Investigation of Copper Speciation in Pig Slurry by a Multitechnique

Pigs assimilate very little Cu, so 80−90% Cu is excreted and found in pig slurry (7). ... Forty-five μXRF spectra were recorded for each sample wit...
1 downloads 0 Views 513KB Size
Download PDF
Environ. Sci. Technol. 2010, 44, 6926–6932

Investigation of Copper Speciation in Pig Slurry by a Multitechnique Approach SAMUEL LEGROS,† PERRINE CHAURAND,‡ ´ RÔME ROSE,‡ ARMAND MASION,‡ JE ´ RIE BRIOIS,§ JEAN-HENRY FERRASSE,| VALE ´ SAINT MACARY,⊥ HERVE JEAN-YVES BOTTERO,‡ AND E M M A N U E L D O E L S C H * ,⊥ CIRAD, UPR Recyclage et risque, F-97408 Saint-Denis, Re´union, France, CEREGE UMR 6635 CNRS/Aix-Marseille University, Europoˆle Me´diterrane´en de l’Arbois, BP 80, 13545 Aix-en-Provence Cedex 04, France, Synchrotron SOLEIL, L’Orme des Merisiers, Saint-Aubin - BP 48, Gif-sur-Yvette F-91192, MSNM-GP UMR 6181 CNRS Universite´ Paul Ce´zanne, Europoˆle Me´diterrane´en de l’Arbois Pavillon Lae¨nnec, BP 80, 13545 Aix en Provence Cedex 4, France, and CIRAD, UPR Recyclage et risque, F-34398 Montpellier, France

Received February 18, 2009. Revised manuscript received July 19, 2010. Accepted August 2, 2010.

It is now well-known that copper (Cu) can accumulate on the surface of soils upon which pig slurry has been applied. This is due to the high quantity of Cu in pig slurry resulting from its use as a growth promoter in animal feeds. The mobility and bioavailability of Cu from pig slurry spreading can be better predicted by determining the speciation of this element in addition to its total concentration. The aim of this study was to present a multitechnique approach to investigate Cu speciation in pig slurry. First, size fractionation and chemical characterization of each size fraction were performed to complement results obtained in raw samples. Micro X-ray fluorescence spectroscopy (µXRF) highlighted the colocalization of Cu and sulfur (S). Finally, X-ray absorption near-edge structure spectroscopy (XANES) showed that Cu speciation in raw pig slurry and size fractions could be described by Cu2S and that its oxidation state is Cu(I). In addition, geochemical calculation demonstrated that chalcocite (Cu2S) was the major Cu species present under pig slurry lagoon physical-chemical conditions. This Cu speciation in pig slurry may be the main reason for the observed Cu accumulation at the soil surface.

Introduction According to the Food and Agriculture Organization of the United Nations (FAO) (1), world pig production rose from around 830.6 million heads in 1997 to some 989.9 million heads in 2007, which represents a 19% increase, with a concomitant increase in the quantity of pig slurry produced. For instance, French pig breeding farms (14 million pigs * Corresponding author phone: +33 (0)442 971 775; fax: +33 442 971 559; e-mail: [email protected]; current address: CEREGE, Europole de l’Arbois, BP 80, 13545 Aix en Provence cedex 04, France. † CIRAD, UPR Recyclage et risque, Re´union, France. ‡ CEREGE UMR 6635 CNRS/Aix-Marseille University. § Synchrotron SOLEIL. | MSNM-GP UMR 6181 CNRS Universite´ Paul Ce´zanne. ⊥ CIRAD, UPR Recyclage et risque, Montpellier, France. 6926

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 18, 2010

produced) generate about 24 million t of pig slurry annually (2). Pig slurry spreading in fields is the most common way of managing this waste (3). As elements essential for plant life (i.e., nitrogen (N), phosphorus (P), and potassium (K)) are found in pig slurry, it can thus be used as a fertilizer in crop fields (4). Unfortunately, pig slurry also contains heavy metals (e.g., zinc and copper). It has been demonstrated that intensive pig slurry spreading can have impacts on field ecosystems. Heavy metals can accumulate at the soil surface (5), and high concentrations are phytotoxic. This accumulation, through the enhancement of heavy metal mobility, can therefore ultimately lead to surface water and/or groundwater contamination (6). Hence, it is essential to assess heavy metal speciation in pig slurry, i.e., determine the bearing phases, atomic coordination, and oxidation state of heavy metals, to evaluate their mobility. This is essential to determine the overall environmental impact of pig slurry spreading on agricultural soils. Among the heavy metals found in pig slurry, there is a high quantity of copper (Cu), i.e., 360-800 mg kg-1 dry matter (dm) (7, 8). Cu addition (in CuO and/or CuSO4 form) is currently authorized, even at very high concentration, in pig feeds (20-220 mg kg-1 dm), because Cu is a growth promoter. Pigs assimilate very little Cu, so 80-90% Cu is excreted and found in pig slurry (7). Pig slurry is a complex matrix composed of liquid (urine + water) and solid (feces) phases in which organic (C, N) and inorganic (N, P, K) chemical elements are present (4). To date, only a few studies (based on conventional sequential extraction protocols) have focused on Cu speciation in pig slurry, and the results are contradictory. Robel and Ross (8) detected copper sulfide, whereas Miller et al. (9) identified Cu bound to organic matter, and L’Herroux et al. (10) observed a mixture of Cu bound to organic matter along with Cu bound to Fe-Mn oxides. One could expect that these different copper species have different fates in the environment. For example, while copper sulfide (which is very insoluble in the soil environment (11) would accumulate in soil, copper bound to organic matter could be more bioavailable (12). These inconsistencies among the different studies could be attributed to sequential extraction shortcomings, including incomplete dissolution of the target phase, and/or dissolution of the nontarget species (13). In these studies, sequential extraction fractions were therefore only operationally determined through the extraction procedure (14). Studying Cu speciation in pig slurry is thus a scientific challenge that can be addressed through an in situ multitechnique approach. The unique feature of the present study is that it involved a combination of techniques to investigate Cu speciation in pig slurry. This study was conducted in two steps: characterization analyses first focused on raw pig slurry, and then, due to the great complexity of the sample, size fractionation was performed in an attempt to separate individual Cu species. Laboratory chemical microanalyses were performed using micro X-ray fluorescence spectroscopy (µXRF) to locate Cu within pig slurry and identify spatial correlations between elements. The results were compared with those obtained through geochemical calculations of the pig slurry lagoon geochemical system. These calculations were performed to determine the major Cu species in pig slurry. Finally, more detailed insight into Cu speciation was obtained using synchrotron-based X-ray absorption spectroscopy (XAS). XAS is one of the best known structural in situ techniques for direct determination of the speciation of chemical elements in complex matrices. This technique includes X-ray absorp10.1021/es101651w

 2010 American Chemical Society

Published on Web 08/24/2010

tion near-edge structure spectroscopy (XANES), which provides global information on the oxidation state, threedimensional geometry, and coordination environment of elements under investigation.

Experimental Section Sampling, Sample Preparation, and Chemical Analysis. Pig slurry (PS) was sampled during the pig fattening stage. This specific PS was chosen because Cu supplementation in pig feed varies with the pig maturity stage and reaches a maximum during this stage. The raw pig slurry (PS raw) sample was laid-out in a fine layer and dried at 60 °C for 24 h. The size fractionation procedure is described in detail in the Supporting Information (SI-1). Briefly, eight size fractions were obtained: PS sup 1000 (with particle size superior to 1000 µm), PS 630-1000 (with particle size ranging from 630 to 1000 µm), PS 355-630, PS 200-355, PS 50-200, PS 20-50, PS 0.45-20, and PS inf0.45 fraction. All samples were laid-out in a fine layer and dried at 60 °C for 24 h. The coarsest fractions and PS raw were then comminuted (100 µm) before analysis. During the sampling stage and size fractionation procedure, it should be noted that no specific care was taken concerning sample oxidation. Total carbon and total nitrogen were determined by dry combustion with an elemental NC 2100 Analyzer (Thermo Electron Corp.). For major and trace element analyses in each sample, a representative subsample was digested using a mixture of HF, HNO3, and HClO4 (ISO 14869-1). Major and trace element concentrations were then determined with an inductively coupled plasma-optical emission spectrometer (ICP-OES Vista-PRO, Varian, Inc.) with an axially viewed plasma system and a charge coupled device detector. The measurement uncertainty was less than 10%. Laboratory-Based µXRF Coupled with Chemometrics. The measurements were carried out on a HORIBA XGT-5000 microscope equipped with an X-ray tube producing a finely focused and high-intensity beam with a 10-µm spot size. The X-ray beam was generated with an Rh X-ray tube at an accelerating voltage of 30 kV, with a current of 1 mA. X-ray emission from the irradiated sample was detected with an energy-dispersive X-ray (EDX) spectrometer equipped with a liquid nitrogen cooled high-purity Si detector. Elements detected by µXRF ranged from Mg (12) to U (92). The detector resolution was 145 eV at the Mn KR emission line. µXRF analysis detection limits can reach 100 mg kg-1, depending of the nature of the matrix and the atomic weight of the analyzed element. The sample to be analyzed was finely ground and then pressed into a 13-mm-diameter pellet. Forty-five µXRF spectra were recorded for each sample with the incident beam randomly pointed on 45 spots on the pellet surface (total counting time of 1000 s per spectrum). To analyze these spectra properly, a self-modeling mixture analysis technique, called SIMPLISMA (SIMPLe-to-use Interactive Self-modeling Mixture Analysis) was used (15). The spectrum recorded by µXRF is actually a “mixture spectrum” as the incident X-ray beam penetrates through the sample and the collected fluorescence signal is coming from all the phases contained in the probed volume. Thus it could be difficult to clearly establish spatial correlations between elements (trace and major elements). The SIMPLISMA algorithm decomposes the data matrix (i.e., the 45 pig slurry µXRF spectra) into pure contributions due to each of the components in the system, and it can elucidate and characterize components in a mixture (16). As SIMPLISMA is discussed in detail in ref 15, we will only give a brief description here. The algorithm assumes that there is a wavelength with significant intensity contributions from only one of the components (pure component) in the mixture. This wavelength is called the pure variable. If the pure variable

can be found for every component in a mixture, then the intensity at these wavelengths can be used to resolve the corresponding spectra through a least-squares fit. Geochemical Calculations with the CHESS Code. Geochemical calculations were performed using the CHESS code (CHemical Equilibrium of Species and Surfaces) (17) to predict phases that predominate in the pH and Eh range of the studied system (pig slurry). The Pourbaix diagram (or activity diagram) of Cu was thus built. The aim was to simply describe the equilibrium state of the pig slurry lagoon system. The fixed activity of Cu and other components in raw pig slurry were used as input data (see SI-2). The thermodynamic formation constants of the aqueous species and minerals were taken from the eq 3/6 thermodynamic database (18, version 8.6). This standard database was adapted according to the studied system. The formation constant of struvite (MgPO4NH4 · 6H2O) and newberyte (MgHPO4) (two minerals observed in pig slurry by XRD) were added, respectively, logKs ) -13.15 and LogKs ) -5.8 (19). In addition, the stability constant of Cu(II) complexes of humic acids (CuHAs) isolated from pig slurry (20) was added (logK ) 6.20). X-ray Near-Edge Structure Spectroscopy (XANES) at the Cu K Edge. Cu K edge XANES spectra were collected on the SAMBA beamline at the SOLEIL storage ring (France) run at 2.75 GeV with a current of 200 mA. Due to beam time limitation, only PS raw and four size fractions, PS 0.45-20, PS 20-50, PS 50-200, and PS 355-630, were analyzed. These size fractions were chosen because they represent the majority of the copper mass in pig slurry. Spectra acquisition was performed at liquid nitrogen temperature to improve the signal/noise ratio by decreasing the thermal motion of atoms and to limit beam damage. Measurements were carried out in fluorescence mode with a seven-element solid-state germanium detector. Each spectrum is the sum of three recordings ranging from 150 eV below to 400 eV above the absorption edge of Cu at 8979 eV. XANES spectra were acquired from 8959 to 9029 eV with a 0.2 eV monochromator step (i.e., from 20 eV below to 50 eV above the absorption edge) and an integration time of 4.0 s per point. The photon energy of the experimental spectra was calibrated using a Cu foil (threshold energy was taken at the zero-crossing point of the second derivative of the spectrum). The data reduction was accomplished using Athena software (21). To quantitatively compare the intensity of absorption features in various compounds, all experimental Cu K edge XANES spectra were reduced by background subtraction with a linear function in the pre-edge region and with a quadratic polynomial function in the post-edge region, and normalized using the point of inflection of the first EXAFS oscillations as a unit. Quantitative application of XANES to geochemical systems requires determination of the proportions of multiple chemical species that contribute to the measured spectrum. In this study, we used a combination of principal component analysis (PCA), target transformation (TT), and linear combination fitting (LCF) to fit Cu K edge XANES spectra of pig slurry samples. The goodness of fit was assessed with the normalized sum-square (NSS) equation. PCA and TT were done with the SixPack software and LCF was done with Athena software. This procedure has already been used to fit XANES spectra that contain multiple chemical species (22, 23). Indeed, the XANES spectrum of the unknown sample is a weighted sum of all species spectra present; the atomic fraction of each metal species can be obtained by a linear combination fits (LCF) of this spectrum to reference spectra. For further details about this procedure, its prerequisite, and its pitfalls see SI-7, SI-8, and SI-9, and refs 22 and 23. VOL. 44, NO. 18, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6927

TABLE 1. Element Distribution in Size Fractions and Raw Pig Slurry (Concentrations Are Expressed as g kg-1 of Dry Matter) dry matter fraction PS PS PS PS PS PS PS PS

inf 0.45 0.45-20 20-50 50-200 200-355 355-630 630-1000 sup 1000

sum PS raw recovery

g 1.31 2.14 0.33 0.57 0.54 1.09 1.02 1.13

Cu -1

%

g kg

16 26 4.1 7.0 6.6 13 13 14

0.01 1.54 0.06 0.08 0.07 0.08 0.06 0.07

8.13 6.68 122%

1.97 1.82 108%

S -1

%

g kg

0.3 78 3.3 4.1 3.4 4.3 3.1 3.4

1.34 2.64 0.14 0.27 0.26 0.35 0.28 0.27 5.55 4.93 113%

Ca %

g kg

24 48 2.6 4.9 4.7 6.4 4.9 4.9

0.43 8.31 1.5 3.41 3.06 4.52 2.71 1.45 25.4 20.7 123%

XANES spectra analysis is based on comparisons with several reference compounds with a well-known crystal structure. Reference compounds are chosen from a large panel of potential chemical species that could contribute to explaining the coordination, oxidation state, and atomic environment of Cu within pig slurry. Cu K edge XANES spectra for 13 reference compounds (listed in SI-6) were recorded. Inorganic and organic compounds with various Cu oxidation states (Cu(0), Cu(I), and Cu(II)) were in this database. The purity of the mineral reference compounds was checked by XRD. These spectra were all subjected to accurate energy calibration and carefully normalized.

Results and Discussion Chemical Analysis of Raw Pig Slurry and Size Fractions. The raw pig slurry (PS raw) characteristics (pH ) 7.7, Eh ) -150 ( 10 mV, and dry matter ) 8.37%) were close to those reported in the literature (e.g., 24-27) The raw pig slurry also had high P, Mg, K, Cl, and Ca concentrations (Table 1). This could be correlated with the presence of several minerals that we have previously detected in this pig slurry by XRD (28). Indeed, the high P and Mg, K and Cl, and Ca concentrations could be attributed to the presence of struvite (MgPO4NH4 · 6H2O), newberyte (MgHPO4), sylvite (KCl), and calcite (CaCO3). According to the literature, the Cu concentration in pig slurry ranges from 0.4 to 0.8 g kg-1 dm (7, 29). This Cu content concerns pig slurry generally sampled in pig slurry lagoons, which consists of a mixture of pig slurries produced at all pig breeding stages. In the present study, only one of these stages (the fattening stage) was selected because Cu supplements in pig feed were maximum. Thus the pig slurry sampled for this study had a very high Cu concentration (1.82 g kg-1 dm). Each size fraction was also analyzed (Table 1). The Cu content of each size fraction led to recoveries of 108%. Only 0.3% of total Cu was in PS inf0.45, which was potentially the most mobile size fraction. All size fractions with a particle size larger than 20 µm contained roughly 3-4% of total Cu. The majority of the Cu was detected in the PS 0.45-20 fraction, with 78% of total Cu. This is consistent with the findings of Masse et al. (30), who detected 80% Cu in the 0.45-10 µm particle size range. Otherwise, the largest Cu concentrated fraction (PS 0.45-20) also contained the greatest proportion of Fe (51% of total Fe), S (48% of total S), Ca (33% of total Ca), and P (28% of total P). There could be a relationship between these elements and Cu. Localization of Cu in Pig Slurry. Raw pig slurry was analyzed by µXRF to localize Cu and then determine the nature of the potential Cu-bearing phase(s). The initial data set (45 µXRF spectra recorded with a 10µm X-ray beam) was resolved by the SIMPLISMA procedure 6928

9

-1

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 18, 2010

K -1

%

g kg

1.7 33 5.9 13 12 18 11 5.7

36.2 1.95 0.35 0.58 0.59 0.98 1.06 1.21 42.9 31.3 137%

P -1

%

g kg

84 4.5 0.8 1.3 1.3 2.3 2.5 2.8

4.67 8.81 4.28 4.83 2.26 3.06 1.89 1.3 31.1 44 71%

Fe -1

%

g kg

15 28 13 15 7.3 10 6.1 4.2

0.23 2.76 0.4 0.78 0.47 0.36 0.21 0.19 5.4 4.36 124%

Ctot -1

%

g kg

4.2 51 7.4 14 8.6 6.7 3.8 3.6

25 106 7.85 17.8 23.2 49.2 50.7 55.3

% 7.4 32 2.3 5.3 6.9 15 15 17

335 355 94%

into pure components and their contributions in the original spectra. Figure 1 shows the seven resolved spectra obtained (C1-C7). The sum of their contributions led to 98.4% recovery. These resolved spectra showed spatial correlations among elements observed in pig slurry. The main mineral phases present in raw pig slurry (identified by XRD, 28) were also clearly identified here through these chemical correlations. The C3, C4, and C5 spectra could be associated with the presence of calcite (CaCO3), quartz (SiO2), and sylvite (KCl), respectively. Correlating P with Mn (C6) could reveal Mn adsorption on struvite and/or newberyte. Indeed, phosphate can adsorb heavy metals (31, 32). C1 and C2 spectra were associated with Fe-rich and Zn-rich phases, respectively. The most interesting finding concerns the resolved C7 spectrum, which showed colocalization of Cu with S, suggesting the presence of an S-rich Cu-bearing phase. Geochemical Calculations with the CHESS Code. The geochemical calculations enabled us to simulate the balance of the pig slurry system. The speciation calculations were performed at thermodynamic equilibrium and satisfactory results were obtained (see SI-3 for further detail). Because µXRF analyses suggest the presence of an S-rich Cu-bearing phase, only thermodynamic data for Cu-minerals containing S were considered. First, minerals which could precipitate were identified by calculating saturation indices. The saturation indices obtained were consistent with the SIMPLISMA results presented previously in the text and with earlier XRD measurements (28) (see further details in the Supporting Information SI-4). Then, the Pourbaix diagram of Cu was calculated (Figure 2). Figure 2 shows a number of fields where a specific Cu species dominates. In the Eh/pH conditions of PS raw, this diagram shows that chalcocite (Cu2S) with reduced Cu(I) predominates. It should be noted that CuHAs (which stands for Cu(II) bound to humic acid) does not appear in the diagram. Indeed, this is a minor species and masked by other more predominant species (see SI-5 for more details). The predominance of chalcocite (Cu(I)2S) in the pH and Eh range of the pig slurry system is consistent with the Cu correlated with S obtained by µXRF. Nevertheless, Robel and Ross (8) (on the basis of chemical extraction findings) showed the presence of cupric sulfide in pig slurry. Although this is also consistent with the Cu correlated with S, in cupric sulfide, Cu is present in its oxidation state Cu(II). At this point, some doubt could remain on the oxidation state. In addition, a Pourbaix diagram is useful for a first estimate, but the simplifying assumptions of an activity diagram with respect to a full speciation calculation should be kept in mind. Indeed, the Pourbaix diagram predicts the predominance of solids in a pH range, and several other species coexist at quite

FIGURE 1. Seven spectra resolved with the SIMPLISMA procedure (analysis of 45 µXRF spectra in raw pig slurry).

FIGURE 2. Pourbaix diagram of Cu in pig slurry, with the dotted line focusing on the Eh/pH conditions of the raw pig slurry (PS raw) system (pH ) 7.67, Eh ) -150 mV). significant concentrations. Therefore the presence of some secondary species cannot be totally excluded. XANES Cu Speciation on the Molecular Scale. Further insight into the symmetry and oxidation state of Cu was obtained by XANES. XANES spectra analysis is based on comparisons with several reference compounds of wellknown crystal structure. Figure 3 presents the normalized Cu K edge XANES spectra of the reference compounds. Differences were noted among the near-edge spectra (Figure 3a). Inflections in the absorption edge (A, B, C, etc.) corresponded to peaks in the first derivatives (Figure 3b). Some of these inflections provided information on the Cu oxidation state. Indeed, Feature A, from 8977 to 8978 eV, was the pre-edge feature and corresponded to dipoleforbidden electronic transitions 1s f 3d (but hybridized by p orbitals of the ligands). As Cu(0) and Cu(I) did not have

empty 3d initial states, the pre-edge was a signature of Cu(II). The intensity of this feature was weak (Figure 3c) because there was only one empty 3d orbital (out of 10 available) to accept the excited 1s electron. All Cu(II)-bearing reference compounds (reference compounds 1-8) showed pre-edges (feature A) of comparable intensity. Feature B, at 8982 eV, corresponds to the 1s f 4p transitions for Cu(I) compounds (reference compounds 10-12). No Cu(II) reference had a maximum pre-edge between 8980 and 8985 eV. Consequently, the presence of a shoulder between 8980 and 8985 eV in the Cu absorption edge spectrum and the absence of a pre-edge (feature A) from 8977 to 8978 eV indicated the presence of Cu(I) in the sample (33). Some other inflections provided information on the 3D geometry and coordination environment of Cu. Indeed, Features C (8988 eV) and D (8995 eV) correspond respectively VOL. 44, NO. 18, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6929

FIGURE 3. (a) Normalized Cu K edge XANES spectra of reference compounds, (b) corresponding first derivatives of Cu K edge XANES spectra, and (c) zoom on feature A of (a). to the 1s f 4p and 1s f continuum transitions for Cu(II) compounds (reference compounds 3-8) in octahedral symmetry (33). Features C’ (8986 eV), C’’ (8992 eV), and D’ (8997 eV) correspond to 1s f 4p and 1s f continuum transitions for Cu(II) compounds in square planar symmetry (reference compounds 1 and 2) (34). Figure 4 presents the normalized Cu K edge XANES spectra of the pig slurry samples (PS raw and four size fractions) and two selected reference compounds. In all samples, the presence of a shoulder at 8982 eV (feature B) and the absence of a pre-edge from 8977 to 8978 eV (feature A) indicated the presence of Cu(I) in the sample. PS raw and PS 0.45-20 showed a Cu K edge XANES pattern very close to the 6930

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 18, 2010

chalcocite spectrum. However, on the spectra of PS 20-50, PS 50-200, and PS 355-630, a shoulder (C’’) was observed. The peak C’’ resulted from the 1s f 4p transitions in Cu(II) compounds in square planar symmetry like Cu(II)-acetate. The LCF fits were satisfactory for all samples, with very low NSS values (Table 2). All LCF fits showed a recovery of over 100% (see Table SI-9 and Figure SI-9 for non adjusted data), so in Table 2, they were adjusted to 100% to compare the proportions. All size fraction samples were fitted by a high proportion of Cu(I)2S (from 72 to 90%) and a smaller proportion of Cu(II)-acetate (10-28%). The sum of the linear combination fit results for size fractions weighted by the mass of each size fraction reached 91% Cu(I)2S and 9% Cu(II)-

FIGURE 4. (a) Normalized Cu K edge XANES spectra of pig slurry samples and (b) corresponding first derivatives of Cu-K edge XANES spectra.

TABLE 2. Linear Combination Fit Results for Size Fractions and Raw Pig Slurry Cu(I)2S (%) PS PS PS PS PS

355-630 50-200 20-50 0.45-20 raw

a

82 89 72 90 96

Cu(II)-acetate (%) 18 11 28 10 4

a Partial XANES fit percentages. adjusted to reach 100%.

b

Σ (%)

NSS

b

0.05 0.08 0.04 0.04 0.04

100 100 100 100 100

The proportions were

acetate. Taking the uncertainty of the LCF method of 10-15% into account (34), this is consistent with Cu speciation in the raw pig slurry samples (Cu(I)2S ) 96% and Cu(II)-acetate ) 4%). Those percentage values are partial XANES fit percentages which can be highly nonlinear and compounds of nonequivalent valencies may significantly differ in their absorption coefficients (23). Therefore, those percentages are not fully quantitative and must be taken as estimation of the species proportion. The proportion of Cu(II)-acetate (4%) in the raw pig slurry was very low and could be questionable, but it was relatively high in some size fractions (e.g., PS 20-50). The use of Cu(II)acetate in the calculation simulates the presence of Cu bound to carboxyl sites in organic compounds. Therefore, there was only a minority of Cu(II) bound to organic matter in the pig slurry. The presence of Cu(I)2S in great majority in the pig slurry spectra LCF fit was attributed to chalcocite (Cu(I)2S) precipitation. This result was surprising for two reasons. First, the oxidation state of Cu in pig slurry (i.e., oxidation state I) was different from that of the Cu added to pig feed which was in the form of CuSO4 or CuO (oxidation state II). However, the physical-chemical conditions (Eh, pH) within pig slurry lagoons lead to a Cu(I) oxidation state (5), as indicated by the geochemical calculation. Second, the inorganic state of

Cu was also surprising, because the pig slurry was almost completely made up of fresh organic materials. Indeed, it could be argued that the species Cu(I)2S also resulted from complexation with thiol (-SH) functional groups in the organic matter. Nevertheless, this hypothesis was also investigated with the reference Cu(I)-thiophenolate. This reference gave poor results with the LCF. This is why the hypothesis of chalcocite was retained here. Besides, the precipitation of chalcocite was consistent with the µXRF analyses and the geochemical calculations. In pig slurry, Cu is mainly in sulfide form (Cu2S) with a Cu(I) oxidation state. As copper sulfide precipitates in anoxic conditions (36), questions arise concerning the solubility of this mineral phase at the soil surface after pig slurry spreading. Calmano et al. (37) studied the solubilization of metal sulfides. They demonstrated that pH is the key factor even though the redox potential (Eh) influences the solubility of copper sulfides. Indeed, they demonstrated that copper sulfides were soluble in oxidized sediment only at pH < 4.5. Similarly, AlAbed et al. (38) compared the leaching behavior of metals in mineral processing waste via short-term extraction tests. In these experiments, they investigated whether the pH, redox potential (Eh), particle size, and contact time were critical variables in controlling metal stability. They demonstrated that, although cationic heavy metals such as Cu were extracted at acidic pH (pH ) 3), the redox potential (Eh) did not significantly affect the leaching of Cu. Finally, Zhou et al. (11) studied the solubility of copper sulfides in soil under aerobic conditions but in the 6.6-7.6 pH range. They demonstrated that copper sulfide solubility was very low in these conditions. Therefore, we put forward the hypothesis that the solubility of copper sulfide from pig slurry would be very low in soil, even in aerobic conditions, when the soil pH is over 4.5. Cu accumulation at the soil surface resulting from intensive pig slurry spreading (5) could thus be explained by the very low solubility of Cu2S. VOL. 44, NO. 18, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

6931

Acknowledgments This work was financially supported by ADEME under project 0475C0013, the European Community (FEOGA), and the French Re´union Region. We thank B. Devouard who provided us with Cu-rich minerals.

Supporting Information Available Size fractionation procedure (SI-1), data input for calculation with CHESS (SI-2), final equilibrium calculated with CHESS (SI-3), saturation indices calculated with CHESS (SI-4), the Eh/ph domain of predominance of Cu bound to humic acid (SI-5), reference compounds database for XANES (SI-6), PCA analysis for XANES (SI-7), target transformation for XANES (SI-8), and XANES LCF fitting: nonadjusted results and fit evaluation (SI-9). This material is available free of charge via Internet at http://pubs.acs.org

Literature Cited (1) Food and Agriculture Organization of the United Nations. FAO statistical yearbook, 2007; http://faostat.fao.org/ (2) Pauwels, P.; Le Jeannic, F.; De Corlieu, T.; Nanot, B.; CrosnierDavid, M.; Gre´goire, P.; Carme`s, L.; Roudaut, J. P.; Jordan-Meille, L.; Berger, S.; Tre´goue¨t, B.; Nirascou, F.; Belin, M. Les de´chets de l’agriculture en France: essai de quantification; Institut franc¸ais de l’environnement (IFEN): Paris, 2005. (3) Peu, P.; Birgand, F.; Martinez, J. Long term fate of slurry derived nitrogen in soil: A case study with a macro-lysimeter experiment having received high loads of pig slurry (Solepur). Bioresour. Technol. 2007, 98 (17), 3228–3234. (4) Sanchez, M.; Gonzalez, J. L. The fertilizer value of pig slurry. 1. Values depending on the type of operation. Bioresour. Technol. 2005, 96 (10), 1117–1123. (5) Novak, J. M.; Watts, D. W.; Stone, K. C. Copper and zinc accumulation, profile distribution, and crop removal in coastal plain soils receiving long-term, intensive applications of swine manure. Trans. ASAE 2004, 47 (5), 1513–1522. (6) Xue, H.; Phan hong nhat, W.; Sigg, L. Effects of Soil Composition on Zn Speciation in Drainage Waters from Agricultural Soils. Aquat. Geochem. 2005, 11 (3), 303–318. (7) Jondreville, C.; Revy, P. S.; Jaffrezic, A.; Dourmand, J. Y. Le cuivre dans l’alimentation du porc: oligoe´le´ment essentiel, facteur de croissance et risque potentiel pour l’Homme et l’environnement. INRA Prod. Anim. 2002, 15 (4), 247–265. (8) Robel, E. J.; Ross, W. C. Simplified Spectrophotometric Analysis of Copper from Cupric Sulfide Synthesized in Porcine Fecal Matter. J. Agric. Food Chem. 1975, 23 (5), 973–974. (9) Miller, W. P.; Martens, D. C.; Zelazny, L. W.; Kornegay, E. T. Forms of solids phase copper in copper-enriched swine manure. J. Environ. Qual. 1986, 15 (1), 69–72. (10) L’Herroux, L.; Le Roux, S.; Appriou, P.; Martinez, J. Behavior of metals following intensive pig slurry applications to a natural field treatment process in Britany (France). Environ. Pollut. 1997, 97 (1-2), 119–130. (11) Zhou, W.; Hesterberg, D.; Hansen, P. D. H.; Hutchison, K. J.; Sayers, D. E. Stability of copper sulfide in a contaminated soil. J. Synchrotron Radiat. 1999, 6 (3), 630–632. (12) Brandt, K. K.; Holm, P. E.; Nybroe, O. Evidence for bioavailable copper-dissolved organic matter complexes and transiently increased copper bioavailability in manure-amended soils as determined by bioluminescent bacterial biosensors. Environ. Sci. Technol. 2008, 42 (8), 3102–3108. (13) Doelsch, E.; Moussard, G.; Saint-Macary, H. Fractionation of tropical soilborne heavy metals--Comparison of two sequential extraction procedures. Geoderma 2008, 143 (1-2), 168–179. (14) Scheinost, A. C.; Kretzschmar, R.; Pfister, S.; Roberts, D. R. Combining Selective Sequential Extractions, X-ray Absorption Spectroscopy, and Principal Component Analysis for Quantitative Zinc Speciation in Soil. Environ. Sci. Technol. 2002, 36 (23), 5021–5028. (15) Windig, W.; Guilment, J. Interactive self-modeling mixture analysis. Anal. Chem. 1991, 63 (14), 1425–1432. (16) Windig, W.; Gallagher, N. B.; Shaver, J. M.; Wise, B. M. A new approach for interactive self-modeling mixture analysis. Chemometr. Intell. Lab. 2005, 77 (1-2), 85–96. (17) van der Lee, J. Thermodynamic and mathematical concepts of CHESS; CIG-e´cole des Mines de Paris: Fontainebleau, France,

6932

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 18, 2010

1998. (18) Wolery, T. E. Q3/6: A software Package for Geochemical Modelling of Aqueous Systems: Package Overview and Installation Guide; Lawrence Livermore National Laboratory report: Livermore, CA, 1992. (19) Bhuiyan, M. I.H.; Mavinic, D. S.; Koch, F. A. Thermal decomposition of struvite and its phase transition. Chemosphere 2008, 70 (8), 1347–1356. (20) Plaza, C.; Senesi, N.; Garcia-Gil, J. C.; Polo, A. Copper(II) complexation by humic and fulvic acids from pig slurry and amended and non-amended soils. Chemosphere 2005, 61 (5), 711–716. (21) Ravel, B.; Newville, M. ATHENA, ARTHEMIS, HEPHAESTUS: data analysis for X-ray absorption spectroscopy using iFEFFIT. J. Synchrotron Radiat. 2005, 12 (4), 537–541. (22) Beauchemin, S.; Hesterberg, D.; Beauchemin, M. Principal Component Analysis Approach for Modelling Sulfur K-XANES Spectra of Humic Acids. Soil Sci. Soc. Am. J. 2002, 66 (1), 83–91. (23) Ressler, T.; Wong, J.; Roos, J.; Smith, I. L. Quantitative speciation of Mn-bearing particulates emitted from autos burning (methylcyclopentadienyl)manganese tricarbonyl-added gasolines using XANES spectroscopy. Environ. Sci. Technol. 2000, 34 (6), 950–958. (24) Moral, R.; Perez-Murcia, M. D.; Perez-Espinosa, A.; MorenoCaselles, J.; Paredes, C. Estimation of nutrient values of pig slurries in Southeast Spain using easily determined properties. Waste Manage. 2005, 25 (7), 719–725. (25) Hernandez, D.; Plaza, C.; Senesi, N.; Polo, A. Detection of Copper(II) and zinc(II) binding to humic acids from pig slurry and amended soils by fluorescence spectroscopy. Environ. Pollut. 2006, 143 (2), 212–220. (26) De la Torre, A. I.; Jimenez, J. A.; Carballo, M.; Fernandez, C.; Roset, J.; Munoz, M. J. Ecotoxicological evaluation of pig slurry. Chemosphere 2000, 41 (10), 1629–1635. (27) Saviozzi, A.; Levi-Minzi, R.; Riffaldi, R.; Vanni, G. Laboratory studies on the application of wheat straw and pig slurry to soil and the resulting environmental implications. Agric. Ecosyst. Environ. 1997, 61 (1), 35–43. (28) Legros, S.; Doelsch, E.; Masion, A.; Rose, J.; Borschneck, D.; Proux, O.; Hazemann, J.-L.; Saint-Macary, H.; Bottero, J.-Y. Combining Size Fractionation, Scanning Electron Microscopy, and X-ray Absorption Spectroscopy to Probe Zinc Speciation in Pig Slurry. J. Environ. Qual. 2010, 39 (2), 531–540. (29) Nicholson, F. A.; Chambers, B. J.; Williams, J. R.; Unwin, R. J. Heavy metal contents of livestock feeds and animal manures in England and Wales. Bioresour. Technol. 1999, 70 (1), 23–31. (30) Masse, L.; Masse´, D. I.; Beaudette, V.; Muir, M. Size distribution and composition in raw and anaerobically digested swine manure. Trans. ASAE 2005, 48 (5), 1943–1949. (31) Mohan, S.; Sreelakshmi, G. Fixed bed column study for heavy metal removal using phosphate treated rice husk. J. Hazard. Mater. 2008, 153 (1-2), 75–82. (32) Ramessur, R. T.; Parry, S. J.; Ramjeawon, T. The relationship of dissolved Pb to some dissolved trace metals (Al, Cr, Mn, and Zn) and to dissolved nitrate and phosphate in a freshwater aquatic system in Mauritius. Environ. Int. 2001, 26 (4), 223–230. (33) Kau, L. S.; Spira-Solomon, D. J.; Penner-Hahn, J. E.; Hodgson, K. O.; Solomon, E. I. X-ray absorption edge determination of the oxidation state and coordination number of copper. Application to the type 3 site in Rhus vernicifera laccase and its reaction with oxygen. J. Am. Chem. Soc. 1987, 109 (21), 6433– 6442. (34) Manceau, A.; Lanson, B.; Schlegel, M. L.; Harge´, J. C.; Musso, M.; Eybert-Be´rard, L.; Hazemann, J. L.; Chateigner, D.; Lamble, G. M. Quantitative Zn speciation in smelter-contaminated soils by EXAFS spectroscopy. Am. J. Sci. 2000, 300 (4), 289–343. (35) Pourbaix, M. Atlas d’e´quilibres e´lectrochimiques; Gauthier-Villars: Paris, 1963. (36) Zhang, L. Technical note electrochemical equilibrium diagrams for sulphidization of oxide copper minerals. Miner. Eng. 1994, 7 (7), 927–932. (37) Calmano, W.; Hong, J.; Fo¨rstner, U. Binding and mobilization of heavy metals in contaminated sediments affected by pH and redox potential. Water Sci. Technol. 1993, 28 (8-9), 233–235. (38) Al-Abed, S. R.; Hageman, P. L.; Jegadeesan, G.; Madhavan, N.; Allen, D. Comparative evaluation of short-term leach tests for heavy metal release from mineral processing waste. Sci. Total Environ. 2006, 364 (1-3), 14–23.

ES101651W