- Email: [email protected]
Mobility of heavy metals as related to soil chemical and mineralogical characteristics of Brazilian soils
Environmental Pollution 111 (2001) 429±435
www.elsevier.com/locate/envpol
Mobility of heavy metals as related to soil chemical and mineralogical characteristics of Brazilian soils A.T. de Matos a, M.P.F. Fontes b,*, L.M. da Costa b, M.A. Martinez a a
Departamento de Engenharia Agricola, Universidade Federal de VicËosa, 36571-000 VicËosa MG, Brazil b Departamento de Solos, Universidade Federal de VicËosa, 36571-000 VicËosa MG, Brazil Received 1 March 1999; accepted 15 February 2000
``Capsule'': Soil pH greatly in¯uenced adsorption and movement of heavy metals. Abstract In order to better understand the relationship between soil characteristics and mobility of some heavy metals, correlation studies were conducted in samples of unlimed and limed A, B and C horizons of three Brazilian soils, representative of the majority of the tropical soils. A number of chemical and mineralogical characteristics of one Oxisol and two Ultisols were related to the retardation factors (Rf) for zinc (Zn), cadmium (Cd), copper (Cu) and lead (Pb). The retardation factors, obtained in leaching column experiments, were used as an estimate of solute movement in the pro®le. Soil types and soil horizons were found to in¯uence metal retardation factors which, in turn, correlated better with the chemical than the mineralogical soil characteristics. For the unlimed soil samples, the soil characteristics that signi®cantly correlated with Zn-Rf and Cd-Rf were the sum of exchangeable bases (SB), and soil exchangeable (Ca-KCl) and nonexchangeable (Ca-HCl) calcium contents. These results showed the strong in¯uence of the cation exchange phenomenon on the retention and mobility of these two metals. For Cu and Pb, not only SB, cation exchange capacity (CEC) and Ca-KCl and Ca-HCl but also the organic matter correlated well with the Rf, showing that complex or chelate formation may play an important role in the movement of these elements. The important soil chemical characteristics related to the retardation factors in the limed soil samples were SB for Cd, and Ca-HCl for Cu and Pb, suggesting that precipitation may also in¯uence the mobility and retention of the latter two heavy metals in these soil samples. Soil pH in¯uenced the heavy metals adsorption and movement as shown by the signi®cant correlation with the retardation factors when the combined data for the unlimed and limed soil samples was considered. # 2000 Elsevier Science Ltd. All rights reserved. Keywords: Heavy metals; Retardation factors; Soil chemical characteristics; Soil pollution
1. Introduction The impact of contaminants on the environment should be of scienti®c concern, in order to minimize the threat of soil and groundwater pollution. Waste disposal on agricultural lands has been increasingly favored lately and, therefore, it should be scrutinized to diminish the risk of introducing pollutants to soils and waters. In this context, presence of heavy metals in several wastes used in the Brazilian agriculture today has imposed a need for a better understanding of the processes of soil±heavy metal interactions, in particular, their mobility and retention.
* Corresponding author. Tel.: +55-31-899-2630; fax: +55-31-8992648. E-mail address: [email protected] (M.P.F. Fontes).
The waste disposal, whether in sanitary land®ll or, when processed for agricultural use demands understanding of the adsorption phenomena and pollutant mobility in soil pro®les, since these are essential factors that control groundwater contamination (Bittel and Miller, 1974; Elliot et al., 1986). Several laboratory studies, using disturbed soil samples, have been conducted to ®nd soil characteristics which are correlated with adsorption and mobility of heavy metals; however, mixed results have been found due to dierent mineralogical and chemical soil characteristics, and the selectivity of retention sites for each cation. Soil characteristics positively correlated with cadmium (Cd) retention were pH (Tyler and McBride, 1982; Christensen, 1984; King, 1988; Jopony and Young, 1994), organic matter (OM) content and cation exchange capacity (CEC) (Sidle and Kardos, 1977), speci®c surface area (Korte et al., 1976), while free iron oxides was negatively correlated (Amacher et al.,
0269-7491/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved. PII: S0269-7491(00)00088-9
430
A.T. de Matos et al. / Environmental Pollution 111 (2001) 429±435
1986). Zinc (Zn) retention was positively correlated with pH (Harter, 1983; King, 1988), CEC (Shuman, 1977; Sidle and Kardos, 1977) and speci®c surface area (Korte et al., 1976). A number of authors have found positive correlations between copper (Cu) retention and pH (Tyler and McBride, 1982; Harter, 1983; King, 1988), sum of bases (SB) or exchangeable calcium (Ca) (Harter, 1979) OM content (Hickey and Kittrick, 1984) and to the CEC (Sidle and Kardos, 1977). Lead (Pb) retention was better correlated to clay content (King, 1988), pH (Harter, 1983; Jopony and Young, 1994) and SB or exchangeable Ca (Harter, 1979). Korte et al. (1976), using multiple regression analysis, found clay content, speci®c surface area, free iron (Fe) oxide content and free carbonates in¯uencing Cd, Zn, nickel (Ni) and chromium (Cr) adsorption. Some reports found no correlation between soil OM content and metal adsorption (Harter, 1979, 1983). The author credited this lack of correlation to the fact that total OM of a wide variety of soils may not be suciently related to their reactive portions to generate a positive correlation. Many authors found that both free and amorphous Fe oxides had high capacities to adsorb metals because of their ability to make covalent links with them (Korte et al., 1976; Shuman, 1977; Hickey and Kittrick, 1984; Tiller et al., 1984; King, 1988; Slavek and Pickering, 1988). However, Amacher et al. (1986) and Elliot et al. (1986) found negative correlation between free Fe oxide content and retention of Cd, Cu, Zn and Pb. Experiments with disturbed soil samples are important to understand the adsorption of heavy metals; however, estimation of heavy metal cation movement in soils necessitates a better approximation of the chemical retardation processes, which can be done by using soil column experiments. The retardation factor of solutes moving through soils, a parameter that expresses the relationship between the speed of solute movement and the water-leaching front speed (Valocchi, 1984), is an indicator of interactions between the solution and the solid phase. It can give an idea about the mobility of cations throughout the soil pro®le. Estimation of retardation factors of dissolved carbon using soil columns (Li and Shuman, 1997a) and leaching experiments with soil columns to asses mobility of heavy metals (Boyle and Fuller, 1987; Li and Shuman, 1997b) is common practice. But less common is the use of soil column experiments to estimate retardation factors of heavy metals, relate them to their mobility in soils and to the soil chemical and mineralogical characteristics. Therefore, in order to improve the predictive capability of retention and mobility of some heavy metals in soils, the objective of this study was to relate the heavy metals movement, as measured by the Zn, Cd, Cu and Pb retardation factors, to several chemical and
mineralogical soil characteristics for selected Brazilian soils. 2. Materials and methods Samples of the A, B and C horizons of two Ultisols (U1 and U2) and one Oxisol (Ox), from VicËosa, Minas Gerais State, were air dried and passed through a 2-mm sieve. The texture was determined by the pipette method. The clay was analyzed by X-ray diraction, using Cu-Ka monochromatic radiation obtained by a Ni ®lter. Kaolinite and gibbsite contents were obtained by dierential thermal analysis. The pH was determined in water with a 1:2.5 soil:solution ratio. Exchangeable calcium (Ca-KCl) was extracted with KCl 1 mol lÿ1 and non-exchangeable Ca (Ca-HCl) was obtained extracting with boiling HCl 1 mol lÿ1 (Lanyon and Heald, 1982). Crystalline iron (Fed) was determined by dithionite-citrate extraction (Con, 1963) and amorphous iron (Feo) by ammonium oxalate extraction (Schwertmann, 1964). The values of SB and CEC were obtained by calculation from the exchangeable cations contents determined according to Thomas (1982). Organic carbon was determined by Walkley-Black method according to Jackson (1958) and converted to OM. Speci®c surface area was determined according to Carter et al. (1965). Ca and magnesium (Mg) were determined by atomic absorption spectrophotometry, sodium (Na) and potassium (K) by ¯ame photometry and Fe by colorimetry. Chemical and mineralogical characteristics used in correlation analysis are presented in Table 1. Euent breakthrough curves for heavy metals were measured using soil-packed columns. Contaminant solutions were ponded on the soil column surface and 1.5 cm of hydrostatic pressure was maintained during the leaching process. The euent breakthrough curves were measured for solutions containing concentrations of 700 mg lÿ1 of Zn, 20 mg lÿ1 of Cd, 200 mg lÿ1 of Cu and 300 mg lÿ1 of Pb. The retardation factors (Rf) (Table 2) were analytically determined, from regression equations, as the necessary number of pores to make the euent concentration equal to 50% of the concentration of the contaminant solution (Nielsen and Biggar, 1961; Wierenga et al., 1975; Baes and Sharp, 1983; Christensen, 1985; Van Genuchten and Wierenga, 1986). A complete description on how the retardation factors were obtained can be found in de Matos et al. (1999). Sub-samples of each soil were treated with CaCO3, cp, in a dose of 8 t haÿ1, allowed to equilibrate for a period of 15 days and submitted to the same procedure to calculate the retardation factors. Four replicate columns of each soil, horizon, lime treatment and controls were set up.
A.T. de Matos et al. / Environmental Pollution 111 (2001) 429±435
431
Table 1 Chemical and mineralogical characteristics of unlimed and limed soil samplesa Soil
Horizon pH
Ca2+ (Cmol kgÿ1) KCl
SB CEC Al3+ OM Fed Feo SSA Ka Gb Clay ÿ1 ÿ1 (Cmol kg ) (Cmol kg ) (Cmol kgÿ1) (g kgÿ1) (g kgÿ1) (g kgÿ1) (m2 gÿ1) (g kgÿ1) (g kgÿ1) (g kgÿ1)
HCl
U1
A B C
6.02 3.09 6.48 1.46 6.85 0.83
4.62 2.67 1.83
4.37 1.95 1.16
4.67 2.15 1.16
0.30 0.20 0.00
55.1 16.8 5.4
55.4 74.7 33.8
0.12 0.13 0.07
20.1 16.8 8.7
350 460 90
767 863 905
36 54 51
U2
A B C
6.26 3.94 5.36 0.45 5.88 0.02
6.20 1.68 1.29
6.09 0.85 0.17
6.19 1.35 0.57
0.10 0.50 0.40
52.4 24.2 11.4
64.1 82.8 77.4
0.07 0.12 0.12
22.0 21.2 19.0
520 730 380
819 765 785
28 16 43
Ox
A B C
4.61 0.25 5.06 0.08 5.38 0.00
1.54 1.34 0.53
0.52 0.14 0.07
2.32 0.94 0.72
1.80 0.80 0.65
57.1 30.2 2.7
63.8 75.6 29.0
0.38 0.17 0.10
26.2 21.5 12.6
630 720 40
582 709 853
26 34 21
U1 Lb A B C
8.34 6.41 8.53 5.73 8.70 4.21
11.78 10.09 8.29
9.92 4.06 2.22
9.92 4.06 2.22
n.d. n.d. n.d.
U2 L
A B C
8.45 7.98 8.31 6.46 8.45 6.17
14.32 9.36 9.94
10.04 5.46 3.50
10.04 5.46 3.50
n.d. n.d. n.d.
Ox L
A B C
6.49 6.27 7.78 6.07 8.53 4.32
8.32 8.49 8.81
8.61 6.69 2.46
10.51 6.69 2.46
n.d. n.d. n.d.
a SB, sum of bases; CEC, cation exchange capacity; OM, organic matter; Fed, Fe extracted by dithionite; Feo, Fe by oxalate; SSA, speci®c surface area; Ka, kaolinite; Gb, gibbsite; n.d., not detected. b L, limed soils.
3. Results and discussion 3.1. Retardation factors The retardation factors for Zn, Cd, Cu and Pb obtained for the three soils and horizons are shown in Table 2. Zn generally has the lowest values followed closely by Cd whereas Cu and Pb reached the highest values, especially Pb. It means that Zn and Cd have the highest mobility associated to the lowest adsorption and Cu and Pb present the opposite behavior. The Ultisols (U1 and U2) reveal, as a general trend, a decrease of the retardation factors with depth. This showed that all of the heavy metals, for these soils, are more adsorbed in the A horizon, than in the B and C horizons. For the Oxisol (Ox), the B horizon had higher retardation factors as compared to A and C horizons, indicating that in the B horizon of this soil all the heavy metals have less mobility. In general, the values for Rf increased dramatically when the soils were limed. This shows that the mobility of all the heavy metals decreased with the pH increase. The phenomenon can be related to both the increasing of the speci®c and non-speci®c adsorption and to the precipitation reaction. McBride (1994, p. 320) reports that an experiment working with accelerated leaching in
a column of an acid mineral soil have produced some movement of trace metal ions initially applied to the column surface. He also relates that the same soil after liming to pH 6.5 retained all metals more strongly. The author concludes that liming reduces both mobility and plant availability of toxic metals. 3.2. Correlation studies Using the retardation factor values (Table 2) and soil chemical and mineralogical characteristics (Table 1) a linear correlation study was conducted. Table 3 present the signi®cant simple linear correlation coecients between retardation factors and the soil characteristics for Zn, Cd, Cu and Pb, respectively. Because this study used three soils and three horizons making a total of nine observations, the correlation coecients considered signi®cant were only those with probability level smaller than 1 and 0.1% (P<0.01 and P<0.001). For the linear correlation study, the mineralogical soil characteristics studied gave no signi®cant correlation coecients with retardation factors for any metal. The reasons could be: for the amorphous Fe oxides (Feo) their very low amounts in these highly weathered tropical soils (Fontes and Weed, 1991); for the crystalline Fe oxides (Fed), it can be related to the fact that the in¯uence of the
432
A.T. de Matos et al. / Environmental Pollution 111 (2001) 429±435
Table 2 Retardation factors (Rf) for Zn, Cd, Cu and Pb of unlimed and limed (L) soil samples Soil
Horizon
Rf Zn
Cd
Cu
Pb
U1
A B C
2.308 1.566 1.727
3.375 1.841 1.732
6.445 3.393 2.742
12.571 6.691 4.933
U2
A B C
2.398 0.932 1.272
3.151 1.146 1.410
5.676 2.287 2.792
11.264 4.830 5.032
Ox
A B C
0.902 1.347 1.244
1.226 1.466 1.106
3.238 3.880 1.781
5.901 6.413 3.750
U1 L
A B C
5.830 3.864 3.363
8.845 4.528 4.895
15.664 10.317 11.981
27.795 15.786 19.710
U2 L
A B C
5.204 2.687 2.214
7.410 3.884 2.807
15.301 8.093 6.587
36.528 13.248 8.610
Ox L
A B C
3.164 3.911 3.231
4.787 5.612 3.849
8.353 9.341 8.486
14.744 13.158 10.956
Fe oxides in certain soil reactions depends not only on the amounts but it also depends on the relative proportions of hematite and goethite, their crystallinity, their preferential exposed faces, etc. (Fontes and Weed, 1996) and for kaolinite and gibbsite contents the same reasoning probably applies, although there are no studies relating their contents directly to adsorption of heavy metals. The correlation observed between mineralogical characteristics and adsorption of heavy metals in laboratory-disturbed soil samples, well documented in the literature, and the lack of correlation between these variables and retardation factors for the heavy metals in
our study is suggesting that there are other variables which can also play important roles in leaching column experiments. In this case, it seems that hydraulic conductivity, porosity, etc., for example, would in¯uence the residence time of the solution in the column which, in turn, would aect the mobility and retention of the metals. In the limed soils most of the characteristics did not show signi®cant correlation with the retardation factors for the metals studied. With the pH increase brought about by the liming there was an increase in the metals adsorption and there was also an increase of the metals precipitation. It seems that, at such high pHs reached by the strong lime application, the reaction that governed the metals behavior was the precipitation. These results agree with McBride (1994) who states that liming reduces both mobility and plant availability of toxic metals and reinforces the observations that overliming can decrease heavy metals movement in soils. It can be seen in Table 1 that the soils pH, with the exception of the A horizon of the Oxisol, reached values around 8.0 or more, which is well above the recommended pH range for agricultural uses in these soils. The purpose of liming highly weathered tropical soils like these, is primarily to neutralize the exchangeable aluminum (Al), and this is normally accomplished by raising the pH to 5.5 (Sanchez, 1976, p. 240; Uehara and Gillman, 1981, p. 66). 3.3. Zn In general, retardation factors for Zn, were smaller than those for the other metals (Table 2), showing that its mobility in these soils is the highest. The CEC, SB and soil Ca content (Ca-KCl, Ca-HCl) were signi®cantly correlated with Zn retardation factor (Zn-Rf) in the unlimed soil samples (Table 3). The signi®cance of the relationship between Zn-Rf and CEC, SB and Ca contents show that, in this competitive system, Zn has its adsorption to the soil solid particles more related to the non-speci®c, i.e. electrostatic adsorption.
Table 3 Correlation coecients between Zn, Cd, Cu and Pb retardation factors and some soil chemical and mineralogical characteristicsa Characteristic chemical
pH OM SB CEC Ca-KCl Ca-HCl a
Zn
Cd
Cu
Pb
Unlimed
Limed
Unlimed
Limed
Unlimed
Limed
Unlimed
Limed
0.657n.s. 0.348n.s. 0.799** 0.793** 0.907*** 0.873**
0.052n.s. 0.598n.s. 0.587n.s. 0.587n.s. 0.310n.s. 0.559n.s.
0.491n.s. 0.581n.s. 0.941*** 0.894*** 0.956*** 0.939***
0.001n.s. 0.721n.s. 0.783** 0.707n.s. 0.450n.s. 0.639n.s.
0.229n.s. 0.763** 0.852** 0.867** 0.864** 0.879***
0.274n.s. 0.513n.s. 0.581n.s. 0.489n.s. 0.337n.s. 0.756**
0.213n.s. 0.764** 0.896*** 0.911*** 0.904*** 0.910***
0.168n.s. 0.614n.s. 0.680n.s. 0.602n.s. 0.549n.s. 0.835**
OM, organic matter; SB, sum of exchangeable bases; CEC, cation exchange capacity; n.s., not signi®cant. **Signi®cant to 1%. ***Signi®cant to 0.1%.
A.T. de Matos et al. / Environmental Pollution 111 (2001) 429±435
The exchangeable Ca content showed the highest simple linear correlation coecient (r=0.907, P<0.001) with Zn-Rf variation suggesting that Zn adsorption in exchange sites may be the preferential form of this metal retention in these soils. These results agree with many authors (Sidle and Kardos, 1977; Harter, 1979; Hickey and Kittrick, 1984; de Matos et al., 1994, 1996). In the limed soils, none of the characteristics presented signi®cant correlation with Zn-Rf, at the probability levels tested. 3.4. Cd The values for Cd-Rf were slightly higher than those of Zn, but still much lower than Cu and Pb also indicating a high mobility of Cd in these pro®les. The Cd retardation factors (Cd-Rf) in unlimed soil samples, as already observed for Zn, were signi®cantly correlated with the SB, CEC and the soil Ca contents (Table 3). All of these chemical characteristics were highly correlated with CdRf as seen by the probability values (P<0.001). In the limed soil materials, the behavior was similar, with the correlation coecients showing no signi®cance with the Cd-Rf, except for SB which showed r=0.783 and P<0.01. 3.5. Cu Cu presented higher values for Rf, as compared to Zn and Cd, showing that it is more adsorbed and has less mobility than these two metals, in these soil pro®les. In the unlimed soil samples, signi®cant correlation of Cu retardation factor (Cu-Rf) were obtained with CEC, OM and soil Ca contents (Ca-KCl and Ca-HCl) (Table 3). In a dierent way of what was observed for Zn and Cd, the OM presented a signi®cant correlation coecient with Cu-Rf (P<0.01), which demonstrates the strong association of OM with this metal already observed by several authors (Sidle and Kardos, 1977; Hickey and Kittrick, 1984; de Matos et al., 1996). The Cu retardation factor in the limed soil samples was signi®cantly correlated with Ca-HCl (r=0.756, P<0.01). It seems to indicate that precipitation is the reaction which is favored and probably accounts for the majority of the Cu retention in these soils when limed. The correlation coecient for OM decreased, becoming non-signi®cant, what could be due to the lowering of the more reactive OM content which always accompanies the liming. 3.6. Pb Pb presented, consistently, the highest values for retardation factors showing clearly that it is highly immobilized by these soils.
433
For the unlimed soil samples, Pb-Rf presented similar correlation (Table 3) with the same soil characteristics as observed for Cu, i.e. the OM content, SB, CEC and soil Ca contents. The coecient for OM was signi®cant at a P value<0.01, demonstrating of the importance of this soil fraction on the reactions that control the mobility of Pb. In the limed soil samples, all of the coecients had their signi®cance decreased, reaching no signi®cance, except for Ca-HCl, which showed a strong association to Pb-Rf (r=0.835, P<0.01). This association, as discussed before for Cu, can be due to the precipitation reaction. 3.7. Combined analysis In order to emphasize the dierences between the behavior of the retardation factor in the soil samples with liming as compared to the samples with no liming, a combined analysis was conducted, putting together all the data from the unlimed and the limed soil samples. The correlation coecients are presented in Table 4. The additional and maybe the most important aspect in the combined analysis is the inclusion of soil pH as a variable of signi®cant correlation to all the metals retardation factors. This correlation can be explained by considering that pH is the main soil characteristic to in¯uence the CEC of highly weathered soils and the dominant ionic forms in solution. The pH and the CEC are, almost always, reported as soil characteristics to show good association to soil adsorption of elements such as Cd and Zn (Tyler and McBride, 1982; Christensen, 1984; King, 1988). Also, the speci®c adsorption and or precipitation reactions participate increasingly in all metals adsorption (Harter, 1979). For the individual elements in the combined analysis, it still may be seen the strong association between SB, CEC and soil Ca contents and Zn-Rf and Cd-Rf. The initial observations showed that a large part of the Zn and Cd retention was due to the exchange sites, but the combined Table 4 Correlation coecients between retardation factors and some soil chemical and mineralogical characteristics for the combined analysisa Characteristic chemical
Zn
Cd
Cu
Pb
PH OM SB CEC Ca-KCl Ca-HCl
0.688*** 0.341n.s. 0.729*** 0.773*** 0.769*** 0.819***
0.722*** 0.403n.s. 0.889*** 0.841*** 0.851*** 0.886***
0.786*** 0.319n.s. 0.807*** 0.753*** 0.857*** 0.922***
0.636** 0.441n.s. 0.824** 0.782*** 0.776*** 0.849***
a
OM, organic matter; SB, sum of exchangeable bases; CEC, cation exchange capacity; n.s., not signi®cant. **Signi®cant to 1%. ***Signi®cant to 0.1%.
434
A.T. de Matos et al. / Environmental Pollution 111 (2001) 429±435
analysis suggests that part of it may be associated with speci®c adsorption (Garcia-Miragaya and DaÂvalos, 1986) and with precipitation (Harter, 1979). By looking at the relationship between the Cu-Rf and the soil characteristics in the combined analysis, it can be observed the same trend as showed for Zn and Cd. The SB (r=0.807, P<0.001) and the Ca-HCl (0.922, P<0.001) were soil characteristics which better associated to the Cu-Rf. Very similar results were obtained by Harter (1979), who found high signi®cance to the correlation between Cu adsorption, SB and the soil Ca content. The same author considered that the Cu hydrolysis, which forms CuOH+, under pH values above 6.0, was the factor of greater importance to the soil Cu retention, since it would make it possible adsorption of larger amounts of these species in the exchange complex. Also for Pb-Rf, there was a reiteration of the same trend veri®ed to all other heavy metals, with the signi®cant correlation being with pH, SB, CEC, Ca-KCl and Ca-HCl. Harter (1979) found similar correlation with Pb adsorption, attributing it to the hydrolysis of Pb2+ to PbOH+, making possible more ionic adsorption in the exchange complex, as well as metal precipitation. 4. Conclusions From the results the following conclusions can be drawn. Soil types and soil horizons in¯uenced the metals retardation factors. The Ultisols showed, consistently, higher retardation factors as compared to the Oxisol. Also, the Ultisols showed higher retardation factors for the A horizons, whereas the Oxisol showed, generally, higher retardation factors for the B horizon. The soil chemical characteristics showed signi®cant correlation to the metals retardation factors whereas the mineralogical characteristics did not. In the unlimed soil samples, the SB and the exchangeable Ca content (CaKCl) were the characteristics of better association to the metals retardation factors. In the limed soil samples, the signi®cant correlation coecients were observed with the SB, OM content, and non-exchangeable Ca content (Ca-HCl). The soil chemical characteristics were better estimators than the soil mineralogical characteristics to predict heavy metals mobility and retention in these soils. Soil pH was a variable that in¯uenced heavy metals adsorption, retention and movement, as measured by the retardation factors, showing high signi®cant correlation in the combined analysis of unlimed and limed soil samples. Liming the soils was very eective in controlling heavy metals mobility in the studied soils and, therefore, should be considered as an aid in environmental protection management practices.
References Amacher, M.C., Kotuby-Amacher, J., Selim, H.M., Iskandar, I.K., 1986. Retention and release of metals by soils: evaluation of several models. Geoderma 38, 131±154. Baes, C.F., Sharp, R.D., 1983. A proposal for estimation of soil leaching constants for use in assessment models. Journal of Environmental Quality 12, 17±28. Bittel, J.E., Miller, R.J., 1974. Lead, cadmium, and calcium selectivity coecients on montmorillonite, illite, and kaolinite. Journal of Environmental Quality 3, 250±253. Boyle, M., Fuller, W.H., 1987. Eect of municipal solid waste leachate composition on zinc migration through soils. Journal of Environmental Quality 16, 357±360. Carter, D.L., Heilman, M.D., Gonzalez, C.L., 1965. Ethylene glycol monoethyl ether for determining surface area of silicate minerals. Soil Science 100, 356±360. Christensen, T.H., 1984. Cadmium soil sorption at low concentrations, II: reversibility, eect of changes in solute composition, and eect of soil aging. Water, Air and Soil Pollution 21, 115±125. Christensen, T.H., 1985. Cadmium soil sorption at low concentrations, III: prediction and observation of mobility. Water, Air and Soil Pollution 26, 255±264. Con, D.E., 1963. A method for determination of free iron oxides in soils and clays. Can. Journal of Soil Science 43, 9±17. de Matos, A.T., Fontes, M.P.F., JordaÄo, C.P. 1994. Mobility and retention of Cd, Zn, Cu and Pb in a Brazilian Oxisol pro®le. In: XV World Congress of Soil Science Acapulco. Transactions, Vol. 3b; pp. 193±194. de Matos, A.T., Fontes, M.P.F., JordaÄo, C.P., da Costa, L.M., 1996. Mobility and retention forms of heavy metals in a Red-Yellow Latosol pro®le. Revista Brasileira de CieÃncia do Solo 20, 379±386. de Matos, A.T., da Costa, L.M., Fontes, M.P.F., Martinez, M.A., 1999. Retardation factors and the dispersion diusion coecients of Zn, Cd, Cu and Pb in soils from VicËosa- MG, Brazil. Transactions of the American Society of Agricultural Engineers (ASAE) 42 (4), 903±910. Elliot, H.A., Liberati, M.R., Huang, C.P., 1986. Eect of iron oxide removal on heavy metal sorption by acid subsoils. Water, Air and Soil Pollution 37, 379±389. Fontes, M.P.F., Weed, S.B., 1991. Iron oxides in selected Brazilian Oxisols: I. Mineralogy. Soil Science Society of America Journal 55 (4), 1143±1149. Fontes, M.P.F., Weed, S.B., 1996. Phosphate adsorption in clays from Brazilian Oxisols: relationships with speci®c surface area and mineralogy. Geoderma 72, 37±51. Garcia-Miragaya, J., DaÂvalos, M., 1986. Sorption and desorption of Zn on Ca-kaolinite. Water, Air and Soil Pollution 27, 217±224. Harter, R.D., 1979. Adsorption of copper and lead by Ap and B2 horizons of several northeastern United States soils. Soil Science Society of America Journal 43, 679±683. Harter, R.D., 1983. Eect of soil pH on adsorption of lead, copper, zinc, and nickel. Soil Science Society of America Journal 47, 47±51. Hickey, M.G., Kittrick, J.A., 1984. Chemical partitioning of cadmium, copper, nickel and zinc in soils and sediments containing high levels of heavy metals. Journal of Environmental Quality 13, 372±376. Jackson, M.L., 1958. Soil Chemical Analysis. Prentice-Hall, Englewood Clis, NJ. Jopony, M., Young, S.D., 1994. The solid±solution equilibria of lead and cadmium in polluted soils. European Journal of Soil Science 45, 59±70. King, L.D., 1988. Retention of metals by several soils of the southeastern United States. Journal of Environmental Quality 17, 239±246. Korte, N.E., Skopp, J., Fuller, W.H., Niebla, E.E., Alesii, B.A., 1976. Trace element movement in soils: in¯uence of soil physical and chemical properties. Soil Science 122, 350±359.
A.T. de Matos et al. / Environmental Pollution 111 (2001) 429±435 Lanyon, L.E., Heald, W., 1982. Magnesium, calcium, strontium and barium. In: Page, A.L., Miller, R.H., Keeney, D.R. (Eds.), Methods of Soil Analysis Part 2: Chemical and Biological Properties. ASA, SSSA, Madison, pp. 247±262. Li, Z., Shuman, L.M., 1997a. Estimation of retardation factor of dissolved organic carbon in sandy soils using batch experiments. Geoderma 78, 197±206. Li, Z., Shuman, L.M., 1997b. Mobility of Zn, Cd, and Pb in soils as aected by poultry litter extract Ð I. Leaching in soil columns. Environmental Pollution 95, 219±226. McBride, M.B., 1994. Environmental Chemistry of Soils. Oxford University Press, New York. Nielsen, D.R., Biggar, J.W., 1961. Miscible displacement in soils. I: experimental information. Soil Science Society of America Journal 25, 1±5. Sanchez, P.A., 1976. Properties and Management of Soils in the Tropics. John Wiley, New York. Schwertmann, U., 1964. The dierentiation of iron oxide in soils by a photochemical extraction with acid ammonium oxalate. Z. P¯anzenernahr. Dueng. Bodenk. 105, 194±201. Shuman, L.M., 1977. Adsorption of Zn by Fe and Al hydrous oxides as in¯uenced by aging and pH. Soil Science Society of America Journal 41, 703±706. Sidle, R.C., Kardos, L.T., 1977. Adsorption of copper, zinc, and cadmium by a forest soil. Journal of Environmental Quality 6, 313±317.
435
Slavek, J., Pickering, W.F., 1988. Metal ion interaction with the hydrous oxides of aluminium. Water, Air and Soil Pollution 39, 201±216. Thomas, G.W., 1982. Exchangeable cations. In: Page, A.L., Miller, R.H., Keeney, D.R. (Eds.), Methods of Soil Analysis. Part 2: Chemical and Biological Properties. ASA, SSSA, Madison, pp. 159±165. Tiller, L.D., Gerth, J., Brummer, G., 1984. The relative anities of Cd, Ni and Zn, for dierent soil clay fractions and goethite. Geoderma 34, 17±35. Tyler, L.D., McBride, M.B., 1982. Mobility and extractability of cadmium, copper, nickel, and zinc in organic and mineral soil columns. Soil Science 134, 198±204. Uehara, G., Gillman, G., 1981. The Mineralogy, Chemistry, and Physics of Tropical Soils with Variable Charge Clays. Westview Press, Boulder. Valocchi, A.J., 1984. Describing the transport of ion-exchanging contaminants using an eective Kd approach. Water Resource Research 20, 499±503. Van Genuchten, M.T., Wierenga, P.J., 1986. Solute dispersion: coecients and retardation factors. Methods of Soil Analysis. Part 1: Physical and Mineralogical Methods, ASA, SSSA Madison, 1025± 1031. Wierenga, P.J., Van Genuchten, M.T., Boyle, F.W., 1975. Transfer of boron and tritiated water through sandstone. Journal of Environmental Quality 4, 83±87.
www.elsevier.com/locate/envpol
Mobility of heavy metals as related to soil chemical and mineralogical characteristics of Brazilian soils A.T. de Matos a, M.P.F. Fontes b,*, L.M. da Costa b, M.A. Martinez a a
Departamento de Engenharia Agricola, Universidade Federal de VicËosa, 36571-000 VicËosa MG, Brazil b Departamento de Solos, Universidade Federal de VicËosa, 36571-000 VicËosa MG, Brazil Received 1 March 1999; accepted 15 February 2000
``Capsule'': Soil pH greatly in¯uenced adsorption and movement of heavy metals. Abstract In order to better understand the relationship between soil characteristics and mobility of some heavy metals, correlation studies were conducted in samples of unlimed and limed A, B and C horizons of three Brazilian soils, representative of the majority of the tropical soils. A number of chemical and mineralogical characteristics of one Oxisol and two Ultisols were related to the retardation factors (Rf) for zinc (Zn), cadmium (Cd), copper (Cu) and lead (Pb). The retardation factors, obtained in leaching column experiments, were used as an estimate of solute movement in the pro®le. Soil types and soil horizons were found to in¯uence metal retardation factors which, in turn, correlated better with the chemical than the mineralogical soil characteristics. For the unlimed soil samples, the soil characteristics that signi®cantly correlated with Zn-Rf and Cd-Rf were the sum of exchangeable bases (SB), and soil exchangeable (Ca-KCl) and nonexchangeable (Ca-HCl) calcium contents. These results showed the strong in¯uence of the cation exchange phenomenon on the retention and mobility of these two metals. For Cu and Pb, not only SB, cation exchange capacity (CEC) and Ca-KCl and Ca-HCl but also the organic matter correlated well with the Rf, showing that complex or chelate formation may play an important role in the movement of these elements. The important soil chemical characteristics related to the retardation factors in the limed soil samples were SB for Cd, and Ca-HCl for Cu and Pb, suggesting that precipitation may also in¯uence the mobility and retention of the latter two heavy metals in these soil samples. Soil pH in¯uenced the heavy metals adsorption and movement as shown by the signi®cant correlation with the retardation factors when the combined data for the unlimed and limed soil samples was considered. # 2000 Elsevier Science Ltd. All rights reserved. Keywords: Heavy metals; Retardation factors; Soil chemical characteristics; Soil pollution
1. Introduction The impact of contaminants on the environment should be of scienti®c concern, in order to minimize the threat of soil and groundwater pollution. Waste disposal on agricultural lands has been increasingly favored lately and, therefore, it should be scrutinized to diminish the risk of introducing pollutants to soils and waters. In this context, presence of heavy metals in several wastes used in the Brazilian agriculture today has imposed a need for a better understanding of the processes of soil±heavy metal interactions, in particular, their mobility and retention.
* Corresponding author. Tel.: +55-31-899-2630; fax: +55-31-8992648. E-mail address: [email protected] (M.P.F. Fontes).
The waste disposal, whether in sanitary land®ll or, when processed for agricultural use demands understanding of the adsorption phenomena and pollutant mobility in soil pro®les, since these are essential factors that control groundwater contamination (Bittel and Miller, 1974; Elliot et al., 1986). Several laboratory studies, using disturbed soil samples, have been conducted to ®nd soil characteristics which are correlated with adsorption and mobility of heavy metals; however, mixed results have been found due to dierent mineralogical and chemical soil characteristics, and the selectivity of retention sites for each cation. Soil characteristics positively correlated with cadmium (Cd) retention were pH (Tyler and McBride, 1982; Christensen, 1984; King, 1988; Jopony and Young, 1994), organic matter (OM) content and cation exchange capacity (CEC) (Sidle and Kardos, 1977), speci®c surface area (Korte et al., 1976), while free iron oxides was negatively correlated (Amacher et al.,
0269-7491/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved. PII: S0269-7491(00)00088-9
430
A.T. de Matos et al. / Environmental Pollution 111 (2001) 429±435
1986). Zinc (Zn) retention was positively correlated with pH (Harter, 1983; King, 1988), CEC (Shuman, 1977; Sidle and Kardos, 1977) and speci®c surface area (Korte et al., 1976). A number of authors have found positive correlations between copper (Cu) retention and pH (Tyler and McBride, 1982; Harter, 1983; King, 1988), sum of bases (SB) or exchangeable calcium (Ca) (Harter, 1979) OM content (Hickey and Kittrick, 1984) and to the CEC (Sidle and Kardos, 1977). Lead (Pb) retention was better correlated to clay content (King, 1988), pH (Harter, 1983; Jopony and Young, 1994) and SB or exchangeable Ca (Harter, 1979). Korte et al. (1976), using multiple regression analysis, found clay content, speci®c surface area, free iron (Fe) oxide content and free carbonates in¯uencing Cd, Zn, nickel (Ni) and chromium (Cr) adsorption. Some reports found no correlation between soil OM content and metal adsorption (Harter, 1979, 1983). The author credited this lack of correlation to the fact that total OM of a wide variety of soils may not be suciently related to their reactive portions to generate a positive correlation. Many authors found that both free and amorphous Fe oxides had high capacities to adsorb metals because of their ability to make covalent links with them (Korte et al., 1976; Shuman, 1977; Hickey and Kittrick, 1984; Tiller et al., 1984; King, 1988; Slavek and Pickering, 1988). However, Amacher et al. (1986) and Elliot et al. (1986) found negative correlation between free Fe oxide content and retention of Cd, Cu, Zn and Pb. Experiments with disturbed soil samples are important to understand the adsorption of heavy metals; however, estimation of heavy metal cation movement in soils necessitates a better approximation of the chemical retardation processes, which can be done by using soil column experiments. The retardation factor of solutes moving through soils, a parameter that expresses the relationship between the speed of solute movement and the water-leaching front speed (Valocchi, 1984), is an indicator of interactions between the solution and the solid phase. It can give an idea about the mobility of cations throughout the soil pro®le. Estimation of retardation factors of dissolved carbon using soil columns (Li and Shuman, 1997a) and leaching experiments with soil columns to asses mobility of heavy metals (Boyle and Fuller, 1987; Li and Shuman, 1997b) is common practice. But less common is the use of soil column experiments to estimate retardation factors of heavy metals, relate them to their mobility in soils and to the soil chemical and mineralogical characteristics. Therefore, in order to improve the predictive capability of retention and mobility of some heavy metals in soils, the objective of this study was to relate the heavy metals movement, as measured by the Zn, Cd, Cu and Pb retardation factors, to several chemical and
mineralogical soil characteristics for selected Brazilian soils. 2. Materials and methods Samples of the A, B and C horizons of two Ultisols (U1 and U2) and one Oxisol (Ox), from VicËosa, Minas Gerais State, were air dried and passed through a 2-mm sieve. The texture was determined by the pipette method. The clay was analyzed by X-ray diraction, using Cu-Ka monochromatic radiation obtained by a Ni ®lter. Kaolinite and gibbsite contents were obtained by dierential thermal analysis. The pH was determined in water with a 1:2.5 soil:solution ratio. Exchangeable calcium (Ca-KCl) was extracted with KCl 1 mol lÿ1 and non-exchangeable Ca (Ca-HCl) was obtained extracting with boiling HCl 1 mol lÿ1 (Lanyon and Heald, 1982). Crystalline iron (Fed) was determined by dithionite-citrate extraction (Con, 1963) and amorphous iron (Feo) by ammonium oxalate extraction (Schwertmann, 1964). The values of SB and CEC were obtained by calculation from the exchangeable cations contents determined according to Thomas (1982). Organic carbon was determined by Walkley-Black method according to Jackson (1958) and converted to OM. Speci®c surface area was determined according to Carter et al. (1965). Ca and magnesium (Mg) were determined by atomic absorption spectrophotometry, sodium (Na) and potassium (K) by ¯ame photometry and Fe by colorimetry. Chemical and mineralogical characteristics used in correlation analysis are presented in Table 1. Euent breakthrough curves for heavy metals were measured using soil-packed columns. Contaminant solutions were ponded on the soil column surface and 1.5 cm of hydrostatic pressure was maintained during the leaching process. The euent breakthrough curves were measured for solutions containing concentrations of 700 mg lÿ1 of Zn, 20 mg lÿ1 of Cd, 200 mg lÿ1 of Cu and 300 mg lÿ1 of Pb. The retardation factors (Rf) (Table 2) were analytically determined, from regression equations, as the necessary number of pores to make the euent concentration equal to 50% of the concentration of the contaminant solution (Nielsen and Biggar, 1961; Wierenga et al., 1975; Baes and Sharp, 1983; Christensen, 1985; Van Genuchten and Wierenga, 1986). A complete description on how the retardation factors were obtained can be found in de Matos et al. (1999). Sub-samples of each soil were treated with CaCO3, cp, in a dose of 8 t haÿ1, allowed to equilibrate for a period of 15 days and submitted to the same procedure to calculate the retardation factors. Four replicate columns of each soil, horizon, lime treatment and controls were set up.
A.T. de Matos et al. / Environmental Pollution 111 (2001) 429±435
431
Table 1 Chemical and mineralogical characteristics of unlimed and limed soil samplesa Soil
Horizon pH
Ca2+ (Cmol kgÿ1) KCl
SB CEC Al3+ OM Fed Feo SSA Ka Gb Clay ÿ1 ÿ1 (Cmol kg ) (Cmol kg ) (Cmol kgÿ1) (g kgÿ1) (g kgÿ1) (g kgÿ1) (m2 gÿ1) (g kgÿ1) (g kgÿ1) (g kgÿ1)
HCl
U1
A B C
6.02 3.09 6.48 1.46 6.85 0.83
4.62 2.67 1.83
4.37 1.95 1.16
4.67 2.15 1.16
0.30 0.20 0.00
55.1 16.8 5.4
55.4 74.7 33.8
0.12 0.13 0.07
20.1 16.8 8.7
350 460 90
767 863 905
36 54 51
U2
A B C
6.26 3.94 5.36 0.45 5.88 0.02
6.20 1.68 1.29
6.09 0.85 0.17
6.19 1.35 0.57
0.10 0.50 0.40
52.4 24.2 11.4
64.1 82.8 77.4
0.07 0.12 0.12
22.0 21.2 19.0
520 730 380
819 765 785
28 16 43
Ox
A B C
4.61 0.25 5.06 0.08 5.38 0.00
1.54 1.34 0.53
0.52 0.14 0.07
2.32 0.94 0.72
1.80 0.80 0.65
57.1 30.2 2.7
63.8 75.6 29.0
0.38 0.17 0.10
26.2 21.5 12.6
630 720 40
582 709 853
26 34 21
U1 Lb A B C
8.34 6.41 8.53 5.73 8.70 4.21
11.78 10.09 8.29
9.92 4.06 2.22
9.92 4.06 2.22
n.d. n.d. n.d.
U2 L
A B C
8.45 7.98 8.31 6.46 8.45 6.17
14.32 9.36 9.94
10.04 5.46 3.50
10.04 5.46 3.50
n.d. n.d. n.d.
Ox L
A B C
6.49 6.27 7.78 6.07 8.53 4.32
8.32 8.49 8.81
8.61 6.69 2.46
10.51 6.69 2.46
n.d. n.d. n.d.
a SB, sum of bases; CEC, cation exchange capacity; OM, organic matter; Fed, Fe extracted by dithionite; Feo, Fe by oxalate; SSA, speci®c surface area; Ka, kaolinite; Gb, gibbsite; n.d., not detected. b L, limed soils.
3. Results and discussion 3.1. Retardation factors The retardation factors for Zn, Cd, Cu and Pb obtained for the three soils and horizons are shown in Table 2. Zn generally has the lowest values followed closely by Cd whereas Cu and Pb reached the highest values, especially Pb. It means that Zn and Cd have the highest mobility associated to the lowest adsorption and Cu and Pb present the opposite behavior. The Ultisols (U1 and U2) reveal, as a general trend, a decrease of the retardation factors with depth. This showed that all of the heavy metals, for these soils, are more adsorbed in the A horizon, than in the B and C horizons. For the Oxisol (Ox), the B horizon had higher retardation factors as compared to A and C horizons, indicating that in the B horizon of this soil all the heavy metals have less mobility. In general, the values for Rf increased dramatically when the soils were limed. This shows that the mobility of all the heavy metals decreased with the pH increase. The phenomenon can be related to both the increasing of the speci®c and non-speci®c adsorption and to the precipitation reaction. McBride (1994, p. 320) reports that an experiment working with accelerated leaching in
a column of an acid mineral soil have produced some movement of trace metal ions initially applied to the column surface. He also relates that the same soil after liming to pH 6.5 retained all metals more strongly. The author concludes that liming reduces both mobility and plant availability of toxic metals. 3.2. Correlation studies Using the retardation factor values (Table 2) and soil chemical and mineralogical characteristics (Table 1) a linear correlation study was conducted. Table 3 present the signi®cant simple linear correlation coecients between retardation factors and the soil characteristics for Zn, Cd, Cu and Pb, respectively. Because this study used three soils and three horizons making a total of nine observations, the correlation coecients considered signi®cant were only those with probability level smaller than 1 and 0.1% (P<0.01 and P<0.001). For the linear correlation study, the mineralogical soil characteristics studied gave no signi®cant correlation coecients with retardation factors for any metal. The reasons could be: for the amorphous Fe oxides (Feo) their very low amounts in these highly weathered tropical soils (Fontes and Weed, 1991); for the crystalline Fe oxides (Fed), it can be related to the fact that the in¯uence of the
432
A.T. de Matos et al. / Environmental Pollution 111 (2001) 429±435
Table 2 Retardation factors (Rf) for Zn, Cd, Cu and Pb of unlimed and limed (L) soil samples Soil
Horizon
Rf Zn
Cd
Cu
Pb
U1
A B C
2.308 1.566 1.727
3.375 1.841 1.732
6.445 3.393 2.742
12.571 6.691 4.933
U2
A B C
2.398 0.932 1.272
3.151 1.146 1.410
5.676 2.287 2.792
11.264 4.830 5.032
Ox
A B C
0.902 1.347 1.244
1.226 1.466 1.106
3.238 3.880 1.781
5.901 6.413 3.750
U1 L
A B C
5.830 3.864 3.363
8.845 4.528 4.895
15.664 10.317 11.981
27.795 15.786 19.710
U2 L
A B C
5.204 2.687 2.214
7.410 3.884 2.807
15.301 8.093 6.587
36.528 13.248 8.610
Ox L
A B C
3.164 3.911 3.231
4.787 5.612 3.849
8.353 9.341 8.486
14.744 13.158 10.956
Fe oxides in certain soil reactions depends not only on the amounts but it also depends on the relative proportions of hematite and goethite, their crystallinity, their preferential exposed faces, etc. (Fontes and Weed, 1996) and for kaolinite and gibbsite contents the same reasoning probably applies, although there are no studies relating their contents directly to adsorption of heavy metals. The correlation observed between mineralogical characteristics and adsorption of heavy metals in laboratory-disturbed soil samples, well documented in the literature, and the lack of correlation between these variables and retardation factors for the heavy metals in
our study is suggesting that there are other variables which can also play important roles in leaching column experiments. In this case, it seems that hydraulic conductivity, porosity, etc., for example, would in¯uence the residence time of the solution in the column which, in turn, would aect the mobility and retention of the metals. In the limed soils most of the characteristics did not show signi®cant correlation with the retardation factors for the metals studied. With the pH increase brought about by the liming there was an increase in the metals adsorption and there was also an increase of the metals precipitation. It seems that, at such high pHs reached by the strong lime application, the reaction that governed the metals behavior was the precipitation. These results agree with McBride (1994) who states that liming reduces both mobility and plant availability of toxic metals and reinforces the observations that overliming can decrease heavy metals movement in soils. It can be seen in Table 1 that the soils pH, with the exception of the A horizon of the Oxisol, reached values around 8.0 or more, which is well above the recommended pH range for agricultural uses in these soils. The purpose of liming highly weathered tropical soils like these, is primarily to neutralize the exchangeable aluminum (Al), and this is normally accomplished by raising the pH to 5.5 (Sanchez, 1976, p. 240; Uehara and Gillman, 1981, p. 66). 3.3. Zn In general, retardation factors for Zn, were smaller than those for the other metals (Table 2), showing that its mobility in these soils is the highest. The CEC, SB and soil Ca content (Ca-KCl, Ca-HCl) were signi®cantly correlated with Zn retardation factor (Zn-Rf) in the unlimed soil samples (Table 3). The signi®cance of the relationship between Zn-Rf and CEC, SB and Ca contents show that, in this competitive system, Zn has its adsorption to the soil solid particles more related to the non-speci®c, i.e. electrostatic adsorption.
Table 3 Correlation coecients between Zn, Cd, Cu and Pb retardation factors and some soil chemical and mineralogical characteristicsa Characteristic chemical
pH OM SB CEC Ca-KCl Ca-HCl a
Zn
Cd
Cu
Pb
Unlimed
Limed
Unlimed
Limed
Unlimed
Limed
Unlimed
Limed
0.657n.s. 0.348n.s. 0.799** 0.793** 0.907*** 0.873**
0.052n.s. 0.598n.s. 0.587n.s. 0.587n.s. 0.310n.s. 0.559n.s.
0.491n.s. 0.581n.s. 0.941*** 0.894*** 0.956*** 0.939***
0.001n.s. 0.721n.s. 0.783** 0.707n.s. 0.450n.s. 0.639n.s.
0.229n.s. 0.763** 0.852** 0.867** 0.864** 0.879***
0.274n.s. 0.513n.s. 0.581n.s. 0.489n.s. 0.337n.s. 0.756**
0.213n.s. 0.764** 0.896*** 0.911*** 0.904*** 0.910***
0.168n.s. 0.614n.s. 0.680n.s. 0.602n.s. 0.549n.s. 0.835**
OM, organic matter; SB, sum of exchangeable bases; CEC, cation exchange capacity; n.s., not signi®cant. **Signi®cant to 1%. ***Signi®cant to 0.1%.
A.T. de Matos et al. / Environmental Pollution 111 (2001) 429±435
The exchangeable Ca content showed the highest simple linear correlation coecient (r=0.907, P<0.001) with Zn-Rf variation suggesting that Zn adsorption in exchange sites may be the preferential form of this metal retention in these soils. These results agree with many authors (Sidle and Kardos, 1977; Harter, 1979; Hickey and Kittrick, 1984; de Matos et al., 1994, 1996). In the limed soils, none of the characteristics presented signi®cant correlation with Zn-Rf, at the probability levels tested. 3.4. Cd The values for Cd-Rf were slightly higher than those of Zn, but still much lower than Cu and Pb also indicating a high mobility of Cd in these pro®les. The Cd retardation factors (Cd-Rf) in unlimed soil samples, as already observed for Zn, were signi®cantly correlated with the SB, CEC and the soil Ca contents (Table 3). All of these chemical characteristics were highly correlated with CdRf as seen by the probability values (P<0.001). In the limed soil materials, the behavior was similar, with the correlation coecients showing no signi®cance with the Cd-Rf, except for SB which showed r=0.783 and P<0.01. 3.5. Cu Cu presented higher values for Rf, as compared to Zn and Cd, showing that it is more adsorbed and has less mobility than these two metals, in these soil pro®les. In the unlimed soil samples, signi®cant correlation of Cu retardation factor (Cu-Rf) were obtained with CEC, OM and soil Ca contents (Ca-KCl and Ca-HCl) (Table 3). In a dierent way of what was observed for Zn and Cd, the OM presented a signi®cant correlation coecient with Cu-Rf (P<0.01), which demonstrates the strong association of OM with this metal already observed by several authors (Sidle and Kardos, 1977; Hickey and Kittrick, 1984; de Matos et al., 1996). The Cu retardation factor in the limed soil samples was signi®cantly correlated with Ca-HCl (r=0.756, P<0.01). It seems to indicate that precipitation is the reaction which is favored and probably accounts for the majority of the Cu retention in these soils when limed. The correlation coecient for OM decreased, becoming non-signi®cant, what could be due to the lowering of the more reactive OM content which always accompanies the liming. 3.6. Pb Pb presented, consistently, the highest values for retardation factors showing clearly that it is highly immobilized by these soils.
433
For the unlimed soil samples, Pb-Rf presented similar correlation (Table 3) with the same soil characteristics as observed for Cu, i.e. the OM content, SB, CEC and soil Ca contents. The coecient for OM was signi®cant at a P value<0.01, demonstrating of the importance of this soil fraction on the reactions that control the mobility of Pb. In the limed soil samples, all of the coecients had their signi®cance decreased, reaching no signi®cance, except for Ca-HCl, which showed a strong association to Pb-Rf (r=0.835, P<0.01). This association, as discussed before for Cu, can be due to the precipitation reaction. 3.7. Combined analysis In order to emphasize the dierences between the behavior of the retardation factor in the soil samples with liming as compared to the samples with no liming, a combined analysis was conducted, putting together all the data from the unlimed and the limed soil samples. The correlation coecients are presented in Table 4. The additional and maybe the most important aspect in the combined analysis is the inclusion of soil pH as a variable of signi®cant correlation to all the metals retardation factors. This correlation can be explained by considering that pH is the main soil characteristic to in¯uence the CEC of highly weathered soils and the dominant ionic forms in solution. The pH and the CEC are, almost always, reported as soil characteristics to show good association to soil adsorption of elements such as Cd and Zn (Tyler and McBride, 1982; Christensen, 1984; King, 1988). Also, the speci®c adsorption and or precipitation reactions participate increasingly in all metals adsorption (Harter, 1979). For the individual elements in the combined analysis, it still may be seen the strong association between SB, CEC and soil Ca contents and Zn-Rf and Cd-Rf. The initial observations showed that a large part of the Zn and Cd retention was due to the exchange sites, but the combined Table 4 Correlation coecients between retardation factors and some soil chemical and mineralogical characteristics for the combined analysisa Characteristic chemical
Zn
Cd
Cu
Pb
PH OM SB CEC Ca-KCl Ca-HCl
0.688*** 0.341n.s. 0.729*** 0.773*** 0.769*** 0.819***
0.722*** 0.403n.s. 0.889*** 0.841*** 0.851*** 0.886***
0.786*** 0.319n.s. 0.807*** 0.753*** 0.857*** 0.922***
0.636** 0.441n.s. 0.824** 0.782*** 0.776*** 0.849***
a
OM, organic matter; SB, sum of exchangeable bases; CEC, cation exchange capacity; n.s., not signi®cant. **Signi®cant to 1%. ***Signi®cant to 0.1%.
434
A.T. de Matos et al. / Environmental Pollution 111 (2001) 429±435
analysis suggests that part of it may be associated with speci®c adsorption (Garcia-Miragaya and DaÂvalos, 1986) and with precipitation (Harter, 1979). By looking at the relationship between the Cu-Rf and the soil characteristics in the combined analysis, it can be observed the same trend as showed for Zn and Cd. The SB (r=0.807, P<0.001) and the Ca-HCl (0.922, P<0.001) were soil characteristics which better associated to the Cu-Rf. Very similar results were obtained by Harter (1979), who found high signi®cance to the correlation between Cu adsorption, SB and the soil Ca content. The same author considered that the Cu hydrolysis, which forms CuOH+, under pH values above 6.0, was the factor of greater importance to the soil Cu retention, since it would make it possible adsorption of larger amounts of these species in the exchange complex. Also for Pb-Rf, there was a reiteration of the same trend veri®ed to all other heavy metals, with the signi®cant correlation being with pH, SB, CEC, Ca-KCl and Ca-HCl. Harter (1979) found similar correlation with Pb adsorption, attributing it to the hydrolysis of Pb2+ to PbOH+, making possible more ionic adsorption in the exchange complex, as well as metal precipitation. 4. Conclusions From the results the following conclusions can be drawn. Soil types and soil horizons in¯uenced the metals retardation factors. The Ultisols showed, consistently, higher retardation factors as compared to the Oxisol. Also, the Ultisols showed higher retardation factors for the A horizons, whereas the Oxisol showed, generally, higher retardation factors for the B horizon. The soil chemical characteristics showed signi®cant correlation to the metals retardation factors whereas the mineralogical characteristics did not. In the unlimed soil samples, the SB and the exchangeable Ca content (CaKCl) were the characteristics of better association to the metals retardation factors. In the limed soil samples, the signi®cant correlation coecients were observed with the SB, OM content, and non-exchangeable Ca content (Ca-HCl). The soil chemical characteristics were better estimators than the soil mineralogical characteristics to predict heavy metals mobility and retention in these soils. Soil pH was a variable that in¯uenced heavy metals adsorption, retention and movement, as measured by the retardation factors, showing high signi®cant correlation in the combined analysis of unlimed and limed soil samples. Liming the soils was very eective in controlling heavy metals mobility in the studied soils and, therefore, should be considered as an aid in environmental protection management practices.
References Amacher, M.C., Kotuby-Amacher, J., Selim, H.M., Iskandar, I.K., 1986. Retention and release of metals by soils: evaluation of several models. Geoderma 38, 131±154. Baes, C.F., Sharp, R.D., 1983. A proposal for estimation of soil leaching constants for use in assessment models. Journal of Environmental Quality 12, 17±28. Bittel, J.E., Miller, R.J., 1974. Lead, cadmium, and calcium selectivity coecients on montmorillonite, illite, and kaolinite. Journal of Environmental Quality 3, 250±253. Boyle, M., Fuller, W.H., 1987. Eect of municipal solid waste leachate composition on zinc migration through soils. Journal of Environmental Quality 16, 357±360. Carter, D.L., Heilman, M.D., Gonzalez, C.L., 1965. Ethylene glycol monoethyl ether for determining surface area of silicate minerals. Soil Science 100, 356±360. Christensen, T.H., 1984. Cadmium soil sorption at low concentrations, II: reversibility, eect of changes in solute composition, and eect of soil aging. Water, Air and Soil Pollution 21, 115±125. Christensen, T.H., 1985. Cadmium soil sorption at low concentrations, III: prediction and observation of mobility. Water, Air and Soil Pollution 26, 255±264. Con, D.E., 1963. A method for determination of free iron oxides in soils and clays. Can. Journal of Soil Science 43, 9±17. de Matos, A.T., Fontes, M.P.F., JordaÄo, C.P. 1994. Mobility and retention of Cd, Zn, Cu and Pb in a Brazilian Oxisol pro®le. In: XV World Congress of Soil Science Acapulco. Transactions, Vol. 3b; pp. 193±194. de Matos, A.T., Fontes, M.P.F., JordaÄo, C.P., da Costa, L.M., 1996. Mobility and retention forms of heavy metals in a Red-Yellow Latosol pro®le. Revista Brasileira de CieÃncia do Solo 20, 379±386. de Matos, A.T., da Costa, L.M., Fontes, M.P.F., Martinez, M.A., 1999. Retardation factors and the dispersion diusion coecients of Zn, Cd, Cu and Pb in soils from VicËosa- MG, Brazil. Transactions of the American Society of Agricultural Engineers (ASAE) 42 (4), 903±910. Elliot, H.A., Liberati, M.R., Huang, C.P., 1986. Eect of iron oxide removal on heavy metal sorption by acid subsoils. Water, Air and Soil Pollution 37, 379±389. Fontes, M.P.F., Weed, S.B., 1991. Iron oxides in selected Brazilian Oxisols: I. Mineralogy. Soil Science Society of America Journal 55 (4), 1143±1149. Fontes, M.P.F., Weed, S.B., 1996. Phosphate adsorption in clays from Brazilian Oxisols: relationships with speci®c surface area and mineralogy. Geoderma 72, 37±51. Garcia-Miragaya, J., DaÂvalos, M., 1986. Sorption and desorption of Zn on Ca-kaolinite. Water, Air and Soil Pollution 27, 217±224. Harter, R.D., 1979. Adsorption of copper and lead by Ap and B2 horizons of several northeastern United States soils. Soil Science Society of America Journal 43, 679±683. Harter, R.D., 1983. Eect of soil pH on adsorption of lead, copper, zinc, and nickel. Soil Science Society of America Journal 47, 47±51. Hickey, M.G., Kittrick, J.A., 1984. Chemical partitioning of cadmium, copper, nickel and zinc in soils and sediments containing high levels of heavy metals. Journal of Environmental Quality 13, 372±376. Jackson, M.L., 1958. Soil Chemical Analysis. Prentice-Hall, Englewood Clis, NJ. Jopony, M., Young, S.D., 1994. The solid±solution equilibria of lead and cadmium in polluted soils. European Journal of Soil Science 45, 59±70. King, L.D., 1988. Retention of metals by several soils of the southeastern United States. Journal of Environmental Quality 17, 239±246. Korte, N.E., Skopp, J., Fuller, W.H., Niebla, E.E., Alesii, B.A., 1976. Trace element movement in soils: in¯uence of soil physical and chemical properties. Soil Science 122, 350±359.
A.T. de Matos et al. / Environmental Pollution 111 (2001) 429±435 Lanyon, L.E., Heald, W., 1982. Magnesium, calcium, strontium and barium. In: Page, A.L., Miller, R.H., Keeney, D.R. (Eds.), Methods of Soil Analysis Part 2: Chemical and Biological Properties. ASA, SSSA, Madison, pp. 247±262. Li, Z., Shuman, L.M., 1997a. Estimation of retardation factor of dissolved organic carbon in sandy soils using batch experiments. Geoderma 78, 197±206. Li, Z., Shuman, L.M., 1997b. Mobility of Zn, Cd, and Pb in soils as aected by poultry litter extract Ð I. Leaching in soil columns. Environmental Pollution 95, 219±226. McBride, M.B., 1994. Environmental Chemistry of Soils. Oxford University Press, New York. Nielsen, D.R., Biggar, J.W., 1961. Miscible displacement in soils. I: experimental information. Soil Science Society of America Journal 25, 1±5. Sanchez, P.A., 1976. Properties and Management of Soils in the Tropics. John Wiley, New York. Schwertmann, U., 1964. The dierentiation of iron oxide in soils by a photochemical extraction with acid ammonium oxalate. Z. P¯anzenernahr. Dueng. Bodenk. 105, 194±201. Shuman, L.M., 1977. Adsorption of Zn by Fe and Al hydrous oxides as in¯uenced by aging and pH. Soil Science Society of America Journal 41, 703±706. Sidle, R.C., Kardos, L.T., 1977. Adsorption of copper, zinc, and cadmium by a forest soil. Journal of Environmental Quality 6, 313±317.
435
Slavek, J., Pickering, W.F., 1988. Metal ion interaction with the hydrous oxides of aluminium. Water, Air and Soil Pollution 39, 201±216. Thomas, G.W., 1982. Exchangeable cations. In: Page, A.L., Miller, R.H., Keeney, D.R. (Eds.), Methods of Soil Analysis. Part 2: Chemical and Biological Properties. ASA, SSSA, Madison, pp. 159±165. Tiller, L.D., Gerth, J., Brummer, G., 1984. The relative anities of Cd, Ni and Zn, for dierent soil clay fractions and goethite. Geoderma 34, 17±35. Tyler, L.D., McBride, M.B., 1982. Mobility and extractability of cadmium, copper, nickel, and zinc in organic and mineral soil columns. Soil Science 134, 198±204. Uehara, G., Gillman, G., 1981. The Mineralogy, Chemistry, and Physics of Tropical Soils with Variable Charge Clays. Westview Press, Boulder. Valocchi, A.J., 1984. Describing the transport of ion-exchanging contaminants using an eective Kd approach. Water Resource Research 20, 499±503. Van Genuchten, M.T., Wierenga, P.J., 1986. Solute dispersion: coecients and retardation factors. Methods of Soil Analysis. Part 1: Physical and Mineralogical Methods, ASA, SSSA Madison, 1025± 1031. Wierenga, P.J., Van Genuchten, M.T., Boyle, F.W., 1975. Transfer of boron and tritiated water through sandstone. Journal of Environmental Quality 4, 83±87.