- Email: [email protected]
Use of sequential extraction to assess metal partitioning in soils
Environmental Pollution 126 (2003) 225–233 www.elsevier.com/locate/envpol
Use of sequential extraction to assess metal partitioning in soils Marika Kaasalainen, Markku Yli-Halla* MTT Agrifood Research Finland, Environmental Research, 31600 Jokioinen, Finland Received 1 November 2002; accepted 14 May 2003
‘‘Capsule’’: Sequential extraction is most useful with soils with low metal pollutant levels. Abstract The state of heavy metal pollution and the mobility of Cd, Cu, Ni, Cr, Pb and Zn were studied in three texturally different agricultural soil profiles near a Cu–Ni smelter in Harjavalta, Finland. The pseudo-total concentrations were determined by an aqua regia procedure. Metals were also determined after division into four fractions by sequential extraction with (1) acetic acid (exchangeable and specifically adsorbed metals), (2) a reducing agent (bound to Fe/Mn hydroxides), (3) an oxidizing agent (bound to soil organic matter) and (4) aqua regia (bound to mineral structures). Fallout from the smelter has increased the concentrations of Cd, Cu and Ni in the topsoil, where 75–90% of Cd, 49–72% of Cu and 22–52% of Ni occurred in the first two fractions. Slight Pb and Zn pollution was evident as well. High proportions of mobile Cd, Cu and Ni also deeper in the sandy soil, closest to the smelter, indicated some downward movement of metals. The hydroxide-bound fraction of Pb dominated in almost all soils and horizons, while Ni, Cr and Zn mostly occurred in mineral structures. Aqua regia extraction is usefully supplemented with sequential extraction, particularly in less polluted soils and in soils that exhibit substantial textural differences within the profiles. # 2003 Elsevier Ltd. All rights reserved. Keywords: Soil profile; Sequential extraction; Soil texture; Smelter; Heavy metal pollution
1. Introduction Heavy metal contamination of arable soils is attributable to the application of sewage sludge, manure and phosphatic fertilizers and to atmospheric deposition from anthropogenic sources such as ore refining industry and traffic (Grant et al., 1996; Kabata-Pendias and Pendias, 2000). Total metal concentrations in soil are also highly influenced by the parent material. Native concentrations of heavy metals are relatively high in shales and clays, and usually lower in sands and limestones. Total heavy metal concentrations may give insufficient information on soil pollution and metal leaching when soils comprise horizons of different textures. This, according to Mokma et al. (2000), is a common feature of soils of Finland. Metals from anthropogenic sources tend to be more mobile than pedogenic or lithogenic ones (Kuo et al., * Corresponding author. Tel.: +358-3-4188-3140; fax: +358-34188-3196. E-mail address: markku.yli-halla@mtt.fi (M. Yli-Halla). 0269-7491/03/$ - see front matter # 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0269-7491(03)00191-X
1983; Chlopecka et al., 1996; Karczewska, 1996). Speciation of the metals can help assess how strongly they are retained in soil and how easily they may be released into soil solution. Different sequential extraction techniques, such as the five-step procedure of Tessier et al. (1979), are commonly applied (Chlopecka et al., 1996; Sa´nchez et al., 1999) to evaluate both the actual and the potential mobility of metals in the environment (Rudd et al., 1986; Dang et al., 2002). The speciation of heavy metals by sequential extractions is operational and even slight changes in the procedure may decisively change the outcome (Sahuquillo et al., 1999). Results obtained by different procedures may be difficult to compare as a consequence. It is still common to attribute a fraction dissolved by a particular extractant to a specific soil constituent. Generally, the following forms of metals are distinguished: (1) dissolvable in soil solution, i.e. watersoluble metal cations, (2) exchangeable in inorganic and organic components (oxides/organic matter), (3) structural components of the lattices of soil minerals, and (4) insolubles precipitated with other soil components (Ahumada et al., 1999; Zalidis et al., 1999). The first two
226
M. Kaasalainen, M. Yli-Halla / Environmental Pollution 126 (2003) 225–233
forms can be considered readily available to plants, while the other two may become available in the long term. Harjavalta (61 200 N, 22 090 E) in SW Finland is dominated by its Cu and Ni smelters. Electric smelting of Cu commenced in 1945 and electrolytic refining of Ni in 1960. A sulphuric acid plant has operated in Harjavalta since 1947 using SO2 removed from the smelter flue gases as its raw material. In 1998 altogether 156,000 t of blister Cu, 43,000 t of Ni and 590,000 t of sulphuric acid were produced at the complex. The Harjavalta plants are the major point source of Cu and Ni emissions in Finland (Aunela and Larjava, 1990). Concentrations of Cu and Ni in mosses of Finland are by far the highest in and around Harjavalta (Kubin et al., 2000). During the growing season of 1988, the deposition of Cu, Ni and Cd measured 0.8 km from the smelter was 36, 20 and 18 times that measured in an unpolluted area (Yla¨ranta, 1996). According to Derome and Nieminen (1998), the annual bulk deposition in 1993–1998 comprised 1500 g Cu ha1 and 640 g Ni ha1 at a distance of 0.5 km from the smelter complex, while the amounts were only 27 g Cu ha1 and 19 g Ni ha1 at 8 km. Even though the emissions from the complex decreased to just one tenth their previous value towards the end of the 1990s (Melanen et al., 1999), the earlier load has mostly accumulated in the soil. The litter decomposition rate has decreased in the near-by forest as a result (Pennanen et al., 1996), and forest vegetation is suffering. Indeed, extremely high concentrations of total Cu (5800 mg kg1) and other heavy metals have been determined in the thin organic layer of the forest floor 0.5 km from the smelter, and in the acidic (pH about 4) sandy soils the metals are in a fairly mobile state (Derome and Lindroos, 1998). The smelters are situated near the Kokema¨enjoki river and are surrounded by cultivated fields. Cu, Ni and Cd accumulation into nearby agricultural soils has been demonstrated (Sippola and Ervio¨, 1986; Yla¨ranta, 1996). Cd is not produced by the smelter of Harjavalta, but it is present as a minor metallic constituent in the copper sulphide ore or other raw materials (Scoullos, 2000). Even though the agricultural soils are less acidic than the forest soils, abnormally high concentrations of metals, partly taken up from soil and partly from the airborne deposition, have been measured in crops at Harjavalta (Yla¨ranta, 1996). Farmers are aware of the metal deposition and want to know to what extent heavy metals accumulated in the soil endanger the quality of their farmland and crops. The main objective of this study was to document the speciation of Cd, Cu, Ni, Cr, Pb and Zn in three agricultural soils near the Cu–Ni smelter at Harjavalta, western Finland. The vertical distribution of the elements in these profiles was also of interest, to establish the possible downward movement of the metals. The metals were studied by an aqua regia procedure that
gives the pseudo-total metal concentration, and by a sequential extraction procedure that divides the metals into four operationally defined fractions. The extraction procedure was originally developed for sediment materials (Sahuquillo et al., 1999; Rauret et al., 1999), and considerable within-laboratory and between-laboratory variability in the heavy metal concentrations has been observed upon its application (Rauret et al., 2000). Soils are usually less homogeneous than sediments, having a more variable particle size distribution, and the suitability of the procedure for soil samples needs to be verified.
2. Materials and methods 2.1. Soil sampling Three cultivated soils at a distance of 0.8–2.2 km from the metal smelter were sampled by horizon (first three uppermost horizons; 0–70 cm), in autumn 1999 from soil pits. The soils were in agricultural use despite the Cu, Ni and Cd contamination observed in earlier studies (Sippola and Ervio¨, 1986; Yla¨ranta, 1996). Soil profiles were characterized (Table 1) according to Soil Taxonomy (Soil Survey Staff, 1998) and sampled according to visible horizon boundaries. The rainfall in 1999 was 662 mm and the average temperature 5.1 C. 2.2. Soil analyses The soil samples were air dried (35 C), screened to pass a 2-mm sieve. After digestion with H2O2 and dispersion with Na4P2O7.10 H2O, silt and clay were determined by a pipette method and sand by wet sieving (Elonen, 1971). Soil carbon (C) was determined with a LECO CN-2000 Elemental Analyzer. All C was assumed to be organic. Soil pH(H2O) was determined in a 1:2.5 soil/water suspension with a Beckman Instruments, Inc. FuturaTM refillable combination electrode. The soil suspension was allowed to stand overnight before pH determination. The cation exchange capacity (CEC) at pH 7.0 was calculated by summation of Ca, Mg, K, Na and titratable acidity, all extracted with 1 M ammonium acetate (CH3COONH4). Base saturation was calculated as Ca+Mg+K+Na expressed as percentage of CEC. Heavy metals were fractionated with the three-step extraction procedure developed within the Standards, Measurements and Testing Programme (formerly the Community Bureau of Reference, BCR) of the European Commission. This BCR procedure was originally developed for and used in the certification of heavy metal concentrations of a sediment reference material CRM 701 (Rauret et al., 1999; Pueyo et al., 2001). In the procedure, a sample is sequentially extracted with
227
M. Kaasalainen, M. Yli-Halla / Environmental Pollution 126 (2003) 225–233 Table 1 Characterization of selected physicochemical properties in the three horizons of the soilsa Horizon
Soil 1, Oxyaquic Cryorthent I
Plant cultivated Distance from the smelter complex/km Horizon Depth/cm Texture pH (H2O) CEC/cmol(+)kg1 Base saturation/% Organic C/% Clay ( <0.002 mm)/% Silt (0.002–0.06 mm)/% Sand (0.06–2 mm)/% a
II
III
Soil 2, Typic Cryaquept
Soil 3, Oxyaquic Cryorthent
I
I
Sugar beet; potato 0.8
II
III
Barley 1.2
II
III
Barley 2.3
Ap 0–25 Loamy sand
BC 25–63 Sand
2C 63–67 Silt loam
Ap 0–30 Silt loam
Bw1 30–47 Clay
Bw2 47–70 Silty clay (loam)
Ap 0–30 Silt loam
Bw1 30–45 Silt
Bw2 47–55 Silt loam
6.34 9.7 66 1.4 3 16 81
6.27 3.2 24 0.1 0 4 96
6.40 7.1 30 0.3 11 52 37
6.49 11.3 63 0.9 17 65 18
6.06 25.6 78 0.4 69 25 6
6.73 14.0 70 0.3 40 59 1
7.08 2.6 83 1.3 9 58 33
6.58 6.0 64 0.2 11 84 5
6.36 6.6 61 0.2 13 83 5
Soils were classified according to Soil Survey Staff (1998).
dilute acetic acid, a reducing agent and an oxidizing agent. As an additional step, the aqua regia procedure ISO 11466 (1995) was used to extract metals from the residue remaining after the three extraction steps. The four steps divide the trace metals into the following fractions: (1) exchangeable and specifically adsorbed (EX), (2) Fe and Mn (hydr)oxide-bound (OXI), (3) organically bound (ORG) and (4) residual (i.e. mineralbound; Res). The extraction procedure, chemical reagents and the experimental conditions are presented in detail in the flow chart (Fig. 1). It needs to be pointed out that the 0.11 M (about 0.5%) acetic acid that was used is less concentrated than that often used (2.5% in Elsokkary and La˚g, 1978; Iyengar et al., 1981) to extract specifically adsorbed metals. Three subsamples of every soil sample were extracted. The original soil samples were also analysed by the aqua regia procedure for their pseudo-total heavy metal concentrations. In Finnish soils, this aqua regia procedure dissolves about 95% of total Cu, Ni and Zn, about 80% of total Cd and about 60% of total Cr and Pb (Baghdady and Sippola, 1983a,b). 2.3. Elemental analysis and instrumentation The EX and ORG fractions of Cd, Cu, Ni, Cr, Pb and Zn were analysed with a Perkin-Elmer Sciex ELAN 6000 ICP-MS. The OXI and the aqua regia extractable Cu, Ni, Cr and Zn were analysed by TJA/Baird Division IRIS (II) ICP-AES and the corresponding fractions of Cd and Pb by Varian SpectrAA-400 with a GTA-96 graphite tube atomiser and Zeeman background correction. All element concentrations are presented on a dry matter basis.
2.4. Statistical methods Arithmetic means (AM) of the triplicate extraction results were calculated as well as the standard deviations (SD) and the coefficients of variation (CV, i.e. RSD relative standard deviation that is a ratio of the SD to AM value, expressed in %). The relationships between selected soil characteristics (clay, organic C, CEC, pH) and the metal fractions were studied by correlation analysis. Adequate information for the correlation analysis was obtained by expanding the data of the three soil profiles (n=9) with the results of an additional profile (n=4) from the same area. A detailed description of this soil will be presented in a later paper. All the mathematical operations were conducted with Microsoft1 Excel 2000.
3. Results and discussion 3.1. Soil properties As shown in Table 1, the soil profiles represented three different textural classes, varying from clay to sand. The texture differed between the horizons of the soil profiles because the horizons had been sedimented at different stages of the retreat of the continental ice at the end of the Weichselian glaciation ( 10,000 years ago). The clay content (0–69%) did not show any trend with depth. Because of the intensive liming, the pH in the plow layers was rather high (6.3–7.1) as compared to the average (< 6) in Finnish fields, and it was notably higher than in the forest soils of the area (about 4, Derome and Lindroos, 1998).
228
M. Kaasalainen, M. Yli-Halla / Environmental Pollution 126 (2003) 225–233
3.2. Pseudo-total concentrations of Cd, Cu, Ni, Cr, Pb and Zn The pseudo-total (aqua regia extractable) concentrations of the metals in the various soils and horizons are
set out in Table 2. The sandy soil 1, which was closest (0.8 km) to the smelter, exhibited the highest pseudototal concentrations of Cd, Cu, Ni and Pb in the plow layer even tough, among mineral soils, sandy soils tend to be lowest in heavy metals. There was a steady decline in the pseudo-total concentrations, especially of Cu and Ni, with increasing distance from the smelter. Also around a metal smelter in Ruston near Tacoma, Washington, USA, the concentration maxima of Cd, Cu and Zn in the surface soils were within 2–3 km of the smelter (Kuo et al., 1983). In Sudbury, Canada, however, Dudka et al. (1995) found that the spatial distribution of total Cu and Ni in soil was sometimes inconsistent with respect to distance from the smelter due to unevenness of the terrain, the dispersion effect of the stack and the influence of other sources of metal contamination in the area. The soils of our study probably represent the most loaded area, however, because in the study by Derome and Lindroos (1998) the Cu, Ni and Zn concentrations of the forest soils of Harjavalta showed a consistent decline along a 8-km transect from the smelter. In a recent survey (Ma¨ntylahti and Laakso, 2002), the mean aqua regia extractable metal concentrations in 182 coarse agricultural soils of Finland were as follows (mg kg1): Cd 0.10, Cu 13.9, Ni 6.7, Cr 17.8, Pb 7.8 and Zn 35.6. Measured against these values, the plow layers of all our soils were high in Cd, Cu, Ni and Zn, and soil 1 also in Pb. Compared with the values of another study on 10 diverse agricultural soils in Finland (Baghdady and Sippola, 1983a,b), however, the plow layers of our soils exhibited elevated concentrations of only Cu, while soil 1 was high in Ni and Cd as well. These differing results for Finnish soils mean that only the strongest contamination with Cu, Cd and Ni could be confirmed with certainty in our soils on the basis of pseudo-total analysis. In all three soils the concentrations of pseudo-total Cd and Cu were high in the plow layer relative to the deeper horizons. In soil 1 the same trend was observed in Ni, Pb and Zn. Also in soil 1, the decrease of Cd, Cu, and Ni from the plow layer to the second horizon was abrupt; most notably, the Cu concentration was 44-fold in the plow layer relative to the second horizon. In contrast to this, in soil 2 the concentrations of Cr, Ni and Zn were highest in the second and third horizons, which contained more clay than the loamy topsoil. 3.3. Fractionation of metals
Fig. 1. Schematic presentation of the extraction procedures applied to the soil samples. The extract was separated by centrifuging (3000g, 20 min) and dectantation. Steps 1, 2 and 3 all terminated with washing of the residue with 20 ml H2O.
Fig. 2 shows the distribution of the metals into the four fractions of the sequential extraction. In the plough layer, the proportions of the metals in the acetic acid extractable fraction (EX) followed the order Cr4Pb < Zn < Ni < Cu < Cd. Cadmium was particularly abundant in this most mobile fraction (66, 44 and 53% of the
0.085 0.011 97.6 11.7 0.744 103.3 13.5 0.525 102.3 27.0 1.010 104.2 7.12 0.289 103.4 46.7 1.904 100.5 0.0890.009 88.2 11.4 1.253 87.1 13.6 0.174 92.9 27.8 1.175 95.0 6.78 0.497 94.5 53.9 1.087 93.3 0.2740.002 96.5 52.40.411 93.8 16.70.924 95.6 18.70.960 99.6 9.790.488 96.2 61.00.318 95.8 0.1620.003 96.0 28.10.636 100.2 32.41.108 98.8 58.01.306 99.5 12.60.351 87.6 1051.789 97.9 0.0990.008 106.1 39.30.500 99.1 42.81.623 104.9 71.61.702 104.3 19.20.616 94.8 1321.730 100.3 0.2500.011 94.8 73.20.976 101.0 25.90.910 104.2 31.60.566 108.5 11.60.579 102.6 65.11.735 102.4 0.062 0.006 93.1 8.24 0.433 100.3 10.2 0.644 102.2 23.0 0.843 100.5 9.59 0.597 98.8 50.2 2.964 98.4 0.0440.007 97.7 6.07 0.134 107.2 4.07 0.158 93.4 5.03 0.392 104.7 3.97 0.217 89.9 13.6 0.570 92.7 0.6550.031 89.5 263 5.663 98.1 54.2 4.835 86.0 9.09 0.233 101.5 15.8 0.588 90.7 67.3 9.127 79.0 Zn
Pb
Cr
Ni
Cu
Cd
Pseudo-total concentration/mg kg (BCR/pseudo-total)/% Pseudo-total concentration/mg kg1 (BCR/pseudo-total)/% Pseudo-total concentration/mg kg1 (BCR/pseudo-total)/% Pseudo-total concentration/mg kg1 (BCR/pseudo-total)/% Pseudo-total concentration/mg kg1 (BCR/pseudo-total)/% Pseudo-total concentration/mg kg1 (BCR/pseudo-total)/%
III II I I
1
II
III I
II
III
Soil 3 Soil 2 Soil 1 Horizon
Table 2 The pseudo-total (aqua regia extractable) concentration of the metals (mg kg1) in the horizons (I–III) of the three soils (meanS.D., n=3)a and sum of the metals extracted by BCR procedure (EX+OXI+ORG+residual fraction) expressed as percentage of the pseudo-total concentration of the metals.
M. Kaasalainen, M. Yli-Halla / Environmental Pollution 126 (2003) 225–233
229
pseudo-total in the plow layer of soils 1, 2 and 3, respectively), in agreement with findings at other anthropogenically loaded sites (Elsokkary and La˚g, 1978; Kuo et al., 1983; Chlopecka et al., 1996; Sa´nchez et al., 1999; Li and Thorton, 2001). Although the absolute CdEX concentration was much lower in the deeper horizons, as proportion of the pseudo-total concentration it was still 56% in the second horizon of soil 1 closest to the smelter (in contrast to 14% in soil 2 and 22% in soil 3). In soil 1, the CuEX and NiEX fractions each represented about 30% in the plow layer and were at least equally as large in the horizon below; in the other two soils these fractions comprised 11–16% in the plow layer and only 1–6% deeper in the soil. Proportions of ZnEX were < 8% except in the first horizon of soil 1 (24%). Deposition can thus be considered the source of the elevated concentrations of the acetic acid extractable fraction of Cd, Cu and Ni all three soils and of Zn in soil 1. The large fractions of CdEX, CuEX and NiEX in horizon II show that these metals have been transported downwards in the sandy soil, a feature described earlier for heavy metals of anthropogenic origin (Karczewska, 1996). The CrEX and PbEX fractions were < 2% throughout the profiles. In all soils, the concentrations of CdOXI and CuOXI, likely associated with Fe/Mn oxides, were much higher in the plow layer than in the deeper horizons, showing that accumulated metals are found in this fraction as well. In soil 1 the concentration of PbOXI was 10 times as high in the plow layer as in the horizon below. This finding confirms our tentative conclusion of some contamination of soil 1 with Pb, made on the basis of the pseudo-total concentrations in our soils and concentrations in nearby forest soils (Derome and Lindroos, 1998). The accumulation of anthropogenic Pb in oxidebound form is in accordance with earlier findings (Chlopecka et al., 1996; Sa´nchez et al., 1999; Sahuquillo et al., 1999). It is noteworthy that, even in the deeper horizons where Pb likely was native, this element largely was bound by Fe/Mn (hydr)oxides. In contrast to the two more coarse-textured soils, the highest concentrations of PbOXI, NiOXI and ZnOXI in soil 2 occurred in the second horizon, where they probably are associated with the high clay content. All samples were low in CrOXI (5–7%). The fraction extracted after oxidizing with hydrogen peroxide may in some degree be associated with soil organic matter. The concentrations of CuORG in all soils and those of NiORG and ZnORG in soils 1 and 3 were substantially higher in the plow layer than in the deeper horizons, while the concentrations of CdORG and PbORG were more or less the same in all horizons, irrespective of the loading and the concentration of organic matter. The results for this fraction yielded little additional information to what had already been found on the basis of the two preceding fractions. Possibly part of the
230
M. Kaasalainen, M. Yli-Halla / Environmental Pollution 126 (2003) 225–233
organically bound metals had already been extracted in the second step as is suggested by findings made when this BCR procedure was developed (Sahuquillo et al., 1999): decrease of the pH (2.9 0.3!1.5 0.1) and increase in the concentration of NH2O.HCl (0.1!0.5 M) in the extraction of the OXI fraction increased the extraction of Cr by 647%, Cu by 825% and Pb by 450%, and the amounts in the subsequent fractions were decreased accordingly. These results strongly argue against the physical interpretation that metals extracted at this step truly represent the fraction associated with soil organic matter. The proportion of the residual metal fraction (Res), reflecting the native metal concentration in soil, increased with depth for every metal studied, in contrast to the proportion of the EX fraction. In the deepest horizon, the average proportions of the residual metals were, within a narrow range, as follows: Cd 25%, Cr 88%, Cu 70%, Ni 83%, Pb 45% and Zn 84%. The
somewhat low percentages of CdRes and PbRes in the subsoil probably do not reflect pollution with these metals but are native characteristics of the soil. 3.4. Correlations between Cd, Cu, Ni, Cr, Pb and Zn fractions and soil properties Significant correlations between Cd, Cu, Ni, Cr, Pb and Zn fractions and soil properties are presented in Table 3. Individual metal fractions did not correlate with pH and they correlated poorly with soil organic C, possibly because the ranges of these soil properties were narrow. The pseudo-total concentrations of Cu, Cd and Ni, which were elevated markedly by human activities, did not correlate with clay, while pseudo-total Cr, Pb and Zn, which are less affected by anthropogenic activities, correlated positively with the clay content. Residual fractions of all metals except Cd correlated positively with clay content.
Fig. 2. Distribution of Cd, Cu, Ni, Cr, Pb and Zn in sequentially extracted fractions of soils near Harjavalta Cu–Ni smelter. Abbreviations for the fractions and extractants: EX (exchangeable and specifically adsorbed) 0.11 M HOAc; OXI (bound to the oxides of Fe and Mn) 0.5 M NH2O HCl, pH 1.5; ORG (bound to soil organic matter) 8.8 M H2O2+1 M NH4Ac and Res (bound to mineral structures) aqua regia. I, II and III refer to the three uppermost horizons of each soil.
M. Kaasalainen, M. Yli-Halla / Environmental Pollution 126 (2003) 225–233 Table 3 All significant correlations (r values) between selected trace metal fractions (pseudo-total, EX, OXI, ORG, residual) and soil properties (n=13) Metal fraction
Clay
Organic C
CEC
Cdtot CdOXI
– –
0.59* 0.89***
– 0.62*
CuORG
–
0.68*
–
NiOXI Nires
– 0.91***
0.75** –
0.69** –
Crtot CrEX CrOXI CrORG Crres
0.97*** – 0.97*** 0.67* 0.96***
– – – – –
– – – 0.84* –
Pbtot PbORG Pbres
0.66* 0.69* 0.84***
– – –
– – –
Zntot ZnOXI ZnORG Znres
0.93*** 0.86*** – 0.96***
– – 0.65* –
– 0.66* 0.79** –
*, **, ***Significant correlations at the P40.1, 0.01 and 0.001 level, respectively.
231
mobility of heavy metals. In spite of a near-neutral pH in the studied soils, a substantial part of the Cd, Cu and Ni, and in soil 1 also Zn, occurred in the most mobile (exchangeable and specifically adsorbed) form. Thus, liming and tillage do not completely eliminate the risk of higher heavy metal uptake of crops in the area of Harjavalta despite the now low level of the airborne loading. Pseudo-total analysis indicated clear Cu, Ni and Cd pollution of the most loaded soil 1, closest to the smelter, and Cu pollution in the other two soils. The results of the sequential extraction showed that acetic acid extractable Cd and Ni were high in soils 2 and 3 as well as in soil 1. Pollution with Zn, accumulating in the acetic acid extractable form, and pollution with Pb, bound to Fe/Mn hydr(oxides), were demonstrated in soil 1 but not in soils further away from the smelter. Sequential extraction also revealed elevated concentrations of acetic acid extractable Cd, Cu and Cd below the plow layer in the most sandy soil 1, but quantitatively the risk of downward movement of these metals seems to be small. High pseudo-total concentrations of Cr, Ni, Pb and Zn below the plow layer in soil 2 were native, relating to the high clay content of the subsoil, and did not indicate downward movement of the metals.
3.5. Analytical accuracy 4. Conclusions No marked differences were observed between the pseudo-total metal concentrations extracted with aqua regia and the sum of metals extracted by the sequential extraction procedure (Table 2). Indicative of the good quality of the analytical work, the overall recovery rates (the sum of the four fractions expressed as percent of the independent pseudo-total concentration) ranged from 87 to 106% in all soils and horizons, except in a few cases where the concentrations in the subsoil were very low. 3.6. Metal status of cultivated soils at Harjavalta The concentrations of heavy metals in agricultural soils at Harjavalta were much lower than the extremely high values measured in a thin surface layer of forest soils (Derome and Lindroos, 1998). The explanation of the lower values is the nearly annual plowing of the soils, which has mixed the metals deposited onto the soil surface through the 25–30 cm thick layer. In forest soils, the undisturbed top layer is subjected to metal deposition year after year. Quantitatively, loading of forest and field has probably been the same, but in arable land, the metals have been diluted into a larger volume of soil. Agricultural soils are commonly less acidic, finer in texture, and thus richer in chemically active soil constituents, all of these being factors that reduce the
1. In the plow layers of the studied soils the mobility of the metals increased approximately in the order Cr < Zn < Pb < Ni4Cu < Cd. Cadmium occurred more abundantly in soluble form than did the other elements. Cd, Cu and Ni appeared in mobile forms to a particularly large extent in the surface layer and also in the second horizon of the most polluted, sandy soil (soil 1). 2. The BCR procedure gave reproducible results in surface soil high in several metals and in subsoil low in metals. Application of the procedure allowed reliable assessment of pollution in soils differing in particle size distribution. Some doubt was raised about the physical nature of the fraction extracted with hydrogen peroxide, however. 3. The BCR sequential extraction can assist in identifying heavy metal contamination of soil and possible downward movement of metals. It is most useful in soils with texturally different horizons at sites where metal pollution is only minimal. If remediation is carried out in time, the detrimental effects of heavy metals can be minimized. Aqua regia extractable concentrations indicate metal pollution at a more advanced stage.
232
M. Kaasalainen, M. Yli-Halla / Environmental Pollution 126 (2003) 225–233
Acknowledgements We thank Ms. Leila Lindstedt and Ms. Pa¨ivi Alle´n for technical assistance and carrying out the laboratory work. Outokumpu Harjavalta Metals Oy funded the experimental phase of the study and the Finnish Cultural Foundation and the Outokumpu Foundation supported the publication of the results. We are grateful for this assistance. Our thanks, too, to the farmers of Harjavalta who participated in the study.
References Ahumada, I., Mendoza, J., Navarrete, E., Ascar, L., 1999. Sequential extraction of heavy metals in soils irrigated with wastewater. Communications in Soil Science and Plant Analysis 30, 1507–1519. Aunela, L., Larjava, K., 1990. Raskasmetallipa¨a¨sto¨t Suomessa. (Heavy metal emissions in Finland), Technical Research Centre of Finland. Research Notes 111, 1–78 (in Finnish). Baghdady, N.H., Sippola, J., 1983a. Total heavy metal recovery by aqua regia in soils of different origins. Annales Agriculturae Fenniae 22, 175–185. Baghdady, N.H., Sippola, J., 1983b. Efficiency of aqua regia in extracting Cd, Cr, Hg, Ni and Pb from soils of different origins. Annales Agriculturae Fenniae 22, 240–244. Chlopecka, A., Bacon, J.R., Wilson, M.J., Kay, J., 1996. Forms of cadmium, lead, and zinc in contaminated soils from southwest Poland. Journal of Environmental Quality 25, 69–79. Dang, Z., Liu, C., Haigh, M.J., 2002. Mobility of heavy metals associated with the natural weathering of coal mine spoils. Environmental Pollution 118, 419–426. Derome, J., Lindroos, A.-J., 1998. Effects of heavy metal contamination on macronutrient availability and acidification parameters in forest soil in the vicinity of the Harjavalta Cu–Ni smelter, SW Finland. Environmental Pollution 99, 225–232. Derome, J., Nieminen, T., 1998. Metal and macronutrient fluxes in heavy-metal polluted Scots pine ecosystems in SW Finland. Environmental Pollution 103, 219–228. Dudka, S., Ponce-Hernandez, R., Hutchinson, T.C., 1995. Current level of total element concentrations in the surface layer of Sudbury’s soils. The Science of the Total Environment 162, 161–171. Elonen, P., 1971. Particle size analysis of soil. Acta Agralia Fennica 122, 1–122. Elsokkary, I.H., La˚g, J., 1978. Distribution of different fractions of Cd, Pb, Zn, and Cu in industrially polluted and non-polluted soils of Odda region, Norway. Acta Agriculturae Scandinavica 28, 262– 268. Grant, C.A., Bailey, L.D., Therrien, M.C., 1996. The effect of N, P and KCl fertilizers on grain yield and Cd concentration of malting barley. Fertilizer Research 45, 153–161. ISO 11466, 1995. International Standard, Soil Quality—extraction Of Trace Elements Soluble in aqua regia. Iyengar, S.S., Martens, D.C., Miller, W.P., 1981. Distribution and plant availability of soil zinc fractions. Soil Science Society of America Journal 45, 735–739. Kabata-Pendias, A., Pendias, H., 2000. Trace Elements in Soil and Plants, third ed. CRC Press, Boca Raton, FL. Karczewska, A., 1996. Metal species distribution in top- and sub-soil in an area affected by copper smelter emissions. Applied Geochemistry 11, 35–42. Kubin, E., Lippo, H., Poikolainen, J., 2000. Heavy metal loading. In: Ma¨lko¨nen, E. (Ed.), Forest Condition in a Changing Environment— the Finnish Case. Kluwer Academic, The Netherlands, pp. 60–71.
Kuo, S., Heilman, P.E., Baker, A.S., 1983. Distribution and forms of copper, zinc, cadmium, iron, and manganese in soils near a copper smelter. Soil Science 135, 101–109. Li, X., Thornton, I., 2001. Chemical partitioning of trace and major elements in soils contaminated by mining and smelting activities. Applied Geochemistry 16, 1693–1706. Ma¨ntylahti, V., Laakso, P., 2002. Arsenic and heavy metal concentrations in agricultural soils in South Savo province. Agricultural and Food Science in Finland 11, 285–300. Melanen, M., Ekqvist, M., Mukherjee, A.B., Aunela-Tapola, L., Verta, M., Salmikangas, T., 1999. Raskasmetallien pa¨a¨sto¨t ilmaan Suomessa 1990-luvulla (Emissions of heavy metals in Finland in the 1990s). Suomen ympa¨risto¨ 329, 1–92 (in Finnish). Mokma, D.L., Yli-Halla, M., Hartikainen, H., 2000. Soils in a young landscape on the coast of southern Finland. Agricultural and Food Science in Finland 9, 291–302. Pennanen, T., Frostega˚rd, A˚., Fritze, H., Ba˚a˚th, E., 1996. Phospholipid fatty acid composition and heavy metal tolerance of soil microbial communities along two heavy metal polluted gradients in coniferous forests. Applied Environmental Microbiology 62, 420– 428. Pueyo, M., Rauret, G., Lu¨ck, D., Yli-Halla, M., Muntau, H., Quevauviller, Ph., Lo´pez-Sa´nchez, J.F., 2001. Certification of the extractable contents of Cd, Cr, Cu, Ni, Pb and Zn in a freshwater sediment following a collaboratively tested and optimised three-step sequential extraction procedure. Journal of Environmental Monitoring 3, 243–250. Rauret, G., Lo´pez-Sa´nchez, J.F., Sahuquillo, A., Rubio, R., Davidson, C., Ure, A., Quevauviller, Ph., 1999. Improvement of the BCR three step sequential extraction procedure prior to the certification of new sediment and soil reference materials. Journal of Environmental Monitoring 1, 57–61. Rauret, G., Lo´pez-Sa´nchez, J.F., Sahuquillo, A., Barahona, E., Lachica, M., Ure, A.M., Davidson, C.M., Gomez, A., Lu¨ck, D., Bacon, J., Yli-Halla, M., Muntau, H., Quevauviller, Ph., 2000. Application of a modified BCR sequential extraction (three-step) procedure for the determination of extractable trace metal contents in sewage sludge amended reference material (CRM 483), complemented by a three-year stability study of acetic acid and EDTA extractable metal content. Journal of Environmental Monitoring 2, 228–233. Rudd, T., Campbell, J.A., Lester, J.N., 1986. Characterization of metal forms in sewage sludges by chemical extraction. In: Lester, J.N., Perry, R., Sterrit, R.M. (Eds.), Chemicals in the Environment. Selper Ltd, London, pp. 756–771. Sahuquillo, A., Lo´pez-Sa´nchez, J.F., Rubio, R., Rauret, G., Thomas, R.P., Davidson, C.M., Ure, A.M., 1999. Use of a certified reference material for extractable trace metals to assess sources of uncertainty in the BCR three-stage sequential extraction procedure. Analytica Chimica Acta 382, 317–327. Sa´nchez, G., Moyano, A., Mun˜ez, C., 1999. Forms of cadmium, lead, and zinc in polluted mining soils and uptake by plants (Soria Province, Spain). Communications in Soil Science and Plant Analysis 30, 1385–1402. Scoullos, M., 2000. Data and trends on cadmium. Synopsis prepared as a background document for the International EUPHEMET’s Workshop (Toward an Integrated EU Policy for Heavy Metals), 17–18 April 2000, Athens, Greece. Sippola, J., Ervio¨, R., 1986. Raskasmetallit maapera¨ssa¨ ja viljelykasveissa Harjavallan tehtaiden ympa¨risto¨ssa¨ (Heavy metals in soils and cultivated plants in the Harjavalta smelter area). Ympa¨risto¨ ja terveys 5, 270–275 (in Finnish). Soil Survey Staff. 1998. Keys to Soil Taxonomy, eighth ed. USDA– NRCS. Tessier, A., Campbell, P.G.C., Bisson, M., 1979. Sequential extraction procedure for the speciation of particulate trace metals. Analytical Chemistry 51, 844–851.
M. Kaasalainen, M. Yli-Halla / Environmental Pollution 126 (2003) 225–233 Yla¨ranta, T., 1996. Uptake of heavy metals by plants from airborne depositions and polluted soils. Agricultural and Food Science in Finland 5, 431–447.
233
Zalidis, G., Barbayiarinis, N., Matsi, T., 1999. Forms and distribution of heavy metals in soils of the Axios Delta of Northern Greece. Communications in Soil Science and Plant Analysis 30, 817–827.
Use of sequential extraction to assess metal partitioning in soils Marika Kaasalainen, Markku Yli-Halla* MTT Agrifood Research Finland, Environmental Research, 31600 Jokioinen, Finland Received 1 November 2002; accepted 14 May 2003
‘‘Capsule’’: Sequential extraction is most useful with soils with low metal pollutant levels. Abstract The state of heavy metal pollution and the mobility of Cd, Cu, Ni, Cr, Pb and Zn were studied in three texturally different agricultural soil profiles near a Cu–Ni smelter in Harjavalta, Finland. The pseudo-total concentrations were determined by an aqua regia procedure. Metals were also determined after division into four fractions by sequential extraction with (1) acetic acid (exchangeable and specifically adsorbed metals), (2) a reducing agent (bound to Fe/Mn hydroxides), (3) an oxidizing agent (bound to soil organic matter) and (4) aqua regia (bound to mineral structures). Fallout from the smelter has increased the concentrations of Cd, Cu and Ni in the topsoil, where 75–90% of Cd, 49–72% of Cu and 22–52% of Ni occurred in the first two fractions. Slight Pb and Zn pollution was evident as well. High proportions of mobile Cd, Cu and Ni also deeper in the sandy soil, closest to the smelter, indicated some downward movement of metals. The hydroxide-bound fraction of Pb dominated in almost all soils and horizons, while Ni, Cr and Zn mostly occurred in mineral structures. Aqua regia extraction is usefully supplemented with sequential extraction, particularly in less polluted soils and in soils that exhibit substantial textural differences within the profiles. # 2003 Elsevier Ltd. All rights reserved. Keywords: Soil profile; Sequential extraction; Soil texture; Smelter; Heavy metal pollution
1. Introduction Heavy metal contamination of arable soils is attributable to the application of sewage sludge, manure and phosphatic fertilizers and to atmospheric deposition from anthropogenic sources such as ore refining industry and traffic (Grant et al., 1996; Kabata-Pendias and Pendias, 2000). Total metal concentrations in soil are also highly influenced by the parent material. Native concentrations of heavy metals are relatively high in shales and clays, and usually lower in sands and limestones. Total heavy metal concentrations may give insufficient information on soil pollution and metal leaching when soils comprise horizons of different textures. This, according to Mokma et al. (2000), is a common feature of soils of Finland. Metals from anthropogenic sources tend to be more mobile than pedogenic or lithogenic ones (Kuo et al., * Corresponding author. Tel.: +358-3-4188-3140; fax: +358-34188-3196. E-mail address: markku.yli-halla@mtt.fi (M. Yli-Halla). 0269-7491/03/$ - see front matter # 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0269-7491(03)00191-X
1983; Chlopecka et al., 1996; Karczewska, 1996). Speciation of the metals can help assess how strongly they are retained in soil and how easily they may be released into soil solution. Different sequential extraction techniques, such as the five-step procedure of Tessier et al. (1979), are commonly applied (Chlopecka et al., 1996; Sa´nchez et al., 1999) to evaluate both the actual and the potential mobility of metals in the environment (Rudd et al., 1986; Dang et al., 2002). The speciation of heavy metals by sequential extractions is operational and even slight changes in the procedure may decisively change the outcome (Sahuquillo et al., 1999). Results obtained by different procedures may be difficult to compare as a consequence. It is still common to attribute a fraction dissolved by a particular extractant to a specific soil constituent. Generally, the following forms of metals are distinguished: (1) dissolvable in soil solution, i.e. watersoluble metal cations, (2) exchangeable in inorganic and organic components (oxides/organic matter), (3) structural components of the lattices of soil minerals, and (4) insolubles precipitated with other soil components (Ahumada et al., 1999; Zalidis et al., 1999). The first two
226
M. Kaasalainen, M. Yli-Halla / Environmental Pollution 126 (2003) 225–233
forms can be considered readily available to plants, while the other two may become available in the long term. Harjavalta (61 200 N, 22 090 E) in SW Finland is dominated by its Cu and Ni smelters. Electric smelting of Cu commenced in 1945 and electrolytic refining of Ni in 1960. A sulphuric acid plant has operated in Harjavalta since 1947 using SO2 removed from the smelter flue gases as its raw material. In 1998 altogether 156,000 t of blister Cu, 43,000 t of Ni and 590,000 t of sulphuric acid were produced at the complex. The Harjavalta plants are the major point source of Cu and Ni emissions in Finland (Aunela and Larjava, 1990). Concentrations of Cu and Ni in mosses of Finland are by far the highest in and around Harjavalta (Kubin et al., 2000). During the growing season of 1988, the deposition of Cu, Ni and Cd measured 0.8 km from the smelter was 36, 20 and 18 times that measured in an unpolluted area (Yla¨ranta, 1996). According to Derome and Nieminen (1998), the annual bulk deposition in 1993–1998 comprised 1500 g Cu ha1 and 640 g Ni ha1 at a distance of 0.5 km from the smelter complex, while the amounts were only 27 g Cu ha1 and 19 g Ni ha1 at 8 km. Even though the emissions from the complex decreased to just one tenth their previous value towards the end of the 1990s (Melanen et al., 1999), the earlier load has mostly accumulated in the soil. The litter decomposition rate has decreased in the near-by forest as a result (Pennanen et al., 1996), and forest vegetation is suffering. Indeed, extremely high concentrations of total Cu (5800 mg kg1) and other heavy metals have been determined in the thin organic layer of the forest floor 0.5 km from the smelter, and in the acidic (pH about 4) sandy soils the metals are in a fairly mobile state (Derome and Lindroos, 1998). The smelters are situated near the Kokema¨enjoki river and are surrounded by cultivated fields. Cu, Ni and Cd accumulation into nearby agricultural soils has been demonstrated (Sippola and Ervio¨, 1986; Yla¨ranta, 1996). Cd is not produced by the smelter of Harjavalta, but it is present as a minor metallic constituent in the copper sulphide ore or other raw materials (Scoullos, 2000). Even though the agricultural soils are less acidic than the forest soils, abnormally high concentrations of metals, partly taken up from soil and partly from the airborne deposition, have been measured in crops at Harjavalta (Yla¨ranta, 1996). Farmers are aware of the metal deposition and want to know to what extent heavy metals accumulated in the soil endanger the quality of their farmland and crops. The main objective of this study was to document the speciation of Cd, Cu, Ni, Cr, Pb and Zn in three agricultural soils near the Cu–Ni smelter at Harjavalta, western Finland. The vertical distribution of the elements in these profiles was also of interest, to establish the possible downward movement of the metals. The metals were studied by an aqua regia procedure that
gives the pseudo-total metal concentration, and by a sequential extraction procedure that divides the metals into four operationally defined fractions. The extraction procedure was originally developed for sediment materials (Sahuquillo et al., 1999; Rauret et al., 1999), and considerable within-laboratory and between-laboratory variability in the heavy metal concentrations has been observed upon its application (Rauret et al., 2000). Soils are usually less homogeneous than sediments, having a more variable particle size distribution, and the suitability of the procedure for soil samples needs to be verified.
2. Materials and methods 2.1. Soil sampling Three cultivated soils at a distance of 0.8–2.2 km from the metal smelter were sampled by horizon (first three uppermost horizons; 0–70 cm), in autumn 1999 from soil pits. The soils were in agricultural use despite the Cu, Ni and Cd contamination observed in earlier studies (Sippola and Ervio¨, 1986; Yla¨ranta, 1996). Soil profiles were characterized (Table 1) according to Soil Taxonomy (Soil Survey Staff, 1998) and sampled according to visible horizon boundaries. The rainfall in 1999 was 662 mm and the average temperature 5.1 C. 2.2. Soil analyses The soil samples were air dried (35 C), screened to pass a 2-mm sieve. After digestion with H2O2 and dispersion with Na4P2O7.10 H2O, silt and clay were determined by a pipette method and sand by wet sieving (Elonen, 1971). Soil carbon (C) was determined with a LECO CN-2000 Elemental Analyzer. All C was assumed to be organic. Soil pH(H2O) was determined in a 1:2.5 soil/water suspension with a Beckman Instruments, Inc. FuturaTM refillable combination electrode. The soil suspension was allowed to stand overnight before pH determination. The cation exchange capacity (CEC) at pH 7.0 was calculated by summation of Ca, Mg, K, Na and titratable acidity, all extracted with 1 M ammonium acetate (CH3COONH4). Base saturation was calculated as Ca+Mg+K+Na expressed as percentage of CEC. Heavy metals were fractionated with the three-step extraction procedure developed within the Standards, Measurements and Testing Programme (formerly the Community Bureau of Reference, BCR) of the European Commission. This BCR procedure was originally developed for and used in the certification of heavy metal concentrations of a sediment reference material CRM 701 (Rauret et al., 1999; Pueyo et al., 2001). In the procedure, a sample is sequentially extracted with
227
M. Kaasalainen, M. Yli-Halla / Environmental Pollution 126 (2003) 225–233 Table 1 Characterization of selected physicochemical properties in the three horizons of the soilsa Horizon
Soil 1, Oxyaquic Cryorthent I
Plant cultivated Distance from the smelter complex/km Horizon Depth/cm Texture pH (H2O) CEC/cmol(+)kg1 Base saturation/% Organic C/% Clay ( <0.002 mm)/% Silt (0.002–0.06 mm)/% Sand (0.06–2 mm)/% a
II
III
Soil 2, Typic Cryaquept
Soil 3, Oxyaquic Cryorthent
I
I
Sugar beet; potato 0.8
II
III
Barley 1.2
II
III
Barley 2.3
Ap 0–25 Loamy sand
BC 25–63 Sand
2C 63–67 Silt loam
Ap 0–30 Silt loam
Bw1 30–47 Clay
Bw2 47–70 Silty clay (loam)
Ap 0–30 Silt loam
Bw1 30–45 Silt
Bw2 47–55 Silt loam
6.34 9.7 66 1.4 3 16 81
6.27 3.2 24 0.1 0 4 96
6.40 7.1 30 0.3 11 52 37
6.49 11.3 63 0.9 17 65 18
6.06 25.6 78 0.4 69 25 6
6.73 14.0 70 0.3 40 59 1
7.08 2.6 83 1.3 9 58 33
6.58 6.0 64 0.2 11 84 5
6.36 6.6 61 0.2 13 83 5
Soils were classified according to Soil Survey Staff (1998).
dilute acetic acid, a reducing agent and an oxidizing agent. As an additional step, the aqua regia procedure ISO 11466 (1995) was used to extract metals from the residue remaining after the three extraction steps. The four steps divide the trace metals into the following fractions: (1) exchangeable and specifically adsorbed (EX), (2) Fe and Mn (hydr)oxide-bound (OXI), (3) organically bound (ORG) and (4) residual (i.e. mineralbound; Res). The extraction procedure, chemical reagents and the experimental conditions are presented in detail in the flow chart (Fig. 1). It needs to be pointed out that the 0.11 M (about 0.5%) acetic acid that was used is less concentrated than that often used (2.5% in Elsokkary and La˚g, 1978; Iyengar et al., 1981) to extract specifically adsorbed metals. Three subsamples of every soil sample were extracted. The original soil samples were also analysed by the aqua regia procedure for their pseudo-total heavy metal concentrations. In Finnish soils, this aqua regia procedure dissolves about 95% of total Cu, Ni and Zn, about 80% of total Cd and about 60% of total Cr and Pb (Baghdady and Sippola, 1983a,b). 2.3. Elemental analysis and instrumentation The EX and ORG fractions of Cd, Cu, Ni, Cr, Pb and Zn were analysed with a Perkin-Elmer Sciex ELAN 6000 ICP-MS. The OXI and the aqua regia extractable Cu, Ni, Cr and Zn were analysed by TJA/Baird Division IRIS (II) ICP-AES and the corresponding fractions of Cd and Pb by Varian SpectrAA-400 with a GTA-96 graphite tube atomiser and Zeeman background correction. All element concentrations are presented on a dry matter basis.
2.4. Statistical methods Arithmetic means (AM) of the triplicate extraction results were calculated as well as the standard deviations (SD) and the coefficients of variation (CV, i.e. RSD relative standard deviation that is a ratio of the SD to AM value, expressed in %). The relationships between selected soil characteristics (clay, organic C, CEC, pH) and the metal fractions were studied by correlation analysis. Adequate information for the correlation analysis was obtained by expanding the data of the three soil profiles (n=9) with the results of an additional profile (n=4) from the same area. A detailed description of this soil will be presented in a later paper. All the mathematical operations were conducted with Microsoft1 Excel 2000.
3. Results and discussion 3.1. Soil properties As shown in Table 1, the soil profiles represented three different textural classes, varying from clay to sand. The texture differed between the horizons of the soil profiles because the horizons had been sedimented at different stages of the retreat of the continental ice at the end of the Weichselian glaciation ( 10,000 years ago). The clay content (0–69%) did not show any trend with depth. Because of the intensive liming, the pH in the plow layers was rather high (6.3–7.1) as compared to the average (< 6) in Finnish fields, and it was notably higher than in the forest soils of the area (about 4, Derome and Lindroos, 1998).
228
M. Kaasalainen, M. Yli-Halla / Environmental Pollution 126 (2003) 225–233
3.2. Pseudo-total concentrations of Cd, Cu, Ni, Cr, Pb and Zn The pseudo-total (aqua regia extractable) concentrations of the metals in the various soils and horizons are
set out in Table 2. The sandy soil 1, which was closest (0.8 km) to the smelter, exhibited the highest pseudototal concentrations of Cd, Cu, Ni and Pb in the plow layer even tough, among mineral soils, sandy soils tend to be lowest in heavy metals. There was a steady decline in the pseudo-total concentrations, especially of Cu and Ni, with increasing distance from the smelter. Also around a metal smelter in Ruston near Tacoma, Washington, USA, the concentration maxima of Cd, Cu and Zn in the surface soils were within 2–3 km of the smelter (Kuo et al., 1983). In Sudbury, Canada, however, Dudka et al. (1995) found that the spatial distribution of total Cu and Ni in soil was sometimes inconsistent with respect to distance from the smelter due to unevenness of the terrain, the dispersion effect of the stack and the influence of other sources of metal contamination in the area. The soils of our study probably represent the most loaded area, however, because in the study by Derome and Lindroos (1998) the Cu, Ni and Zn concentrations of the forest soils of Harjavalta showed a consistent decline along a 8-km transect from the smelter. In a recent survey (Ma¨ntylahti and Laakso, 2002), the mean aqua regia extractable metal concentrations in 182 coarse agricultural soils of Finland were as follows (mg kg1): Cd 0.10, Cu 13.9, Ni 6.7, Cr 17.8, Pb 7.8 and Zn 35.6. Measured against these values, the plow layers of all our soils were high in Cd, Cu, Ni and Zn, and soil 1 also in Pb. Compared with the values of another study on 10 diverse agricultural soils in Finland (Baghdady and Sippola, 1983a,b), however, the plow layers of our soils exhibited elevated concentrations of only Cu, while soil 1 was high in Ni and Cd as well. These differing results for Finnish soils mean that only the strongest contamination with Cu, Cd and Ni could be confirmed with certainty in our soils on the basis of pseudo-total analysis. In all three soils the concentrations of pseudo-total Cd and Cu were high in the plow layer relative to the deeper horizons. In soil 1 the same trend was observed in Ni, Pb and Zn. Also in soil 1, the decrease of Cd, Cu, and Ni from the plow layer to the second horizon was abrupt; most notably, the Cu concentration was 44-fold in the plow layer relative to the second horizon. In contrast to this, in soil 2 the concentrations of Cr, Ni and Zn were highest in the second and third horizons, which contained more clay than the loamy topsoil. 3.3. Fractionation of metals
Fig. 1. Schematic presentation of the extraction procedures applied to the soil samples. The extract was separated by centrifuging (3000g, 20 min) and dectantation. Steps 1, 2 and 3 all terminated with washing of the residue with 20 ml H2O.
Fig. 2 shows the distribution of the metals into the four fractions of the sequential extraction. In the plough layer, the proportions of the metals in the acetic acid extractable fraction (EX) followed the order Cr4Pb < Zn < Ni < Cu < Cd. Cadmium was particularly abundant in this most mobile fraction (66, 44 and 53% of the
0.085 0.011 97.6 11.7 0.744 103.3 13.5 0.525 102.3 27.0 1.010 104.2 7.12 0.289 103.4 46.7 1.904 100.5 0.0890.009 88.2 11.4 1.253 87.1 13.6 0.174 92.9 27.8 1.175 95.0 6.78 0.497 94.5 53.9 1.087 93.3 0.2740.002 96.5 52.40.411 93.8 16.70.924 95.6 18.70.960 99.6 9.790.488 96.2 61.00.318 95.8 0.1620.003 96.0 28.10.636 100.2 32.41.108 98.8 58.01.306 99.5 12.60.351 87.6 1051.789 97.9 0.0990.008 106.1 39.30.500 99.1 42.81.623 104.9 71.61.702 104.3 19.20.616 94.8 1321.730 100.3 0.2500.011 94.8 73.20.976 101.0 25.90.910 104.2 31.60.566 108.5 11.60.579 102.6 65.11.735 102.4 0.062 0.006 93.1 8.24 0.433 100.3 10.2 0.644 102.2 23.0 0.843 100.5 9.59 0.597 98.8 50.2 2.964 98.4 0.0440.007 97.7 6.07 0.134 107.2 4.07 0.158 93.4 5.03 0.392 104.7 3.97 0.217 89.9 13.6 0.570 92.7 0.6550.031 89.5 263 5.663 98.1 54.2 4.835 86.0 9.09 0.233 101.5 15.8 0.588 90.7 67.3 9.127 79.0 Zn
Pb
Cr
Ni
Cu
Cd
Pseudo-total concentration/mg kg (BCR/pseudo-total)/% Pseudo-total concentration/mg kg1 (BCR/pseudo-total)/% Pseudo-total concentration/mg kg1 (BCR/pseudo-total)/% Pseudo-total concentration/mg kg1 (BCR/pseudo-total)/% Pseudo-total concentration/mg kg1 (BCR/pseudo-total)/% Pseudo-total concentration/mg kg1 (BCR/pseudo-total)/%
III II I I
1
II
III I
II
III
Soil 3 Soil 2 Soil 1 Horizon
Table 2 The pseudo-total (aqua regia extractable) concentration of the metals (mg kg1) in the horizons (I–III) of the three soils (meanS.D., n=3)a and sum of the metals extracted by BCR procedure (EX+OXI+ORG+residual fraction) expressed as percentage of the pseudo-total concentration of the metals.
M. Kaasalainen, M. Yli-Halla / Environmental Pollution 126 (2003) 225–233
229
pseudo-total in the plow layer of soils 1, 2 and 3, respectively), in agreement with findings at other anthropogenically loaded sites (Elsokkary and La˚g, 1978; Kuo et al., 1983; Chlopecka et al., 1996; Sa´nchez et al., 1999; Li and Thorton, 2001). Although the absolute CdEX concentration was much lower in the deeper horizons, as proportion of the pseudo-total concentration it was still 56% in the second horizon of soil 1 closest to the smelter (in contrast to 14% in soil 2 and 22% in soil 3). In soil 1, the CuEX and NiEX fractions each represented about 30% in the plow layer and were at least equally as large in the horizon below; in the other two soils these fractions comprised 11–16% in the plow layer and only 1–6% deeper in the soil. Proportions of ZnEX were < 8% except in the first horizon of soil 1 (24%). Deposition can thus be considered the source of the elevated concentrations of the acetic acid extractable fraction of Cd, Cu and Ni all three soils and of Zn in soil 1. The large fractions of CdEX, CuEX and NiEX in horizon II show that these metals have been transported downwards in the sandy soil, a feature described earlier for heavy metals of anthropogenic origin (Karczewska, 1996). The CrEX and PbEX fractions were < 2% throughout the profiles. In all soils, the concentrations of CdOXI and CuOXI, likely associated with Fe/Mn oxides, were much higher in the plow layer than in the deeper horizons, showing that accumulated metals are found in this fraction as well. In soil 1 the concentration of PbOXI was 10 times as high in the plow layer as in the horizon below. This finding confirms our tentative conclusion of some contamination of soil 1 with Pb, made on the basis of the pseudo-total concentrations in our soils and concentrations in nearby forest soils (Derome and Lindroos, 1998). The accumulation of anthropogenic Pb in oxidebound form is in accordance with earlier findings (Chlopecka et al., 1996; Sa´nchez et al., 1999; Sahuquillo et al., 1999). It is noteworthy that, even in the deeper horizons where Pb likely was native, this element largely was bound by Fe/Mn (hydr)oxides. In contrast to the two more coarse-textured soils, the highest concentrations of PbOXI, NiOXI and ZnOXI in soil 2 occurred in the second horizon, where they probably are associated with the high clay content. All samples were low in CrOXI (5–7%). The fraction extracted after oxidizing with hydrogen peroxide may in some degree be associated with soil organic matter. The concentrations of CuORG in all soils and those of NiORG and ZnORG in soils 1 and 3 were substantially higher in the plow layer than in the deeper horizons, while the concentrations of CdORG and PbORG were more or less the same in all horizons, irrespective of the loading and the concentration of organic matter. The results for this fraction yielded little additional information to what had already been found on the basis of the two preceding fractions. Possibly part of the
230
M. Kaasalainen, M. Yli-Halla / Environmental Pollution 126 (2003) 225–233
organically bound metals had already been extracted in the second step as is suggested by findings made when this BCR procedure was developed (Sahuquillo et al., 1999): decrease of the pH (2.9 0.3!1.5 0.1) and increase in the concentration of NH2O.HCl (0.1!0.5 M) in the extraction of the OXI fraction increased the extraction of Cr by 647%, Cu by 825% and Pb by 450%, and the amounts in the subsequent fractions were decreased accordingly. These results strongly argue against the physical interpretation that metals extracted at this step truly represent the fraction associated with soil organic matter. The proportion of the residual metal fraction (Res), reflecting the native metal concentration in soil, increased with depth for every metal studied, in contrast to the proportion of the EX fraction. In the deepest horizon, the average proportions of the residual metals were, within a narrow range, as follows: Cd 25%, Cr 88%, Cu 70%, Ni 83%, Pb 45% and Zn 84%. The
somewhat low percentages of CdRes and PbRes in the subsoil probably do not reflect pollution with these metals but are native characteristics of the soil. 3.4. Correlations between Cd, Cu, Ni, Cr, Pb and Zn fractions and soil properties Significant correlations between Cd, Cu, Ni, Cr, Pb and Zn fractions and soil properties are presented in Table 3. Individual metal fractions did not correlate with pH and they correlated poorly with soil organic C, possibly because the ranges of these soil properties were narrow. The pseudo-total concentrations of Cu, Cd and Ni, which were elevated markedly by human activities, did not correlate with clay, while pseudo-total Cr, Pb and Zn, which are less affected by anthropogenic activities, correlated positively with the clay content. Residual fractions of all metals except Cd correlated positively with clay content.
Fig. 2. Distribution of Cd, Cu, Ni, Cr, Pb and Zn in sequentially extracted fractions of soils near Harjavalta Cu–Ni smelter. Abbreviations for the fractions and extractants: EX (exchangeable and specifically adsorbed) 0.11 M HOAc; OXI (bound to the oxides of Fe and Mn) 0.5 M NH2O HCl, pH 1.5; ORG (bound to soil organic matter) 8.8 M H2O2+1 M NH4Ac and Res (bound to mineral structures) aqua regia. I, II and III refer to the three uppermost horizons of each soil.
M. Kaasalainen, M. Yli-Halla / Environmental Pollution 126 (2003) 225–233 Table 3 All significant correlations (r values) between selected trace metal fractions (pseudo-total, EX, OXI, ORG, residual) and soil properties (n=13) Metal fraction
Clay
Organic C
CEC
Cdtot CdOXI
– –
0.59* 0.89***
– 0.62*
CuORG
–
0.68*
–
NiOXI Nires
– 0.91***
0.75** –
0.69** –
Crtot CrEX CrOXI CrORG Crres
0.97*** – 0.97*** 0.67* 0.96***
– – – – –
– – – 0.84* –
Pbtot PbORG Pbres
0.66* 0.69* 0.84***
– – –
– – –
Zntot ZnOXI ZnORG Znres
0.93*** 0.86*** – 0.96***
– – 0.65* –
– 0.66* 0.79** –
*, **, ***Significant correlations at the P40.1, 0.01 and 0.001 level, respectively.
231
mobility of heavy metals. In spite of a near-neutral pH in the studied soils, a substantial part of the Cd, Cu and Ni, and in soil 1 also Zn, occurred in the most mobile (exchangeable and specifically adsorbed) form. Thus, liming and tillage do not completely eliminate the risk of higher heavy metal uptake of crops in the area of Harjavalta despite the now low level of the airborne loading. Pseudo-total analysis indicated clear Cu, Ni and Cd pollution of the most loaded soil 1, closest to the smelter, and Cu pollution in the other two soils. The results of the sequential extraction showed that acetic acid extractable Cd and Ni were high in soils 2 and 3 as well as in soil 1. Pollution with Zn, accumulating in the acetic acid extractable form, and pollution with Pb, bound to Fe/Mn hydr(oxides), were demonstrated in soil 1 but not in soils further away from the smelter. Sequential extraction also revealed elevated concentrations of acetic acid extractable Cd, Cu and Cd below the plow layer in the most sandy soil 1, but quantitatively the risk of downward movement of these metals seems to be small. High pseudo-total concentrations of Cr, Ni, Pb and Zn below the plow layer in soil 2 were native, relating to the high clay content of the subsoil, and did not indicate downward movement of the metals.
3.5. Analytical accuracy 4. Conclusions No marked differences were observed between the pseudo-total metal concentrations extracted with aqua regia and the sum of metals extracted by the sequential extraction procedure (Table 2). Indicative of the good quality of the analytical work, the overall recovery rates (the sum of the four fractions expressed as percent of the independent pseudo-total concentration) ranged from 87 to 106% in all soils and horizons, except in a few cases where the concentrations in the subsoil were very low. 3.6. Metal status of cultivated soils at Harjavalta The concentrations of heavy metals in agricultural soils at Harjavalta were much lower than the extremely high values measured in a thin surface layer of forest soils (Derome and Lindroos, 1998). The explanation of the lower values is the nearly annual plowing of the soils, which has mixed the metals deposited onto the soil surface through the 25–30 cm thick layer. In forest soils, the undisturbed top layer is subjected to metal deposition year after year. Quantitatively, loading of forest and field has probably been the same, but in arable land, the metals have been diluted into a larger volume of soil. Agricultural soils are commonly less acidic, finer in texture, and thus richer in chemically active soil constituents, all of these being factors that reduce the
1. In the plow layers of the studied soils the mobility of the metals increased approximately in the order Cr < Zn < Pb < Ni4Cu < Cd. Cadmium occurred more abundantly in soluble form than did the other elements. Cd, Cu and Ni appeared in mobile forms to a particularly large extent in the surface layer and also in the second horizon of the most polluted, sandy soil (soil 1). 2. The BCR procedure gave reproducible results in surface soil high in several metals and in subsoil low in metals. Application of the procedure allowed reliable assessment of pollution in soils differing in particle size distribution. Some doubt was raised about the physical nature of the fraction extracted with hydrogen peroxide, however. 3. The BCR sequential extraction can assist in identifying heavy metal contamination of soil and possible downward movement of metals. It is most useful in soils with texturally different horizons at sites where metal pollution is only minimal. If remediation is carried out in time, the detrimental effects of heavy metals can be minimized. Aqua regia extractable concentrations indicate metal pollution at a more advanced stage.
232
M. Kaasalainen, M. Yli-Halla / Environmental Pollution 126 (2003) 225–233
Acknowledgements We thank Ms. Leila Lindstedt and Ms. Pa¨ivi Alle´n for technical assistance and carrying out the laboratory work. Outokumpu Harjavalta Metals Oy funded the experimental phase of the study and the Finnish Cultural Foundation and the Outokumpu Foundation supported the publication of the results. We are grateful for this assistance. Our thanks, too, to the farmers of Harjavalta who participated in the study.
References Ahumada, I., Mendoza, J., Navarrete, E., Ascar, L., 1999. Sequential extraction of heavy metals in soils irrigated with wastewater. Communications in Soil Science and Plant Analysis 30, 1507–1519. Aunela, L., Larjava, K., 1990. Raskasmetallipa¨a¨sto¨t Suomessa. (Heavy metal emissions in Finland), Technical Research Centre of Finland. Research Notes 111, 1–78 (in Finnish). Baghdady, N.H., Sippola, J., 1983a. Total heavy metal recovery by aqua regia in soils of different origins. Annales Agriculturae Fenniae 22, 175–185. Baghdady, N.H., Sippola, J., 1983b. Efficiency of aqua regia in extracting Cd, Cr, Hg, Ni and Pb from soils of different origins. Annales Agriculturae Fenniae 22, 240–244. Chlopecka, A., Bacon, J.R., Wilson, M.J., Kay, J., 1996. Forms of cadmium, lead, and zinc in contaminated soils from southwest Poland. Journal of Environmental Quality 25, 69–79. Dang, Z., Liu, C., Haigh, M.J., 2002. Mobility of heavy metals associated with the natural weathering of coal mine spoils. Environmental Pollution 118, 419–426. Derome, J., Lindroos, A.-J., 1998. Effects of heavy metal contamination on macronutrient availability and acidification parameters in forest soil in the vicinity of the Harjavalta Cu–Ni smelter, SW Finland. Environmental Pollution 99, 225–232. Derome, J., Nieminen, T., 1998. Metal and macronutrient fluxes in heavy-metal polluted Scots pine ecosystems in SW Finland. Environmental Pollution 103, 219–228. Dudka, S., Ponce-Hernandez, R., Hutchinson, T.C., 1995. Current level of total element concentrations in the surface layer of Sudbury’s soils. The Science of the Total Environment 162, 161–171. Elonen, P., 1971. Particle size analysis of soil. Acta Agralia Fennica 122, 1–122. Elsokkary, I.H., La˚g, J., 1978. Distribution of different fractions of Cd, Pb, Zn, and Cu in industrially polluted and non-polluted soils of Odda region, Norway. Acta Agriculturae Scandinavica 28, 262– 268. Grant, C.A., Bailey, L.D., Therrien, M.C., 1996. The effect of N, P and KCl fertilizers on grain yield and Cd concentration of malting barley. Fertilizer Research 45, 153–161. ISO 11466, 1995. International Standard, Soil Quality—extraction Of Trace Elements Soluble in aqua regia. Iyengar, S.S., Martens, D.C., Miller, W.P., 1981. Distribution and plant availability of soil zinc fractions. Soil Science Society of America Journal 45, 735–739. Kabata-Pendias, A., Pendias, H., 2000. Trace Elements in Soil and Plants, third ed. CRC Press, Boca Raton, FL. Karczewska, A., 1996. Metal species distribution in top- and sub-soil in an area affected by copper smelter emissions. Applied Geochemistry 11, 35–42. Kubin, E., Lippo, H., Poikolainen, J., 2000. Heavy metal loading. In: Ma¨lko¨nen, E. (Ed.), Forest Condition in a Changing Environment— the Finnish Case. Kluwer Academic, The Netherlands, pp. 60–71.
Kuo, S., Heilman, P.E., Baker, A.S., 1983. Distribution and forms of copper, zinc, cadmium, iron, and manganese in soils near a copper smelter. Soil Science 135, 101–109. Li, X., Thornton, I., 2001. Chemical partitioning of trace and major elements in soils contaminated by mining and smelting activities. Applied Geochemistry 16, 1693–1706. Ma¨ntylahti, V., Laakso, P., 2002. Arsenic and heavy metal concentrations in agricultural soils in South Savo province. Agricultural and Food Science in Finland 11, 285–300. Melanen, M., Ekqvist, M., Mukherjee, A.B., Aunela-Tapola, L., Verta, M., Salmikangas, T., 1999. Raskasmetallien pa¨a¨sto¨t ilmaan Suomessa 1990-luvulla (Emissions of heavy metals in Finland in the 1990s). Suomen ympa¨risto¨ 329, 1–92 (in Finnish). Mokma, D.L., Yli-Halla, M., Hartikainen, H., 2000. Soils in a young landscape on the coast of southern Finland. Agricultural and Food Science in Finland 9, 291–302. Pennanen, T., Frostega˚rd, A˚., Fritze, H., Ba˚a˚th, E., 1996. Phospholipid fatty acid composition and heavy metal tolerance of soil microbial communities along two heavy metal polluted gradients in coniferous forests. Applied Environmental Microbiology 62, 420– 428. Pueyo, M., Rauret, G., Lu¨ck, D., Yli-Halla, M., Muntau, H., Quevauviller, Ph., Lo´pez-Sa´nchez, J.F., 2001. Certification of the extractable contents of Cd, Cr, Cu, Ni, Pb and Zn in a freshwater sediment following a collaboratively tested and optimised three-step sequential extraction procedure. Journal of Environmental Monitoring 3, 243–250. Rauret, G., Lo´pez-Sa´nchez, J.F., Sahuquillo, A., Rubio, R., Davidson, C., Ure, A., Quevauviller, Ph., 1999. Improvement of the BCR three step sequential extraction procedure prior to the certification of new sediment and soil reference materials. Journal of Environmental Monitoring 1, 57–61. Rauret, G., Lo´pez-Sa´nchez, J.F., Sahuquillo, A., Barahona, E., Lachica, M., Ure, A.M., Davidson, C.M., Gomez, A., Lu¨ck, D., Bacon, J., Yli-Halla, M., Muntau, H., Quevauviller, Ph., 2000. Application of a modified BCR sequential extraction (three-step) procedure for the determination of extractable trace metal contents in sewage sludge amended reference material (CRM 483), complemented by a three-year stability study of acetic acid and EDTA extractable metal content. Journal of Environmental Monitoring 2, 228–233. Rudd, T., Campbell, J.A., Lester, J.N., 1986. Characterization of metal forms in sewage sludges by chemical extraction. In: Lester, J.N., Perry, R., Sterrit, R.M. (Eds.), Chemicals in the Environment. Selper Ltd, London, pp. 756–771. Sahuquillo, A., Lo´pez-Sa´nchez, J.F., Rubio, R., Rauret, G., Thomas, R.P., Davidson, C.M., Ure, A.M., 1999. Use of a certified reference material for extractable trace metals to assess sources of uncertainty in the BCR three-stage sequential extraction procedure. Analytica Chimica Acta 382, 317–327. Sa´nchez, G., Moyano, A., Mun˜ez, C., 1999. Forms of cadmium, lead, and zinc in polluted mining soils and uptake by plants (Soria Province, Spain). Communications in Soil Science and Plant Analysis 30, 1385–1402. Scoullos, M., 2000. Data and trends on cadmium. Synopsis prepared as a background document for the International EUPHEMET’s Workshop (Toward an Integrated EU Policy for Heavy Metals), 17–18 April 2000, Athens, Greece. Sippola, J., Ervio¨, R., 1986. Raskasmetallit maapera¨ssa¨ ja viljelykasveissa Harjavallan tehtaiden ympa¨risto¨ssa¨ (Heavy metals in soils and cultivated plants in the Harjavalta smelter area). Ympa¨risto¨ ja terveys 5, 270–275 (in Finnish). Soil Survey Staff. 1998. Keys to Soil Taxonomy, eighth ed. USDA– NRCS. Tessier, A., Campbell, P.G.C., Bisson, M., 1979. Sequential extraction procedure for the speciation of particulate trace metals. Analytical Chemistry 51, 844–851.
M. Kaasalainen, M. Yli-Halla / Environmental Pollution 126 (2003) 225–233 Yla¨ranta, T., 1996. Uptake of heavy metals by plants from airborne depositions and polluted soils. Agricultural and Food Science in Finland 5, 431–447.
233
Zalidis, G., Barbayiarinis, N., Matsi, T., 1999. Forms and distribution of heavy metals in soils of the Axios Delta of Northern Greece. Communications in Soil Science and Plant Analysis 30, 817–827.