Thermodynamics of Dimethylarsinic Acid and Arsenate Interactions

ARTICLE pubs.acs.org/est

Thermodynamics of Dimethylarsinic Acid and Arsenate Interactions with Hydrated Iron-(Oxyhydr)oxide Clusters: DFT Calculations Adrian Adamescu, I. P. Hamilton, and Hind A. Al-Abadleh* Chemistry Department, Wilfrid Laurier University, Waterloo, Ontario N2L 3C5 Canada

bS Supporting Information ABSTRACT: Dimethylarsinic Acid (DMA) belongs to an important class of organoarsenical compounds commonly detected in arsenic speciation studies of environmental samples and pyrolysis products of fossil fuels. Transformation of DMA under certain conditions leads to the formation of other forms of arsenic, which could be more toxic than DMA to biota, and more efficient in deactivating catalysts used in petrochemical refining. Published surface sensitive X-ray and infrared spectroscopic work suggested that DMA simultaneously forms inner- and outer-sphere complexes with iron-(oxyhydr)oxides. Computational work on the complexation of arsenicals with various surfaces of environmental and industrial interest provides useful information that aids in the interpretation of experimental spectroscopic data as well as predictions of thermodynamic favorability of surface interactions. We report herein Gibbs free energies of adsorption, ΔGads, for various ligand exchange reactions between hydrated complexes of DMA and Fe-(oxyhydr)oxide clusters calculated using density functional theory (DFT) at the B3LYP/6-311+G(d,p) level. Calculations using arsenate were also performed for comparison. Calculated As-(O,Fe) distances and stretching frequencies of As
O bonds are also reported for comparison with experimental spectroscopic data. Gibbs free energies of desorption, ΔGdes, due to reactions with phosphorus species at pH 7 are reported as well. Our results indicate that the formation of both inner- and outersphere DMA complexes is thermodynamically favorable, with the former having a more negative ΔGads. Values of ΔGdes indicate that desorption favorability of DMA complexes increases in this order: bidentate < mondentate < outersphere. The significance of our results for the overall surface complexation mechanism of DMA is discussed.

’ INTRODUCTION Methylated arsenicals, such as monomethylarsonic acid (MMA) and dimethylarsinic acid (DMA), are produced naturally in biomethylation processes of inorganic arsenic.1,2 They also accumulate in petroleum, shale, oil sands, and coal deposits as a result of biogeochemical processes,3,4 and in biomass grown in arsenic-rich soils.5 Methylated arsenicals were used historically as pesticides and herbicides on large agricultural fields,6 and in Canada they were used to control the mountain pine beetle infestations that devastated forests in British Columbia.7 MMA and DMA have also been detected in the leachates of landfills rich in waste containing arsenic such as glass, alloys, and semiconductors,8 and in the urine of animals as metabolic products of inorganic arsenic.2 The toxicity of arsenic varies greatly with its chemical form and oxidation state. The toxicity order of arsenic compounds in several cell lines is: DMA(III), MMA(III) > iAs(III) > iAs(V) > DMA(V), MMA(V).1 In particular, DMA— which is the focus of this study—has been shown to have multiorgan tumor promoting activity in rodents.9 Its trivalent form, produced under certain soil conditions, can be highly cytotoxic and genotoxic, directly interacting with the genetic material.9 Overall, arsenic compounds in their organic and inorganic forms present a challenge to the energy industry and pose both a health and an environmental risk.10 Experimental studies using extended X-ray absorption fine structure (EXAFS) and attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR) have investigated r 2011 American Chemical Society

the surface interaction of DMA with different soils 11,12 and metal-(oxyhydr)oxides that include TiO2,13 Al2O3,14 Fe2O3, 15,16 and FeOOH particles.11,16 Results reported in these studies include binding thermodynamics, kinetics and structural parameters of surface complexes. Sorption studies have shown that MMA and DMA adsorb mostly to Fe-(oxyhydr)oxides in the soil and with EXAFS it was determined that for both DMA and MMA, the interatomic As
Fe distance was 3.30 Å which is indicative of bidentate binuclear complexation.11 We earlier reported using ATR-FTIR and quantum chemical calculations the simultaneous formation of both outer-sphere and innersphere complexes of DMA on goethite and hematite.15 Calculated As
Fe distances using density functional theory (DFT) with the B3LYP functional and the 6-31G(d) basis set are found to be 3.16
3.34 Å for the inner sphere complexes, and 4.95
6.42 Å for the outer sphere complexes.15 Further quantum chemical calculations of thermodynamic quantities are required to determine which of the outer-sphere or monodentate or bidentate inner-sphere complexes are more favorable under neutral, basic, and acidic conditions. Computational work on the complexes of arsenicals with various surfaces found in soil provides useful information that Received: August 7, 2011 Accepted: October 27, 2011 Revised: October 18, 2011 Published: October 27, 2011 10438

dx.doi.org/10.1021/es202749h | Environ. Sci. Technol. 2011, 45, 10438–10444

Environmental Science & Technology aids in the interpretation of experimental spectroscopic data as well as the prediction of thermodynamically favorable surface interactions. For example, Kubicki et al.17 performed quantum chemical calculations using DFT on surface complexes of inorganic arsenic, sulfate, carbonate, and phosphate with Fe-(oxyhydr) oxides. Calculated geometries, vibrational frequencies, and Gibbs free energies of adsorption (ΔGads) were compared with published experimental data.17 However, to date, there has been no computational work done on the thermodynamics of potential ligand exchange reactions of organic forms of arsenic, such as MMA and DMA, with Fe-(oxyhydr)oxides. Performing thermodynamic calculations of reactant and product species in simulated bulk and surface environments will provide mechanistic details on the energetics of DMA surface chemistry that are challenging to obtain experimentally. In the present study, we report values of ΔGads for various ligand exchange reactions between hydrated complexes of DMA on Fe-(oxyhydr)oxide clusters using DFT with the B3LYP functional and the 6-311+G(d,p) basis set. Reactions forming inner- and outer-sphere complexes are constructed to simulate neutral, slightly basic, and acidic aqueous environments. For comparison, complementary thermodynamic calculations were performed on ligand exchange reactions of iAs(V) with the same Fe-(oxyhydr)oxide clusters that form bidentate binuclear clusters. Thermodynamic favorability of the desorption of adsorbed DMA due to phosphate was also modeled by calculating Gibbs free energies of desorption, ΔGdes, under conditions that simulate neutral environments.

’ COMPUTATIONAL SECTION Calculations were performed using Gaussian 09 running on SHARCNET 18 typically on 8 or 16 processors. Structures were energy-minimized without symmetry constraints using DFT with the B3LYP functional and the 6-311+G(d,p) basis set. This basis set is larger and more accurate than the one we used earlier.15 The calculations were performed on hydrated clusters with a net charge of zero or +1, at 298.15 K (25 °C) and 1 atm. Hydration was simulated both explicitly by adding four water molecules and implicitly by using the Integral-Equation-Formalism Polarizable Continuum Model (IEFPCM) 19 which is referred to as “hydrated environment” throughout the text. Clusters containing Fe 3+ were calculated at high spin (Multiplicity 11) which yielded the lowest electronic energy. It was observed that for high spin Fe calculations, having more than three strong OH ligands on the Fe3+ atom causes a change in the coordination geometry and that adding two additional water molecules converts the 4-fold coordinated species into a 5-fold coordinated species. This is consistent with the findings of Kubicki,20 where it was observed that a 4-fold (tetrahedral) coordination geometry forms in less hydrated environments with four or fewer explicit water molecules surrounding the iron hydroxide [Fe(OH)3(OH2)] 3 (H2O)2, and a 5-fold (trigonal bipyramidal) coordination geometry formed when six extra water molecules were added to form the [Fe(OH)3(OH2)2] 3 (H2O)8 complex. For all other complexes with Fe3+ having three or fewer OH ligands, the octahedral coordination is found to be stable. The stability of the wave function for all high-spin Fe3+ complexes A through F (Table1, Figure S2 of the Supporting Information, SI) was checked using the Stable=Opt command in Gaussian. Motivated by a recent study of Chan et al.21 which examined alternatives to the B3LYP functional, the reactants and

ARTICLE

products involved in the bidentate DMA reactions 3, 4, 8, and 9 (Table 2) and iAs reactions 13 and 15 (Table 3) were also optimized with the BMK functional.22 It may be seen that the BMK/6-311+G(d,p) results show differences in ΔG values of 15
30% compared to the B3LYP/6-311+G(d,p) results but, because these differences are systematic, the order of the ΔG values is unchanged. We favor the B3LYP functional since it has proven to give results that parallel our own experimental findings very well, and is the method of choice in computational studies 17,23 that we use for comparison with ours.

’ RESULTS AND DISCUSSION Calculations of ΔGads of DMA and iAs(V) on Iron Oxide Clusters. The Gibbs Energy, G, is calculated by adding the total

electronic energy, E0, to the thermal correction to the Gibbs free energy, Gcorr (which also includes the zero-point correction), as shown in eq 1 G ¼ E0 þ Gcorr

ð1Þ

Then ΔGads is given by the following: ΔGads ð298:15KÞ ¼

24

0 þ Gcorr Þ
∑ ðE0 þ Gcorr Þ ∑ ðEproducts reactants

ð2Þ Since DMA has a pKa = 6.2, we simulated adsorption under neutral/slightly basic conditions (6.2 < pH < 9) where most of the reactant DMA is deprotonated (negatively charged). Calculations to simulate acidic conditions (pH < 6.2) were also carried out, where most of the reactant DMA is protonated (neutral). Since most Fe-(oxyhydr)oxides have isoelectric points around pH 9, surface sites of these materials are mostly positively charged with some neutral ones. Figures S1 and S2 of the SI show energy-minimized equilibrium structures of the reactants (hydrated DMA, Fe-(oxyhydr)oxide) and the products (innerand outer-sphere DMA-Fe-(oxyhydr)oxide clusters) used in constructing these ligand exchange reactions. Table 1 lists values for the calculated E0, Gcorr, and G, for all reactants and products considered at 298.15K and 1 atm (Figures S1 and S2 of the SI). The calculated ΔGads for various outer- and inner-sphere complexation reactions are listed in Table 2. From the negative sign of ΔGads it is clear that the formation of inner- and outersphere complexes is thermodynamically favorable. Under simulated neutral/slightly basic environments, formation of neutral surface clusters (reactions 1, 3, and 5) is more favorable than positively charged ones (reaction 2, 4) with reaction 3, corresponding to bidentate binuclear cluster formation, being the most favorable. Under simulated acidic environments, formation of positively charged inner-sphere complexes (reactions 7 and 9) is more favorable than neutral ones (reactions 6 and 8) with reaction 9, corresponding to bidentate binuclear cluster formation, being the most favorable. Formation of neutral and positively charged outer-sphere complexes (reactions 10 and 11) is also favorable. It is clear from Table 2 that electrostatic attraction between oppositely charged reactants to form neutral surface clusters and leaving groups is the main driving force for ligand exchange under neutral/basic conditions. Under acidic conditions, fully protonated DMA favors interaction with positively charged sites to form inner-sphere complexes. These results are consistent with experimental data that showed DMA adsorption onto goethite to be independent of solution ionic strength which 10439

dx.doi.org/10.1021/es202749h |Environ. Sci. Technol. 2011, 45, 10438–10444

Environmental Science & Technology

ARTICLE

Table 1. Calculated Electronic Energy, E0, the Thermal Correction to Gibbs Free Energy, Gcorr, and the Gibbs Free Energy, G, For All Reactants and Products Considered at 298.15 K and 1 atm E0 (a.u.)a

Gcorr (a.u.)

G (a.u.)


2772.29436

0.13093


2772.16343


2772.75445
2925.24839

0.14596 0.16658


2772.60849
2925.08181


3594.56486

0.21521


3594.34965


3595.01157

0.22837


3594.78320


949.29560

0.07723


949.21837


949.76309

0.08763


949.67545

reactants in SI Figure S1


b c

DMA 3 (H2O)4 DMA-H 3 (H2O)4 DMA
3 (H2O)6

d

Fe2(OH)6(OH2)2 3 (H2O)6

a

Fe2(OH)5(OH2)5+ 3 (H2O)4

e

HPO42‑ 3 (H2O)4

f

H2PO4
3 (H2O)4

g

products in SI Figure S2 A

DMA-Fe2(OH)5(OH2)3 3 (H2O)7 [DMA-Fe2(OH)4(OH2)5 3 (H2O)4]+ DMA-Fe2(OH)5(OH2)3 3 (H2O)4

B C

[DMA-Fe2(OH)4(OH2)4 3 (H2O)4]+ DMA 3 (H2O)5 Fe2(OH)5(OH2)4

D E

[DMA-H 3 (H2O)4 Fe2(OH)5(OH2)5]+

F G

Fe2(OH)4(OH2)4-PO4
3 (H2O)4

H

Fe2(OH)4(OH2)4-HPO4 3 (H2O)4

Gcorr (a.u.)

G (a.u.)


6137.91267

0.31177


6137.60090


5985.41196
5908.47915

0.28823 0.25455


5985.12373
5908.22459


5908.93386

0.26262


5908.67124


6061.43201

0.29520


6061.13681


6061.87090

0.30479


6061.56611


4085.49776

0.20814


4085.28962


4085.95797

0.21479


4085.74318

E0 (a.u.)

Gcorr (a.u.)

G (a.u.)

AsO4 3 (H2O)4 HAsO4
2 3 (H2O)4


2843.20656
2843.69798

0.06596 0.07512


2843.14060
2843.62286

3

H2AsO4
3 (H2O)4


2844.17010

0.08290


2844.08720

4

H3AsO4 3 (H2O)4


2844.61780

0.09617


2844.52163

reactants in SI Figure S3
3

1 2

products in SI Figure S4 A B C

a

E0 (a.u.)




[AsO4
Fe2(OH)4(OH2)4 3 (H2O)4] HAsO4
Fe2(OH)4(OH2)4 3 (H2O)4

[H2AsO4Fe2(OH)4(OH2)4 3 (H2O)4]+

E0 (a.u.)

Gcorr (a.u.)

G (a.u.)


5979.89659

0.20605


5979.69055


5980.35740

0.21149


5980.14591


5980.79753

0.21904


5980.57849

leaving groups

E0 (a.u.)

Gcorr (a.u.)

G (a.u.)

H2O OH
3 (H2O)2


76.46641
228.93031

0.00354 0.02629


76.46287
228.90402

1 a.u. = 2625.5 kJ/mol.

is indicative of an inner-sphere adsorption mechanism.16,25 At the same time, DMA adsorption did not shift goethite’s point of zero charge, which is indicative of an outer-sphere adsorption mechanism.25 Also, ATR-FTIR spectra of DMA adsorption as function of pH showed that the intensity of spectral components characteristic of outer-sphere complexation increased with decreasing pH along with those assigned to inner-sphere complexes.15 This is evidence that the reactions considered in this study are realistic representations of what takes place in aqueous environments at the molecular level. For completeness, calculations on ligand exchange reactions between iAs(V) and the Fe-(oxyhydr)oxide clusters were performed as listed in Table 3. Figure S3 of the SI shows energyminimized equilibrium structures of the hydrated iAs(V) (reactants) for different protonation states: AsO43‑ 3 (H2O)4, HAsO42‑ 3 (H2O)4, H2AsO4
3 (H2O)4, and H3AsO4 3 (H2O)4. The dominant species under normal environmental conditions (5 < pH < 9) are HAsO42‑ and H2AsO4
, while H3AsO4 may be found in very acidic environments and AsO43‑ in very basic environments. Figure S4 of the SI shows energy-minimized equilibrium structures of the products (bidentate binuclear

iAs(V)-Fe-(oxyhydr)oxide clusters) used in constructing the ligand exchange reactions in Table 3. Under normal environmental conditions with pH ranges from 5 to 9 the most relevant processes are reactions 13
16 in Table 3. Of these, the most thermodynamically favorable are reactions 14 and 15, with a ΔGads of
162 and
138 kJ/mol, respectively. In these reactions, the anions adsorb to the positively charged surface while forming a negatively charged complex and a neutral complex for reactions 14 and 15, respectively, with waters as the leaving groups. Kubicki et al.17 performed similar ligand exchange reaction calculations for iAs(V) at the B3LYP/6-311 +G(d,p)//B3LYP/6-31G(d) level of theory with the IEFPCM solvation model and reported very similar results. A reaction comparable to our reaction 15 in Table 3 (except having 8 waters surrounding the H2AsO4
instead of 4) yielded a ΔGads of
140 kJ/mol compared to
138 kJ/mol for our reaction 15. In the following section, analysis of geometrical parameters and vibrational modes in simulated surface clusters of DMA and iAs(V) is presented for comparison with experimental spectroscopic data. Correlation of Calculated As-(O, Fe) Bond Distances with v(As
O). Tables S1 and S2 of the SI list the calculated As
O 10440

dx.doi.org/10.1021/es202749h |Environ. Sci. Technol. 2011, 45, 10438–10444

Environmental Science & Technology

ARTICLE

Table 2. Calculated ΔGads for Ligand Exchange Reactions of DMA and Fe-(Oxyhydr)oxides ΔGads (kJ/mol)

ligand exchange reactions under simulated neutral and slightly basic environments (6.2 < pH < 9) monodentate A and B 1

DMA
3 (H2O)6 + Fe2(OH)5(OH2)5+ 3 (H2O)4 f DMA-Fe2(OH)5(OH2)3 3 (H2O)7 + 5(H2O)

2

DMA
3 (H2O)4 + Fe2(OH)5(OH2)5+ 3 (H2O)4 f [DMA-Fe2(OH)4(OH2)5 3 (H2O)4]+ + OH
3 (H2O)2 + 2(H2O)

3

DMA
3 (H2O)4 + Fe2(OH)5(OH2)5+ 3 (H2O)4 f DMA-Fe2(OH)5(OH2)3 3 (H2O)4 + 6(H2O) DMA
3 (H2O)4 + Fe2(OH)5(OH2)5+ 3 (H2O)4 f [DMA-Fe2(OH)4(OH2)4 3 (H2O)4]+ + OH
3 (H2O)2 + 3(H2O)


132
18

bidentate C and D 4


145 (
170)a
45 (
65)a

outer-sphere E 5

DMA
3 (H2O)4 + Fe2(OH)5(OH2)5+ 3 (H2O)4 f DMA 3 (H2O)5 Fe2(OH)5(OH2)4 + 5(H2O) under simulated acidic environments (pH < 6.2)

6

DMA-H 3 (H2O)4 + Fe2(OH)6(OH2)2 3 (H2O)6 f DMA-Fe2(OH)5(OH2)3 3 (H2O)7 + 3(H2O)


109

monodentate A and B 7

DMA-H 3 (H2O)4 + Fe2(OH)5(OH2)5+ 3 (H2O)4 f [DMA-Fe2(OH)4(OH2)5 3 (H2O)4]+ + 5(H2O)

8

DMA-H 3 (H2O)4 + Fe2(OH)6(OH2)2 3 (H2O)6 f DMA-Fe2(OH)5(OH2)3 3 (H2O)4 + 6(H2O)


82
122

bidentate C and D Fe2(OH)5(OH2)5+ 3 (H2O)4

+

9

DMA-H 3 (H2O)4 +

f [DMA-Fe2(OH)4(OH2)4 3 (H2O)4] + 6(H2O)

10

DMA-H 3 (H2O)4 + Fe2(OH)6(OH2)2 3 (H2O)6 f DMA 3 (H2O)5 Fe2(OH)5(OH2)4 + 4(H2O)


115 (
159)a
149 (
181)a

outer-sphere E and F 11 a

DMA-H 3 (H2O)4 + Fe2(OH)5(OH2)5+ 3 (H2O)4 f [DMA-H 3 (H2O)4 Fe2(OH)5(OH2)5]++ 4(H2O)


79
68

Calculated with BMK/6-311+G(d,p) for comparison (the values for ΔGads for the BMK and B3LYP reactions differ by 15
30%).

Table 3. Calculated ΔGAds for Ligand Exchange Reactions of iAs(V) and Fe-(Oxyhydr)oxides ΔGads (kJ/mol)

ligand exchange reactions 12 13 14 15 16 17 18

AsO43
3 (H2O)4 + Fe2(OH)5(OH2)5+ 3 (H2O)4 f [Fe2(OH)4(OH2)4-AsO4 3 (H2O)4]
+ OH
3 (H2O)2 + 3(H2O)

HAsO42
3 (H2O)4 + Fe2(OH)5(OH2)5+ 3 (H2O)4 f Fe2(OH)4(OH2)4-HAsO4 3 (H2O)4 + OH
3 (H2O)2 + 3(H2O)

HAsO42
3 (H2O)4 + Fe2(OH)5(OH2)5+ 3 (H2O)4 f [Fe2(OH)4(OH2)4-AsO4 3 (H2O)4]
+ 6(H2O) H2AsO4
3 (H2O)4 + Fe2(OH)5(OH2)5+ 3 (H2O)4 f Fe2(OH)4(OH2)4-HAsO4 3 (H2O)4 + 6(H2O) H2AsO4
3 (H2O)4 + Fe2(OH)6(OH2)2 3 (H2O)6 f [Fe2(OH)4(OH2)4-AsO4 3 (H2O)4]
+ 6(H2O) H3AsO4 3 (H2O)4 + Fe2(OH)5(OH2)5+ 3 (H2O)4 f [Fe2(OH)4(OH2)4-H2AsO4 3 (H2O)4]+ + 6(H2O) H3AsO4 3 (H2O)4 + Fe2(OH)6(OH2)2 3 (H2O)6 f Fe2(OH)4(OH2)4-HAsO4 3 (H2O)4 + 6(H2O)


156
85 (
105)a
162
138 (
167)a
81
134
136

Calculated with BMK/6-311+G(d,p) for comparison (the values for ΔGads for the BMK and B3LYP reactions differ by 19% and 17% for reactions 13 and 15, respectively.) a

bond lengths and interatomic distances between As and Fe in DMA-Fe-(oxyhydr)oxide and iAs(V)-Fe-(oxyhydr)oxide clusters, respectively. The inner-sphere clusters of DMA (A-D) have As
O bond lengths ranging from 1.70
1.74 Å, in good agreement with recent EXAFS studies reporting bond distances to be 1.71 Å 11 for bidentate DMA. The broad range of As
O bond lengths in inner-sphere complexes reflects how binding to Fe, degree of hydrogen bonding, and protonation affect this distance. The interatomic As
Fe distance was calculated to be 3.30 to 3.43 Å for the bidentate DMA clusters, compared to 3.3 Å from EXAFS studies.11 The monodentate DMA clusters also have As
Fe distances (3.36, 3.45 Å) similar to those calculated for the bidentate DMA clusters. In the monodentate clusters, however, the As atom is located about 5 Å away from the second Fe atom in the cluster. Distances between As and Fe greater than 5 Å are usually undetectable using EXAFS, and a coordination number of 2 (rather than 1) results in a better fit to the data as reported by Shimizu et al.11 The interatomic distances for outer-sphere

complexes depend on whether the deprotonated DMA
or protonated DMA-H was used. For the deprotonated DMA
, the As
Fe distances were calculated to be 4.84 and 5.28 Å, while for protonated species DMA-H, the As
Fe distances were calculated to be 5.75 and 6.84 Å. Again, these distances are too long to be detected using EXAFS, which makes this technique inconclusive regarding the existence of outer-sphere DMA complexes, even if their formation is thermodynamically favorable. X-ray techniques such as in situ resonant surface scattering utilized by Catalano et al. 26,27 for studying arsenate adsorption on hematite and corundum single crystal surfaces can measure outer-sphere arsenate complexes with As-metal distances >5 Å. Hence, studying the surface complexation of DMA using these techniques will provide invaluable insight into their geometries. EXAFS studies summarized in ref 11 for iAs(V) adsorption to goethite report As
Fe interatomic distances between 3.23 and 3.37 Å. These results are in excellent agreement with our calculated As
Fe distances, which are between 3.23 and 3.36 Å 10441

dx.doi.org/10.1021/es202749h |Environ. Sci. Technol. 2011, 45, 10438–10444

Environmental Science & Technology

ARTICLE

Table 4. Assignment of Calculated v(As
O) Frequencies (cm
1) in DMA- and iAs(V)-Fe-(Oxyhydr)oxides Complexesa v(As
OFe)

DMA-Fe-(oxyhydr)oxides clusters (Figure S2)

v(AsdO)

v(As
OH)

monodentate A. DMA-Fe2(OH)5(OH2)3 3 (H2O)7

741b,c,d

772d,c

c,d

+

B. [DMA-Fe2(OH)4(OH2)5 3 (H2O)4]

c,d

755 , 774b,c,d

724

bidentate 723b,c,d, 775c,d

C. DMA-Fe2(OH)5(OH2)3 3 (H2O)4

c,d

710 , 731c,d, 741c,d, 769b,c,d

+

D. [DMA-Fe2(OH)4(OH2)4 3 (H2O)4]

outer-sphere 727c,d, 748c,d, 757b,c,d, 771c,d, 773c,d

E. DMA 3 (H2O)5 Fe2(OH)5(OH2)4

797c,d, 800c,d, 820c,d, 826b,c,d

+

F. [DMA-H 3 (H2O)4 Fe2(OH)5(OH2)5]

699c,d

iAs(V)-Fe-(oxyhydr)oxides Bidentate Clusters (Figure S4) A. [Fe2(OH)4(OH2)4 AsO4]
(H2O)4

661b,c,d, 692c,d

B. Fe2(OH)4(OH2)4
HAsO4 (H2O)4

c,d

C. [Fe2(OH)4(OH2)4
H2AsO4] (H2O)4 +

784 , 803 b,c

791c,d, 883c,d

b,c,d

, 815

c,d

c,d

886c,d

c,d

621c,d c

690 , 722c,d, 743c,d

832 , 849 , 888

a

A scaling factor of 0.9787 was used to correct calculated frequencies for anharmonicity for the organoarsenical clusters. A scaling factor of 1.0199 was used to correct calculated frequencies for anharmonicity for the iAs(V) clusters. b Most intense spectral component. c Coupled with water rocking/ wagging. d Coupled with OH bending.

Table 5. Calculated ΔGdes of Adsorbed DMA by P(V) Species under Simulated Neutral Conditions ΔGdes (kJ/mol)

ligand exchange reactions corresponding to the desorption of .... monodentate A and B 19 20 21 22

DMA-Fe2(OH)5(OH2)3 3 (H2O)7 + H2PO4
3 (H2O)4 f Fe2(OH)4(OH2)4-HPO4 3 (H2O)4 + DMA
3 (H2O)4 + 3(H2O) DMA-Fe2(OH)4(OH2)5+ 3 (H2O)4 + H2PO4
3 (H2O)4 f Fe2(OH)4(OH2)4-HPO4 3 (H2O)4 + DMA-H 3 (H2O)4 + H2O

DMA-Fe2(OH)5(OH2)3 3 (H2O)7 + HPO42
3 (H2O)4 f Fe2(OH)4(OH2)4-PO4
3 (H2O)4 + DMA
3 (H2O)4 + 3(H2O) DMA-Fe2(OH)4(OH2)5+ 3 (H2O)4 + HPO42
3 (H2O)4 f Fe2(OH)4(OH2)4-HPO4 3 (H2O)4 + DMA
3 (H2O)4 + H2O


50
40
59
72

bidentate C and D 23 24 25 26

DMA-Fe2(OH)5(OH2)3 3 (H2O)4 + H2PO4
3 (H2O)4 f Fe2(OH)4(OH2)4-HPO4 3 (H2O)4 + DMA
3 (H2O)4 DMA-Fe2(OH)4(OH2)4+ 3 (H2O)4 + H2PO4
3 (H2O)4 f Fe2(OH)4(OH2)4-HPO4 3 (H2O)4 + DMA-H 3 (H2O)4 DMA-Fe2(OH)5(OH2)3 3 (H2O)4 + HPO42
3 (H2O)4 f Fe2(OH)4(OH2)4-PO4
3 (H2O)4 + DMA
3 (H2O)4 DMA-Fe2(OH)4(OH2)4+ 3 (H2O)4 + HPO42
3 (H2O)4 f Fe2(OH)4(OH2)4-HPO4 3 (H2O)4 + DMA
3 (H2O)4


17
13
26
45

outer-sphere E and F 27 28 29 30

DMA-H 3 (H2O)4 Fe2(OH)5(OH2)5+ + H2PO4
3 (H2O)4 f Fe2(OH)4(OH2)4-HPO4 3 (H2O)4 + DMA-H 3 (H2O)4 + 2(H2O) DMA 3 (H2O)5 Fe2(OH)5(OH2)4 + H2PO4
3 (H2O)4 f Fe2(OH)4(OH2)4-HPO4 3 (H2O)4 + DMA
3 (H2O)4 + 2(H2O)

DMA-H 3 (H2O)4 Fe2(OH)5(OH2)5+ + HPO42
3 (H2O)4 f Fe2(OH)4(OH2)4
HPO4 3 (H2O)4 + DMA
3 (H2O)4 + 2(H2O) DMA 3 (H2O)5 Fe2(OH)5(OH2)4 + HPO42
3 (H2O)4 f Fe2(OH)4(OH2)4-PO4
3 (H2O)4 + DMA
3 (H2O)4 + 2(H2O)

(Table S2 of the SI). Sherman and Randall 28 also observed from EXAFS studies that when iAs(V) is adsorbed to iron (hydr)oxides, the oxygen shell is distorted to give two oxygens at a short distance and two oxygens at a longer distance, namely 1.62, 1.67, 1.71, and 1.71 Å. This is similar to our calculated results for similar complexes yielding As
O distances between 1.66 and 1.70 Å, As
OH distances between 1.74 and 1.81 Å and As
O(Fe) distances between 1.68 and 1.78 Å as seen in Complexes A, B, and C (Figure S4 and Table S2 of the SI). Kubicki et al.17 also reported calculated As
O distances for the complex Fe2(OH)4(OH2)4
HAsO4 3 (H2O)4 to be 1.66, As
OH 1.77 and As
O(Fe) 1.71 Å. This structure is similar to Complex B (Table S2 of the SI), where calculated As
O distances are found to be 1.66 and 1.81 Å for uncomplexed As
O and As
OH bonds, respectively, and 1.71 and 1.72 Å for As
O(Fe) bonds. Table 4 shows the values of v(As
O) in the various DMA(ads) complexes considered in this study. We reported earlier


94
53
126
62

ATR-FTIR adsorption data of DMA(ads) on hematite and goethite where distinct spectral components were assigned to both inner- and outer-sphere complexes.15 The IR signature of DMA(ads) on hematite yielded v(As
O) at 877, 840, 816, 793, and 775 cm
1, and on goethite at 876, 837, 787, and 768 cm
1. Surface coverage of DMA(ads) in these experiments was about 0.5 monolayer. Desorption experiments of DMA(ads) from hematite analyzed from the temporal behavior of three of the above peaks due to flowing chloride (Cl
) and hydrogen phosphate ions (HPO42‑) at pH 7 are shown in Figure 2 in ref 15. These experimental results suggested that the 775 cm
1 peak belongs to strongly adsorbed DMA, while the 793 and 840 cm
1 peaks, which reduced in intensity relatively faster, belong to more weakly DMA(ads). From the calculated v(As
O) in Table 4, we can offer a more detailed explanation of these experimental results. The 775 cm
1 peak belongs to a strongly bound innersphere complex such as Complexes B, C, or D (Table 4), while 10442

dx.doi.org/10.1021/es202749h |Environ. Sci. Technol. 2011, 45, 10438–10444

Environmental Science & Technology

ARTICLE

the 793 and 840 cm
1 peaks belong to a weakly adsorbed outersphere complex such as Complex F. Moreover, spectra collected using phosphate as desorbing agent showed the growth of peaks assigned to inner-sphere P(V)-Fe-(oxyhydr)oxide similar to those observed at pH 7 by Elzinga and Sparks29 and in our earlier control experiments.30 The temporal behavior of one of these peaks at 1045 cm
1 was shown in Figure 2 in ref 15. In these studies, adsorption of phosphate takes place due to ligand exchange with DMA(ads) and coordinated water/hydroxyl groups. To explore the thermodynamic desorption favorability of DMA(ads) by phosphate under neutral condition, ligand exchange reactions were constructed between hydrated HPO42‑ and H2PO4
clusters and all DMA-Fe-(oxyhydr)oxide complexes (Table 5). The aforementioned phosphate species are the most dominant based on the pKa values of phosphoric acid (2.2, 7.2 and 12.3).31 Figure S5 of the SI shows minimum energy equilibrium geometries of HPO42‑ 3 (H2O)4 and H2PO4
3 (H2O)4 calculated using DFT[B3LYP/6-311+G(d,p)] and the IEFPCM solvation model in Gaussian 09. The formation of a mixture of protonated and nonprotonated bidentate and monodentate P(V)-Fe-(oxyhydr)oxide complexes was reported on goethite32 and hematite29 particles depending on pH and surface coverage. In this study, we considered only the formation of protonated and nonprotonated bidentate P(V)-Fe-(oxyhydr)oxide complexes (Figure S5 of the SI) from reactions with adsorbed DMA (Table 5). For comparison, eqs 3
6 list ΔGads of hydrated phosphate with Fe-(oxyhydr)oxide clusters: Fe2 ðOHÞ6 ðOH2 Þ2 3 ðH2 OÞ6

þ HPO4 2
3 ðH2 OÞ4 f Fe2 ðOHÞ4 ðOH2 Þ4
PO4
3 ðH2 OÞ4 þ OH
3 ðH2 OÞ2 þ 3ðH2 OÞ, ΔGads ¼
37 kJ=mol

ð3Þ

Fe2 ðOHÞ5 ðOH2 Þ5 þ 3 ðH2 OÞ4

þ HPO4 2
3 ðH2 OÞ4 f Fe2 ðOHÞ4 ðOH2 Þ4
PO4
3 ðH2 OÞ4 þ 6ðH2 OÞ, ΔGads ¼
171 kJ=mol

ð4Þ

Fe2 ðOHÞ6 ðOH2 Þ2 ðH2 OÞ6 þ H2 PO4
3 ðH2 OÞ4 f Fe2 ðOHÞ4 ðOH2 Þ4
HPO4 3 ðH2 OÞ4 þ OH
3 ðH2 OÞ2 þ 3ðH2 OÞ, ΔGads ¼
28 kJ=mol

ð5Þ

Fe2 ðOHÞ5 ðOH2 Þ5 þ ðH2 OÞ4 þ H2 PO4
3 ðH2 OÞ4 f Fe2 ðOHÞ4 ðOH2 Þ4
HPO4 3 ðH2 OÞ4

þ 6ðH2 OÞ, ΔGads ¼
162 kJ=mol

ð6Þ

Calculated energy values and thermal corrections of all phosphate species are listed in Table 1. These values were used to calculated the Gibbs energy of desorption, ΔGdes, listed in Table 5 for the desorption of monodentate (reactions 19
22), bidentate (reactions 23
26), and outersphere (reactions 27
30) DMA-Fe-(oxyhydr)oxide complexes, respectively. The main findings from Table 5 are that (a) values of ΔGdes are more negative for the desorption of outersphere < monodentate < bidentate DMA-Fe-(oxyhydr)oxide complexes; and (b) The species HPO42‑ 3 (H2O)4 desorbs positively charged DMA complexes more favorably than neutral ones, irrespective of the coordination suggesting the large contribution of electrostatic attraction to the desorption process. The latter finding can also

be inferred from eq 3-6 where reaction of HPO42‑ 3 (H2O)4 with positively charged Fe-(oxyhydr)oxide clusters is thermodynamically more favorable than with neutral ones. Detailed kinetic analysis using ATR-FTIR of DMA desorption and phosphate adsorption due to flowing phosphate species is currently underway in our lab as a function of observed spectral components. Significance. Results from quantum chemical calculations reported herein are the first to quantify the thermodynamic favorability of the formation of inner- and outer-sphere complexes of DMA with Fe-(oxyhydr)oxides. Ligand exchange reactions were constructed between solvated bulk and surface reactant clusters simulating species under acidic, neutral, and basic conditions. The structure of these clusters was guided by the speciation of DMA and surface sites of Fe-(oxyhydr)oxides as a function of pH. The formation of inner- and outer-sphere surface complexes was considered in these calculations based on the mechanisms reported earlier from spectroscopic measurements.11,15 While thermodynamic binding constants are usually extracted from fitting isotherm data to empirical models, these constants represent average values and cannot be attributed to a specific ligand exchange reaction. Surface complexation models (SCM) are superior to empirical ones for modeling data collected at equilibrium because they are based on understanding the surface chemistry of oxyanions.33,34 When a triple layer SCM was applied to DMA adsorption isotherm and pH-envelope data obtained using ATR-FTIR, it only converged with either inneror outer-sphere ligand exchange reactions with the latter producing a better fit to the data.16 This has been attributed to a limitation in the SCM model. Using quantum chemical calculations, however, energies of these reactions are calculated from total energies of optimized structures of reactants and products, which have calculated stretching vibrations of As
O bonds that correlate very well with experimental values.15 Hence, relative comparison of reaction energies obtained from quantum chemical calculations on realistic model clusters provides an accurate way of quantifying the relative thermodynamic favorability of multiple reactions. Our results show that while simultaneous formation of inner- and outer-sphere DMA surface complexes is spontaneous, the formation of the former is energetically more favorable. This suggests that the adsorption and desorption kinetics of these complexes are different. These findings have implications regarding the mobility of DMA and its relative partitioning between the aqueous and solid phases containing Fe-(oxyhydr)oxides.

’ ASSOCIATED CONTENT

bS

Supporting Information. Optimized structures of reactants and products in Table 1, and their calculated geometrical parameters. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*Phone: (519)884-0710, ext. 2873; fax: (519)746-0677; e-mail: [email protected].

’ ACKNOWLEDGMENT The authors acknowledge funding from NSERC and thank the donors of the American Chemical Society Petroleum Research 10443

dx.doi.org/10.1021/es202749h |Environ. Sci. Technol. 2011, 45, 10438–10444

Environmental Science & Technology Fund for partial support of this research. We also thank Gregory Wentworth for useful discussion during group meetings.

’ REFERENCES (1) Cullen, W. R. Is Arsenic an Aphrodisiac? The Sociochemistry of an Element; RSC Publishing: Cambridge, 2008. (2) Hayakawa, T.; Kobayashi, Y.; Cui, X.; Hirano, S. A new metabolic pathway of arsenite: Arsenic
glutathione complexes are substrates for human arsenic methyltransferase Cyt19. Arch. Toxicol. 2005, 79, 183–191. (3) Cramer, S. P.; Siskin, M.; Brown, L. D.; George, G. N. Characterization of arsenic in oil shale and oil shale derivatives by X-ray absorption spectroscopy. Energy Fuels 1988, 2, 175–180. (4) Li, F.; Fan, L.-S. Clean coal conversion processes—Progress and challenges. Energy Environ. Sci. 2008, 1, 248–267. (5) Meharg, A. A.; Hartley-Whitaker, J. Arsenic uptake and metabolism in arsenic resistant and nonresistant plant species. New Phytol. 2002, 154. (6) Pelley, J. Common arsenical pesticide under scrutiny. Environ. Sci. Technol. 2005, 39, 122A–123A. (7) Reimer, K. J.; Cullen, W. R. “Arsenic concentrations in wood, soil and plant tissue collected around trees treated with monosodium methanearsonate (MSMA) for bark beetle control.” Report for the BC Ministry of Forests and Range, 2009. (8) Ponthieu, M.; Pinel-Raffaitin, P.; Le Hecho, I.; Mazeas, L.; Amouroux, D.; Donard, O.; Pontin-Gautier, M. Speciation analysis of arsenic in landfill leachate. Water Res. 2007, 41, 3177–3185. (9) Kenyon, E. M.; Hughes, M. F. A concise review of the toxicity and carcinogenicity of dimethylarsinic acid. Toxicology 2001, 160, 227–236. (10) Datta, R.; Sarkar, D.; Sharma, S.; Sand, K. Arsenic biogeochemistry and human health risk assessment in organo-arsenical pesticideapplied acidic and alkaline soils: An incubation study. Sci. Total Environ. 2006, 372, 39–48. (11) Shimizu, M.; Arai, Y.; Sparks, D. L. Multiscale assessment of methylarsenic reactivity in soil. 1. Sorption and desorption on soils. Environ. Sci. Technol. 2011, 45, 4293–4299. (12) Shimizu, M.; Arai, Y.; Sparks, D. L. Multiscale assessment of methylarsenic reactivity in soil. 2. Distribution and speciation in soil. Environ. Sci. Technol. 2011, 45, 4300–4306. (13) Jing, C.; Xiaogunag, M.; Liu, S.; Baidas, S.; Patraju, R.; Christodoulatos, C.; Korfiatis, G. P. Surface complexation of organic arsenic on nanocrystalline titanium oxide. J. Colloid Interface Sci. 2005, 290, 14–21. (14) Shimizu, M.; Ginder-Vogel, M.; Parikh, S. J.; Sparks, D. L. Molecular scale assessment of methylarsenic sorption on aluminum oxide. Environ. Sci. Technol. 2010, 44, 612–617. (15) Adamescu, A.; Mitchell, W.; Hamilton, I. P.; Al-Abadleh, H. A. Insights into the surface complexation of dimethylarsinic acid on iron (oxyhydr)oxides from ATR-FTIR studies and quantum chemical calculations. Environ. Sci. Technol. 2010, 44, 7802–7807. (16) Mitchell, W.; Goldberg, S.; Al-Abadleh, H. A. In-situ ATRFTIR and surface complexation modeling studies on the adsorption thermodynamics of mono- and di-substituted organoarsenicals on iron-(oxyhydr)oxides. J. Colloid Interface Sci. 2011, 358, 534–540. (17) Kubicki, J. D.; Kwon, K. D.; Paul, K. W.; Sparks, D. L. Surface complex structures modelled with quantum chemical calculations: carbonate, phosphate, sulphate, arsenate, and arsenite. Europ. J. Soil Sci. 2007, 58, 932–944. (18) Shared hierarchical academic research computing network (SHARCNET). http://www.sharcnet.ca (19) Tomasi, J.; Mennucci, B.; Cammi, R. Quantum mechanical continuum solvation models. Chem. Rev. 2005, 105, 2999–3093. (20) Kubicki, J. D. Self-consistent reaction field calculations of aqueous Al3+, Fe3+, and Si4+: Calculated aqueous-phase deprotonation energies correlated with experimental ln(Ka) and pKa. J. Phys. Chem. A. 2001, 105, 8756–8762. (21) Chan, B.; Den, J.; Radom, L. G4(MP2)-6x: A cost-effective improvement to G4(MP2). J. Chem. Theory Comput. 2011, 7, 112–120.

ARTICLE

(22) Boese, A. D.; Martin, J. M. L. Development of density functionals for thermochemical kinetics. J. Chem. Phys. 2004, 121, 3405–3416. (23) Bi, S.; Jin, X.; Yan, Y.; Shi, W., Density functional theory studies on the structures and water exchange reactions of aqueous Al(III)-oxalate complexes. Environ. Sci. Technol. 2011, Just Accepted (Oct 5, 2011). (24) Ochterski, J. W. Thermochemistry in Gaussian [online]. http://www.gaussian.com/g_whitepap/thermo.htm (25) Zhang, J. S.; Stanforth, R. S.; Pehkonen, S. O. Effect of replacing a hydroxyl group with a methyl group on arsenic (V) species adsorption on goethite (α-FeOOH). J. Colloid Interface Sci. 2007, 306, 16–21. (26) Catalano, J. G.; Park, C.; Fenter, P.; Zhang, Z. Simultaneous inner- and outer-sphere arsenate adsorption on corundum and hematite. Geochem. Cosmochim. Acta 2008, 72, 1986–2004. (27) Catalano, J. G.; Zhang, Z.; Park, C.; Fenter, P.; Bedzyk, M. J. Bridging arsenate surface complexes on the hematite (012) surface. Geochem. Cosmochim. Acta 2007, 71, 1883–1897. (28) Sherman, D. M.; Randall, S. R. Surface complexation of arsenic(V) to iron(III) (hydr)oxides: Structural mechanism from ab initio molecular geometries and EXAFS spectroscopy. Geochem. Cosmochim. Acta 2003, 67, 4223–4230. (29) Elzinga, E. J.; Sparks, D. L. Phosphate adsorption onto hematite: An in situ ATR-FTIR investigation of the effects of pH and loading level on the mode of phosphate surface complexation. J. Colloid Interface Sci. 2007, 308, 53–70. (30) Chabot, M.; Hoang, T. N.; Al-Abadleh, H. A. ATR-FTIR studies on the nature of surface complexes and desorption efficiency of p-arsanilic acid on iron-(oxyhydr)oxides. Environ. Sci. Technol. 2009, 43, 3142–3147. (31) Cotton, F. A.; Wilkinson, G. Advanced Inorganic Chemistry, 5th ed.; Wiley-Interscience: New York, 1988. (32) Tejedor-Tejedor, M. I.; Anderson, M. A. Protonation of phosphate on the surface of goethite as studied by CIR-FTIR and electrophoretic mobility. Langmuir 1990, 6, 602–611. (33) Goldberg, S.; Hyun, S.; Lee, L. S. Chemical modeling of arsenic(III, V) and selenium(IV, VI) adsorption by soils surrounding ash disposal facilities. Vadose Zone J. 2008, 7, 1231–1238. (34) Goldberg, S.; Johnston, C. T.; Suarez, D. L.; Lesch, S. M. Mechanism of molybdenum adsorption on soils and soil minerals evaluated using vibrational spectroscopy and surface complexation modeling. In Adsorption of Metals by Geomedia II. Variables, Mechanisms, And Model Applications; Barnett, M. O., Kent, D. B., Eds.; Elsevier Press: New York, 2008; Chapter 9, pp 235
266.

10444

dx.doi.org/10.1021/es202749h |Environ. Sci. Technol. 2011, 45, 10438–10444