Environ. Sci. Technol. 2005, 39, 3048-3055
Extended X-ray Absorption Fine Structure Spectroscopy Evidence for the Complexation of Cadmium by Reduced Sulfur Groups in Natural Organic Matter TORBJO ¨ R N K A R L S S O N , * ,† PER PERSSON,‡ AND ULF SKYLLBERG† Department of Forest Ecology, Swedish University of Agricultural Sciences, S-901 83 Umeå, and Inorganic Chemistry, Department of Chemistry, Umeå University, S-901 87 Umeå, Sweden
It is widely accepted that the bioavailability, toxicity, and mobility of trace metals are highly dependent on complexation reactions with functional groups in natural organic matter (NOM). In this study, the coordination chemistry of Cd in NOM was investigated by extended X-ray absorption fine structure spectroscopy. Soil organic matter (SOM) from different types of organic soils and dissolved organic matter (DOM) from an organic and a mineral soil horizon of a Spodosol and aquatic DOM from Suwannee River were investigated. In SOM samples (1000-25000 µg of Cd g-1, pH 4.6-6.6), Cd was coordinated by 1.0-2.5 S atoms at a distance of 2.49-2.55 Å and by 3.0-4.5 O/N atoms at a distance of 2.22-2.25 Å. In DOM samples (1750-4250 µg of Cd g-1, pH 5.4-6.3), Cd was coordinated by 0.3-1.8 S atoms at a distance of 2.51-2.56 Å and 3.6-4.5 O/N atoms at a distance of 2.23-2.26 Å. In both SOM and DOM samples a second coordination shell of 1.7-6.0 carbon atoms was found at an average distance of 3.12 Å. This is direct evidence for inner-sphere complexation of Cd by functional groups in NOM. Furthermore, ion activity measurements showed that less than 1% of total Cd was in the form of free Cd2+ in our samples. Bond distances and coordination numbers suggest that Cd complexed in SOM and DOM is a mixture of a 4-coordination with S (thiols) and 4- and 6-coordinations with O/N ligands. Given that Cd-S associations on average are stronger than Cd-O/N associations, our results strongly indicate that reduced S ligands are involved in the complexation of Cd by NOM also at native concentrations of metal in oxidized organicrich soils and in humic streams.
Introduction Cadmium is highly toxic to humans, animals, and plants (1). It has properties partly similar to those of Zn, which is an essential micronutrient, is easily taken up by plants, and bioaccumulates in the food web. The toxicity, and in humans carcinogenic effect, is believed to be caused by competition with Zn for metal-binding sites in proteins (2). Due to * Corresponding author phone: +46 (0) 090-786 86 38; fax: +46 (0) 090-786 81 63; e-mail:
[email protected]. † Swedish University of Agricultural Sciences. ‡ Umeå University. 3048
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 9, 2005
increased anthropogenic emissions the concentration of Cd in the humus layer of Swedish forest soils has increased considerably during the 20th century (3), and in agricultural soils a widespread use of phosphate fertilizers has led to elevated concentrations in crops. The worldwide average Cd concentration in soils is 0.06-1.1 µg g-1 (1), while phosphate fertilizers made from phosphate rock of sedimentary origin can have Cd concentrations as high as 75 µg g-1 (4). The total concentration of Cd in soil or water is in most cases poorly related to biouptake by organisms. Instead the specific chemical forms of the metal, i.e., the chemical speciation of Cd, controls bioavailability and toxicity. In general, soluble Cd species are considered to be bioavailable, and the free metal ion (Cd2+) and its soluble inorganic complexes are assumed to be the most toxic and bioavailable forms (5). The role of dissolved organic matter (DOM) is ambiguous, and studies have shown DOM to both decrease and increase the toxicity and bioavailability of Cd to aquatic organisms (6). Of special importance for the speciation are chemical interactions with surfaces. It has been shown that Cd binds to functional groups of natural organic matter (NOM), clay minerals, and iron, manganese, and aluminum oxyhydroxides (7-10). In contrast to the inorganic chemistry of Cd, which is fairly well-known, associations of Cd with NOM and its functional groups are poorly described, and there are only a few spectroscopic studies on the coordination chemistry of Cd in NOM. Otto et al. (11) and Li et al. (12) used 113Cd NMR to study the complexation of Cd to fulvic acids (FAs) and aquatic NOM from Suwannee River, respectively. Both studies reported that Cd was primarily coordinated by O, and partly by N, containing functional groups. Relatively high concentrations of Cd in relation to organic C were applied in these experiments, and therefore, possible contribution from S groups may have been difficult to detect, given an approximate ratio of 1:100 between S and O sites in NOM. Another powerful technique, synchrotron-based X-ray absorption spectroscopy (XAS), has been used to obtain information on the local chemical environment of trace metals in a variety of environmental materials, including NOM. Using extended X-ray absorption fine structure (EXAFS) spectroscopy, Xia et al. (13, 14) reported Co, Ni, and Cu to form inner-sphere complexes in octahedral geometry involving one, two, and four O/N (carboxyl/amine) functional groups, respectively, in humic substances (Suwannee River FA, humic acid (HA), and a soil humic substance) at 90% metal saturation of the total binding capacity. In the same study Zn was reported to form inner-sphere complexes with two S groups in an octahedral geometry. Other EXAFS studies have shown that Hg and CH3Hg bind to reduced organic S groups in NOM in a linear 2-coordination (15, 16). To our knowledge there are only two published EXAFS studies on the binding of Cd to organic matter in soils and waters. In one of the studies, Liu et al. (17) reported Cd to be coordinated by six O atoms in a soil humic acid. Data were, however, not modeled in k-space, and therefore, the contribution from other ligands such as sulfur cannot with certainty be ruled out. In another study, Collins et al. (8) reported 7-8 O atoms to be involved in the complexation of Cd by sodium humate. In this study the Cd loading to samples was large enough to more than saturate possible strong ligands such as thiols. In a combined XANES (X-ray absorption near edge structure spectroscopy) and EDX (energy-dispersive X-ray analysis) study, Martinez et al. (18) found that low solubility of Cd in a cultivated peat soil, despite high total Cd concentrations, was caused by associations of Cd to reduced organic or 10.1021/es048585a CCC: $30.25
2005 American Chemical Society Published on Web 03/19/2005
inorganic S. From studies of plants and bacteria, Salt et al. (19) reported Cd to be coordinated by 6 O atoms in xylem sap and by 4 S atoms in root tissue of Indian mustard (Brassica juncea L.), Ku ¨ pper et al. (20) reported 0.3-3.1 S atoms to be involved together with O/N ligands in a 6-coordination of Cd in the tissue of the Cd/Zn hyperaccumulator Thlaspi caerulescens, and Boyanov et al. (21) reported Cd to be coordinated by 6 O (phosphoryl and carboxyl groups) atoms in Bacillus subtilis cell walls. On the basis of these results, and given that Zn, Cd, and Hg are group IIB metals (chalcophiles), it is plausible that reduced organic S (Org-Sred) groups are involved in the complexation of Cd in NOM, especially if the ratio of Cd to Org-Sred is kept to a minimum, to avoid complete saturation of strong complexing sites. In this paper we present the first EXAFS results for the coordination chemistry of Cd in NOM from organic soils and in DOM from soils and a stream at low Cd concentrations (down to 1000 µg g-1), corresponding to Cd/Org-Sred molar ratios of 2% and above.
Material and Methods Sampling, Sample Preparation, and Chemical Analysis. Soil organic matter (SOM) samples were collected at four different sites. A subalpine fen peat (SFP) dominated by Carex spp. was sampled at Ifjord, northern Norway, situated within 5 km from the Atlantic Ocean (70°5′N, 27°1′E). A boreal forest peat soil (BFP), covered by Picea abies and Sphagnum and Polytricum mosses, and a Spodosol (22) with a histic organic (O) horizon (POH) and an organic-rich spodic (Bhs, illuvial) horizon (PBH) covered by P. abies, Vaccinium shrubs, and Hylocomnium splendens and Pleurozium schreberi mosses were sampled at Svartberget Research station, Vindeln, Sweden (64°14′N, 19°46′E). Finally a cultivated Phragmites fen peat (AFP) was sampled at Majnegården, Falko¨ping, in southwest Sweden (58°0′N, 14°5′E). All samples were treated following protocols for a clean sampling procedure. Soil samples were sealed in double plastic bags and stored at 4 °C until freeze-dried (Edwards Modulyo 4K freeze-dryer) and homogenized by a tungsten carbide ball mill (Retsch, S2, Germany). Three different types of DOM were also used, Suwannee River NOM (processed by the International Humic Substance Society, IHSS) and DOM extracts from the POH sample (designated DOH) and from the PBH sample (DBH). Suwannee River NOM was collected from the Suwannee River near Fargo, GA, and concentrated by reverse osmosis. The DOM extracts from the Spodosol were obtained using a modified method of Adams and Byrne (23). A mass of 50100 g of soil and 6-18 g of technical grade ion-exchange resin in Na form (Chelex 20, Bio-Rad) were weighed into 250 mL centrifuge bottles (Nalgene), and 75-150 mL of Milli-Q water was added. The soil and ion-exchange resin were left to equilibrate for 1-16 h. After equilibration the bottles were centrifuged at 5000 rpm (Beckman J2-21M/E model, JA.14 rotor) for 5 min, and the supernatant was collected by decantation. Additional amounts of Milli-Q water (75-150 mL) were added to the mixture of soil and ion-exchange resin, and the extraction procedure was repeated 4-5 times for the DOH sample and 2-3 times for the DBH sample. The sum of all DOM extracts was finally separated from soil particles by a prewashed filter paper (Munktell 3, pore size >10 µm, STORA, Sweden). The pH was determined to be 6.5 and 6.6 in the DOH and DBH extracts, respectively. Finally the DOM samples were freeze-dried (Edwards Modulyo 4K freeze-dryer) and stored in darkness. Total sulfur in SOM and DOM samples was determined on a LECO sulfur analyzer (LECO Corp., Michigan). In agreement with Xia et al. (24) and Skyllberg et al. (25) OrgSred was determined as the sum of sulfur species showing absorption peak maxima in the energy range 2472-2474 eV,
TABLE 1. Chemical Composition of the Different Types of SOM and DOM Samples Used in This Study: Organic Carbon (Org-C), Total Sulfur (Stot), and Reduced Organic Sulfur (Org-Sred) NOM type
[Org-C] (g kg-1)
[Stot] (g kg-1)
[Org-Sred] (% of Stot)
SFPa BFPa AFPa POHa PBHa DOHb DBHb SRNb
410 493 427 565 7.0 431 329 525
20 4.1 8.8 3.0 0.98 3.0 ndf 6.5
74.5c 61c 61.5d 62.5e 50e 26e nd 28e
a SOM. b DOM. c Determined by XANES (reported as fen peat and organic soil, respectively, by Qian et al. (16). d Determined by XANES. e Determined by XPS. f nd ) not determined.
TABLE 2. Chemical Composition of SOM and DOM Samples Prepared for EXAFS Analysis and Measurement of Cd Ion Activity: Reduced Organic Sulfur (Org-Sred), Organic Carbon (Org-C), and Total Cd (Cdtot) sample
[Cd] (µg g-1)
pH
SFP.1 SFP.2 SFP.3 BFP.1 BFP.2 BFP.3 AFP POH DOH.1 DOH.2 DBH SRN
1000 10000 25000 1000 4500 11000 2000 1000 1750 4250 4250 2500
5.5 5.5 5.5 6.1 6.1 6.1 4.6 6.6 6.3 6.3 6.1 5.4
a
[Cd/Org-Sred] (mol/mol)
[Cd/Org-C] (mol/mol)
% free Cd2+ of Cdtot
0.02 0.19 0.48 0.11 0.51 1.26 0.11 0.15 0.64 1.55
0.0003 0.0026 0.0065 0.0002 0.001 0.0024 0.0005 0.0002 0.0004 0.0011 0.0014 0.0005
