 for Yingtan soil. The PZCs were shifted to lower pH values in the presence of As(V). A shift in PZC with ion sorption was another meaningful characteristic of inner-sphere sorption because inner-sphere complexes generally cause shifts in both PZC and zeta potential because of specific ion adsorption inside the shear plane [8,13,30]. In addition, the shifts in PZCs of the soils to a more acid value also indicated that ligand exchange dominates the As(V) sorption reaction in the two soils [31].

Arsenate K-edge EXAFS analyses

Figure 5a shows the experimental x(k) spectra (solid lines) along with their best fit (open circles) of As(V) sorbed on the

two soils at two different initial As(V) concentrations and the scorodite standard. A sinusoidal oscillation of a strong first neighbor O-shell backscattering atom was present in all of the sorption samples. The x structures were distinctly different in phase, wavelength, and shape, particularly at higher k. The uniqueness of the x structures was due to the differences in backscattering from the atoms residing in the second coordination shells surrounding the sorbed As. Fourier transformation of the x(k) function led to a radial structure function in which the peak positions corresponded to the interatomic distances within the samples (Fig. 5b). The representative frequencies of the individual components in the x function were more clearly illustrated in the radial structure function and thus could be more readily detected in the radial structure function than in the x function. The radial structure functions derived from the EXAFS data for the samples, and the reference compound showed the quantitative estimation of interatomic distances between As(V) and the next neighboring atoms in the samples (Fig. 5b). These peak positions, however, were uncorrected for phase shift; therefore, the positions were slightly shifted from the true interatomic distances. The structural parameters (e.g., CN, R, ct2) obtained from the linear least squares fits of the EXAFS data are given in Table 3.

The best fit was obtained when the first and second shells were regarded as the contributions of As-O pairs and As-Al and As-Fe pairs, respectively. The first shell centered at 1.30 A (uncorrected for phase shift) can be fit with about 4.8 O backscatterers at approximately 1.68 A from the central As atom, which is similar to the reported values of As(V) sorbed on metal oxides [13,32]. The high CN values relative to the actual CN value of 4 for the AsO4 tetrahedral molecular structure might be because of the high correlation between CN and ct2 [26]. However, the first shell information was of minor importance in this study. Our main objective was to determine the surroundings of As(V), focusing on the backscatterers from the second shell around As(V). Because of their similar distances, the peaks in the second shell were overlapped to a certain extent and difficult to isolate from each other in some samples. The peaks, ranging from 2.10 to 3.40 A (uncorrected for phase shift), could be fit with radial distances of the As-Al shells (3.17 ± 0.03 A) and of the As-Fe shells (3.28 ± 0.04 A) from the central atom. The accuracies for the higher shells of As-Al or As-Fe pairs (i.e., CN, R, and ct2) were

Fig. 6. A comparison between single-shell fit (As-Fe pair at 3.35 A) and the double-shell fit (As-Fe pairs at 2.85 and 3.35 A) for the second coordination sphere of scorodite. k (A41) is the photoelectron wave vector unit, x(k)k3 is the normalized k-weighed EXAFS signal, and FEFF fit means the data were fitted with the multiple-scattering computer code FEFF 8.20 [22].

slightly poorer than those reported on pure or synthesized minerals. However, the fits produced excellent agreement of R and CN with the values obtained previously from single-scattering atoms of Al or Fe, and the values statistically matched well with the interatomic distances reported in the As(V) bidentate-binuclear complex structures on aluminum and iron minerals [11,14,32,33]. Either As-Al or As-Fe could not provide satisfying fit results separately.

Regardless of the different surface loadings (0.2 and 2.0 mmol/L), the As-Al and As-Fe shells were consistently presented, suggesting that Al and Fe oxides were the predominant sorbents for As(V) in the soils. However, the CN and R values varied with different surface loadings. In Guangzhou soil, the proportion of the As-Al pair at approximately 3.17 A was a little more pronounced compared with the As-Fe pair at approximately 3.28 A, and the predominance was enhanced at

higher surface loading, indicating that a higher contribution for As(V) sorption was from aluminum oxides in the soil, which was also supported by Livesey and Huang [3]. In Ying-tan soil, the predominance of the As-Al backscattering pair decreased, and the As-Fe pair became pronounced, possibly because of the relatively low content of aluminum minerals. An As(V) bidentate-binuclear complex at approximately 3.28 A was even dominant in the backtransformed curve, with the fits improving slightly by adding the shell of As-Al at the lower surface loading (0.2 mmol/L). In addition to the biden-tate-binuclear complex at approximately 3.29 A for As-Fe and 3.18 A for As-Al complexes, the other two kinds of complexes, the monodentate complex (>3.50 A) and the bidentate-mononuclear complex (<3.05 A) have been suggested between As and Fe on iron mineral surfaces [7,34] and a bidentate-mononuclear complex (3.02 A) on aluminum minerals [13,14] according to the parameters (i.e., CN and R) from fitted EXAFS spectra. However, we did not observe the monodentate complexes in our system, and monodentate complexes were usually believed to associate with very low surface sorption densities [7,34]. Fits of all the samples were slightly improved with the addition of either an Fe shell at 2.85 A or an Al shell at 3.02 A (characteristic of bidentate-mononuclear complexes), and the improvement was also substantiated in the fit of the EXASF spectrum of the scorodite standard (Fig. 6), although there is no edge-sharing structure between FeO6 and AsO4 polyhedra in the scorodite. Fitting the data with a model including a multiple scattering showed that the shells seemed to match edge-sharing complexes, which could be accounted for by multiple scattering involving O-O pairs within the AsO4 tetrahedron [9,35]. Again, both bidentate-mononuclear and mon-odentate complexing were considered energetically unfavorable relative to the bidentate corner-sharing complexing for As(V) (H2AsO4-) in the conditions of this study [13,34,35]. Overall, on the basis of the above results, it is reasonable to suggest that the bidentate-binuclear AsO4 coordination environments predominate on the Al and Fe oxide surfaces in the two soils.

It might be possible for As(V) to form precipitates with Al or Fe in the soils, as mentioned in the Arsenate sorption isotherms section. The presence of a single As-Al or As-Fe shell at 3.15 or 3.36 A was previously reported in the minerals of mansfieldite (AlAsO4-2H2O) by Arai et al. [14] and of scorodite (FeAsO4-2H2O) by O'Reilly et al. [32], respectively. Although the As-Al and As-Fe bonds we obtained were slightly different from those reported for mansfieldite and scorodite, we still cannot exculpate the possibility of Al/Fe-AsO4 precipitate formation in the As(V) sorption system considering the errors in the radial distances resulted from the fits.

In addition, because of the complicated components in the soils, there must be other metal oxides or minerals (e.g., TiO2, MnO2) that could contribute to As(V) sorption and account for the remaining discrepancies between the data and reported fits (Fig. 5). However, inclusion of these constituents into the data fitting proved extremely difficult because of their low contents in the soils or lack of reference compounds, and the issue requires further investigation.


The results of this study from both macroscopic and microscopic experiments provided consistent evidences for the mechanisms of As(V) sorption on the two Chinese red soils. Arsenate sorption occurred mainly via ligand exchange reac-

tions, and inner-sphere complexes were formed predominantly with bidentate-binuclear structures at the conditions. The results would improve the understanding of the mechanisms of As(V) sorption on soils. Although the formation of bidentate-binuclear, monodenate, and bidentate-mononuclear complexes have been suggested between As and Fe/Al on mineral surfaces, the results of this study confirmed that bidentate-binuclear coordination between As(V) and Fe/Al was dominant in the two soils. However, much work is still needed to validate the relevant As(V) sorption mechanisms with the use of different soils.

Acknowledgement—This work was funded by the National Natural

Science Foundation of China (Project 20377049 and 20237010).


1. Smith E, Naidu R, Alston AM. 1998. Arsenic in the soil environment: A review. Adv Agron 64:149-195.

2. Nordstrom DK. 2002. Worldwide occurrences of arsenate in ground water. Science 296:2143-2145.

3. Livesey NT, Huang PM. 1981. Adsorption of arsenate by soils and its relation to selected chemical properties and anions. Soil Sci 131:88-94.

4. Violante A, Pigna M. 2002. Competitive sorption of arsenate and phosphate on different clay minerals and soils. Soil Sci Soc Am J 66:1788-1796.

5. Anderson MA, Ferguson JF, Gavis J. 1976. Arsenate adsorption on amorphous aluminum hydroxide. J Colloid Interface Sci 54: 391-399.

6. Lumsdon DG, Fraser AR, Russell JD, Livesey NT. 1984. New infrared band assignments for the arsenate ion adsorbed on synthetic goethite (a-FeOOH). J Soil Sci 35:381-386.

7. Fendorf S, Eick MJ, Grossl P, Sparks DL. 1997. Arsenate and chromate retention mechanisms on goethite. 1. Surface structure. Environ Sci Technol 31:315-326.

8. Goldberg S, Johnston CT. 2001. Mechanisms of arsenate adsorption on amorphous oxides evaluated using macroscopic measurements, vibrational spectroscopy, and surface complexation modeling. J Colloid Interface Sci 234:204-216.

9. Pandya KI. 1994. Multiple-scattering effects in X-ray adsorption fine structure: Chromium in a tetrahedral configuration. Phys Rev B 50:15509-15515.

10. Fuller CC, Davis JA, Waychunas GA. 1993. Surface chemistry of ferrihydrite: Part 2. Kinetics of arsenate adsorption and co-precipitation. Geochim Cosmochim Acta 57:2271-2282.

11. Waychunas GA, Rea BA, Fuller CC, Davis JA. 1993. Surface chemistry of ferrihydrite: Part 1. EXAFS studies of the geometry of coprecipitated and adsorbed arsenate. Geochim Cosmochim Acta 57:2251-2269.

12. Sun X, Doner HE. 1996. An investigation of arsenate and arsenite bonding structures on goethite by FTIR. Soil Sci 161:865-872.

13. Ladeira ACQ, Ciminelli VST, Duarte HA, Alves MC, Ramos AY. 2001. Mechanism of anion retention from EXAFS and density functional calculations: Arsenic (V) adsorbed on gibbsite. Geo-chim Cosmochim Acta 65:1211-1217.

14. Arai Y, Sparks DL, Davis JA. 2005. Arsenate adsorption mechanisms at the allophone-water interface. Environ Sci Technol 39: 2537-2544.

15. Ladeira ACQ, Ciminelli VST 2004. Adsorption and desorption

of arsenic on an oxisol and its constituents. Water Res 38:20872094.

16. O'Day PA, Rehr JJ, Zabinsky SI, Brown GE Jr. 1994. Extended X-ray absorption fine structure (EXAFS) analysis of disorder and multiple-scattering in complex crystalline solids. J Am Chem Soc 116:2938-2949.

17. Jiang W, Zhang S, Shan X, Feng M, Zhu Y, McLaren RG. 2005. Adsorption of arsenate on soils. Part 2: Modeling the relationship between adsorption capacity and soil physiochemical properties using 16 Chinese soils. Environ Pollut 138:285-289.

18. Manning BA, Goldberg S. 1997. Arsenic (III) and arsenic (V) adsorption on three California soils. Soil Sci 162:886-895.

19. Nelson DW, Sommers LE. 1996. Total carbon, organic carbon, and organic matter. In Sparks DL, Page AL, Helmke PA, Loeppert RH, Soltanpour PN, Tabatabai MA, Summer ME, eds, Methods of Soil Analysis, Part 3—Chemical Methods. SSSA Book Series 5. Madison, WI, USA, pp 961-1010.

20. Duquette M, Hendershot M. 1993. Soil surface charge evaluation by back-titration: Theory and method development. Soil Sci Soc Am J 57:1222-1228.

21. Ressler T. 1997. WinXAS: A new software package not only for the analysis of energy-dispersive XAS data. J Phys IV 7:269270.

22. Ankudinov AL. 1999. Real-space multiple-scattering approach to XANES. J Synchrotron Radiat 6:236-238.

23. Kitahama K, Kiriyama R, Baba Y. 1975. Refinement of the crystal structure of scorodite. Acta Crystallogr B 31:322-324.

24. Harrison WTA. 2000. Synthetic mansfieldite, AlAsO42H2O. Acta Crystallogr C 56:E421.

25. Paktunc D, Foster A, Heald S, Laflamme G. 2004. Speciation and characterization of arsenic in gold ores and cyanidation tailings using X-ray absorption spectroscopy. Geochim Cosmochim Acta 68:969-983.

26. Foster AL, Brown GE Jr, Tingle TN, Parks GA. 1998. Quantitative arsenic speciation in mine tailings using X-ray absorption spec-troscopy. Am Mineral 83:553-568.

27. Johnson DW, Cole DW, Miegroet HV, Horng FW. 1986. Factors affecting anion movement and retention in four forest soils. Soil Sci Soc Am J 50:776-783.

28. Redman AD, Macalady DL, Ahmann D. 2002. Natural organic matter affects arsenic speciation and sorption on hematite. Environ Sci Technol 36:2889-2896.

29. Bohn HL, McNeal BL, O'Connor GA. 1979. Soil Chemistry. John Wiley, New York, NY, USA.

30. Arai Y, Elzinga EJ, Sparks DL. 2001. X-ray adsorption spectroscopic investigation of arsenite and arsenate adsorption at the aluminum oxide-water interface. J Colloid Interface Sci 235:8088.

31. Sparks DL. 2003. Environmental Soil Chemistry. Academic, San Diego, CA, USA.

32. O'Reilly SE, Strawn DG, Sparks DL. 2001. Residence time effects on arsenate adsorption/desorption mechanisms on goethite. Soil Sci Soc Am J 65:67-77.

33. Beaulieu BT, Savage KS. 2005. Arsenate adsorption structures on aluminum oxide and phyllosilicate mineral surfaces in smelter-impacted soils. Environ Sci Technol 39:3571-3579.

34. Waychunas GA, Davis JA, Fuller CC. 1995. Geometry of sorbed arsenate on ferrihydrite and crystalline FeOOH: Re-evaluation of EXAFS results and topological factors in predicting sorbate geometry, and evidence for monodentate complexes. Geochim Cos-mochim Acta 59:3655-3661.

35. Sherman DM, Randall SR. 2003. Surface complexation of ar-senic(V) to iron(III) (hydr)oxides: Structural mechanism from ab initio molecular geometries and EXAFS spectroscopy. Geochim Cosmochim Acta 67:4223-4230.