pH-Dependent Release of Cadmium, Copper, and Lead from Natural ...

27 downloads 0 Views 123KB Size Report
pH-Dependent Release of Cadmium, Copper, and Lead from Natural and Sludge-Amended Soils. Orathai Sukreeyapongse, Peter E. Holm, Bjarne W. Strobel, ...
pH-Dependent Release of Cadmium, Copper, and Lead from Natural and Sludge-Amended Soils Orathai Sukreeyapongse, Peter E. Holm, Bjarne W. Strobel, Supamard Panichsakpatana, Jakob Magid, and Hans Christian Bruun Hansen* ABSTRACT

organically complexed forms of heavy metals could account for migration of substantial fractions of heavy metals out of the surface soil several decades after sludge application. Several studies have demonstrated that pH is the most influential factor controlling sorption–desorption of heavy metals in soils (e.g., Gerritse and van Driel, 1984; Anderson and Christensen, 1988; Buchter et al., 1989; Naidu et al., 1994; Filius et al., 1998). Sauve´ et al. (2000) recently reviewed and compiled data from more than 70 such studies and found by multiple linear regression analysis that approximately 50% of the overall variation in the distribution coefficient for Cd and Pb could be explained by variations in pH. For Cu only 30% of the total variation could be ascribed to pH, probably due to the influence of Cu complexation by dissolved organic matter (Sauve´ et al., 2000). The steep increase in heavy metal sorption with rising pH is usually illustrated by means of sorption edges (Stumm, 1992). The location of the edges depends on the tendency of the cation to hydrolyze and the type of sorbent. For the cations studied here the affinity sequence of Pb ⬎ Cu ⬎ Cd is usually observed (McBride, 1989; Yong and Phadungchewit, 1993; Gao et al., 1997). Selective extractions have been used to characterize the distribution of heavy metal cations between different pools in soils (Beckett, 1989; Fo¨rstner, 1991). However, selective fractionations do not quantify the actual lability or leachability of heavy metals in the different pools. A useful approach for comparison of heavy metal bonding and availability, circumventing the problem of attaining equilibrium and the often difficult interpretations of selective fractionation data, is to compare how fast metal cations can be released at constant pH, a procedure often adopted in weathering studies. Release of metal cations due to weathering of soil minerals often follows simple rate laws. In acid solutions, the change in dissolution rates as a function of pH is usually described by the expression (Wieland et al., 1988; Drever, 1994; Lasaga et al., 1994):

The pH-dependent release of cadmium, copper, and lead from soil materials was studied by use of a stirred flow cell to quantify their release and release rates, and to evaluate the method as a test for the bonding strength and potential mobility of heavy metals in soils. Soil materials from sludge-amended and nonamended A horizons from a Thai coarse-textured Kandiustult and a Danish loamy Hapludalf were characterized and tested. For each soil sample, release experiments with steady state pH values in the range 2.9 to 7.1 and duration of 7 d were performed. The effluent was continuously collected and analyzed. Release rates and total releases were higher for the Hapludalf than the Kandiustult and higher for the sludge-amended soils than the nonamended soils. With two exceptions the relative release rates (release rate/total content of metal in soil) plotted vs. steady state pH followed the same curves for each metal, indicating similar bonding strengths. These curves could be described by a rate expression of the form: relative release rate ⫽ k[Hⴙ]a, with specific a (empirical constant) and k (rate constant) parameters for each metal demonstrating that metal release in these systems can be explained by proton-induced desorption and dissolution reactions. With decreasing pH, pronounced increases in release rates were observed in the sequence cadmium ⬎ lead ⬎ copper, which express the order of metal lability in the soils. The flow cell system is useful for comparison of metal releases as a function of soil properties, and can be used as a test to rank soils with respect to heavy metal leaching.

T

he use of phosphate fertilizers, application of sewage sludge and manure, and atmospheric deposition are the main inputs of heavy metals to agricultural soils. Losses due to leaching and assimilation by plants and microorganisms are key issues in assessing the environmental risk of heavy metals in the terrestrial environment (Dowdy et al., 1991; Lorenz et al., 1998; Berti and Jacobs, 1998; Giller et al., 1998; Holm et al., 1998). The mass balances of heavy metals in soil are strongly dependent on biogeochemical processes such as sorption and dissolution. These processes are affected by the type of sorbents and heavy metal pools present, pH, redox, concentrations of coadsorbates, and the concentration of complexing ligands in solution. For soils receiving sewage sludge, the high content of humic material and amorphous oxides in the sludge may be important for controlling metal availability in the soil (Jing and Logan, 1992; Hyun et al., 1998, Luo and Christie, 1998). McBride et al. (1997) suggested that mobile

release rate ⫽ k[H⫹]a

[1]

or in logarithmic form: log (release rate) ⫽ ⫺a pH ⫹ log k

[2]

where k is a rate constant and a an empirical constant. When pH is kept constant, release rates are also constant, corresponding to zero-order kinetics. Generally, the techniques to test metal solubility, leaching, and relative mobility comprise extraction–

O. Sukreeyapongse and S. Panichsakpatana, Dep. of Soil Science, Kasetsart University, Phaholyathin, Chatujak, Bangkok, 10900, Thailand. P.E. Holm, B.W. Strobel, and H.C.B. Hansen, Chemistry Dep., and J. Magid, Dep. of Agricultural Sciences, The Royal Veterinary & Agricultural University, Thorvaldsensvej 40, DK-1871 Frederiksberg C, Denmark. Received 16 Oct. 2001. *Corresponding author (haha@ kvl.dk).

Abbreviations: CEC, cation exchange capacity; DOC, dissolved organic carbon.

Published in J. Environ. Qual. 31:1901–1909 (2002).

1901

1902

J. ENVIRON. QUAL., VOL. 31, NOVEMBER–DECEMBER 2002

batch experiments and flow cell systems (van der Sloot et al., 1997; Raulund-Rasmussen et al., 1998; Pang and Close, 1999; Jensen et al., 2000). Stirred flow cells are well suited for continuous monitoring of metal release reactions (Knauss and Wolery, 1986; Strobel et al., 2001). In flow cell systems, the solution composition changes gradually and continuously until steady state is reached, in contrast to shaking flask techniques where abrupt changes in solution composition may take place. Furthermore, the continuous renewal and dilution of the solution in flow cells prevents formation of secondary precipitates, allowing for measurement of maximum release rates (Kraemer and Hering, 1997). Finally, the effects of change in influent composition, such as change in pH, and type and concentration of cations and ligands, are easily studied with flow-through systems. In this work, release experiments conducted at constant pH were used to characterize the release of Cd, Cu, and Pb in two widely different sewage sludge– amended and nonamended soils. Stirred flow cells were used to quantify the release and relative release rates of Cd, Cu, and Pb, and to characterize the strength of metal bonding. Relationships between metal release rates and pH for the two different soil types are used to test if mineral dissolution kinetics apply. Finally, we discuss how stirred flow cell experiments can be used as a test to rank soils with respect to heavy metal leaching. MATERIALS AND METHODS Soil Samples Four topsoil (0- to 20-cm depth) samples were used in this study. Two soil samples classified as fine-loamy, siliceous, isohyperthermic Typic Kandiustults (Soil Survey Staff, 1997) were collected from lysimeter experiments in Bangkok, Thailand. Thai Soil Sample 1 (T1) was from a control lysimeter experiment, which had not received sludge. Thai Soil Sample 2 (T2) was from a lysimeter having received one amendment of sewage sludge, equivalent to 9.9 Mg (dry matter) per ha (104 m2). The sludge treatment was performed in January 1999 and the soils sampled April 1999 with stainless steel tubes. Two soil samples classified as sandy loam, mixed, mesic Typic Hapludalf (Soil Survey Staff, 1997) were collected from a long-term field experiment at Askov Research Station, Vejen, Denmark (Larsen and Petersen, 1993). Danish Soil Sample 1 (D1) was from a reference plot, which had not received sludge. Danish Soil Sample 2 (D2) had received industrial sewage sludge at a yearly rate of 21 Mg (dry matter) per ha during the period from 1974 to 1979. In December 1997, 18 yr after end of the yearly application of sludge, the soils were sampled to a depth of 0.25 m with a stainless steel auger. The Danish soil samples (D1 and D2) were air-dried and stored in a soil archive until December 1999, where subsamples of 500 g were obtained by passing the soil samples through an 8-mm stainless steel sieve.

Soil Characterization Soil samples were air-dried, gently crushed, and passed through a 2-mm stainless steel sieve. The pH in water and 10 mM CaCl2 was measured potentiometrically in 1:2.5 (w/w) soil to solution suspensions after equilibration for 1 h. Particle size distribution was determined by the pipette method (Andreasen, 1939). Cation exchange capacity (CEC) was deter-

mined by extraction with 1 M NH4NO3 (Stuanes et al., 1984). The oxalate (ox)-extractable Al, Fe, and Mn were determined according to Schwertmann (1964), and the citrate–bicarbonate–dithionite (CBD)-extractable Al, Fe, and Mn were determined according to Mehra and Jackson (1960). A subsample of 10 g of each soil sample was ground to a particle diameter less than 0.5 ␮m in a McCrone (London, UK) micronizing mill, and total organic carbon determined by dry combustion with Eltra (Neuss, Germany) CS500 carbon sulfur determinator. The total content of Cd, Cu, and Pb was determined by digesting ground soil samples with aqua regia and hydrofluoric acid followed by determination of the metals by graphite furnace atomic absorption spectroscopy (GFAAS) (Perkin Elmer [Wellesley, MA] 5100, Zeeman 5100) (Hossner, 1996).

Flow Cell Release Experiments Release of Cd, Cu, and Pb from the four soil samples was investigated in a series of stirred flow cell experiments. The stirred flow cells are made of polycarbonate and hold a volume of 94 mL. The inlet is placed at the bottom of the reactor and the outlet at the top where a 0.45-␮m RC membrane filter (Sartorius, Goettingen, Germany), supported by a teflon packing, retains particles inside the reactor. In all flow cell experiments, 5.0 g of soil was used and kept suspended by magnetic stirring at 180 rpm. The design is shown and details given in Strobel et al. (2001). Influent solutions were 10⫺3 M CaCl2 adjusted to the desired levels of pH in the range 2.9 to 7.1 with small amounts of HCl and NaOH. The CaCl2 electrolyte establishes an almost constant ionic strength and keeps an approximately constant concentration of Ca2⫹ comparable with the concentrations in the natural soil. The Cl⫺ concentrations of 2 ⫻ 10⫺3 M are higher than usually observed for soil solutions and may result in about 10% of Cd being present as CdCl⫹ complexes in solution as determined by a MINTEQA2 equilibrium computation (Allison et al., 1991). The slightly increased solution concentration of Cd due to complexation by chloride was not thought to affect the release kinetics of cadmium to any measurable extent. Chloride was preferred as the electrolyte anion compared with NO3⫺ or ClO4⫺, which may result in either enhanced biological activity in the flow cell experiments, or in clogging of salt bridge junctions of pH electrodes. The influent solution was continuously pumped through the reactor with a flow rate of approximately 1.6 mL h⫺1 applied with a peristaltic pump (Alitea-XV) and Aliprene tubings (Alitea, Stockholm, Sweden). The flow cell experiments were run continuously for 7 d at ambient room temperature of 22⬚C. The outlet was directed through teflon tubings to a fraction collector with 10-mL polypropylene vials. Ten milliliters of effluent was collected at time zero and successive effluent fractions of approximately 2.7 mL were collected every 100 min, resulting in 100 fractions for each experimental run. The actual amount of effluent collected in each vial was determined by weighing. Sufficient solution volume for heavy metal analysis was achieved by mixing two subsequent vials. For each set of five successive vials, the first two (e.g., Vials 1 and 2 or 6 and 7) were combined and analyzed for metals, and the next three (e.g., Vials 3–5 or 8–10) were combined and pH and dissolved organic carbon (DOC) determined. The vials for collection of samples for metal analysis had 0.2 mL of 2% (v/v) HNO3 added beforehand. The acidified effluent samples were analyzed for Cd, Cu, and Pb by GFAAS. The detection limits of the heavy metals were 0.02 ␮g Cd L⫺1, 0.09 ␮g Cu L⫺1, and 0.16 ␮g Pb L⫺1. The pH in the effluent samples was determined potentiometrically with a Metrohm (Herisau, Switzerland) electrode

1903

SUKREEYAPONGSE ET AL.: pH-DEPENDENT METAL RELEASE FROM SOILS

(6.0222.100) and Metrohm 691 pH meter. The DOC in the effluent was quantified with a total organic carbon analyzer (Shimadzu [Kyoto, Japan] TOC-500). In total, each of the 26 flow cell experiment comprised analysis of five analytes (pH, TOC, Cd, Cu, Pb) in 20 solution samples, giving a total of 100 determinations per experiment. All chemicals were suprapure grade. From the measurements of the effluent metal concentrations and the flow rates, accumulated releases, Racc (mol g⫺1) of the metals were calculated based on the expression developed by Strobel et al. (2001). For n ⱖ 6 the following formula is used:

Racc ⫽

96



n⫽6,11,16,... 4

冤兺 V i⫽0

1 5.00

n⫺1

·

(Cn;n⫹1 ⫹ Cn⫺5;n⫺4) ⫹ 2

(Cn;n⫹1 ⫺ Cn-5;n⫺4) · 92

[3]



In this expression, Vn⫺i is the actual volume (L) of each fraction (2.8 mL ⫾ 10%), and Cn,n⫹1 (mol L⫺1) is the cation concentration in the combined vials number n and n ⫹ 1. In the bracket, the first term refers to the effluent content and the second term to the change in the amount of metal cation of the solution in the flow cell (92 mL, i.e., a total cell volume of 94 mL minus approximately 2 mL taken up by soil solids). Another formula is used for the initial sample from each experiment (Strobel et al., 2001). The accumulated release of a metal is the total amount of metal released per mass of soil at a specified time, and plots of accumulated release versus time can be produced. In the calculations, it is assumed that the reactor is fully stirred and that the concentrations determined in the effluent represent the concentrations in the reactor. The initially released amounts and the average release rates of the metals were estimated from the intercepts and slope of linear regressions of accumulated release vs. time data.

RESULTS Soil Characteristics The Thai soil samples represent old, highly weathered soils with coarse texture, low content of organic matter, low CEC, and a moderate content of Fe and Al oxides (Table 1). In contrast, the Danish soils are young, containing three to five times as much clay and organic matter, and the CEC and content of Fe and Al oxides are much higher than in the Thai soils. The pH of the Danish soils is approximately 0.5 units higher than the Thai soils. Table 1 shows that the nonamended Thai soil (T1) has a low content of heavy metals (Cd: 0.80, Cu: 73, Pb: 35 ␮mol kg⫺1) compared with the nonamended Danish soil (D1) (Cd: 2.3, Cu: 108, Pb: 67 ␮mol kg⫺1), but both soils have heavy metal contents within the range reported for nonpolluted soils (Tjell and Hovmand, 1978; Alloway, 1995). Sludge amendment of the Thai soil (T2) caused the Cd content to double and the content of Cu to increase by 50%, whereas the content of Pb only increased slightly (Table 1). The contents of Cd, Cu, and Pb in the Danish sludge-amended soil (D2) increased 2.5 to 5 times compared with the nonamended soil (D1), reflecting the high doses of sludge applied.

Table 1. Selected soil characteristics for A horizons of the Thai (T1, T2) and Danish (D1, D2) soils used in the release experiments. Parameter Texture ⬍2 ␮m, % 2–50 ␮m, % 50–250 ␮m, % 250–2000 ␮m, % Total carbon, % pH (in water suspension) CECpH 7†, cmol kg⫺1 AlCBD‡, mmol kg⫺1 AlOx§, mmol kg⫺1 FeCBD‡, mmol kg⫺1 FeOx§, mmol kg⫺1 MnCBD‡, mmol kg⫺1 MnOx§, mmol kg⫺1 CdTotal¶, ␮mol kg⫺1 CuTotal¶, ␮mol kg⫺1 PbTotal¶, ␮mol kg⫺1

T1

T2

D1

D2

4 12 67 17 0.25 6.0 0.9 7.0 4.1 12.8 4.1 1.7 1.5 0.80 73 35

3 9 71 17 0.36 6.3 1.1 8.4 5.1 13.4 5.1 1.7 1.5 1.6 110 39

13 21 40 26 1.5 6.8 7.3 51 46 52 35 2.9 2.4 2.3 108 67

15 20 40 25 1.8 6.8 8.5 52 50 58 48 3.6 3.0 7.0 513 165

† Cation exchange capacity determined at pH 7 (Stuanes et al., 1984). ‡ CBD, citrate–bicarbonate–ditionite extractable (Mehra and Jackson, 1960). § Ox, oxalate extractable (Schwertmann, 1964). ¶ Total, total content extracted with aqua regia and hydrofluoric acid (Hossner, 1996).

Release Experiments Examples of the effluent composition with respect to pH, flow rate, and concentrations of DOC, Cd, Cu, and Pb vs. time are shown in Fig. 1. A steady state pH measured in the effluent was established in the flow cells after a period of 1.5 to 30 h, longest for the Danish soils, which have the highest buffering capacities. The steady state pH values for all experiments are summarized in Table 2. Another measure of the soil buffering capacity, the difference between influent pH and effluent steady state pH, is also much higher for the Danish than the Thai soils (data not shown). The flow rate during an experiment was almost constant and varied within the range of 1.25 to 1.89 mL h⫺1 between individual experiments. The DOC concentrations in the effluents from the Danish soils typically were in the range 0.1 to 0.2 mM, which was two to three times higher than observed for the Thai soils (Fig. 1). The concentration of DOC in the effluent decreased slightly over time, although this decrease was not statistically significant. The Thai soils showed a tendency toward higher DOC concentrations in experiments run at higher pH whereas the Danish soils tended to have a minimum DOC concentration around pH 4 (data not shown). Effluents from the sludge-treated samples did not contain higher DOC concentrations than those from untreated samples. Metal Concentrations in Effluent Effluent metal concentrations were in general higher for the Danish soils than for the Thai soils, which also had more observations below the detection limits, especially for Pb. The highest effluent concentrations were all obtained in sample D2 run at the lowest pH (3.1) with initial Cd, Cu, and Pb concentrations of 0.88, 12, and 10 ␮g L⫺1, respectively, and maximum concentrations of 6.07, 28.0, and 12.4 ␮g L⫺1, respectively (data not shown). The lowest concentrations were observed

1904

J. ENVIRON. QUAL., VOL. 31, NOVEMBER–DECEMBER 2002

Table 2. Experimental conditions and results for the release experiments with Thai (T1, T2) and Danish soils (D1, D2). Cd release Soil

T1

T2

D1

D2

pH†

Rate¶

3.0 3.3 3.9 5.0 6.2 6.5 3.0 3.4 4.1 5.7 6.5 6.6 2.9 3.5 3.7 4.3 5.0 6.1 7.0 3.1 3.3 3.6 4.0 5.7 6.1 7.1

fmol g ⫺ 1 s⫺ 1 0.009 –# 0.043 0.029 – – 0.034 0.046 0.018 0.069 – – 1.65 0.90 0.92 0.35 0.087 0.042 0.030 5.54 3.03 2.10 0.41 0.27 0.15 0.12

Initial¶ nmol g–1 0.044 – 0.024 0.002 – – 0.11 0.11 0.03 0.01 – – 0.28 0.06 0.01 0 0.017 0.028 0.009 ⫺0.14 ⫺0.18 0.07 0.06 0.05 0.04 0.02

Cu release r

2

0.87 – 0.82 0.95 – – 0.51 0.44 0.96 0.97 – – 0.93 0.99 0.99 0.98 0.99 0.94 0.80 0.98 0.98 0.96 0.98 0.99 0.97 0.96

Rate¶

Initial¶

fmol g ⫺ 1 s⫺ 1 4.64 3.67 0.74 – 0.50 – 22.80 20.60 6.77 0.73 1.62 0.66 5.88 0.96 0.94 0.53 0.18 0.19 0.35 32.2 16.3 13.7 5.2 4.9 3.1 3.7

nmol g⫺1 0.59 1.01 0.22 – 0.25 – 4.99 1.66 ⫺0.14 0.33 0.37 0.21 ⫺0.11 0.24 1.33 0.51 0.20 0.27 0.33 ⫺0.3 1.7 5.6 2.2 1.7 1.9 1.4

Pb release r

2

0.96 0.93 0.82 – 0.77 – 0.94 0.99 0.95 0.95 0.97 0.87 0.99 0.95 0.97 0.96 0.61 0.63 0.53 0.98 0.99 0.99 0.97 0.98 0.97 0.98

Rate¶

Initial¶

fmol g⫺1 s⫺1 0.29 0.17 – – – – 0.79 0.42 – – – – 3.37 0.39 0.31 0.11 – – – 3.46 1.14 0.91 0.50 – – –

nmol g⫺1 ⫺0.02 0.04 – – – – 0.03 0.01 – – – – ⫺0.32 ⫺0.01 0.07 0.03 – – – 0.01 0.20 0.71 0.19 – – –

Accumulated release‡ r

2

Cd

Cu nmol

0.92 0.84 – – – – 0.97 0.93 – – – – 0.98 0.90 0.95 0.91 – – – 0.93 0.97 0.95 0.97 – – –

0.045 – 0.044 0.018 – – 0.13 0.12 0.12 0.049 – – 1.1 0.56 0.54 0.21 0.069 0.054 0.025 3.0 1.7 1.4 0.29 0.21 0.12 0.087

Pb

g ⫺1

3.4 3.1 0.60 – 0.50 – 17.3 13.1 4.3 0.75 1.32 0.53 3.4 0.81 1.9 0.81 0.30 0.36 0.50 19.9 11.3 13.9 5.0 4.3 3.6 3.5

0.17 0.16 – – – – 0.42 0.28 – – – – 1.7 0.26 0.27 0.086 – – – 2.5 0.94 1.3 0.48 – – –

Initial release§ Cd

Cu

Pb

% of accumulated release at 160 h 90 18 – – 32 29 49 34 – 11 – – – 46 – – – – 85 28 6 81 12 4 23 – – 20 44 – – 28 – – 36 – 23 – – 10 30 – 2 71 28 – 63 32 25 66 – 54 71 – 34 62 – – – 0 – 15 23 6 42 58 20 42 40 24 38 – 32 52 – 22 40 –

† pH in effluent solution at steady state. ‡ Accumulated release observed after 160 h. § Initial release (Columns 4, 7, 10) as percentage of accumulated release calculated from release parameter (Columns 3 ⫹ 4, 6 ⫹ 7, and 9 ⫹ 10). ¶ Rate and initial release determined from slope and intercept of regression line in accumulated metal vs. time curves. # Concentration is below detection limit.

Fig. 1. Examples of change in pH (䉫), flow rate (⫹), and dissolved organic carbon (DOC) concentration (⫻) vs. time and the corresponding variation in effluent concentrations for Cd (䉱), Cu (䊉), and Pb (䊐) in release experiments with Danish (a, c ) and Thai soils (b, d ) (Experiment D1 run at pH 3.5 and T1 at pH 3.0).

1905

SUKREEYAPONGSE ET AL.: pH-DEPENDENT METAL RELEASE FROM SOILS

for the Thai soils at high pH. The initial concentrations of Cd generally decreased with increasing pH, whereas the highest initial Cu concentrations were observed at pH between 3.0 and 3.5. Initial concentrations of Pb were always low. The change in effluent concentrations of Cd, Cu, and Pb over time varied with the type of soil sample, the pH, and the metal studied (data not shown). Generally, effluent concentrations increased with time in experiments run at low pH and decreased with time in experiments at higher pH. An exception to this pattern was observed for Cd in Thai soils where solution concentrations decreased over time at low pH but became almost constant or increased with time in experiments at pH above 3.5. Constant metal concentrations vs. time were only obtained in experiments at higher pH and were most often observed for Cd. Increases in released metal concentrations due to sludge amendment were clearly expressed for Cd and Cu in the Danish soils and for Cu in Thai soils. Accumulated Release and Release Rates The accumulated release vs. time curves became more or less linear after an initial period of 1.5 to 40 h, a period that was shorter for the Thai than the Danish soils. Linear regressions of the linear segments in the accumulated release vs. time diagrams were calculated to estimate the release rates and the initial releases represented by the slope and the intercepts of the regression lines, respectively (Fig. 2). Linear regressions were satisfactory with regression coefficients (r2) in the range 0.80 to 0.99, except for a few experiments where the slopes were very small (Table 2). Table 2, which summarizes the results for the release experiments, shows that Cd release rates were 10 to 100 times higher for Danish soils than Thai soils, and only for the Danish soils, a significant increase due to sludge amendment could be observed. The Cu release rates were of similar magnitudes for D1 and T1 with a five-times increase in the corresponding sludge-amended soils. Release rates for Pb were 5 to 10 times higher for Danish than for Thai soils; an increase in release rate due to sludge amendment was indicated only in the latter.

The amounts of the heavy metals initially released were not generally higher for the Danish than the Thai soils, except at higher pH (Table 2). At the end of the experiments, 5 to 30 times more Cd, Cu, and Pb had been released from Danish soils compared with the Thai soils (Table 2). A different picture was obtained by analyzing the relative accumulated release qi,rel, that is, the accumulated release of a metal (qi) (t ⫽ 7 d) relative to the total content of the respective metal in the soil (mi): qi,rel ⫽ qi/mi

[4]

with i ⫽ Cd, Cu, and Pb. The Thai soils showed a maximum qCd,rel of 8%, in strong contrast with Danish soils with qCd,rel values close to 50% (Table 3). The Thai samples showed higher qCu,rel values than the Danish samples; the highest value of qCu,rel ⫽ 16% was observed for T2. Lead was the metal showing lowest relative accumulated releases with maximum values of 2.5% in D1 (Table 3). The contribution of the initial release to the accumulated release after 160 h was relatively high (⬎25%) and reflects a fast initial release followed by a slow release. Such patterns were most evident for Cd in Thai soils at low pH and for Cu in both soil types at higher pH (Table 2).

Release versus pH Both the initial and accumulated release at 160 h decrease with increasing pH (Table 2). The ratio of the initial release to the accumulated release does not change in a systematic way with pH, but is different for the different soil materials. As seen for Cd in the Thai samples, the initial release almost equaled the accumulated release (160 h) at low pH values. For Cu and Pb, the ratio of initial release to the accumulated release (160 h) was in the range of 10 to 70% and generally was highest for the Danish soils at pH ⬎ 3.5, which demonstrates the presence of Cu and Pb fractions with very fast release kinetics. Relative release rates (ri, rel) derived by dividing the rates obtained from the linear regressions for each metal (ri) (Table 2) with the total content of the metal in the soil samples (mi) were estimated: ri,rel ⫽ ri/mi

[5]

Table 3. Relative accumulated release of Cd, Cu and Pb (% of total content) after 160h from the Thai (T1, T2) and Danish (D1, D2) soils at low and high pH. Relative accumulated release‡ Soil

pH†

Cd

Cu

Pb

T1

3.0 6.5 3.0 6.6 2.9 7.0 3.1 7.1

5.6 2.3 8.1 3.0 47.8 1.1 42.9 1.2

% 4.7 0.7 15.7 0.5 3.2 0.5 3.9 0.7

0.01 0 0.1 0 2.5 0 1.5 0

T2 D1 D2

Fig. 2. Example of linear regression of accumulated release vs. time data and the estimates of initial release and release rate obtained from this analysis. Sample D1 at pH 3.7. Cd 䉱, Cu 䊉, Pb 䊐.

† pH recorded in the effluent (compare with Table 2). ‡ Accumulated release as fraction of the total content of the respective metal (Eq. [4]).

1906

J. ENVIRON. QUAL., VOL. 31, NOVEMBER–DECEMBER 2002

(Eq. [4]) was similar to the variation of the relative release rates with pH.

DISCUSSION Time Dependence of Effluent Metal Concentrations The effluent metal concentrations reveal different C ⫺ t curve forms with concentrations either increasing, decreasing, or being constant with time (Fig. 1). In a few experiments, effluent metal concentrations showed a maximum during the experimental period. Different C ⫺ t curve shapes result from differences in initially released amounts, and because of variations in magnitude and constancy of metal release rates. In most cases release rates are high at low pH, which gives rise to effluent concentrations increasing over time. However, when initial release is high and is followed by slow release, as for Cu in Danish soils and Cd in Thai soils at low pH, effluent concentrations decrease over time. The considerable variation in effluent concentrations with time for individual experiments at constant pH demonstrates that the release is kinetically controlled. Instantaneous or very fast release reactions would produce effluent concentrations at a more or less constant level over time.

The pH Effect on Metal Release Rates

Fig. 3. Relative release rates (ri, rel) vs. pH for Cd (a ), Cu (b ), and Pb (c ). D1 䊊, D2 䊉, T1 䉭, and T2 䉱, where D is Danish and T is Thai. Relative release rates are calculated by use of Eq. [5]. Solid lines are drawn as a guide to the eye.

with i ⫽ Cd, Cu, and Pb. The relative release rate has the dimension s⫺1. With a few exceptions, the three metals show similar and almost coincident relative release rates vs. pH patterns for all soil samples with a distinct increase in relative release rates at decreasing pH (Fig. 3). At comparable pH, the magnitudes of the relative release rates follow the sequence rCd,rel ⬎ rPb,rel ⬎ rCu,rel. When pH decreases the relative release rates are seen to increase markedly, with increase in rCd,rel starting at a higher pH than the increases of rCu,rel and rPb,rel. Two deviations from the general patterns were evident: (i) a low relative release rate for Cd in the Thai soils at low pH, and (ii) an exceptionally high relative release rate of Cu for the sludge-amended T2 soil. For all metals, the pH variation of the relative accumulated release

The sludge-amended soils are higher in Cd, Cu, and Pb contents than the nontreated soils (Table 1), and in general, the sludge-amended soils show the highest release rates (Table 2). Thus, the heavy metals added with sludge are not immobile. To test if the metals added with sludge are less or more mobile than the native metals in the soils, we suggest making comparisons by calculation of the relative release rates (Eq. [5]). If heavy metals added with sludge are more labile than the native metals in the soils, the relative release rates of the metals in the sludge-amended soil will be higher than for the nontreated soil. It is important that comparisons of relative release rates are performed at the same pH, as this variable strongly influences metal bonding and mineral stability. We have extended that approach by plotting relative release rates vs. pH to demonstrate differences in metal lability–bonding among different soil materials. In most cases the relative release rates versus pH data follow the same curves for each of Cd, Cu, and Pb, indicating that the metal cations are bonded with almost equal strength irrespective of soil type and sludge or no sludge treatment (Fig. 3). There are two deviations from the general release patterns. First, in the Thai soils, there is a fast initial release of Cd followed by a slow release (Fig. 3a), a feature that is most prominent at low pH and that gives rise to much lower accumulated releases from the Thai than from the Danish soils (Tables 2, 3). This release pattern shows that the Thai soils contain a small fraction of rapidly leachable Cd and a large fraction of slowly accessible Cd, in contrast to the Danish soils, with large pools of easily accessible Cd. In Fig. 3a, the Thai soils

1907

SUKREEYAPONGSE ET AL.: pH-DEPENDENT METAL RELEASE FROM SOILS

clearly separate from the general pattern by showing very low Cd release rates at low pH. All experimental data clearly reflect a strong difference in bonding of Cd among the Thai and Danish soils. In the Danish soils, which have relatively high CEC values due to the contents of clay and organic matter (Table 1), most Cd is assumed to be bonded in easy available pools (e.g., cation exchange) as found for similar soils by Andersen et al. (2002). This is not the case for the Thai Ultisols, which have low contents of clay and organic matter, and consequently low CECs (Table 1). Furthermore, the Ultisols are highly leached. Hence, the major fraction of Cd is expected to be present in rather resistant pools, for example, as substitutions within silicates and metal oxides (Naidu et al., 1997). The second deviation from the general release patterns is seen for Cu in the sludge-amended Thai soil that shows four to six times higher release rates than the other soils, including the Danish sludge-amended soil (Fig. 3b). The high release of Cu from T2 at low pH is further substantiated by high initial and accumulated releases; more than 15% of the soil’s total content of Cu has been released after 160 h at pH 3 (Tables 2, 3). Hence, it is evident that the T2 soil contains a pool of weakly bound Cu. There are two likely reasons for this feature. First, because Cu is known to bind strongly to soil organic matter (McBride, 1989), the five-timeslower content of organic carbon in the sludge-amended Thai soil than the Danish soil will result in weaker Cu bonding in the former soil. This gives rise to a higher relative release of Cu. Second, the recent sludge amendment of the Thai soil has not allowed the sludge to age and the Cu to move to less accessible pools (Hogg et al., 1993; Lombi et al., 1998). If the bonding strength of the Cu added with the Thai sludge was equal to the bonding strength of the Cu in the native soil, no increases in relative release rates or relative accumulated releases would occur (Table 3, Fig. 3). The substantial increases in released amounts from the T2 soil compared with the T1 soil demonstrate that Cu in the sludge is present in a rather labile form, and that this labile bonding has continued after mixing with the soil. This is in strong contrast to the Danish sludge-amended soil, which was given a very high amount of Cu (Table 1) but without increasing the relative releases. However, the Danish soil has a higher sorption capacity due to much higher contents of Al- and Fe-oxides, clay, and organic matter (Table 1), and furthermore the sludge has equilibrated with the soil for more than 20 yr. In conclusion, by use of stirred flow cells the release properties of soils are easily quantified and compared. Estimates of relative accumulated releases and relative release rates are efficient tools for characterizing the strength of metal bonding in soils, in particular for identification of soil materials with release characteristics deviating from the general release patterns.

Rate Laws for Proton-Induced Heavy Metal Release When pH decreases in the range from 5 to 3, the relative release rates show pronounced increases (re-

lease edge), demonstrating that protons cause release of the heavy metal cations from the soil material (Fig. 3). The increase in release rates of Cd occurs at a higher pH than the increases for Pb and Cu, so that the order is Cd ⬎ Pb ⬎ Cu. In analogy with sorption edges, we may use the term release edges for the steep increases in release by lowering of pH. The location of the release edges with respect to pH is kinetically controlled and they occur at a pH about one unit lower than the corresponding sorption edges (Harter, 1983; McBride, 1989; Yong and Phadungchewit, 1993). However, the analogy between sorption and release edges suggests that the mechanism of metal cation release is due to proton sorption at bonding sites of minerals and organic matter schematically summarized in the following reaction schemes (Stumm, 1992): ⬅S–O–M⫹ ⫹ H⫹ → ← ⬅S–OH ⫹ M2⫹ ⬅L–M⫹ ⫹ H⫹ → ← ⬅LH ⫹ M2⫹

[6]

where ⬅S–O and ⬅L represents bonding sites at hydroxylated mineral surfaces and in organic matter, respectively, and M2⫹ is a divalent metal cation. Proton exchange with metal cations at permanent charge cation exchange sites in phyllosilicates will further contribute to metal release, in particular for the Danish soils with high contents of clay (Table 1). In the acid range, sorption of protons to hydroxylated sites may also cause mineral dissolution by weakening of bonds between surface metal cations and the bulk of the crystals (Lasaga et al., 1994): ⬅Al–OH ⫹ H⫹ → ← ⬅Al–OH2⫹ → ← 3⫹ dissolution (Al , fresh surfaces)

[7]

The transition from sorption–desorption to dissolution is thought to be gradual. We have tested if the observed release rates of heavy metals may confirm with simple proton-induced release kinetics according to Eq. [1]. This seems to be the case at least for Cd and Pb, where plots of logarithmized relative release rates vs. pH produce good linear correlations (Table 4). Copper does not show a similar good linear correlation; in fact, the release rate tends to show a minimum at pH 4 to 5. A likely explanation is that Cu in contrast to Cd and Pb is more strongly bound to organic matter both in the solid state and in solution, and hence mineral dissolution kinetics fail to explain metal release kinetics. A minimum in Cu release between pH 4 and 5 has been observed previously and was attributed to low solution Table 4. Linear regression equations for log (relative release rate) vs. pH for release of Cd, Cu, and Pb (see Fig. 3). Element log(rCd,rel)‡ log(rCu,rel)¶ log(rPb,rel)¶

Regression equation

R2†

⫽ ⫺0.41 (⫾0.05) pH ⫺ 5.15 (⫾0.22)§ 0.83*** ⫽ ⫺0.28 (⫾0.06) pH ⫺ 6.53 (⫾0.29) 0.49*** ⫽ ⫺0.84 (⫾0.15) pH ⫺ 5.23 (⫾0.50) 0.77***

Equation [8] [9] [10]

*** Significant at the 0.001 level. † R2 denotes the correlation coefficient of the regression equation. ‡ Expression of the relative release rate (s⫺1) of Cd based on Eq. [2] and [5], exclusive of experiments with pH ⬍ 3.5 for T1 and T2. § Numbers in parentheses represent the standard error. ¶ Expression of the relative release rate vs. pH for Cu and Pb calculated by use of Eq. [2] and [5].

1908

J. ENVIRON. QUAL., VOL. 31, NOVEMBER–DECEMBER 2002

complexation and a strong sorption of Cu at that pH (Strobel et al., 2001). The fact that rates of heavy metal dissolution conform with rate expressions for protoninduced mineral dissolution indicates, but does not prove, that heavy metal release may be treated as a mineral dissolution reaction. More work is needed to evaluate if such rate expressions apply for description of heavy metal release from minerals and soils in general. The rate expressions may be useful in integrated models of heavy metal leaching.

CONCLUSIONS The steady state flow cell approach for testing of soil metal release properties gave essential information regarding the release properties of Cd, Pb, and Cu in solute–soil systems. Effluent metal concentrations decreased with increasing pH and were generally higher for the Danish soils than the Thai soils. The initially released amounts of the metals, the release rates, and the accumulated releases were higher for the Alfisol than the Ultisol and higher for sludge-amended soils compared with the nonamended soils. The relative release rates for Cd, Cu, and Pb showed uniform patterns when presented as a function of pH, except for two soil samples with different bonding strength of Cd and Cu. The relative release rates could be described with pHdependent kinetics previously used to describe metal release due to proton-induced mineral dissolution. At a given pH, the relative release rates were highest for Cd, followed by Cu and Pb. The test system provides information regarding the short-term (initial release), long-term (release rates), and pH-dependent release controlled by the kinetics of the release processes. Relative release rates vs. pH diagrams are especially useful for characterization of the strength of metal bonding in soil materials. The test system is well suited for comparative studies of metal release as a function of soil types, treatments, and metal contents, and can be used as a test to rank soils with respect to the heavy metal releases. ACKNOWLEDGMENTS

This research was supported by grants of the Thailand Research Fund (TRF) and DANCED. The authors are grateful to Erik Damgaard at Askov Experimental Station and senior scientist Jens Petersen, Dep. of Crop Physiology and Soil Science, Danish Institute of Agricultural Sciences, Research Centre Foulum, who provided us with soil samples and information regarding the Danish soil samples. Hanne NanckeKrogh provided technical assistance in conducting the metal determinations. We are indebted to the reviewers for improvement of the manuscript.

REFERENCES Allison, J.D., D.S. Brown, and K.J. Novo-Gradac. 1991. MINTEQA2/ PRODEFA2, A geochemical assessment model for environmental systems: Version 3.0 user’s manual. USEPA, Athens, GA. Alloway, B.J. (ed.) 1995. Heavy metals in soils. 2nd ed. Chapman and Hall, Glasgow. Andersen, M.K., K. Raulund-Rasmussen, H.C.B. Hansen, and B.W. Strobel. 2002. Distribution and fractionation of heavy metals in

pairs of arable and afforested soils in Denmark. Eur. J. Soil Sci. 53:491–502. Anderson, P.R., and T.H. Christensen. 1988. Distribution coefficients of Cd, Co, Ni and Zn in soils. J. Soil Sci. 39:15–22. Andreasen, A.H.M. 1939. The fineness of solids and the technological importance of fineness. Ingeniørvidenskabelige Skrifter 3:1–71. Beckett, P.H.T. 1989. The use of extractants in studies on trace metals in soils, sewage sludges, and sludge treated soils. Adv. Soil Sci. 10:143–176. Berti, W.R., and L.W. Jacobs. 1998. Distribution of trace elements in soil from repeated sewage sludge application. J. Environ. Qual. 27:1280–1286. Buchter, B., B. Davidoff, M.C. Amacher, C. Hinz, I.K. Iskandar, and H.M. Selim. 1989. Correlation of Freundlich Kd and n retention parameters with soils and elements. Soil Sci. 148:370–379. Dowdy, R.H., J.J. Latterell, T.D. Hinesly, R.B. Grossman, and D.L. Sullivan. 1991. Trace metal movement in an Aeric Ochraqualf following 14 years of annual sludge application. J. Environ. Qual. 20:119–123. Drever, J.I. 1994. The effect of land plants on weathering rates of silicate minerals. Geochim. Cosmochim. Acta 58:2325–2332. Filius, A., T. Streck, and J. Richter. 1998. Cadmium sorption and desorption in limed topsoils as influenced by pH: Isotherms and simulated leaching. J. Environ. Qual. 27:12–18. Fo¨rstner, U. 1991. Soil pollution phenomena—Mobility of heavy metals in contaminated soil. p. 543–582. In G.H. Bolt et al. (ed.) Interactions at the soil colloid–soil solution interface. Kluwer Academic Publ., Dordrecht, the Netherlands. Gao, S.A., W.J. Walker, R.A. Dahlgren, and J. Bold. 1997. Simultaneous sorption of Cd, Cu, Ni, Zn, Pb, and Cr on soils treated with sewage sludge supernatant. Water Air Soil Pollut. 93:331–345. Gerritse, R.G., and W. van Driel. 1984. The relationship between adsorption, organic matter and pH in temperate soils. J. Environ. Qual. 13:197–203. Giller, K.E., E. Witter, and S.P. McGrath. 1998. Toxicity of heavy metals to microorganisms and microbial processes in agricultural soils: A review. Soil Biol. Biochem. 30:1389–1414. Harter, R.D. 1983. Effect of soil pH on adsorption of lead, copper, zinc, and nickel. Soil Sci. Soc. Am. J. 47:47–51. Holm, P.E., T.H. Christensen, S.E. Lorenz, R.E. Hamon, H.C. Domingues, E.M. Sequeira, and S.P. McGrath. 1998. Measured soil water concentrations of cadmium and zinc in plant pots and estimated leaching outflows from contaminated soils. Water Air Soil Pollut. 102:105–115. Hogg, D.S., R.G. McLaren, and R.S. Swift. 1993. Desorption of copper from some New Zealand soils. Soil Sci. Soc. Am. J. 57:361–366. Hossner, L.R. 1996. Dissolution for total element analysis. p. 49–64. In D.L. Sparks et al. (ed.) Methods of soil analysis. Part 3. Chemical methods. SSSA Book Ser. 5. SSSA, Madison, WI. Hyun, H., A.C. Chang, D.R. Parker, and A.L. Page. 1998. Cadmium solubility in sludge treated soil: Effects of soil organic carbon. J. Environ. Qual. 27:329–334. Jensen, D.L., P.E. Holm, and T.H. Christensen. 2000. Leachability of heavy metals from scrap dirt sampled at two scrap iron and metal recycling facilities. Waste Manage. Res. 18:367–379. Jing, J., and T.J. Logan. 1992. Effects of sewage sludge cadmium concentrations on chemical extractability and plant uptake. J. Environ. Qual. 21:73–81. Knauss, K.G., and T.J. Wolery. 1986. Dependence of albite dissolution kinetics on pH and time at 25⬚C and 70⬚C. Geochim. Cosmochim. Acta 50:2481–2497. Kraemer, S.M., and J.G. Hering. 1997. Influence of solution saturation state on the kinetics of ligand-controlled dissolution of oxide phases. Geochim. Cosmochim. Acta 61:2855–2866. Larsen, K.E., and J. Petersen. 1993. Long-term field- and lysimeter experiments with large annual dressings of heavy metal loaded sewage sludge. (In Danish with summary, tables and figures in English.) Rep. 3. Statens Planteavlsforsøg, Landbrugsministeriet, Copenhagen. Lasaga, A.C., J.M. Soler, J. Ganor, T.E. Burch, and K.L. Nagy. 1994. Chemical weathering rate laws and global geochemical cycles. Geochim. Cosmochim. Acta 58:2361–2386. Lombi, E., M.H. Gerzabek, and O. Horak. 1998. Mobility of heavy

SUKREEYAPONGSE ET AL.: pH-DEPENDENT METAL RELEASE FROM SOILS

metals in soil and their uptake by sunflowers grown at different contamination levels. Agronomie 18:361–371. Lorenz, S.E., R.E. Hamon, P.E. Holm, H.C. Domingues, E.M. Sequeira, T.H. Christensen, and S.P. McGrath. 1998. Cadmium and zinc in plants and soil solutions from contaminated soils. Plant Soil 189:21–31. Luo, Y.M., and P. Christie. 1998. Bioavailability of copper and zinc in soils treated with alkaline stabilized sewage sludges. J. Environ. Qual. 27:335–342. McBride, M.B. 1989. Reactions controlling heavy metal solubility in soils. Adv. Soil Sci. 10:1–56. McBride, M.B., B.K. Richards, T. Steenhuis, J.J. Russo, and S. Sauve´. 1997. Mobility and solubility of toxic metals and nutrients in soil fifteen years after sludge application. Soil Sci. 162:487–500. Mehra, O.P., and M.L. Jackson. 1960. Iron oxide removal from soils and clays by a dithionite–citrate system buffered with sodium bicarbonate. Clays Clay Miner. 5:317–327. Naidu, R., N.S. Bolan, R.S. Kookana, and K.G. Tiller. 1994. Ionicstrength and pH effects on the sorption of cadmium and the surface charge of soils. Eur. J. Soil Sci. 45:419–429. Naidu, R., R.S. Kookana, M.E. Sumner, R.D. Harter, and K.G. Tiller. 1997. Cadmium sorption and transport in variable charge soils: A review. J. Environ. Qual. 26:602–617. Pang, L., and M.E. Close. 1999. Non-equilibrium transport of Cd in alluvial gravels. J. Contam. Hydrol. 36:185–206. Raulund-Rasmussen, K., O.K. Borggaard, H.C.B. Hansen, and M. Olsson. 1998. Effect of natural organic soil solutes on weathering rates of soil minerals. Eur. J. Soil Sci. 49:397–406.

1909

Sauve´, S., W. Hendershot, and H.E. Allen. 2000. Solid–solution partitioning of metals in contaminated soils: Dependence of pH, total metal burden, and organic matter. Crit. Rev. Environ. Sci. Technol. 34:1125–1131. Schwertmann, U. 1964. Differenzierung der Eisenoxide des Bodens durch Extraktion mit Ammoniumoxalat-lo¨sung. Z. Pflanzenerna¨hr. Bodenkd. 105:194–202. Soil Survey Staff. 1997. Keys to soil taxonomy. 7th ed. Pocahontas Press, Blacksburg, VA. Strobel, B.W., H.C.B. Hansen, O.K. Borggaard, M.K. Andersen, and K. Raulund-Rasmussen. 2001. Cadmium and copper release kinetics in relation to afforestation of cultivated soils. Geochim. Cosmochim. Acta 65:1233–1242. Stuanes, A.O., G. Ogner, and M. Opem. 1984. Ammonium nitrate as extractant for soil exchangeable cations, exchangeable acidity and aluminium. Commun. Soil Sci. Plant Anal. 15:773–778. Stumm, W. 1992. Chemistry at the solid–water interface. John Wiley & Sons, New York. Tjell, J.C., and M.F. Hovmand. 1978. Metal concentrations in Danish arable soils. Acta Agric. Scand. 28:81–89. Van der Sloot, H.A., L. Heasman, and Ph. Quevauviller. 1997. Harmonization of leaching/extraction tests. Studies in Environ. Sci. 70. Elsevier Sci., Amsterdam. Wieland, E., B. Wehrli, and W. Stumm. 1988. The coordination chemistry of weathering: III. A generalization on the dissolution rates of minerals. Geochim. Cosmochim. Acta 52:1969–1981. Yong, R.N., and Y. Phadungchewit. 1993. pH influence on selectivity and retention of heavy metals in some clay soils. Can. Geotechnol. J. 30:821–833.