SUPPLEMENTAL MATERIAL

Transcription

1 SUPPLEMENTAL MATERIAL Zerovalent Iron with High Sulfur Content Enhances the Formation of Cadmium Sulfide in Reduced Paddy Soils Yohey Hashimoto 1*, Mitsuhiro Furuya 1, Noriko Yamaguchi 2*, and Tomoyuki Makino 2 1) Tokyo University of Agriculture and Technology, Nakacho, Koganei, Tokyo, , Japan 2) National Institute for Agro-environmental Sciences Kan-nondai, Tsukuba, , Japan *CORRESPONDING AUTHORS N.Y. nyamag@affrc.go.jp Y.H. yhashim@cc.tuat.ac.jp Number of pages: 11 Number of figures: 5 Number of tables: 4 Cadmium Sulfide in Reduced Paddy Soils Soil Sci. Soc. Am. J. -Page 1-

2 Soil incubation and the soil thin-section for the μ-xrf and μ-xafs study Fig. S1 Alignment of soil, filter paper and ZVI layers on the soil thin-section for the µ-xrf and µ-xanes study Cadmium Sulfide in Reduced Paddy Soils Soil Sci. Soc. Am. J. -Page 2-

3 Eh-pH diagrams for Cd speciation in the soil solution The equilibrium reactions and equilibrium constants (log K) of Cd species were based on Lindsay (1979): Soil-Cd Cd 2+ (-7.00) Cd 2+ + S 2- CdS (27.07) SO e - + 8H + S H 2 O (20.74) Cd 2+ + CO 2 + H 2 O CdCO H + (-14.06) (a) Andisol Control Cd 2+ CdS H 2 O 25 o C, SO 4 = 10-3 M, CO 2 = atm LS-ZVI HS-ZVI Eh (mv) H 2 (b) Entisol Control LS-ZVI Cd 2+ CdS H 2 O H 2 HS-ZVI ph Fig. S2. Redox speciation of Cd in Andisol (a) and Entisol (b) with and without LS-ZVI and HS-ZVI. Black and gray plots indicate the data collected at 3 and 30-day incubation periods, respectively. Error bars represent standard deviations of mean (n = 3) Cadmium Sulfide in Reduced Paddy Soils Soil Sci. Soc. Am. J. -Page 3-

4 Soil solution SO 4 concentrations Soil solution SO 4 (mg L -1 ) Andisol bc c bc a c ab Entisol a ab ab b ab ab 0 Control LS-ZVI HS-ZVI Control LS-ZVI HS-ZVI Fig. S3. Soil solution SO 4 concentrations in Andisol and Entisol amended with LS-ZVI and HS-ZVI and without ZVI (Control). Black and gray bars represent the data collected at 3 and 30-day incubation period, respectively. Different letters on bars indicate statistically different values at p < Cadmium Sulfide in Reduced Paddy Soils Soil Sci. Soc. Am. J. -Page 4-

5 Cd K-edge EXAFS spectroscopy Table S1. Target transformation SPOIL values of selected reference spectra computed by PCA. Reference SPOIL value Cd birnessite 9.01 Cd ferrihydrite 3.48 Cd gibbsite 9.37 Cd humus 8.47 Cd kaolinite 5.38 CdCl CdCO Cd(CH 3 COO) CdS 1.25 References of the SPOIL values with underline were used for the LCF procedure. SPOIL values < 1.5 indicate an excellent fit, good, fair, poor, and > 6 are unacceptable. Table S2 The best and second-best (in parentheses) results of LCF on Cd K-edge EXAFS spectra of soil samples Soil Cd-ferrihydrite Cd-carboxyl Cd-phyllosilicate Cd-humus CdS R % Andisols Day 3 (equilibrium) Control (88) (12) LS-ZVI (92) (8) HS-ZVI (89) (11) Day 30 (reduced) Control (87) (13) LS-ZVI (96) (4) HS-ZVI (87) (13) (0.044) (0.074) (0.062) (0.043) (0.056) (0.035) Entisols Day 3 (equilibrium) Control (82) (18) LS-ZVI (86) (14) HS-ZVI Day 30 (reduced) Control 65 (93) LS-ZVI 92 (88) (12) 35 (7) (72) (28) HS-ZVI 84 (81) R: residual value for fitting; R = Σ(μ exp μ model ) 2 / Σ(μ exp ) 2 References used for the LCF procedure are selected based on the SPOIL score (Table S1) (0.037) (0.049) (0.048) (0.035) (0.072) (19) (0.071) Cadmium Sulfide in Reduced Paddy Soils Soil Sci. Soc. Am. J. -Page 5-

6 XANES-LCF on soil samples The XANES spectroscopy of Cd analysis was performed on the soil samples at the beamline BL01B1 at SPring-8 in Hyogo, Japan. The XANES data were collected in transmission mode for BN diluted samples or in fluorescent mode (synthesized references and soil samples) with 19-element Ge semiconductor detector at ambient temperature across the Cd K absorption edge at ev using Si(311) monochromator crystals. The data were collected with a stepwise in 0.25 ev with 0.4 s/point (dwell time). The data of Cd K-edge XANES were processed background correction and normalization using a software program, Athena ver (Ravel and Newville, 2005). The relative proportions of Cd species in the multi-component XANES spectra of soil samples were obtained by the liner-combination fitting (LCF) procedure using all possible binary-combinations of Cd references. The quality of LCF result was quantified a residual (R) value defined by the following equation: R = Σ(μ exp μ model ) 2 / Σ(μ exp ) 2 where μ is normalized absorption. Among all references, the binary-component fit resulting in the best and second-best R value was reported. This LCF procedure can avoid non-unique fitting solution since the use of a large number of references can inappropriately improve the R value (Hashimoto and Yamaguchi, 2013; Hashimoto et al., 2014). All LCF was performed in the energy range between and ev (i.e., relative energy between -20 and 50 ev). The XANES-LCF determined that Cd-ferrihydrite was the primary Cd species in all soils, regardless of redox states and ZVI additions. The overall trend of Cd-ferrihydrite proportion accounting for the total Cd was consistent with that of EXAFS-LCF. The Cd-humus was identified as a secondary Cd phases in most of the soils instead of Cd-carboxyl that was determined as a secondary species by the EXAFS-LCF. The structure of XANES spectra of Cd-humus and Cd-carboxyl were similar, which hindered distinguishing these species in the soil using the XANES-LCF. As well as the Cd-caroboxyl reference, Cd in humus may partially be bound with the carboxyl functional group, suggesting that Cd in these compounds possess a common binding structure. Cadmium sulfide was only found in the reduced Entisol samples with HS-ZVI (19%), which was in agreement with the result of EXAFS-LCF. Cadmium Sulfide in Reduced Paddy Soils Soil Sci. Soc. Am. J. -Page 6-

7 Cd-birnessite Cd-ferrihydrite Cd-gibbsite Normalized absorption (a. u.) Cd-kaolinite Cd-humus CdCO 3 CdCl 2 Cd(OH) 2 Cd(NO) 3 Cd(CH 3 COO) 2 CdSO 4 CdS Relative energy (ev) Energy (kev) Fig. S4. Normalized Cd K-edge X-ray absorption near edge structure spectra for the selected references. Cadmium Sulfide in Reduced Paddy Soils Soil Sci. Soc. Am. J. -Page 7-

8 Table S3 The best and second-best (in parentheses) results of LCF on Cd K-edge XANES spectra of soil samples Soil Cd-ferrihydrite Cd-humus Cd-carboxyl CdS R (10-4 ) % Andisol Day 3 (equilibrium) Control (40) (60) (6.15) LS-ZVI (22) (78) (4.95) HS-ZVI (49) (51) (4.90) Day 30 (reduced) Control (57) (43) (4.38) LS-ZVI (52) (48) (3.51) HS-ZVI (93) (7) (4.72) Entisol Day 3 (equilibrium) Control 62 (39) 38 LS-ZVI (44) HS-ZVI (50) Day 30 (reduced) Control 63 (85) (15) LS-ZVI (70) HS-ZVI 81 R: residual value for fitting; R = Σ(μ exp μ model ) 2 / Σ(μ exp ) 2 (61) (56) (50) 4.68 (6.43) 4.15 (5.97) 3.60 (4.88) (3.82) 3.20 (30) (6.34) (73) (27) (3.15) Cadmium Sulfide in Reduced Paddy Soils Soil Sci. Soc. Am. J. -Page 8-

9 EXAFS shell fitting The EXAFS shell fitting was performed using the software code Artemis (Ravel and Newville, 2005). EXAFS oscillation was extracted from the original spectrum by a spline smoothing method. The k 3 -weighted EXAFS spectra were Fourier-transformed over the k-range of Å -1 for reference spectra and Å -1 for soil samples using a Hanning window. Shell-fits were performed in r-space over the r-range 1.0 to 3.6 Å, depending on samples. Theoretical scattering paths were obtained using FEFF based on the structure of greenockite (CdS) for the Cd-S path (Graf, 1961) and otavite (CdCO 3 ) for Cd-O path (Wyckoff, 1963). A similar fitting procedure is also used by Fulda et al. (2013). Values reported include the product of coordination number (CN), interatomic distance (R): threshold E 0 shift ( E 0 ), and the Debye Waller factor (σ 2 ). The amplitude correction factors (S 2 0 ) for Cd-O and for Cd-S single-scattering paths were fixed to the value of 1. The uncertainty in the analysis was evaluated by R factor given by the following equation: R = Σ{k n χ exp (k) k n χ model (k)} 2 / Σ{k n χ exp (k)} 2 The EXAFS shell-fit on soil Cd was performed on selected samples showing the notable difference in the proportion of CdS by EXAFS-LCF. The fitted CN and R values of Cd-O, Cd-S, and Cd-Cd for CdCO 3 and CdS were in agreement with those reported in the crystallographic data (Graf, 1961; Wyckoff, 1963) and other studies (Fulda et al., 2013). The R values computed by the shell-fit for the soil Cd were not well matched with those computed for reference compounds due probably to a narrow data range (up to k~7.5 Å -1 ). The result for Cd in Andisol without ZVI at day 3 exhibited the presence of Cd-O coordination with interatomic distances of ~2.22 Å. In the HS-ZVI treated Andisol at day 30 th (reducing period), the first shell of EXAFS was better fit by Cd-O than Cd-S. This is in agreement with the result of EXAFS-LCF demonstrating that CdS was a minor compound in all Andisols, regardless of ZVI additions and redox conditions (Fig. 2, and Table S2). In contrast, the first shell of reduced HS-ZVI Entisol sample has a broader right tail than other spectra, and a small shell indicative of Cd-Cd bond for CdS. Such features represent the presence of CdS in these soils. For this spectrum, the first shell fit was performed using a binary combination of Cd-O and Cd-S (Fulda et al., 2013), yielding a better fit than the single-component fit using either Cd-O or Cd-S. This is in agreement with the result of EXAFS-LCF that determined CdS as a second major Cd compounds in the ZVI added Entisols at day 30 th. Although the results of R values for soil Cd were not directly comparable to the Cd references, the EXAFS shell fit suggests the occurrence of CdS in ZVI-added soils. Cadmium Sulfide in Reduced Paddy Soils Soil Sci. Soc. Am. J. -Page 9-

10 Table S4. Structural data on the local environment of Cd derived from the simulation of EXAFS data of soils Samples Shell CN * R (Å) E 0 (ev) σ 2 Literature values# CN R (Å) Andisol Control Day 3 (equilibrium) Cd-O (0.07) Andisol HS-ZVI Day 30 (reduced) Cd-O (0.02) Entisol HS-ZVI Day 30 (reduced) Cd-O (0.02) Cd-S (0.09) Reference CdCO 3 Cd-O (0.005) (0.01) Cd-C (0.022) (0.03) Cd-O (0.044) (0.04) Cd-Cd (0.014) (0.02) CdS Cd-S (0.01) (0.01) Cd-Cd (0.05) (0.02) CN: coordination number R: interatomic distance E 0 : threshold E 0 shift σ 2 : Debye-Waller factor # Fulda et al. (2013) Values of standard deviation for a least-squares fit are given in parentheses. * The value is fixed. EXAFS fitting curves are shown in Fig. S5. S O Fourier transform magnitude CdS CdCO 3 C O Cd Cd Andisol Ctrl Day3 Andisol HS-ZVI Day30 Entisol HS-ZVI Day R + delta R (A) o Fig. S5. Experimental (circle) and fitting (solid line) data for Fourier transformed spectra for Cd reference compounds and selected soil samples. Vertical lines are included as an eye guide. A vertical dash line at 2.8 Å indicates the edge of right tail of Cd-S. The spectrum of Cd in Entisol HS-ZVI at day 30 has a longer right tail than other soils, and a small shell indicative of Cd-Cd bond at 4.2 Å, indicating the occurrence of CdS. Cadmium Sulfide in Reduced Paddy Soils Soil Sci. Soc. Am. J. -Page 10-

11 References Fulda, B., A. Voegelin, and R. Kretzschmar Redox-Controlled Changes in Cadmium Solubility and Solid-Phase Speciation in a Paddy Soil As Affected by Reducible Sulfate and Copper. Environ. Sci. Technol. 47: Graf, D.L Crystallographic tables for the rhombohedral carbonates. Am. Mineral. 46: Hashimoto, Y., and N. Yamaguchi Chemical Speciation of Cadmium and Sulfur K-Edge XANES Spectroscopy in Flooded Paddy Soils Amended with Zerovalent Iron. Soil Sci. Soc. Am. J. 77: Hashimoto, Y., A. Takamoto, R. Kikkawa, K. Murakami, and N. Yamaguchi Formations of Hydroxyapatite and Inositol Hexakisphosphate in Poultry Litter during the Composting Period: Sequential Fractionation, P K-edge XANES and Solution 31P NMR Investigations. Environ. Sci. Technol. 48: Lindsay, W.L Chemical equilibria in soils John Wiley, New York, NY. Ravel, B., and M. Newville Athena, artemis, hephaestus: data analysis for X-ray absorption spectroscopy using IFEFFIT. J. Synchr. Radiat. 12: Wyckoff, R.W.G Crystal Structures Second edition. Interscience Publishers, New York, New York. Cadmium Sulfide in Reduced Paddy Soils Soil Sci. Soc. Am. J. -Page 11-