Abstract : Aqueous solutions of copper, chromium and arsenic (CCA) are extensively used in the
timber preservation industry, and soils at timber treatment site commonly become
contaminated with these three elements. The potential mobility of CCA is of particular
concern since downward leaching of the heavy metals may result in the contamination of
groundwater. The aim of this study is to quantify the leaching of the CCA using laboratory
column and batch experiments and to derive statistic models.
In order to conciliate the field conditions and the laboratory constraints, we have
chosen to experiment the CCA leaching from long-term contaminated and non-destructured
soils. In addition, destructured soils (easier to sample) have been used in batch experiment
(easier to realise than column) in order to compare the results of these two different
From 5 timber treatment sites, 14 contaminated areas have been recognized and
chosen to provide a wide range of physicochemical soil properties. Within each area, of about
800 cm2, 4 soil core replicates (diameter 10 cm, depth 10 cm) were sampled.
The CCA soil contents vary from about 10 to several thousand mg/kg and generally
show the predominance of Cu over Cr and As (Cu > Cr > As). High lateral variations and
strong vertical attenuation of the signal (> 10 times) were observed. Such heterogeneities in
the soil may result of an association of a fast fixation rate along the vertical direction in the
macropores, and a slower diffusion in the micropores acting in all spatial directions.
Rain events were simulated experimentally by a series of leaching cycles in saturated
condition. As CCA concentrations measured in each cycle quickly reach a quasi stationary
state, average concentrations for each pollutant and for each soil have been estimated.
From the measured concentrations of the different CCA aqueous species (speciation),
we have observed that the order of the CCA leachate concentrations is the following : Cr(VI)
< Cutot / Cu2+ < As(V). The majority of the Cr(VI) concentrations are below the 50 μg/l
toxicity level. The Cu2+ proportions are variable and may reach several mg/l from the soil
with the highest Cu content. As(V) concentrations are high and exceed the critical values (50
μg/l) for the soil majority.
We have used statistical analysis with multiple linear regressions to link the CCA
leaching concentrations with the most significant physico-chemical soil characteristics. The
Cu and As concentrations are systematically correlated with the initial soil contents and with
the soil pH. The best model was found for Cu, explaining 96 % of the common variations.
The As model explains 94 % of the common variations but it presents some strong
differences between measured values and modelled values. We could not modelise the Cr
leaching. Relationships are also proposed to link batch with columns results. With care and
verification, such models may be used to assess the CCA leachate concentrations by simple
measures of few soil parameters.