A versatile parameter for comparing the capacities of soils for sorption and retention of heavy metals dumped individually or together: Results for cadmium, copper and lead in twenty soil horizons

https://doi.org/10.1016/j.jcis.2008.08.027Get rights and content

Abstract

Heavy metals can be immobilized by soils and their distribution among the particulate soil components depends on the identity and amount of the metal, the properties of the soil, and other environmental factors. Cd, Cu and Pb are among the most potentially toxic heavy metals, are present—often together—in numerous polluting spills and in agrochemicals. We evaluated the individual and competitive sorption and retention of Cd, Cu and Pb on 20 soil horizons. As is usual, the isotherms constructed were so irregular, especially the retention isotherms, that it was not possible to classify and compare them in terms of the conventional isotherm shapes. Nor could they be compared using Langmuir or Freundlich parameters, since not all could be fitted with either of these equations. They were therefore characterized and compared in terms of several varieties of distribution coefficient, including a novel adimensional parameter Kr which on the basis of correlation and principal components analyses was judged to be the most coherent and generally applicable to all experimental conditions (sorption and desorption starting from single- or multi-metal solutions). Kr proved to be mainly determined by soil pH, effective cation exchange capacity, and Mn oxides content.

Graphical abstract

Correlations for calculation of Kr. Potentially sorbable metal = initial concentration of metal in solution (μmol L−1) divided by concentration of soil in solution (g L−1).

  1. Download : Download full-size image

Introduction

The availability of heavy metals and other soil pollutants for uptake by plants, and the risk of their finding their way into surface or underground waters, depend on their sorption and desorption by soil components (where the term “sorption” is used to encompass adsorption, precipitation on soil particle surfaces, and fixation, and the term “desorption” for the release of sorbed species into the medium surrounding the particles by which they had been sorbed) [1], [2], [3]. The sorption and desorption of cations by soil predominantly involve negatively charged surfaces of organic matter, clays and metallic oxides or hydroxides [4], [5], [6], and the distribution of added metals among these components and others depends on the metal species in question, the properties of the soil, and the amounts of metal added [7], [8], [9].

Cadmium, copper and lead can be introduced into the soil as undesired components of fertilizers, liming agents, sewage sludge and other urban and industrial waste [10], [11], and are often to be found together. Their distribution in the soil profile, and between the solid and liquid phase, depend not only on the factors mentioned above, but also on how the presence of each can affect the behaviour of the others. In spite of this, as in the cases of other metals, most studies of the behaviour of Cd, Cu and Pb in soils have concerned the interaction of soil with solutions containing just one of the metals, and have ignored the effects of mutual competition, which may be decisive in determining which metals occupy less selective binding sites. Since the heterogeneity of soil makes it very difficult to predict the potential mobility and distribution of even single metals, experimental data are essential.

Most experiments have determined the distribution of metals between soil and a solution of metal ion following a contact period during which equilibrium is assumed to be attained. Their results are often reported in terms of distribution coefficients Kd, defined as the ratio, at equilibrium, between the amount of sorbed metal per unit weight of soil and the concentration of the metal in solution [3], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22]. Such coefficients represent the net result of all the various processes by which metal ions can be transferred between soil and solution [23], and are satisfactory for comparing the behaviour of different soils with respect to a given cation under fixed conditions. They are especially useful when the irregularity of empirical sorption and/or desorption isotherms hampers or prevents the fitting of simple empirical curves or theoretical models such as the Freundlich and Langmuir isotherms, as is often the case when the presence of more than one metal results in competition for sorption sites. However, different authors have employed different variants of Kd. Gomes et al. [14] and Covelo et al. [17] reported Kd values (L kg−1) that were obtained using solutions with those initial metal concentrations above which the quantity of metal sorbed or retained tended to level off for some metals and fall for others. For each given metal and soil fraction, Covelo et al. [18] also averaged Kd over all the solution concentrations used, and employed this average to compare the behaviour of various soil fractions for the given metal. For a group of metals, Kaplan et al. [24] defined a compound distribution constant as the ratio between the total amount of sorbed metal per unit weight of soil and the total molar concentration of metal in solution, both totals being calculated by summing over metals. As a measure of the ability of a metal to compete for sorption sites on a soil, rather than of its propensity to be sorbed as such, Fan et al. [25] defined a “competitive coefficient,” Kc, as the ratio between the amount of metal sorbed per unit mass of soil and the initial concentration of that metal in solution, all divided by the sum of such ratios over all metals.

None of the above coefficients is suitable for comparing the sorption and retention capacities of different soils when both single-metal and multi-metal solutions are involved, the coefficients that may be suitable for one situation not being comparable with those that are suitable for another. For example, in the case of Kd as used by Gomes et al. [14] or Covelo et al. [17], the initial concentrations for a multi-metal solution are chosen, as noted above, as those above which the quantity of metal sorbed or retained tends to level off for some metals and fall for others; but at these initial concentrations, when single-metal solutions are used, Kd invariably continues to rise with concentration.

In this study, with a view to comprehensive evaluation and comparison of the capacities of soil horizons withstand pollution by cadmium, copper and lead (singly or jointly), we determined most of the parameters mentioned above for 20 relevant soil horizons and we also developed a dimensionless parameter, Kr, that appears to be a useful indicator of both capacity for sorption and capacity for retention of these metals in both single-metal and multi-metal pollution situations.

Section snippets

Materials and methods

We selected 20 soil horizons (S1–S20) as representative of the most widely soil orders in our region that commonly receive inputs containing Cd, Cu and/or Pb. Six samples of each horizon were collected using an Eijkelkamp Model A sampler and were transported in polyethylene bags to the laboratory, where they were air dried, passed through a 2 mm mesh sieve, pooled, and homogenized in a Fritsch Laborette 27 vibratory solid sample homogenizer. Three subsamples of the homogenized sample were used

Results and discussion

Table 1, Table 2, Table 3 list the measured properties of the soil horizons studied, and Table 4, Table 5 the Giles curve types and Langmuir and Freundlich sorption/retention capacity parameters in those cases in which it was possible to identify them from the experimental isotherm data. The most salient aspect of Table 4, Table 5 is that in many cases it was not possible to identify Giles curve types or to fit Langmuir or Freundlich equations, which makes it impossible to use these parameters

Conclusions

The parameters of the Langmuir and Freundlich equations are not suitable for comparing different soil horizons as regards their capacity for sorption or retention of heavy metals under different conditions (different metals, sorption or retention from single- or multi-metal solutions, etc.), because these models often fail to fit the experimental isotherms. Analogous reasons prevent the use of the commonly employed distribution coefficients (Kds). In this work we defined an adimensional

Acknowledgment

This research was supported by the Spanish Ministry of Education and Science under Project CGL2006-01016/BTE.

References (63)

  • F.A. Vega et al.

    J. Colloid Interface Sci.

    (2006)
  • M.R. Reddy et al.

    Environ. Pollut.

    (1986)
  • E.F. Covelo et al.

    J. Colloid Interface Sci.

    (2004)
  • E.F. Covelo et al.

    J. Colloid Interface Sci.

    (2004)
  • E.F. Covelo et al.

    J. Hazard. Mater.

    (2007)
  • E.F. Covelo et al.

    J. Hazard. Mater.

    (2007)
  • E.F. Covelo et al.

    J. Hazard. Mater.

    (2007)
  • E.F. Covelo et al.

    J. Hazard. Mater.

    (2007)
  • G.W. Thomas
  • C.H. Giles et al.

    J. Colloid Interface Sci.

    (1974)
  • K.W.T. Goulding et al.

    Agr. Ecosyst. Env.

    (1998)
  • C.E. Martínez et al.

    Environ. Pollut.

    (2000)
  • C. Dère et al.

    C. R. Geosci.

    (2006)
  • R.G. McLaren et al.

    Geoderma

    (1983)
  • H.B. Bradl

    J. Colloid Interface Sci.

    (2004)
  • U. Kraus et al.

    Sci. Total Environ.

    (2006)
  • S. Serrano et al.

    Geoderma

    (2005)
  • C.G. Campbell et al.

    Appl. Geochem.

    (2006)
  • D.L. Sparks et al.
  • R. Apak
  • G. Sposito et al.

    Soil Sci. Soc. Am. J.

    (1983)
  • S.P. McGrath et al.

    J. Soil Sci.

    (1992)
  • F.X. Han et al.

    Soil Sci.

    (2001)
  • G.W. Bruemmer et al.

    J. Soil Sci.

    (1988)
  • F.X. Han et al.

    Water Air Soil Pollut.

    (1999)
  • F.X. Han et al.
  • B.J. Alloway

    Heavy Metal in Soils

    (1995)
  • D.C. Adriano

    Trace Elements in the Terrestrial Environment

    (1986)
  • P.R. Anderson et al.

    J. Soil Sci.

    (1988)
  • P.C. Gomes et al.

    Soil Sci. Soc. Am. J.

    (2001)
  • E.F. Covelo et al.

    Commun. Soil Sci. Plant Anal.

    (2004)
  • Cited by (56)

    View all citing articles on Scopus
    View full text