You are on page 1of 14

SEQUENTIALLY EXTRACTED ARSENIC FROM DIFFERENT SIZE FRACTIONS OF CONTAMINATED SOILS

E. LOMBI1∗ , R. S. SLETTEN2 and W. W. WENZEL1 1 Institute of Soil Science, Universität für Bodenkultur, Vienna, Austria; 2 Quaternary Research
Center, University of Washington, Seattle, U.S.A. (∗ author for correspondence, e-mail: enzo.lombi@bbsrc.ac.uk; present address: Department of Soil Science, IACR-Rothamsted, Harpenden, AL5 2JQ, Herts, U.K.)

(Received 28 May 1998; accepted 28 December 1999)

Abstract. Five soils from Austria that have high contents of arsenic (As) due to anthropogenic and/or geogenic sources were separated into sand, silt, and clay fractions. The distribution of As among the particle-size fractions was clay >> silt > sand. The particle-size fractions were analysed according to an 8-step sequential extraction procedure (SEP) designed to assess the primary reservoirs and extractability of As. Steps 1 and 2, (i.e. NH4 NO3 and (NH4 )2 SO4 ) extracted the least As, while considerably greater amounts were extracted in steps 3, 4 and 5 (i.e. NH4 H2 PO4 , NH2 OH-HCl, and NH4 F). The greatest amounts of As were extracted in steps 6 and 7, both NH4 -oxalate extractants targeting the amorphous and crystalline oxide fractions. The residual fraction (step 8) was typically low. The extraction of the clays contained generally more As in steps 3 to 8, while As release in the first two steps was typically highest in the sand and silt fractions. These findings are in accord with factor analysis indicating that the extractants of the SEP are separated into weak (steps 1 and 2) and strong (steps 3 to 8) extractants. SEP data along with microprobe analyses (X-ray microanalyses) indicate that most As is associated with Fe and is therefore not readily mobile. Anyhow, As could be released upon changes in redox potential or pH. The greater extractability of As from sand fractions using weak extractants may explain the higher toxicity of As in coarse-textured soils, as found in previous studies. Keywords: arsenic, sequential extraction, soil size fractions

1. Introduction Arsenic has been long recognized to be a toxic element but recent greater awareness of its toxicity at very low concentrations has renewed our public welfare concern. Primary natural sources of this element in soils are parent materials containing arsenic sulphide minerals and input due to volcanic activity. Important anthropogenic sources are mining and smelter operations, coal burning, and arsenical pesticides. Natural versus anthropogenic sources of arsenic have been estimated at 60:40 (Buat-Menard et al., 1987). The average content of arsenic in uncontaminated soil is 5–6 mg kg−1 (Adriano, 1986). A series of studies have reported relations between mobility and total content of As in soil and soil texture. In soil, arsenic mobility and toxicity to plants are strongly influenced by Fe and Al oxides/hydroxides and clay minerals (Galba,
Water, Air, and Soil Pollution 124: 319–332, 2000. © 2000 Kluwer Academic Publishers. Printed in the Netherlands.

Size limits for the fractions were sand (2000–63 µm). The effectiveness of the clay-silt separation was assessed using a Micrometrics Sedigraph 5000ET. with total As concentrations varying between 240 and 1900 mg kg−1 (Table I). 1994).1.. or enriched in silicon and aluminum. As distribution in sand. M INERALOGICAL AND CHEMICAL COMPOSITION The mineralogical composition of sand and silt fractions was assessed by X-ray diffraction (Philips PW 1130). 1983). 1977. Silt and clay fractions were isolated using a centrifuge technique. These soils have been contaminated with arsenic from various geogenic and anthropogenic sources. Shen et al. One hundred gram of bulk soil samples were dispersed by ultrasonic vibration in water. Methods 3. 1983). (1973) found that arsenic was less toxic in fine-textured soils.320 E.. Many soil properties are related to surface phenomena occurring at liquid–solid interface. in soils with a large content of sand. in regions with soils derived from the same parent material. 1972. and clay and its implication to As mobility has not been comprehensively investigated. Therefore. The separation procedure was repeated 5 times to minimize the amount of clay-size particles remaining in the silt fraction. 1975). A positive correlation between clay and arsenic content in the soil has been reported by various authors (Yang. Jacobs and Keeny (1970) and Woolson et al. and clay) of the soil. In contrast. 1980). the fate of arsenic could be affected by the particle-size distribution. The samples were air-dried and passed through a 2-mm sieve. silt (63–2 µm) and clay (<2 µm). Clay mineralogy was determined using oriented . 3. To our knowledge. Because the chemical and physical characteristics of soils are strongly related to texture. soil texture may be the dominant soil feature to affect the background levels of arsenic in soil (Huang. LOMBI ET AL. Leckie et al. All fractions were oven-dried at 40 ◦ C. 1975. Holm et al. the surface area of particle-size fractions increases in the order sand « silt « clay. Shen et al. Frost and Griffin. 2. silt. Particle size separation was performed according to Genrich and Bremner (1974). As concentrations are typically small (Koyama.. 1979. 1983. silt. Materials Representative bulk soil samples were collected from the B horizon of five Cambisols in Austria (Table I). The sand fraction was separated by wet sieving. Differential contents of oxides and clay minerals in different particle-size fractions suggest that the aresenic distribution also varies among soil texture fractions for a given soil (Wauchope.. The primary aim of this article is to assess the distribution and the major sink of arsenic among various particle size fractions (sand.

and Si as described above. Mn.SEQUENTIALLY EXTRACTED ARSENIC FROM CONTAMINATED SOILS 321 TABLE I Soil characteristics and soil classification after FAO (1988) Sample Sample location and soil type Mitterberg (Salzburg) Eutric Cambisol Mitterberg (Salsburg) Eutric Cambisol Lungau (Salsburg) Dystric Cambisol Brixlegg (Tirol) Dystric Cambisol Kliening (Carinthia) Dystric Cambisol Lavantal (Carinthia) Calcaric Cambisol Horizon pH (CaCl2 ) 5. and Al. Perkin Elmer). Samples (500 mg) of each size fraction were digested HNO3 -H2 O2 in a microwave (MLS Mega 2400). the carbonate content was calculated by difference. 1996). The extractant:soil ratio was 25 mL:1 g. . 3. Fe..62 6. Total carbon in each particle-size fraction was measured using a Carlo Erba CNS Analyzer NA 1500 Series 2. old mining deposit Grassland/geogenic mounts prepared by aspirating a suspension of the clay-size fraction through a porous ceramic plate. Mn and Si were measured using inductive-coupled-plasma emission-spectrometry (ICP Plasmaquent 100. The digest was analyzed for total arsenic by atomic absorption spectrometry (AAS 2100 Perkin Elmer) equipped with a hydride generation system (FIAS 400.2. a semiquantitative estimate of their amount was made using the method of Johns et al. Clay minerals were identified according to Riedmüller (1978). Organic carbon (OC) was measured using the same instrument after pre-treating the sample with 0.30 Total As (mg kg−1 ) 1520 655 214 581 335 1990 Note 951002 951003 960057 960065 960144 960149 Bw1 Bw2 Bw2 B1 Bw Bw Pasture/located near an abandoned smelter near abandoned smelter Forest/geogenic Spruce forest/active smelter Pasture. Fe.20 7.98 4. K. Mg. Extracts from each step were analyzed for As. Na. S EQUENTIAL FRACTIONATION OF ARSENIC A sequential extraction procedure was employed to chemically fractionate arsenic in each of the particle-size fractions (Table II).39 3. Dissolved organic carbon (DOC) in the extracts from steps 2 to 5 was estimated by UV absorption at 254 nm (Brandstetter et al.2 M HCl to remove carbonates. Zeiss). (1954). Ca. Al.67 3. Ca.

0 30 min shaking 2 wash step: NH4 OAc 1 M – 10 min shaking NH4 -oxalate 0. S TATISTICAL ANALYSIS Basic statistical parameters and multivariate methods (i.2 for Windows . 0.0 – 1 hr shaking NH2 OH-HCl 0. TABLE II Sequential soil extraction procedure utilised for arsenic fractionation Step 1 2 3 4 5 Treatment NH4 NO3 1 M – 24 hr shaking (NH4 )2 SO4 0. M ICROPROBE ANALYSIS An ARL SEMQ Electron Microprobe/Scanning Electron Microscope equipment with 6 wavelength dispersive X-ray spectrophotometers (WDS) and a modified Kevex/Oxford/ANS Quantum energy dispersive X-ray spectrophotometer system (EDS) was utilized for imaging and elementeal detection of the soil samples. The SEM/microprobe operating conditions were 25 kV accelerating voltage and 50 nA beam current.4. Saeki and Matsumoto. 1994.2 M – 10 min shaking NH4 -oxalate 0.e factor analysis) were utilised to interpret the data from the sequential extraction procedure using Statistica version 4. LOMBI ET AL.3. EDS spectra were analyzed with software from ANS Quantum (America Nuclear Systems.001 M – 24 hr shaking (NH4 )H2 PO4 0.5 M pH 7.322 E. Knoxville. 3. 1957.1 M pH 3. More detailed analyses were done by EDS and WDS detectors that were set to detect K-α and L-α X-rays of: As.1 M + NH4 OAc 1 M – pH 6. c: Chang and Jackson.2 M – 10 min shaking HNO3 – H2 O2 (microwave digestion) Fraction Exchangeable Specifically sorbed I Specifically sorbed II Al – associated Mn – associated Reference a b b c a 6 Amorphous Fe bound a 7 Crystalline Fe bound a 8 Residual a: Zeien and Brümmer. . Factor analysis was performed to reduce the number of variables and to detect structure in relationships between variables. Cu. Fe. and S.2 M pH 3.25 – 4 hr shaking in the dark wash step: NH4 -oxalate 0.2 M + Ascorbic ac.25 – 30 min shaking in water bath at 96 ◦ C wash step:NH4 -oxalate 0. 3. TN). b: mod. 1989. Zn.05 M – 24 hr shaking NH4 F 0. This method enables to display most of the original variability in a smaller number of dimensions. Mg.

0 43.6 59.3 51.9 125.0 0.0 0.2 53.6 50.3 563.0 449.0 0.8 7. (1998) have found similar amounts of clay-sized particles in the silt fractions.7 As (mg kg−1 ) 559 2002 4935 317 800 2431 70 170 1288 814 495 1774 250 529 1196 818 2367 4400 As (% of total in fine earth) 37.3 51.0 31. Using a similar technique of separation.7 58.0 0.0 0.4 9.0 56.5 760.90 17. C (g kg−1 ) 0.5 430.5 951002 951003 960057 960065 960144 960149 4.5 285. PARTICLE .0 0. Stemmer et al.SIZE SEPARATION Bulk soil samples contain considerable amounts of sand and silt (Table III).0 0.05%) and ultrasonic vibration prior to analysis.9 16. Results and Discussion 4.6 Inorg.5 5228.4 33.4 19.2 Org.0 0. C (g kg−1 ) 5.9 6.6 371.2 31.3 27.3 5. .8 179.0 0.2 305. C (% of total in fine earth) 40.3 352.0 18. probably largely released upon treatment of the samples with Calgon solution (0.50 25.SEQUENTIALLY EXTRACTED ARSENIC FROM CONTAMINATED SOILS 323 TABLE III Total carbon an arsenic in the soil-size fractions Sample Fraction (g kg−1 of fine earth) sand silt clay sand silt clay sand silt clay sand silt clay sand silt clay sand silt clay 698.1 Org.9 2.4 34.0 7.1 8.3 67.7 17.4 62.5 16.0 0.0 0.2 61.8 16.5 23.2 12.1.9 119.9 14.7 35.2 59.7 23.5 55.2 65.6 17.4 23.9 21.0 0.9 15. Particle size analysis using the Sedigraph indicated that the silt fraction still contained 3– 5% clay-sized particles.60 14.9 678.0 0.7 16.9 58.3 46.3 63.5 4.0 0.4 6.2 34.00 218 102 46.5 58.9 48.3 13.0 14.

these fractions are important since they are considered to be the most available to biota and most easily leached to groundwater (Brandstetter et al. Ammonium nitrate (step 1) and sulfate (step 2) extracted the least As. 1999). were found to be present in larger amounts in the clay-size fraction (Table V). In fact.3. is in accord with the large content of this mineral in several soils. 1997) and interacting with organic matter (Harter and Naidu. M INERALOGICAL E. 1981) as found in other studies. The sand and silt fractions released typically equal or larger amounts of arsenic than the clay-sized fraction in the first two steps of extraction. but the clay released more As from all the remaining steps (Table IV). except sample 960065. this is consistent with expactation based on pH data (Table I). The small amount of . oxide surface and the edges of clay minerals are positively charged at pH below. AND CHEMICAL COMPOSITION Quartz and micaceous layer silicates dominate the mineralogy in the majority of the soil. 1995). kaolinite. which. based on the NH4 -oxalate extractions. and chlorite are present in some samples. but more than the first two steps. The most common clay mineral us illite. S EQUENTIAL FRACTIONATION OF ARSENIC The amounts of arsenic extracted during the various steps of the sequential fractionation are shown in Table IV.324 4. LOMBI ET AL. 7 thus attracting anions such as arsenate and arsenite (Sadiq.. The presence of illite and vermiculite. whereas the content of layer soilicates increases. together suggest an association of NH4 F-extractable As with poorlyordered aluminosilicate gels and allophane (Becket. The amounts of OC and total arsenic (Table III) increase considerably in the order sand < silt < clay. The amount of quartz decreases from the sand to the silt fraction. 1989). Shen et al. likely derived from mica. The high amounts of clay minerals and oxides in the clay fraction may explain the accumulation of As and OC in the clay fraction.. Ammonium phosphate (step 3) extracted considerably more As than in steps 1 to 2. 1983) and adsorption (Livesey and Huang. Carbon is present only in organic form except sample 960149 that contains considerable amount of carbonates (Table III). 1973). but vermiculite. 1994). 4. Phosphate competes with As for specific adsorption sites and is more effective than nitrate and sulfate to extract arsenate from soils (Galba and Polacek. These extractants are expected to remove oxyanions from weak adsorption sites (Saeki and Mazumoto. Step 4 (NH4 F) extracted less As than released in step 3. The considerable release of Al and Si but slight dissolution of Fe and Mn (Table V). 1983. The high capacity of clay-sized particles for As may explain the positive correlation between clay and As content (Yang. Two soil samples are quite distinct from the others: sample 960065 contains > 70% feldspar and the sand fraction of sample 960149 presents considerable amounts of calcite (66%) and dolomite (12%). This is consistent with the first two steps being less effective to extract As from oxides.2. in which the smallest amount of OC is present in the silt fraction.

8 26.5 152 472 5.8 35.39 7.3 19.4 73.7 136 239 7. <d.7 491 282 417 681 392 1370 1830 133 97.13 <d.2 98.47 5.8 NH4 -oxalate NH4 -oxalate + ascorbic acid 128 396 969 120 289 773 16.75 10.56 1.6 1.12 <d.39 4.0 122 807 68.72 10.l.47 8.81 4.08 1.l.3 67.24 3.31 27.4 13.l.l.26 1. = < detection limit of 0.88 287 3.99 2.8 46.6 152 439 51.l.8 0.1 38.l.8 6.l.79 3.27 1.TABLE IV Arsenic extracted with the sequential extraction in a given size fraction (data in µg As g−1 soil) Sample/extractant NH4 NO3 a (NH4 )2 SO4 a NH4 H2 PO4 NH4 F NH2 OHHCl + NH4 OAc 12.8 0.26 <d.8 47 30.60 4.l. <d.72 1.84 1.31 951003 950057 960144 960149 960065 325 a <d.77 3.25 9. . 0.4 112 96.90 7.8 35.30 35.3 14.58 2.5 µg As L−1 .40 69.47 5.45 21.1 50.9 82.5 701 Residual SEQUENTIALLY EXTRACTED ARSENIC FROM CONTAMINATED SOILS 951002 Sand Silt Clay Sand Silt Clay Sand Silt Clay Sand Silt Clay Sand Silt Clay Sand Silt Clay 0.46 3. <d.85 4.8 4.3 51.70 33. 0. 0.7 86.87 61.60 2.30 14.37 1.55 1. <d.13 4.51 12. <d.1 29.4 151 337 399 1030 1540 368 195 1120 3.8 202 430 233 107 471 15.13 3.88 3.33 2.8 1.72 0.79 14.29 6.28 6.92 9.7 73.8 2.34 7.00 135 2.8 11.3 175 484 1370 3410 146 420 1020 65.00 7.l.

1994). (NH4 )H2 PO4 4. 1975.2 86.4 108 86.1 Si 0.0 6.d.d.7 13.7 4.7 n. This is supported by the fact that most Mn was . TABLE V Other elements extracted by the sequential extraction for each size fractions of soil 960144 (data in µg g−1 of soil) Step 1.2 243 1. 10. NH4 F 5.5 10.0 21.7 24. 1983).3 4.3 33.6 0.7 28. NH4 NO3 Sample Sand Silt Clay Sand Silt Clay Sand Silt Clay Sand Silt Clay Sand Silt Clay Sand Silt Clay Sand Silt Clay Sand Silt Clay Al 35.6 570 682 1704 1815 88.5 101 153 163 61.5 1. 1972.2 904 880 98.9 53.0 8.7 n.3 164 30.2 0.2 22.7 149 266 1078 4568 57.5 1.1 5.4 12. NH4 -oxalate 7. 64. (NH4 )2 SO4 3.5 85. + NH4 -oxalate 8.8 228 925 230 852 2817 2029 5787 10228 23751 52420 50559 Ca 268 905 2161 17. Ascorbic ac.3 81..6 0. LOMBI ET AL.5 53.5 112 32. Karczewska et al.5 21.8 2. n. NH2 )OH-HCl + NH2 OAc 6.d.d.55 510 1431 50.d.3 73.5 1.2 56.3 174.3 1345 5039 15444 10934 23332 34145 21973 28271 14897 Mn 17.7 125.5 51.2 144 34.5 1. Shen et al. Step 5 (NH2 OH-HCl) extracted As in similar amount as step 4.9 12.0 102 0.9 5. n.9 Fe 0.3 50.3 3.0 3.7 26.4 112 269 n. 8.1 58.7 15. NH2 OH-HCl is considered to be quite selective for easily reducible Mn oxides in soil (Chao.. HNO3 –H2 O2 As in this fraction may explain the findings that soils enriched in Si and Al are typically low in As (Koyama.8 2.8 222 973 26.5 2.2 57.0 46.326 E.3 38.2 0.8 20.6 1.

g.003) typically released in this step (Table V). 2. the variables with major loading (correlation between the original variable and the Factor) are steps 3 to 7 of the sequential extraction. Plotting the factor scores (values of individual observations for the Factor) against the different soil-size fractions. A strong association of arsenic with Fe oxides is also suggested by correlations found between NH4 exalate-extractable Fe and the sorption parameters of Freundlich isotherms for As in 11 soils from the U. especially. Mn oxides may catalyse redox transformations of As in soils. similarly.1 in step 4. whereas steps 1 and 2 have the major loading in Factor 2.64 (0. NH4 -oxalate used in steps 6 and 7 is effective in extracting Fe (Table V).84 (<0. The distribution of DOC among the fractions of the SEP shows that (NH4 )H2 PO4 and. In spite of the study by Thanabalasingam and Pickering (1986) showing humic acids binding relatively large amounts of As (V) and As (III).01) Step 4 0.001) Step 8 0.A. the oxidation of As(III) to As(V) (Oscarson et al. NH4F extract considerable amounts of DOC. In Factor 1. but is significantly correlated with steps 3–8 (Table VI). e.001) Step 7 0. moreover.063) Step 3 0.SEQUENTIALLY EXTRACTED ARSENIC FROM CONTAMINATED SOILS 327 TABLE VI Correlation coefficients between soil arsenic extracted with the sequential extraction and total arsenic (level of significance in parenthesis) Step 1 Total As 0.26 (0. Langmuir adsorption maxima of As in several Canadian soils were correlated to NH4 -extractable Fe and Al (Livesey and Huang.61 (0. Smaller amounts of DOC extracted in steps 1.012) Step 5 0.S. employed to dissolve amorphous and crystalline Fe-oxides. Most arsenic was extracted in steps 6 and 7. unlike metal cations.92 (<0. Arsenic extracted in steps 1 and 2 is not correlated to total As in any of the size fractions. and Factor 2 to weak extractants. 1964). indicates op- .4 (0. and by a mean molar Al:Mn ratio 42:1 in step 5 compared to 336..80 (<0. and 5 suggest that only a minor fraction of As may be bound to soil organic matter. The primary sources of Fe extracted by NH4 -oxalate are considered to derive from Fe oxides (Schwertmann. 1990). Factor 1 could be related to strong extractans. Even though only small amounts of As are extracted by NH2 OH-HCl suggesting that Mn oxides are not a major sink of As in these soils. (Selim et al. arsenate and arsenate anions can hardly accumulate at the predominantly negatively-charged surfaces of SOM. respectively. 1991).909) Step 2 0.. This may be due to the strong competition of Fe oxides for As in soils.89 (<0. 1983).001) Step 6 0. Fractional analysis performed using the sequential fraction data of arsenic reveal that 2 principal factors explain 74% of the total variance (Table VII). there is no evidence in the literature that soil organic matter (SOM) may be a significant sink for As.

404022 0.674751 1. SEM micrographs show marked differences between silt and clay-size fractions (Figure 2). apperently in secondary minerals including oxide deposits and weathering rinds. As in the other extractants is primarily associated with silt and clay.901131a 0. .727629a 0.854195a 0.941956 Cumulative % 58. total Factor 1 0.638916 4.328 E.674751 5.27445 Factor loadings (Varimax normalized) – Extraction: Principal components Variable Step 1 Step 2 Step 3 Step 4 Step 5 Step 6 Step 7 Step 8 Expl. variance Prp.182810 0.144135 0. M ICROPROBE ANALYSIS Microprobe analysis performed on the soil-size fractions were consistent with the results obtained with the SEP: Fe oxides represent the major sink for As in all soil fractions investigated.43439 15.7.326035 0.43439 74.918258a 0.805224a 0. TABLE VII Eigenvalues and factor loading from the factor analysis of arsenic data from the sequential extraction Eigenvalues – Extraction: Principal components Value 1 2 Eigenvalue 4. posite trends (Figure 1).267205 % Total variance 58.169311 0. this area of Austria is well-known for its arsenosiderite deposits (Dana.137731 1.922405a 0. In silt the only detectable As was associated with Fe.073979 0. eigenvalue 4. LOMBI ET AL.84006 Cumul. This agrees with the finding that extractant 1 and 2 are more effective on sand and silt whereas. 1951). There was no evidence for As being associated with primary minerals with the exception of arsenosiderite being found in one sample (960149) from Carinthia.037798 0.521164 Factor 2 0.855244a 0. 4.197970 0.4. This supports the interpretation that As extractable by NH4 -oxlate is mainly associated with Fe oxides.221580 a Marked correlations are bigger than 0.772646 0.

. along with evidence from SEM/EDS analysis. 1999). Plots of factor scores versus soil fractions from factorial analysis of SEP data. This fraction. and hydroxylaminechloride employed before the ammonium oxalate steps. contained the greatest amount of As and this was apparent in the EDS analyses that showed As diffusely disseminated throughout the clay-size material. sulfate. Conclusion The distribution of total arsenic among various particle size fractions of the soil was: clay » silt > sand. as noted above. This pattern of distribution is consistent with a higher sorption capacity of fine-textured fractions due to greater surface area and larger content of Fe oxides.SEQUENTIALLY EXTRACTED ARSENIC FROM CONTAMINATED SOILS 329 Figure 1. indicates that most of the arsenic is associated with iron oxides in all the soils including the calcareous one. phosphate and fluoride. 5. According to . Arsenic is typically less extractable by ammonium salts of nitrate. Arsenic is most abundant in the two acid NH4 -oxalate steps (step 6 and 7) of the SEP employed to estimate arsenic fractions of varying extractability. This finding. SEM micrographs of the clay-size fraction indicate that this fraction is very homogeneous (Figure 2). These fractions may represent ecologically important forms of arsenic that can be mobilized more easily than arsenic associated with iron oxides (Brandstetter et al. Ammonium nitrate and sulfate are more effective at extracting arsenic from sand and silt than from clay-sized fraction.

Figure 2. LOMBI ET AL. SEM micrographs and microprobe analyses (performed using energy dispersive X-ray spectrophotometer) of an iron oxide in sample 951003 silt (A) and clay fraction of sample 960149 (B). .330 E.

Pfanzenernahr Bodenk.). 83. T.. D. Holm. and Wenzel. 53.: 1995.. Hatchinson and K. Scr.. I. in silt. H. and Dana. Chao. Lead. and Griffin. Wenzel. Soil Sci. Advances in Agronomy. FAO: 1988. Proc. Vol. Our results indicate that soil textural characteristics may be useful indicators of arsenic availability and mobility in soil. 2nd Edition. Meena (eds). less homogeneous distribution of Fe oxide coatings on other soil minerals can locally cause extremely high concentrations of As associated with Fe oxides. 1055. We thank F. John Wiley & Sons Ltd.: 2000. Lombi. Acta Fytotech. Soil Sci. Mentler. Dr. S. the AFM for Environment. Soc. J. Buat-Menard. in T. Iverson. D. and Bremner. indicating their enormous capacity to bind As. and by the Austrian National Science Foundation (FWF. Proc. 1970.: 1989. 1983. pp.. in D. R. Galba. Yang. Woolson et al. Unlike these labile pools. Harter.: 1949. Pol’nohospodarstvo. Marcel Dekker Inc. R. R.. Rome. Galba. Z. L. Soil. S. Cannon for their valuable contributions. Brandstetter.: 1987. New York (in press).. Cichon (eds.: 1977. 55. 764. Sparks (eds. D. A. The System of Mineralogy. E. Ottner and B.: 1979. pp. John Wiley. D. F. and Naidu. 953–854. A. M. 133. In finetextured soils As is typically immobilised by Fe oxides that are homogeneously disseminated in the clay fraction. A. Academic Press. R.: 1957. the Federal Status of Salzburg and Tirol. R. Dr.. C. W. These findings may explain the lower toxicity of As found in fine-textured soils (compare Jacobs and Keeney.: 1996. A. A. 1983) our results indicate that coarse textured soils low in Fe oxides are likely to yield the higher amounts of readily As. R. 159. T. 38. and Stanforth. World Resources Report 60. project. W. in D. J. Chang.: 1972. A. R. D. S. Cadmium and Arsenic in the Environment. 605. . R. FAO.).. M. pp. M. A. Acknowledgements This study was funded jointly by the Austrian Federal Ministries (AFM) for Science and Traffic. A. Wise. Frost. T. London. 219–263. Trantolo. P10060 GEO). Shen et al. Amer. 9. S. Brandstetter. 710–739. 143. D.. Anderson. 222. Brandstetter. 28. Soc. and Adriano.: 1973. Mercury. 43–48. M. E. As that is bound more strongly (SEP steps 3–8) is significantly correlated with total As contents.: 1972. C. L.. J. T. Remediation of Hazardous Waste Contaminated Soil. S. Soil Sci. Soil Sci. Sletten. J. P. Inyang and E.: 1974. 41. W. References Beckett. W. These SEP steps are more effective for clay fraction. Amer. and share high loading in the same Factor of the factorial analysis. 187. 18. Sci. Am. and Polacek. 197. Adv. pp. H. and Jackson.SEQUENTIALLY EXTRACTED ARSENIC FROM CONTAMINATED SOILS 331 the factorial analysis these fractions represents the most labile pools of soil arsenic. D. In agreement with other (Koyama. L. P. ACS Symp. D. Genrich. Dana. 36. C. Soc. J. Jockwer. 1973).

Wauchope.. W. Environ. Huang. Jacobs. LOMBI ET AL. M. in Jerome O.: 1994. Circumstantiae. Kaufman. 3. Environ. Y.: 1994. C. Soil Sci. Petr. Soil Sci. Huang. Keys to Soil Taxonomy. U. P. 117. Z. G.: 1964. 88..: 1997. Soil Sci. Plant. K. 4. Mitteilungen der Deutschen Bodenkundlichen Gesellschaft. and Soil Pollut. W. Nriagu. and Brümmer. H. Am. Soil Sci. M. Woolson. Saeki.: 1986. R.: 1978. A. Karczewska. (ed.. 207. Johns.. W. S.: 1975.: 1954. F. 47. and Atman. Common.: 1980. Schwertmann. Arsenic in the Environment. and Iskandar. 440.: 1983. Electric Power Research Institute. Oscarson. P. 12. T.: 1983. E. I. D. 20. M. Grim. R. G. Koyama. K. Qual. 2147. J. Bodenkulndl. O. and Liaw. 254. H. 93.: 1994. J. Acta Sci. 505. W. Selim. 219.. K.: 1970. 156. T. Gerzabek. M. 59. 25. Zeien. Chen. J. Environ.: 1973. 30. and Keeney. W. 6th Edition. Health 16. J. Soc. U. 487. Leckie.: 1981. D. L. H.: 1983. Soc. J. Benjamin. 242. pp.: 1990. 24. Air. Gesellsch. and Bradley. ..: 1998. B. 644. 25. L. 46. Proc. Axley. Soil Biol. 233. Part 1: Cycling and Characterization. 355.. Z. R. 1. 194.. W.332 E. Palo Alto. 85. A. Stemmer. M. Poll. K.: 1989. Yang.-C. H. A. Soil Survey Staff: 1994. T. Amacher. Thanabalasingam. Petr.. Geochem. Z. Water.: 1975. Am. Anal. Soil Science 131. and Pickering. W. and Zhao. and Matzumoto. CRREL Monograph 90–2. G. D. S. R. E. M. M. F. Riedmuller. Nippon Dojo Hiryogaku Zasshi. W. Shen. 17–49. Commun. Bd.. and Hammer. W. Tschermaks Min. Mitt. 9. E. C. 37. and Plant Anal. D. Sed. and Huang. 491. and Mavrodieva. G. John Wiley & Sons. 105. Livesey. Turang Xuebao. and Kandeler. CA.). Wenzel. M. and Kearney. Biochem. P. Hayes. ‘Modelling the Transport of Heavy Metals in Soils’. M. Sadiq. Disch. D. P..