Arsenate and Phosphate Adsorption in Relation to Oxides

Jun 10, 2013 - *Phone: 31-317-482332; fax: 31-317-419000; e-mail: [email protected]. Abstract. Abstract Image. The pH dependent solid-solution distri...
0 downloads 3 Views 1MB Size
Download PDF
Article pubs.acs.org/est

Arsenate and Phosphate Adsorption in Relation to Oxides Composition in Soils: LCD Modeling Yanshan Cui†,‡ and Liping Weng*,‡ †

University of Chinese Academy of Sciences, Beijing 100049, China Department of Soil Quality, Wageningen University, P.O. Box 47, 6700 AA, Wageningen, The Netherlands

‡

S Supporting Information *

ABSTRACT: The pH dependent solid-solution distribution of arsenate and phosphate in five Dutch agricultural soil samples was measured in the pH range 4−8, and the results were interpreted using the LCD (ligand and charge distribution) adsorption modeling. The pH dependency is similar for both oxyanions, with a minimum soluble concentration observed around pH 6−8. This pH dependency can be successfully described with the LCD model and it is attributed mainly to the synergistic effects from Ca adsorption. The solubility of phosphate is much lower than that of arsenate. This big difference cannot be sufficiently explained by the reduction of small amount of As(V) into As(III), neither by slow desorption/adsorption. The difference between phosphate and arsenate in their solid-solution distribution becomes larger with the increase of aluminum (hydr)oxides (Al-oxides) contribution to the total amount of metal (Al and Fe) (hydr)oxides. The influence of Al-oxides is much larger than its relative amount extracted from the soils. When Al-oxides account for >40% of the soil oxides, the whole adsorbents behave apparently similarly to that of pure Al-oxides. These results indicated that surface coating and substitution may have modified significantly oxyanion adsorption to Fe-oxides in soils, and how to account for this complexity is a challenge for geochemical modeling.

■

INTRODUCTION Chemical speciation of many elements in the natural environment is controlled by surface complexation (adsorption) reactions. Combination of advanced analytical techniques and geochemical modeling has led to great progress in predicting ion adsorption to organic and mineral particles. So far, geochemical modeling has been more successful for cationic species than for anionic species in natural samples (water, soil, and sediments).1−5 For oxyanions like phosphate and arsenate, metal (hydr)oxides and edge faces of clay minerals are the most important reactive surfaces in nature. Applying advanced surface complexation models to predict anion adsorption in natural samples is very challenging due to the difficulty in accounting for the competitive effects of natural organic matter (NOM) and due to the presence of different adsorbing surfaces, for example, iron (hydr)oxides (Fe-oxides), aluminum (hydr)oxides (Al-oxides), and edges of clay.6 There are a few examples in applying advanced surface complexation models to describe adsorption of anions in natural samples.6−11 In some of these works, the competitive effects of NOM on anion adsorption were not considered,7−9 whereas in some other studies, effects of NOM were simulated with a simplified approach, in which adsorption of NOM was represented by complexation between a surface group and a carboxylic group.10,11 In a recent work, pH dependent phosphate adsorption to soils was simulated with the LCD (ligand and charge distribution) model.6 In this model, the © XXXX American Chemical Society

effect of NOM on phosphate adsorption was taken into account by considering both site competition and electrostatic interactions.12 The proper prediction of the pH dependent phosphate adsorption in soils with the LCD model indicates that the model representation of factors that are related to pH dependency is reasonable. In the study of Gustafsson9,10 and study of Devau et al.,8 anion adsorption to soils was simulated with various oxides (ferrihydrite, goethite, or gibbsite) and clay minerals (allophane, kaolinite, or illite). In the LCD modeling for phosphate, Fe-oxides (goethite) was used to represent all binding mineral surfaces.6 The robustness of this assumption can be better tested when more than one type of oxyanions are modeled simultaneously. Phosphate (PO43‑) and arsenate (AsO43‑) are two oxyanions that are often studied together for their adsorption behavior in the natural samples, because both are of high environmental relevance and both adsorb strongly. These two anions are interesting candidates in model validation and in identifying binding surfaces in natural samples. The relative binding affinity between PO4 and AsO4 may vary greatly on different soil minerals and on soils characterized by different mineralogy and chemical properties, for example, refs Received: February 4, 2013 Revised: May 29, 2013 Accepted: June 10, 2013

A

dx.doi.org/10.1021/es400526q | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

Table 1. Properties and Some Modeling Parameters of the Soil Samplesa measured property pH soil

code in Hiemstra et al.

2 13 23 26 35

1 5 10 12 16

SOC

4.3 4.9 4.9 4.5 6.0

Clay

%

%

1.8 1.3 8.3 3.3 1.0

5 15 25 13 6

Fe-DCB

Al-DCB

derived property

Fe-ox

Al-ox

P-ox

PO4-ox

−1

2

mmol kg 48 168 318 233 84

41 30 59 29 34

32.4 75.7 252.2 114.4 31.4

A m g

34.4 20.2 46.2 20.9 23.7

19.6 14.9 25.7 30.2 15.6

14.9 13.7 19.5 29.0 12.1

−1

soil

5.9 9.3 26.3 12.9 5.4

fAl

NOMads

%

mg m−2

51 20 16 14 40

0.49 1.04 1.59 0.67 0.86

a SOC: soil organic carbon; Fe-DCB, Al-DCB: concentration of Fe and Al in dithionite-citrate-bicabonate extraction; Fe-ox, Al-ox, P-ox, and PO4-ox: concentration of Fe, Al, total P and phosphate in ammonium oxalate extraction; A: total reactive surface area per gram of soil calculated from the amount of Fe- and Al-(hydr)oxides extracted (see Modeling Section); fAl: contribution of Al to the total reactive surface area; NOMads: fitted amount of natural organic matter adsorbed. Data of pH, SOC, clay, Fe-DCB, Al-DCB, Fe-ox, Al-ox, and P-ox are from ref 11. The rest is from this study.

Table 2. Results of the Kinetic Experimenta Soil2

Soil 13

pH 4.53 ± 0.02 H2SO4 mM KOH mM

Soil 23

pH

6.29 ± 0.03

3.92 ± 0.13

Soil 26

pH

5.26 ± 0.18

4.87 ± 0.06

Soil 35

pH

6.29 ± 0.01

4.70 ± 0.08

pH

6.22 ± 0.03

1.6

4.50 ± 0.22

6.03 ± 0.03

1.5

5.5

15

Day 0 3 7 10 14

2.07 1.20 1.00 0.91 0.89

6.29 1.01 1.00 0.71 0.78

13.3 1.99 1.00 0.82 0.62

14.1 1.84 1.00 0.61 0.48

Day 0 3 7 10 14

0 1.07 1.00 0.97 0.95

0 0.99 1.00 1.62 0.99

0 1.89 1.00 0.84 0.62

0 1.36 1.00 1.20 1.16

As in Solution 111 250 2.72 1.63 1.00 1.00 0.56 0.75 0.50 0.75 P-PO4 in Solution 0 0

8

15.6 1.64 1.00 0.88 0.85

44.4 1.40 1.00 0.91 0.89

3.00 1.27 1.00 0.93 0.75

2.42 1.35 1.00 0.86 0.69

0 1.03 1.00 0.97 0.97

0 0.91 1.00 1.01 1.09

0 1.48 1.00 0.87 0.58

0 0.81 1.00 0.93 0.78

a pH is given as the average with standard deviations for pH measured during the experiment. The soluble concentration of As and P-PO4 were given as ratios to that measured at 7 days.

reactive surface area for phosphate adsorption.11 The soils used in the work of Weng et al.6 in modeling pH dependent phosphate adsorption were also selected from the same sample set, but they are not overlapping with the five soils chosen in the current work. Basic soil properties of the five soils used in the current work can be found in Table 1. Most of the soil properties have been reported in the literature,11,16 except the ortho-phosphate content in ammonium oxalate extraction, which was measured in the current study by conducting the same extraction but measuring not total P but ortho-phosphate using the molybdenum-blue method.17 Due to the interference of oxalate, which can form complexes with molybdate,18 the ammonium oxalate soil extracts were first diluted 10−100 times. Calibrations using standard solutions prepared in the same matrix show a good linear relationship (results not shown). Adsorption/Desorption Experiment. The batch experiment was conducted with the five soil samples chosen (Table 1). For each soil, 10 subsamples of 2.0 g each were put into 50 mL Teflon bottles, to which 10 mL of 0.02 M CaCl2 was added. To nine subsamples, 2 mL of 2.67 mM As(V) (Na2HAsO4·7H2O) solution were added, and pH was adjusted to different values in the range of pH 4−8 by adding appropriate amount of acid and base (0.02 M H2SO4 and 0.1

13−15. Comparative studies show that AsO4 adsorbs similarly or a bit stronger than PO4 on Fe-oxides and Fe-rich phyllosilicates, but PO4 is preferred over AsO4 on Al-oxides, allophane and kaolinite.15 In the present work, the pH dependent competitive adsorption of arsenate and phosphate in five Dutch agricultural soil samples varying in soil properties were measured in a pH range of 4−8. The LCD model was applied to describe simultaneously the solid-solution distribution of both AsO4 and PO4, taking into account the competitive effects of NOM and synergistic effects of major cations (mainly Ca2+). The objectives of this study are to better understand the pH dependent adsorption of both arsenate and phosphate in natural samples, and to use these two anions as probes to identify reactive mineral surfaces for oxyanion adsorption in natural samples.

■

MATERIALS AND METHODS

Soil Samples. Five soil samples were used in this study. These soils were chosen from a large collection of representative agricultural top soils of The Netherlands, known as the Copernicus Series.16 The chosen soils are also part of the samples used by Hiemstra et al. in estimating total B

dx.doi.org/10.1021/es400526q | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

prepared that contain phosphate and arsenate of different ratios in the concentration range relevant to this study. A regression calibration equation was derived from these standard solutions (R2 = 0.991) (results not shown) and concentrations of phosphate in the sample solutions were derived using this regression equation and total soluble As measured with ICPMS, assuming all As is As(V). The results show that the color intensity is about 25% for arsenate compared to phosphate at the same concentration when present simultaneously.

M KOH). The amount of acid or base solution needed was estimated from a pretest using the same soil. The maximum total amount of acid and base added was equivalent to a final concentration of 3.9 mM H2SO4 and 22 mM KOH (in Soil 23). To the tenth subsample (blank), no As was added and pH was not adjusted. Ultrapure water (UPW) was added to all bottles to a final total volume of 20 mL. The final soil-solution ratio (SSR) is thus 1:10, and the final concentration of CaCl2 is 0.01 M. The final As(V) concentration is 0.267 mM or 2.67 mmol kg−1 soil. The soil suspensions were shaken end-over-end at 25 rpm for 7 days in a controlled temperature room at 20 °C, which was most of the time in the dark. The bottle caps were opened about 1 h per day to maintain aerobic conditions. For each soil, to the three subsamples at the high pH end, small extra amount of base was added at day 5. After 7 days shaking, the end pH in the suspension was measured. Thereafter, the suspensions were centrifuged at 3000 rpm for 15 min. Part of the supernatant was filtered through 0.45 μm filter before further analysis of DOC (dissolved organic carbon) and bi(carbonate), total soluble P and ortho-phosphate (P-PO4), total soluble As, and concentration of Ca, Al, and Fe. Kinetic Experiment. Eight soil suspensions were prepared for each soil. The soil suspensions were prepared in a similar way as described above. The eight subsamples were divided into two groups: that is, a low pH and a high pH group. The four suspensions of the low pH group have a pH in the range of 3.9−4.9 (without or with pH adjustment with acid), whereas the other four subsamples in the high pH group have a pH between 5.3−6.3 (without or with pH adjustment with base) (Table 2). The acid or base (Table 2) was added at once in the beginning. After respectively 3, 7, 10, and 14 days of shaking, one subsample of each group was taken out, centrifuged, filtered, and analyzed for concentrations of P-PO4 and total soluble As. Analysis of As(III) and As(V). Two suspensions per soil differing in pH were prepared as described above and the shaking time is 7 days. The supernatant solutions after centrifugation were filtered over 0.45 μm and stored at 4 °C before analysis. Arsenic species in these solutions were separated by high performance liquid chromatography (HPLC, Shimadzu, LC-6A) with a Hamilton PRP-X100 anion exchange column (10 μm; 250 mm × 4.1 mm) and a precolumn. 12.5 mM (NH4)2HPO4 at pH 8.5 was used as the mobile phase and the flow rate was 1 mL min−1. 100 μL sample solution (diluted if necessary) was injected. The HPLC system was coupled online with high resolution ICP-MS (inductively coupled plasma mass spectrometer) (Thermo Scientific, Element2), which measures As concentration every 6 s. Retention time of arsenate (AsO43‑) (13.9 min) and arsenite (AsO33‑) (3.5 min) was determined using standard solutions. Chemical Analysis. pH was measured with a pH meter. Concentrations of DOC and bi(carbonate) were measured with a TOC analyzer (Shimadzu). Concentrations of total soluble P, As, Fe, and Al were measured with high resolution ICP-MS (Thermo Scientific, Element2). Concentration of soluble Ca was measured with ICP-OES (inductively coupled plasma optical emission spectrometer) (Thermo Scientific, Iris Advantage). Concentration of inorganic phosphate (P-PO4) was measured with an optimized molybdate blue colorimetric method using a segmented flow analyzer (Skalar, The Netherlands).17 To correct for the arsenate interference on colorimetric P-PO4 measurement,19 standard solutions were

■

MODELING APPROACH Model Description. The modeling approach followed is similar to that of Weng et al.6 Basically, the framework of the LCD model20 was used to calculate phosphate and arsenate (and arsenite) adsorption to the organo-mineral assemblage in soils. The LCD integrates the NICA model12 and the CDMUSIC model21,22 to describe the surface complexation reactions on soil minerals under the influence of natural organic matter adsorbed (NOMads). Goethite was used as the representative material for reactive surfaces in soils. Adsorption of phosphate and arsenate to these soils is expected to be dominated by metal (hydr)oxides. In Dutch soils, illite is the dominant type of clay. In two studies of Manning and Goldberg,23,24 phosphate adsorption to illite and to goethite was measured, respectively, and the results show a more than 100 times weaker phosphate adsorption on illite than on goethite at the same mass of the adsorbents. In the soil samples of the current work, the mass ratio between clay and Fe- plus Al-oxides is less than 15. In addition, 37−79% of Feoxides (Table 1) in the soils are amorphous and their specific surface area should be larger than the synthetic goethite. Therefore we would think that clay contribution to phosphate and arsenate adsorption will be relatively small. The CDMUSIC model parameters for H+, Na+, Cl−, Ca2+, PO43‑, AsO43‑, and AsO33‑ adsorption to goethite were taken from Stachowicz et al.25,26 (Table SI-1, Supporting Information (SI)). The interaction between NOMads and inorganic ions at the mineral surfaces is simulated using a layer of adsorbed fulvic acid (FA), which is present and equally distributed between the first and second Stern layer.27 The carboxylic groups (RCOO−) of NOMads in the first compact layer can form innersphere complexes (Fe1OOCR−0.5) with the singly coordinated surface sites on goethite. The other carboxylic and phenolic groups on the NOMads can bind H+, Ca2+, Al3+ and Fe3+. The reactions between the NOMads ligands and surface sites, protons and other cations were calculated with the NICA model. It was assumed that the NICA model parameters remain the same as for FA in the solution (Table SI-2, SI). Model Input. The amount of arsenate added to the soil (2.67 mmol kg−1) was used as model input for total arsenate. The initial soluble As in these soils is rather low ( Soil 13 > Soil 26 > Soil 2 > Soil 35. The pH dependency of soluble P-PO4 and As observed is similar to each other, which is also similar to what has been found for phosphate in a previous study using different soils and in the absence of arsenate,6 showing an adsorption maximum (a minimum in soluble concentration) around pH 6−8. Similarly, Goldberg et al.34 found that As(V) adsorption on soil samples generally increased with increasing solution pH, and reached a maximum in adsorption around pH 6−7. Different results have been found by Gustafsson et al.,35 in which the adsorption of arsenate added to the soil decreased with the increase of pH. The total amount of phosphate present in soils (Table 1) is 4.5−10.9 times of the amount of arsenate added. Despite this, the measured soluble phosphate concentration is lower than soluble arsenate in Soil 2, 13, and 35. Only in Soil 26, the soluble phosphate concentration is larger than the soluble arsenate (Figure 1). In this soil, the total phosphate is 10.9 times of the total arsenate. In Soil 23, both soluble phosphate and arsenate concentrations are very low and PO4 data is not E

dx.doi.org/10.1021/es400526q | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

Table 3. Distribution of As between As(V) and As(III) in the Soil Extracts and Recovery Rate of As in the HPLC Analysis Soil 2

As(III) (%) As(V) % As recovery

Soil 13

Soil 23

Soil 26

Soil 35

pH

pH

pH

pH

pH

pH

pH

pH

pH

pH

4.54

6.34

4.04

5.39

4.90

6.29

4.77

6.23

4.75

6.06

8 92 95

30 70 99

13 87 86

30 70 75

43 57 80

41 59 87

1 99 81

1 99 89

3 97 100

0 100

underestimate phosphate adsorption (results not shown). This modeling exercise shows that when Fe-oxides and Al-oxides are modeled separately using the amounts of Fe- and Al-oxides from the ammonium oxalate and DCB extraction, the model can still not explain the big difference in phosphate and arsenate solubility. To quantify the deviation of arsenate adsorption affinity in these soils from that on goethite, another modeling exercise was carried out, in which all the choices were the same as described in the Modeling Approach Section, except that we made the K values for arsenate adsorption to the whole surface adjustable. To keep the fitting simple, we adjusted the K values for all three arsenate surface species (SI Table SI-1) simultaneously with the same scaling factor. A reasonable agreement with the data was obtained (M2-P, M2-As, solid lines, Figure 2) by lowering the

reason and from the kinetic data, we can conclude that the big difference observed in the solubility of phosphate and arsenate cannot be (sufficiently) explained by slow reactions. Further, adsorption to clay minerals is not a likely alternative explanation either for the discrepancy between model and data, because illite is the major clay form in most of Dutch soils, and phosphate and arsenate adsorption to illite is rather similar.15 As(V) and As(III). Another possible reasons for the model overestimation for arsenate adsorption can be attributed to the reduction of As(V) to As(III). Arsenite (As(III)) binds much weaker than arsenate (As(V)) on Fe-oxides (SI Table SI-1). Relative amount of As(III) and As(V) found in soil solutions measured with the HPLC analysis are given in Table 3. Negligible amount of As(III) was found in Soil 26 and Soil 35. In the other soils, significant amount of As(III) was measured and As(III) accounts for maximum 43% of the sum of As(V) and As(III). This indicates that despite the fact that the bottles were kept open frequently during the shaking period (Materials and Methods), a small amount of As(V) was still reduced to As(III) in Soil 2, 13, and 23. The As recovery in the HPLC analysis is >75% (Table 3), which is calculated as the ratio between the sum of As(III) and As(V) measured over the total soluble As injected. The nonrecovered As can be attributed to the presence of colloidal As, which was retained in the precolumn. Colloidal As can be As associated with colloidal organic and mineral particles. The contribution of colloidal As to soluble As is larger when the total soluble As concentration is lower. We can conclude that, overall, more than 45% of soluble As measured is As(V). With the LCD model as above but assuming that a fraction of As(V) added has been reduced to As(III), it is found that to obtain the measured As(III)/As(V) ratio in the soil solutions (Table 3), maximally 1% of the total As can be considered as As(III) (results not shown). This modeling exercise indicates that even when 50% of As in the soil solution is As(III), the adsorbed As is still dominated by As(V). Despite that the presence of less than 1% of As in the form of As(III) has led to a maximum two-times increase of soluble As in these soils, this factor cannot explain the much weaker than expected As adsorption in these soils. Influence of Al-Oxides. To test if the relative weaker arsenate adsorption compared to phosphate on Al-oxides can explain the much higher solubility for arsenate than phosphate, on the basis of the modeling approach described in the beginning, a fraction of the surface sites was made arsenate nonreactive, as an extreme. The fraction of nonreactive site for arsenate equals to the contribution of Al (fAl, Table 1) to the total amount of surface area calculated (Modeling Approach), whereas the arsenate-reactive site equals to the Fe-oxides contribution (1 − fAl). On both types of sites, phosphate can bind with the same affinity. The model calculation shows that even when arsenate adsorption to the “Al-oxides types of sites” was totally turned off, as an extreme, the model improvement is small and it still largely overestimates arsenate adsorption and

Figure 2. Times of reduction of arsenate adsorption affinity constants (KAs) from the original KAs for goethite as a function of aluminum fraction ( fAl) in oxides for soils. For comparison, changes of KAs with increase of fAl for the mixture of Al and Fe oxides synthesized (Masue et al. 2007) was also given in this figure (broken lines).).

K values for arsenate adsorption to goethite (SI Table SI-1) by 6, 9, and 12 times for, respectively, Soil 26, 23, and 13. For Soil 2 and Soil 35, the K values were lowered by 15 times. The remaining discrepancy between the model and the data in Soil 2, 13, and 23 can be partly ascribed to the presence of As(III) in the soil solution (Table 3) and to the influence of As(III) on P-PO4 measurement (All soluble As was assumed as As(V) in correcting As influence on the colorimetric PO4 measurement, see Materials and Methods). The pH dependency of arsenate adsorption can be largely explained by the model (Figure 1). As explained in the last paper regarding phosphate only,6 the observed maximum adsorption for both phosphate and arsenate around the neutral pH can be attributed to the synergistic effect of Ca adsorption to the oxides surface. These fitted reduction factors for arsenate adsorption affinity to soil oxides is positively correlated with fAl (Figure 2), which F

dx.doi.org/10.1021/es400526q | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

smaller particles with surface properties similar to Al-oxides.38 In soils, single metal oxides are difficult to find and the majority of Fe-oxides are Al-substituted.39 The current modeling results show that the surface of soil oxides is probably richer in Al than in the bulk. More evidence should be collected in the future to verify our suggestions that dominance of Al-oxides as the adsorbents leads to higher arsenate solubility compared to phosphate. Further research is needed to look for a suitable methodology to obtain a reliable estimation of surface composition of soil oxides, which is essential for an accurate prediction of oxyanion behavior in soil systems.

means that arsenate adsorption becomes weaker with the increase of relative content of Al in the oxides. This relative content of Al ( fAl) increases with the decrease of the amount of Fe-oxides present in the soils studied (Table 1). Using binary synthesized Al−Fe-oxides, Masue et al. found that arsenate adsorption decreases with the increase of Al fraction.36 The binary Al−Fe-oxides were prepared by adding NaOH to nitrate salts of Al and Fe of various ratios under vigorous stirring. They found that when the material contains less than 20% of Al, Al is mainly present as substitution in the structure of Fe-oxides. At >20% of Al, Al-oxides phases can be identified. The decrease of K values for arsenate adsorption on Al−Fe-oxides measured by Masue et al. at pH 5 and 8 is given in Figure 2, for comparison with the results of this study. The comparison shows that at the same fAl, the adsorption affinity for arsenate is much more reduced for soil oxides than for the Al−Fe mixture prepared by Masue et al.. Arsenate adsorption in soils with 14% and 20% Aloxides is comparable to arsenate adsorption to the synthetic Al−Fe-oxides containing respectively 60% and 87% of Al, whereas arsenate adsorption in soils in which Al-oxides account for >40% of total oxides is comparable to that of pure Al-oxides in the synthetic system (Figure 2). The deviation between the soil oxides and synthetic Al−Fe-oxides would be even larger if we consider the fact that part of Al extracted from soils was initially organic bound, not in oxide minerals. Precipitates. We calculated soluble phosphate concentration under the control of solubility of variscite (AlPO4, logKso = −22.1) and tricalcium phosphate (Ca3(PO4)2, logKso = −28.9) using the measured pH and free Al3+ and Ca2+ concentration calculated from solution speciation (Modeling Approach). At a pH below 5−6, the AlPO4 controlled P-PO4 is mostly lower than that measured and predicted by the adsorption model (Figure 1), which indicates that adsorption is likely the controlling mechanism and the solubility of AlPO4 in soils is probably higher than used in the calculation or nonequilibrium between the P-PO4 and AlPO4 phase. The occurrence and significance of Fe and Al phosphates in acid soil is not well documented and direct evidence is scarce.37 At a pH > 7, Ca3(PO4)2 controlled P-PO4 is mostly close to that measured and lower than the adsorption model predicted (Figure 1). It indicates that in soils with relatively high Ploading, calcium phosphate minerals may be present at high pH and these minerals may be controlling P-PO4 concentration at high pH. Solubility of corresponding arsenate minerals is several orders of magnitude higher than those of phosphate and soluble arsenate concentration controlled by mineral solubility will be higher than those measured. Implications. The “one representative surface” approach (e.g., using only goethite) may become insufficient when the aim is to simulate multioxyanion adsorption simultaneously. For soil containing Al- and Fe-oxides, an obvious choice is to treat the surfaces as a combination of two single metal (hydr)oxides, with their amount proportional to those extracted from soils. Our current research shows that this treatment may sometime not work, due to probably interactions and uneven spatial locations of these oxides. In nature systems, dissolution, coprecipitation, sequential precipitation, and agglomeration of Al- and Fe-oxides are expected to happen constantly. It has been found that in Al and Fe binary system, when Al is present, the surface of oxides formed is enriched in Al compared to the bulk.38 Dissolution of Al-oxides, and consequent sorption or precipitation of soluble A1 species at the Fe-oxides surface results in disaggregation of Fe-oxides particles and creation of

■

ASSOCIATED CONTENT

S Supporting Information *

The results of DOC, soluble Ca, Al, and Fe, model parameters of CD-MUSIC and NICA model can be found in the Supporting Information. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Phone: 31-317-482332; fax: 31-317-419000; e-mail: liping. [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS

■

REFERENCES

We acknowledge the Dutch Science Foundation (ASPASIA program) and National Natural Science Foundation of China (Project number 41271493) for providing part of financial support of this research.

(1) Weng, L. P.; Temminghoff, E. J. M.; Van Riemsdijk, W. H. Contribution of individual sorbents to the control of heavy metal activity in sandy soil. Environ. Sci. Technol. 2001, 35 (22), 4436−4443. (2) Bonten, L. T. C.; Groenenberg, J. E.; Weng, L. P.; Van Riemsdijk, W. H. Use of speciation and complexation models to estimate heavy metal sorption in soils. Geoderma 2008, 146 (1−2), 303−310. (3) Lumsdon, D. G. Partitioning of organic carbon, aluminium and cadmium between solid and solution in soils: application of a mineralhumic particle additivity model. Eur. J. Soil Sci. 2004, 55 (2), 271−285. (4) Cances, B.; Ponthieu, M.; Castrec-Rouelle, M.; Aubry, E.; Benedetti, M. F. Metal ions speciation in a soil and its solution: experimental data and model results. Geoderma 2003, 113 (3−4), 341−355. (5) Dijkstra, J. J.; Meeussen, J. C. L.; Comans, R. N. J. Leaching of heavy metals from contaminated soils: An experimental and modeling study. Environ. Sci. Technol. 2004, 38 (16), 4390−4395. (6) Weng, L. P.; Vega, F. A.; Van Riemsdijk, W. H. Competitive and synergistic effects in pH dependent phosphate adsorption in soils: LCD modeling. Environ. Sci. Technol. 2011, 45 (19), 8420−8428. (7) Devau, N.; Hinsinger, P.; Le Cadre, E.; Colomb, B.; Gerard, F. Fertilization and pH effects on processes and mechanisms controlling dissolved inorganic phosphorus in soils. Geochim. Cosmochim. Acta 2011, 75 (10), 2980−2996. (8) Devau, N.; Le Cadre, E.; Hinsinger, P.; Jaillard, B.; Gerard, F. Soil pH controls the environmental availability of phosphorus: Experimental and mechanistic modelling approaches. Appl. Geochem. 2009, 24 (11), 2163−2174. (9) Gustafsson, J. P. Modelling competitive anion adsorption on oxide minerals and an allophane-containing soil. Environ. Sci. Technol. 2001, 52 (4), 639−653.

G

dx.doi.org/10.1021/es400526q | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

(10) Gustafsson, J. P. Arsenate adsorption to soils: Modelling the competition from humic substances. Geoderma 2006, 136 (1−2), 320−330. (11) Hiemstra, T.; Antelo, J.; Rahnemaie, R.; Van Riemsdijk, W. H. Nanoparticles in natural systems I: The effective reactive surface area of the natural oxide fraction in field samples. Geochim. Cosmochim. Acta 2010, 74 (1), 41−58. (12) Kinniburgh, D. G.; Van Riemsdijk, W. H.; Koopal, L. K.; Borkovec, M.; Benedetti, M. F.; Avena, M. J. Ion binding to natural organic matter: competition, heterogeneity, stoichiometry and thermodynamic consistency. Colloids Surf., A 1999, 151 (1−2), 147−166. (13) Roy, W. R.; Hassett, J. J.; Griffin, R. A. Competitive interactions of phosphate and molybdate on arsenate adsorption. Soil Sci. 1986, 142 (4), 203−210. (14) Melamed, R.; Jurinak, J. J.; Dudley, L. M. Effect of adsorbed phosphate on transport of arsenate through an oxisol. Soil Sci. Soc. Am. J. 1995, 59 (5), 1289−1294. (15) Violante, A.; Pigna, M. Competitive sorption of arsenate and phosphate on different clay minerals and soils. Soil Sci. Soc. Am. J. 2002, 66 (6), 1788−1796. (16) Van Erp, P. J.; Houba, V. J. G.; Van Beusichem, M. L. One hundredth molar calcium chloride extraction procedure. Part I: A review of soil chemical, analytical, and plant nutritional aspects. Commun. Soil Sci. Plant Anal. 1998, 29 (11−14), 1603−1623. (17) Murphy, J.; Riley, J. P. A modified single solution method for the determination of phosphate in natural waters. Anal. Chim. Acta 1962, 26 (1), 31−36. (18) Wolf, A. M.; Baker, D. E. Colorimetric method for phosphorus measurement in ammonium oxalate soil extracts. Commun. Soil Sci. Plant Anal. 1990, 21 (19−20), 2257−2263. (19) Pettersson, A. K.; Karlberg, B. Simultaneous determination of orthophosphate and arsenate based on multi-way spectroscopic-kinetic data evaluation. Anal. Chim. Acta 1997, 354 (1−3), 241−248. (20) Weng, L. P.; Van Riemsdijk, W. H.; Koopal, L. K.; Hiemstra, T. Ligand and Charge Distribution (LCD) model for the description of fulvic acid adsorption to goethite. J. Colloid Interface Sci. 2006, 302 (2), 442−457. (21) Hiemstra, T.; De Wit, J. C. M.; Van Riemsdijk, W. H. Multisite proton adsorption modeling at the solid-solution interface of (hydr)oxides - a new approach. 2. Application to various important (hydr)oxides. J. Colloid Interface Sci. 1989, 133 (1), 105−117. (22) Hiemstra, T.; Van Riemsdijk, W. H. A surface structural approach to ion adsorption: The charge distribution (CD) model. J. Colloid Interface Sci. 1996, 179 (2), 488−508. (23) Manning, B. A.; Goldberg, S. Modeling arsenate competitive adsorption on kaolinite, montmorillonite and Illite. Clays Clay Miner. 1996, 44 (5), 609−623. (24) Manning, B. A.; Goldberg, S. Modeling competitive adsorption of arsenate with phosphate and molybdate on oxide minerals. Soil Sci. Soc. Am. J. 1996, 60 (1), 121−131. (25) Stachowicz, M.; Hiemstra, T.; Van Riemsdijk, W. H. Surface speciation of As(III) and As(V) in relation to charge distribution. J. Colloid Interface Sci. 2006, 302 (1), 62−75. (26) Stachowicz, M.; Hiemstra, T.; Van Riemsdijk, W. H. Multicompetitive interaction of As(III) and As(V) oxyanions with Ca2+, Mg2+, PO43‑, and CO32‑ ions on goethite. J. Colloid Interface Sci. 2008, 320 (2), 400−414. (27) Weng, L. P.; Van Riemsdijk, W. H.; Hiemstra, T. Humic nanoparticles at the oxide−water interface: Interactions with phosphate ion adsorption. Environ. Sci. Technol. 2008, 42 (23), 8747−8752. (28) Keizer, M. G.; Van Riemsdijk, W. H. ECOSAT: Equilibrium Calculation of Speciation and Transport; Manual Program; Agricultural University of Wageningen: Wageningen, 1994. (29) Milne, C. J.; Kinniburgh, D. G.; Tipping, E. Generic NICADonnan model parameters for proton binding by humic substances. Environ. Sci. Technol. 2001, 35 (10), 2049−2059.

(30) Milne, C. J.; Kinniburgh, D. G.; Van Riemsdijk, W. H.; Tipping, E. Generic NICA-Donnan model parameters for metal-ion binding by humic substances. Environ. Sci. Technol. 2003, 37 (5), 958−971. (31) Vega, F. A.; Weng, L. Speciation of heavy metals in River Rhine. Water Res. 2013, 47 (1), 363−372. (32) Weng, L. P.; Van Riemsdijk, W. H.; Hiemstra, T. Factors controlling phosphate interaction with iron oxides. J. Environ. Qual. 2012, 41 (3), 628−635. (33) Meeussen, J. C. L. ORCHESTRA: An object-oriented framework for implementing chemical equilibrium models. Environ. Sci. Technol. 2003, 37 (6), 1175−1182. (34) Goldberg, S.; Lesch, S. M.; Suarez, D. L.; Basta, N. T. Predicting arsenate adsorption by soils using soil chemical parameters in the constant capacitance model. Soil Sci. Soc. Am. J. 2005, 69 (5), 1389− 1398. (35) Gustafsson, J. P.; Mwamila, L. B.; Kergoat, K. The pH dependence of phosphate sorption and desorption in Swedish agricultural soils. Geoderma 2012, 189, 304−311. (36) Masue, Y.; Loeppert, R. H.; Kramer, T. A. Arsenate and arsenite adsorption and desorption behavior on coprecipitated aluminum: iron hydroxides. Environ. Sci. Technol. 2007, 41 (3), 837−842. (37) Hinsinger, P. Bioavailability of soil inorganic P in the rhizosphere as affected by root-induced chemical changes: a review. Plant Soil 2001, 237 (2), 173−195. (38) Anderson, P. R.; Benjamin, M. M. Surface and bulk characteristics of binary oxide suspensions. Environ. Sci. Technol. 1990, 24 (5), 692−698. (39) Schwertmann, U.; Taylor, R. M., Iron oxides. In Minerals in Soil Environments; Dixon, J. B., Weed, S. B., Eds.; Soil Science Society of America: Madison, WI, 1989; pp 379−438.

H

dx.doi.org/10.1021/es400526q | Environ. Sci. Technol. XXXX, XXX, XXX−XXX