Supporting Information to 'Bisulfide reaction with natural organic matter enhances arsenite sorption: Insights from X-ray absorption spectroscopy' Martin Hoffmann, Christian Mikutta* and Ruben Kretzschmar Institute of Biogeochemistry and Pollutant Dynamics, Department of Environmental Systems Science, ETH Zurich, CH-8092 Zurich, Switzerland (14 Pages, 7 Tables, 9 Figures) Table of contents 1. Evaluation of S K-edge XANES spectra of reference compounds... S2 2. Properties of peat and HA... S3 3. Reaction of peat with S(-II): ph dependency... S7 4. Deconvolution of S K-edge XANES spectra... S8 5. Normalized and first derivatives of normalized As K-edge XANES spectra of S(-II)-reacted peat, HA, and reference compounds... S10 6. Shell fits of As K-edge EXAFS spectra of reference compounds... S11 7. Alternative As K-edge EXAFS shell fits of peat and HA samples... S13 8. References... S14 * Corresponding author s e-mail: christian.mikutta@env.ethz.ch Phone: +41-44-6336024; Fax: +41-44-6331118
1. Evaluation of S K-edge XANES spectra of reference compounds Table S1: Electronic oxidation states, white-line energies, and sources of S reference compounds. Compound Electronic oxidation state White-line energy [ev] Source Iron monosulfide -2 2471.1 Commercial (Fluka) Pyrite -1 2472.1 Natural (Paul Scherrer Institute) Elemental S ±0 2472.8 Commercial (Agrikulturchemie) Cystine +0.2 2473.0 Commercial (Merck) Cysteine +0.5 2473.7 Commercial (Merck) Methionine +0.5 2473.6 Commercial (Sigma-Aldrich) Diphenyl sulfoxide +2 2476.4 Commercial (Merck) Cysteic acid +5 2481.2 Commercial (Sigma-Aldrich) Sodium sulfate +6 2482.8 Commercial (Merck) Figure S1: (A) White-line energy and (B) white-line area/mass unit of S reference compounds with different electronic oxidation states: Iron monosulfide (-2), pyrite (-1), elemental S (±0), cystine (+0.2), cysteine (+0.5), methionine (+0.5), diphenyl sulfoxide (+2), cysteic acid (+5), and sodium sulfate (+6). The regression lines are shown as thick solid lines, the 95% confidence bands as dashed lines, and the 95% prediction bands as dotted lines. S2
2. Properties of peat and HA Table S2: Elemental composition of peat and IHSS Elliott soil HA (nd = not determined). [wt.%] Peat HA C 53 58 a N 1.4 4.1 a S 0.12 0.44 a Fe 0.10 0.9 b Si 0.13 nd Na <0.03 nd Mg 0.01 nd Al 0.10 nd P 0.02 0.24 a Cl 0.02 nd K 0.02 nd Ca 0.39 nd Cr <0.001 nd Mn 0.001 nd Co <0.001 nd Ni 0.006 nd Cu 0.002 nd Zn 0.005 nd As <0.001 nd Cd <0.001 nd Pb 0.007 nd molar C/N 44 16 a values provided by IHSS; b ref. 1 S3
Figure S2: Microscopy images of the peat. The laser scanning microscopy images of the peat were taken on a color 3D VK-9710K violet laser scanning microscope (Keyence). The peat is a heterogeneous mixture of organic fibers, mainly plant debris with a low degree of degradation. Particles are in the expected size range of 40-250 μm, but the majority of the particles have a size of less than 100 μm. S4
Figure S3: FT-IR spectra of (A) the peat and (B) IHSS Elliott soil HA. The broad intense band at about 3400-3300 cm -1 results from stretching vibrations of hydroxyl (OH) groups and/or the N-H stretching of secondary amines. Weaker bands at ~2930 and ~2865 cm -1 can be assigned to antisymmetric and symmetric stretching of aliphatic C-H bonds. Stretching vibrations of the C=O group of carboxylic acids or other carbonyl groups lead to the band at 1720 cm -1. The intense band at 1630 cm -1 in combination with that at 1520 cm -1 is an indicator of skeletal vibrations of aromatics. The absorption band at 1630 cm -1 may be intensified by asymmetric C-O stretching vibrations of carboxylates as indicated by a band at 1450-1410 cm -1, which results from a combination of symmetric C-O stretching vibrations and C-H deformations. Alcoholic functionalities are verified by the bands at 1275-1255 cm -1 (phenolic groups) and 1070 cm -1 (aliphatic OH groups). An additional band at ~2600 cm -1 in the Elliott soil HA can be ascribed to stretching vibrations of either thiol or tertiary ammonium groups. Vibrational bands reflecting carbonyl, carboxyl, and aromatic moieties are more pronounced in Elliott soil HA, whereas vibrations of aliphatic structures are dominating in the peat. S5
Figure S4: 13 C CP/MAS-NMR spectrum of the peat. Chemical shifts are given relative to trimethylsilane. The integrated peak intensities are summarized in Table S3. Integration limits are shown as vertical lines. Table S3: 13 C CP/MAS-NMR estimates of the C distribution in peat and IHSS Elliott soil HA. Carbonyl Carboxyl Aromatic Acetal Heteroaliphatic Aliphatic 220-190 ppm 190-165 ppm 165-110 ppm 110-90 ppm 90-60 ppm 0-60 ppm Peat 1 4 11 7 32 45 HA a 6 18 50 4 6 16 a values provided by IHSS 2 S6
3. Reaction of peat with S(-II): ph dependency Figure S5: Time-resolved reaction of peat with S(-II) (10 mm S(-II)/mol C, 0.2 M C) at ph 5, ph 7, and ph 9 (± 0.1) in 30 mm NaCl solution. Error bars indicate the range of sorption (n = 2). Dotted lines serve as a guide for the eye. Equilibrium of S(-II) reaction with peat was attained after approximately 24 hours. S7
4. Deconvolution of S K-edge XANES spectra Figure S6: Sulfur K-edge XANES spectra of peat samples containing (a) 0.84, (b) 1.8, (c) 3.8, (d) 6.2, (e) 6.8, (f) 7.4, and (g) 8.1 mmol S/mol C, and of IHSS Elliott soil HA samples containing (h) 2.5 and (i) 12 mmol S/mol C. Gray lines represent experimental data, thick dotted lines show the overall fits, black lines denote the fitted PseudoVoigt peaks, and thin dotted lines the fitted arctangent functions. The results of the S speciation analysis are given in Table S4. S8
Table S4: Sulfur speciation in the peat and IHSS Elliott soil HA samples. total S [mmol/mol C] reduced S a [%] intermediate oxidized S b [%] oxidized S c [%] S oxidation state -II to +I >+I to +III >+III to +VI Peat HA 0.84 d 70 1 29 1.8 74 <1 26 3.8 70 0 30 6.2 72 0 28 6.8 70 0 30 7.4 68 0 32 8.1 60 0 40 2.5 d 34 3 63 12 55 0 45 a energy-range: 2,475 ev; b energy-range: 2,475 to 2,478 ev; c energy-range: 2,478 ev; d untreated samples S9
5. Normalized and first derivatives of normalized As K-edge XANES spectra of S(-II)-reacted peat, HA, and reference compounds Figure S7: (A) Normalized and (B) first derivatives of normalized As K-edge XANES spectra of (a) Na 2 HAsO 4 7H 2 O, (b) NaAsO 2, (c) arsenopyrite, and peat samples containing (d) 3.8, (e) 6.2, (f) 6.8, (g) 7.4, and (h) 8.1 mmol S/mol C. Energy values indicate the maximum of the first derivatives. Figure S8: (A) Normalized and (B) first derivatives of normalized As K-edge XANES spectra of (a) Na 2 HAsO 4 7H 2 O, (b) NaAsO 2, (c) arsenopyrite, and IHSS Elliott soil HA samples containing (d) 11, (e) 12, (f) 12, and (g) 15 mmol S/mol C. Energy values indicate the maximum of the first derivatives. S10
6. Shell fits of As K-edge EXAFS spectra of reference compounds Figure S9: (A) Arsenic K-edge EXAFS spectra as well as (B) magnitude and (C) real part of the Fourier transform of (a) sodium(meta)arsenite, (b) arsenolite, (c) tris(phenyl)thioarsine, (d) realgar, and (e) orpiment. The Fourier transforms were calculated over a k-range of 2-12.5 Å -1. Shell fits were performed in R-space (fit k-weight = 3) over an R + ΔR-range of 1.0-2.4 Å (sodium(meta)arsenite, arsenolite), 1.0-4.0 Å (tris(phenyl)thioarsine), and 1.0-3.0 Å (realgar, orpiment). Gray lines represent experimental data and dotted lines show the fits of the first (a, b, e) or first and second coordination shells (c, d) of As. The fitted EXAFS parameters are summarized in Table S5. S11
Table S5: EXAFS parameters determined by shell fitting of As K-edge EXAFS spectra of reference compounds. Parameter uncertainties are given in parentheses for the last significant figure. Parameters in bold were fixed to their theoretical values. Compound Path CN a σ 2 [Å 2 ] b R [Å] c S 0 2d ΔE 0 [ev] e R-factor f Redχ 2g Sodium(meta)arsenite As-O 3.0 0.006 (2) 1.81 (2) 0.83 (18) 6.0 (33) 0.05 130 Arsenolite As-O 3.0 0.002 (1) 1.79 (1) 0.84 (15) 4.1 (30) 0.03 770 Tris(phenyl)thioarsine Realgar As-S 3.0 0.003 (1) 2.26 (0) As-C 3.0 0.011 (6) 3.01 (4) As-S 2.0 0.004 (2) 2.25 (2) As-As 1.0 0.006 (2) 2.58 (2) 1.00 (7) 0.0 (10) 0.01 238 0.97 (24) -1.2 (34) 0.09 488 Orpiment As-S 3.0 0.004 (1) 2.28 (1) 0.96 (18) -2.5 (25) 0.06 500 a coordination number (path degeneracy); b Debye-Waller parameter; c mean half path length; d passive amplitude reduction factor; e energy-shift parameter; f, g R-factor = data i-fit i 2 i and reduced χ 2 = N idp data 2 i-fit i N i data N pts ε idp -N var -1 i, i where N idp, N pts, and N var are, respectively, the number of independent points in the model fit, the total number of data points, the number of variables in the fit, and ε i is the uncertainty of the i th data point. 3 N idp /N var = 9/4 (sodium(meta)arsenite, arsenolite), 20/6 (tris(phenyl)thioarsine), 13/6 (realgar), and 13/4 (orpiment). S12
7. Alternative As K-edge EXAFS shell fits of peat and HA samples Table S6: EXAFS parameters determined by shell fitting of As K-edge EXAFS spectra of S(-II)-reacted peat equilibrated with As(III). In these R-space fits (fit range: 1.0-4.0 Å, fit k-weight = 3) we constrained the As-C path degeneracy to that of the respective As-S path. Parameter uncertainties are given in parentheses for the last significant figure. The passive amplitude reduction factor, S 0 2, was set to 0.92 for all samples. Parameters in bold were fixed in the fits. Sulfur [mmol/mol C] a Arsenic [μmol/mol C] b Path CN c σ 2 [Å 2 ] d R [Å] e ΔE 0 [ev] f R-factor g Redχ 2h 3.8 19 6.2 46 6.8 107 7.4 105 8.1 156 As-O 2.2 (1) 0.004 i 1.79 (1) As-S 0.4 (1) 0.003 j 2.31 (2) As-C 0.4 k 0.011 l 3.20 (32) As-O 1.8 (1) 0.004 i 1.79 (1) As-S 0.8 (1) 0.003 j 2.27 (1) As-C 0.8 k 0.011 l 3.17 (13) As-O 1.6 (1) 0.004 i 1.79 (1) As-S 0.9 (1) 0.003 j 2.26 (1) As-C 0.9 k 0.011 l 3.17 (13) As-O 1.5 (1) 0.004 i 1.79 (1) As-S 1.1 (1) 0.003 j 2.25 (1) As-C 1.1 k 0.011 l 3.15 (8) As-O 1.1 (1) 0.004 i 1.81 (1) As-S 1.2 (1) 0.003 j 2.25 (1) As-C 1.2 k 0.011 l 3.16 (7) 2.9 (19) 0.03 19 2.3 (15) 0.03 15 1.6 (16) 0.04 7.5 1.1 (12) 0.02 8.6 1.9 (13) 0.03 15 a conversion factor to [g/kg]: 1.415; b conversion factor to [mg/kg]: 3.306; c coordination number (path degeneracy); d Debye-Waller parameter; e mean half path length; f energy-shift parameter; g, h R-factor = data i-fit i 2 i and reduced χ 2 = N idp data 2 i-fit i N i data N pts ε idp -N var -1 i, where N idp, N pts, and N var are, respectively, i the number of independent points in the model fit, the total number of data points, the number of variables in the fit, and ε i is the uncertainty of the i th data point. 3 N idp /N var = 17/6; i constrained to the mean σ 2 of sodium(meta)arsenite and arsenolite; j constrained to σ 2 of tris(phenyl)thioarsine; k set to CN(As-S) and covaried; l constrained to σ 2 of tris(phenyl)thioarsine. S13
Table S7: EXAFS parameters determined by shell fitting of As K-edge EXAFS spectra of S(-II)-reacted HA equilibrated with As(III) (fit range: 1.0-3.0 Å, fit k-weight = 3). Parameter uncertainties are given in parentheses for the last significant figure. The passive amplitude reduction factor, S 0 2, was set to 0.92 for all samples. Parameters in bold were fixed in the fits. Sulfur [mmol/mol C] a Arsenic [μmol/mol C] b Path CN c σ 2 [Å 2 ] d R [Å] e ΔE 0 [ev] f R-factor g Redχ 2h 11 38 12 53 15 58 As-O 3.4 (3) 0.004 i 1.79 (1) As-S 0.5 (2) 0.004 j 2.36 (3) As-O 3.0 (2) 0.004 i 1.79 (1) As-S 0.5 (2) 0.004 j 2.35 (3) As-O 2.0 (2) 0.004 i 1.78 (1) As-S 1.4 (2) 0.004 j 2.29 (1) 1.7 (29) 0.03 6.0 2.7 (26) 0.02 12 0.8 (21) 0.03 5.1 a conversion factor to [g/kg]: 1.549 (HA); b conversion factor to [mg/kg]: 3.618 (HA); c coordination number (path degeneracy); d Debye-Waller parameter; e mean half path length; f energy-shift parameter; g, h R-factor = data i-fit i 2 i data i and reduced χ 2 = N idp data i-fit i N i pts ε i 2 N idp -N var -1. N idp, N pts, and N var are, respectively, the number of independent points in the model fit, the total number of data points, the number of variables in the fit, and ε i is the uncertainty of the i th data point. 3 N idp /N var = 10/5; i constrained to the mean σ 2 of sodium(meta)arsenite and arsenolite; j constrained to the mean σ 2 of tris(phenyl)thioarsine, realgar, and orpiment. 8. References 1. He, Z. Q.; Ohno, T.; Cade-Menun, B. J.; Erich, M. S.; Honeycutt, C. W. Spectral and chemical characterization of phosphates associated with humic substances. Soil Sci. Soc. Am. J. 2006, 70, 1741-1751. 2. Characterization of the International Humic Substances Society Standard and Reference Fulvic and Humic Acids by Solution State Carbon-13 ( 13 C) and Hydrogen-1 ( 1 H) Nuclear Magnetic Resonance Spectrometry; Water-Resources Investigations Report 89-4196; U.S. Geological Survey: Denver, 1989. 3. Stern, E. A.; Newville, M.; Ravel, B.; Yacoby, Y.; Haskel, D. The UWXAFS analysis package: Philosophy and details. Physica B 1995, 208, 117-120. S14