JEQ Grow Your Career With ASA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (45)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Hammer, D.
Right arrow Articles by Keller, C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Hammer, D.
Right arrow Articles by Keller, C.
Agricola
Right arrow Articles by Hammer, D.
Right arrow Articles by Keller, C.
Related Collections
Right arrow Heavy Metals
Right arrow Soil Pollution
Right arrow Soil Analysis
Right arrow Soil Chemistry
Journal of Environmental Quality 31:1561-1569 (2002)
© 2002 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America

TECHNICAL REPORTS
Heavy Metals in the Environment

Changes in the Rhizosphere of Metal-Accumulating Plants Evidenced by Chemical Extractants

D. Hammer* and C. Keller

Swiss Federal Institute of Technology Lausanne, IATE-Pédologie, 1015 Lausanne, Switzerland

* Corresponding author (daniel.hammer{at}epfl.ch)

Received for publication May 14, 2001.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 SUMMARY AND CONCLUSION
 REFERENCES
 
The plants Salix viminalis L. (common osier) and Thlaspi caerulescens J. Presl & C. Presl have been studied often because of their high potential to extract heavy metals from soils. The soil properties favoring this phytoextraction are not yet fully known. In this study we compared three frequently used single-extracting agents (NaNO3, diethylenetriaminepentaacetic acid [DTPA], and ethylenediaminetetraacetic acid [EDTA]) with a sequential extraction procedure to describe changes in the different Cd, Cu, and Zn pools in the rhizosphere of S. viminalis and T. caerulescens grown on calcareous and acidic Swiss soils in a pot experiment. The sequential extraction was used to assess the chemical affinities of these heavy metals (HM) in the soil whereas the single extractants were used for estimating the bioavailable HM pools in the soils. Cadmium depletion in several pools was most apparent in the acidic soil, with a significant decrease observed in the NaNO3-, DTPA-, and EDTA-extractable fractions following T. caerulescens growth compared with control pots. The sequential extraction showed that most Cd extracted by the plant from the acidic soil originated from the organic pool, which implies that heavy metals bound to organic matter may constitute a significant part of the bioavailable Cd pool in soils. In the calcareous soil only a small amount of Cd was taken up by T. caerulescens, and this came mainly from the carbonate-bound fraction. This study shows that T. caerulescens, and to a lesser extent S. viminalis, can alter the heavy metal distribution in different soil pools within 90 d.

Abbreviations: DTPA, diethylenetriaminepentaacetic acid • EDTA, ethylenediaminetetraacetic acid • HM, heavy metal


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 SUMMARY AND CONCLUSION
 REFERENCES
 
HEAVY METAL (HM) contamination of the environment is a serious concern for agricultural land, ground water, and ultimately animal and human health. An environmentally sustainable approach to remediate moderate, diffuse, but shallow metal contamination is the use of plants that extract heavy metals from soils. This technique, phytoextraction, has already been described by Salt et al. (1995) or McGrath (1998), for example. Many authors have proposed the use of hyperaccumulators, which are plants that can accumulate exceptionally high HM concentrations, for example, >100 mg kg-1 Cd, >1000 mg kg-1 Ni, and >10 000 mg kg-1 Zn (see Brooks et al. [1998] for a review). Others have tested the potential for high-biomass crops to extract similar amounts of HM by compensating low HM concentrations in these plants by a larger biomass production (Felix, 1997). Examples of such crops are tobacco (Nicotiana tabacum L.), S. viminalis, or Indian mustard [Brassica juncea (L.) Czern.]. Besides the intrinsic plant characteristics, which vary between species, the efficiency and therefore the application of phytoextraction on a large scale is often hindered by the availability of a metal for the plant. This availability is a function of the metal to be looked at, its distribution, and its chemical form in the soil. Moreover, plants can change metal availability directly (uptake) and indirectly by different mechanisms (e.g., exudation of complexing agents, respiration of roots, which accounts for pH changes, etc.). The respective effects of these factors have not been quantified, in part because of procedural difficulties.

Several methods have been developed in an attempt to predict phytoavailability. These methods have largely been applied for agricultural purposes (mostly to study the availability of micronutrients) and therefore for agricultural plants and soils. In this context, phytoavailability has often been defined through a one-step soil-extracting procedure involving, for example, sodium nitrate (Gupta and Aten, 1993), DTPA (Lindsay and Norvell, 1978), or EDTA (Lakanen and Ervio, 1971). The amount of HM extracted by such methods gives an idea of the size of a pool that might be depleted by a plant during a growth period, but the method extracting power depends on the soil tested (for example, DTPA was developed for neutral to alkaline soils) and the "equivalence" to plant uptake has been questioned (Miner et al., 1997). Isotopic dilution techniques have also been used to estimate the availability of some elements to plants and include the E- and L-value approach (Echevarria et al., 1997, Hutchinson et al., 2000) and the isotopic exchange kinetic (IEK) (Frossard and Sinaj, 1997; Morel, 1997). Echevarria et al. (1997) showed that red clover (Trifolium pratense L.) took up Ni from the isotopic exchangeable pool and Hutchinson et al. (2000) observed that T. caerulescens was not able to take up "non-radiolabile" Cd. Hamon et al. (1997) compared eight plant species growing on a soil spiked with Cd and Zn isotopes and found that all plants accessed the same pool with the exception of canola (Brassica napus L.), which had a higher specific activity of Cd. Hamon et al. (1997) suggested that canola was unable to access a pool of soil Cd that was available for uptake to the other species. The IEK technique was first developed to estimate bioavailable P in agricultural soils and has been further applied to Ni (Echevarria et al., 1998), Zn (Frossard and Sinaj, 1997), and Cd (Gérard et al., 2000) with a good agreement with plant uptake. Isotopic methods are easy to apply and provide information on the availability of elements to plants. However, in isolation, isotopic methods do not provide information about the chemical form of the metal in question and hence make interpretation of underlying mechanisms difficult.

Beside the one-step soil extracting procedures and the isotopic dilution techniques there is a third approach, which is suggested to be more related to the chemical forms of metals in soils: specific chemical extractants used in a sequence can theoretically access different soil components, which can bind and release HM when subjected to different conditions, including changes in soil conditions induced by plant growth. Although one-step extracting agents can be specific, the application of a subsequent series of extractants of increasing power can be more specifically attributed to certain binding forms, as some of them have already been removed in the previous steps. Many extraction schemes have been proposed depending on the soil characteristics, the soil contamination status, and local regulations and may comprise up to eight steps (Ma and Uren, 1998). Metal distribution among the soil components depends on soil characteristics, metal involved, and type and age of contamination. However, the selectivity and specificity of these extractants are often subject to criticism (for a review see Kennedy et al., 1997). For this reason the pools extracted are usually described as "operationally defined." Nevertheless, the use of a sequential extraction can be justified when the aim is to compare differences in the same soil engendered by different treatments (e.g., the effect of plant growth on metal distribution). Indeed, these heavy metal pools may be selectively affected by plant uptake and changes in their proportions may give ideas on mechanisms responsible for heavy metal uptake. There has been very little research examining plant uptake in comparison with chemical pools in the soil. In the context of phytoextraction it is necessary to assess the respective effects of hyperaccumulators and crop plants on these chemically defined pools and more specifically the potential of these plants to mobilize or deplete metals from these pools. This allows assessment of the potential risks or benefits of using plants to remediate heavy metal–contaminated soils as well as the long-term efficiency of the technique.

In this study we compared three frequently used single-extracting agents with a sequential extraction procedure to describe changes in bioavailability of Cd, Cu, and Zn in the rhizosphere of S. viminalis (chosen as a crop plant) and T. caerulescens (Zn and Cd hyperaccumulator) grown on calcareous and acidic Swiss soils. Sodium nitrate is the method of reference in Switzerland for easily exchangeable and available HM (FAC, 1989). It was compared with DTPA, which is most suitable for calcareous soils, as it is buffered at a pH 7.3 and therefore prevents CaCO3 from dissolution and release of occluded metals (Lindsay and Norvell, 1978). Additionally, DTPA has the potential to strongly chelate Fe, Cu, Mn, and Zn. The third extractant used is EDTA buffered at pH 4.65, which was originally developed for acidic soils. Its extracting power is based on the combined action of ammonium acetate and EDTA, whereby the ammonium ion exchanges with trace elements and EDTA acts as a chelating agent forming stable chelates with many metal ions and preventing secondary precipitation of phosphate compounds during extraction (Cottenie et al., 1982). The six-step sequential extraction we applied was developed by Benitez and Dubois (1998) for calcareous soils and adapted from Tessier et al. (1979) and Keller and Védy (1994).


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 SUMMARY AND CONCLUSION
 REFERENCES
 
Plants
Cuttings of S. viminalis were used (Swedish Clone 78198). This clone has already been described by Landberg and Greger (1996) as heavy metal tolerant and an accumulator of cadmium and zinc in roots and shoots. Thlaspi caerulescens seeds were from a population grown near an ancient Pb and Zn mine at Saint-Laurent-Le-Minier, southern France ("Ganges" population; Robinson et al., 1998). Fanweed (Thlaspi arvense L.) seeds were obtained from the botanical garden of Geneva, Switzerland. Thlaspi arvense was chosen as a reference plant because it is also a Brassicaceae and was described by Lasat et al. (1996) as non-accumulating but tolerant to Zn concentrations up to 6.5 mg L-1 (100 µM) in hydroponic experiments.

Soils
A calcareous topsoil (FAO classification: calcaric Regosol) was collected in the northern part of the Swiss Jura (47°28'47'' N, 7°36'42'' E), exposed during the last century to atmospheric depositions of industrial Cu and Zn emissions from a nearby brass smelter. Kayser et al. (2000) have already described the site at Dornach, Switzerland and the field experiment performed on it. The second soil was sampled from an acidic topsoil (FAO classification: Fluvisol) in the southern part of Switzerland. This site was contaminated by wastes from septic tanks applied regularly between 1960 and 1980. General soil characteristics and total HM concentrations are given in Table 1.


View this table:
[in this window]
[in a new window]
  		Table 1. Selected chemical and physical properties of the soils used in the pots.

 
Pot Experiment and Sample Preparation
The two soils were homogenized, passed through a sieve with a mesh size of 1 cm, and then equilibrated moist during a period of at least 4 wk at 10°C in the dark. Two kilograms of fresh soil were transferred into pots (1.72 and 1.54 kg dry matter for the calcareous and the acidic soils, respectively). The soil was equilibrated again for 10 d. Thlaspi caerulescens was germinated and grown in an uncontaminated garden soil for 47 d and T. arvense for a period of 20 d, as required by their different growth rates. Ten-centimeter-long cuttings of S. viminalis were placed in a 1:4 diluted standard Hoagland solution (Hoagland no. 2 basal salt mixture; Sigma-Aldrich, Irvine, UK) for 10 d. The roots of all plants were then rinsed with deionized water and plants were transferred into the prepared pots. There were four replicates for each plant treatment, and duplicates for controls (pots as described above but without plants).

Plants were grown for 90 d in a controlled-environment growth chamber (16°C for 8 h and 20°C for 16 h) and watered with deionized water. Plants were then removed from the pots and rhizospheric soil was sampled from the surroundings of the roots (soil attached to the roots after shaking and separated from the roots by hand). Soil samples were dried at 35°C and sieved through a 2-mm nylon mesh.

Roots were washed with water and oven-dried at 70°C. Roots were prepared for metal analysis but no exact mass balance was drawn because no total root dry matter weight could be measured. Leaves from T. caerulescens as well as leaves and stalks from S. viminalis were collected, oven-dried at 70°C, and weighed. Roots and shoots were ground in a tungsten Retsch (Haan, Germany) mill.

All material in contact with solutions was decontaminated by soaking into 10% HNO3 (v/v) overnight followed by three rinses with purified water. Purified water (Milli-Q reagent grade water system; Millipore, Bedford, MA) was used throughout.

Soil and Plant Extraction and Analysis
General Soil Characteristics
Soil pH was measured in H2O (soil to water ratio of 1:2.5) after shaking for 1 h. All soil samples were ground to 100 µm and analyzed for organic carbon and nitrogen by a Carlo Erba (Milan, Italy) CHNS-O EA1108 elemental analyzer. Cation exchange capacity was determined by BaCl2 at the soil pH (FAL, 1998). Total heavy metal concentrations in the soils were determined after digestion in a HNO3–HClO4–HF mixture (Ruppert, 1987) as described below in the sixth step of the sequential procedure.

Single Soil Extractions
Metals extracted by 0.1 M NaNO3 (soil to solution = 1:2.5; horizontal shaker for 120 min; centrifugation for 10 min at 4000 rpm; filtration through a 0.45-µm membrane [FAC, 1989]) were compared with concentrations extracted with DTPA-TEA (0.005 M diethylenetriaminepentaacetic acid, 0.1 M triethanolamine, and 0.01 M CaCl2 at pH 7.3 [Lindsay and Norvell, 1978]) and EDTA-NH4OAc (0.02 M ethylenediamine-tetraacetic acid and 0.5 M ammonium acetate at pH 4.65; soil to solution = 1:10; horizontal shaker for 60 min; filtration through a Schleicher & Schuell [Dassel, Germany] 602 1/2 filter [FAC, 1989]).

Sequential Extraction
A sequential extraction was performed after Benitez and Dubois (1998). The extraction scheme consisted of six steps to extract heavy metals bound to the following operationally defined fractions: 1 = easily exchangeable (0.1 M NaNO3); 2 = carbonate compounds (1 M NaOAc buffered at pH 5); 3 = organic matter (0.1 M Na4P2O7); 4 = manganese oxides (0.25 M NH2OH·HCl and 0.05 M HCl); 5 = iron oxides (1 M NH2OH·HCl and 25% acetic acid); and 6 = residual (HF, HNO3, and HClO4) fractions. The procedure was performed as follows: 1 g of air-dried, finely ground sample was introduced in a 50-mL screw-cap centrifuge tube (PTFE). The soil sample was shaken for 90 min with 10 mL of reagent with a rotary shaker at 50 rpm. The sample was centrifuged at 23400 x g for 20 min and the supernatant stored in a polyethylene flask until analysis. This procedure was repeated with 10 mL of the extracting reagent and followed again with 10 mL of purified water. The sum of the three supernatants was transferred to a 50-mL volumetric flask and filled to the mark with purified water. This procedure was used to extract Fractions 1 to 4. Extraction of the carbonate fraction (Step 2) was performed on the calcareous soil only as recommended by Benitez and Dubois (1998). These authors extracted three acidic soils with the sequential procedure and found that acetic acid did not affect the organic pool in the absence of carbonates. However, they suggested that acetic acid attacked the manganese and iron oxides when carbonates were absent, and thus should not be used in acidic soils. Extraction of the iron-oxide fraction (Step 5) was performed as follows: 20 mL of reagent was introduced into the centrifuge tube and, after mixing solution and soil residue with a vortex, the tube was heated in a water bath at 95°C for 5 h. After cooling and centrifugation the supernatant was separated from soil residue. A washing step with 10 mL of purified water followed. The final residue (Step 6) was digested under pressure, with an acid mixture (3 mL HF 40% Merck [Darmstadt, Germany] Suprapur, 5 mL HNO3 65% Merck Suprapur, and 3 mL HClO4 70% Merck pro analysi) with a microwave oven (Mikrowellen-Labor-Systeme [Leutkirch, Germany] 1200 Mega). All solutions were stored at 4°C and analyzed the following day. The sequential extraction was repeated three times for each sample and results presented are the average of the three extractions for the four (with plant) or two (control) pot replicates.

Plant Digestion
Plant material was digested in a microwave as above, with a mixture of 2 mL H2O2 30% (Merck pro analysi), 2 mL purified H2O, and 4 mL HNO3 65% (Merck Suprapur) as extracting solution.

Heavy Metal Analysis
Cadmium content in solutions was always measured by graphite furnace atomic absorption spectrometry (GF–AAS) (PerkinElmer [Wellesley, MA] 5100PC atomic absorption spectrometer equipped with a HGA 600 graphite furnace, a Zeeman-corrected furnace module, and a AS-60 Furnace autosampler). Zinc and other elements (Ca, Cu, Fe, Mg, and Mn for soils; Ca, Cu, Fe, Mg, and P for plants) were measured by inductively coupled plasma atomic emission spectrometry (ICP–AES) (PerkinElmer Plasma, 2000).


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
SUMMARY AND CONCLUSION
 REFERENCES
 
Plant Growth
The yield data shows (Table 2) that on the calcareous soil S. viminalis performed well in comparison with the two Thlaspi species, whereas T. caerulescens produced the highest biomass on the acidic soil. Chlorotic leaves were observed for T. caerulescens and T. arvense on calcareous soil and for S. viminalis on the acidic soil. On this soil none of the four repetitions of T. arvense survived the first week after transplanting. After 90 d in contaminated soil, T. caerulescens was close to maturity although not yet flowering. Salix viminalis is a perennial plant, which was still at a young stage with its longest branches being up to 80 cm in length.


View this table:
[in this window]
[in a new window]
  		Table 2. Dry matter (DM) and concentrations of selected elements in different plant parts of Salix viminalis and Thlaspi caerulescens after the 90-d pot experiment.{dagger}

 
Heavy Metal Concentration in Plants
Cadmium and Zn were only hyperaccumulated by T. caerulescens growing in the acidic soil (Table 2), that is, concentrations in shoots were >0.01% (279 ± 46 mg kg-1) for Cd and >1% (14 400 ± 915 mg kg-1) for Zn, which are the threshold values defined by Reeves and Baker (2000) for hyperaccumulation. However, irrespective of soil type, Cd and Zn concentrations in T. caerulescens were, as is typical, larger in leaves than in roots. In contrast, metal concentrations in roots were always higher than in leaves for T. arvense, a non-accumulator species. On the calcareous soil, concentrations in leaves of T. arvense were, for Cd and Zn respectively, 32 times and 6 times lower than in T. caerulescens. Although S. viminalis is not a hyperaccumulator, concentrations of Cd and Zn in leaves were always larger than in roots, confirming its "accumulating" characteristic as observed by Landberg and Greger (1996). Copper concentrations in the aerial plant parts were between 10 and 70 mg kg-1, which was slightly below the range of Cu concentrations found in plants that were grown on contaminated sites as listed in Kabata-Pendias and Pendias (1992). Concentrations in roots were larger than in shoots for the three species, as was observed for T. caerulescens by Frey et al. (2000). In any case those concentrations were far below the limit of hyperaccumulation of 1000 mg kg-1, the value of interest in phytoextraction. Concentrations of Cd and Zn in the leaves of T. caerulescens growing in calcareous soil were about 10 times lower than of those growing on the acidic one. Salix viminalis had higher Cd and Zn concentrations in leaves than in twigs and concentrations were higher in S. viminalis grown on the acidic soil. These concentrations were thus not related to the total metal concentrations in the soils. Table 2 also shows Ca, Fe, Mg, and P concentrations in plants. Although a relatively low concentration of Fe was found in S. viminalis leaves and twigs compared with the concentration found in T. caerulescens (more than 10 times), there was no difference between the same plant grown on the two soils and no relation could be found between the observed chlorosis and the Fe status in leaves. Calcium, Mg, and P concentrations in T. caerulescens shoots were similar in both soils, with Ca being about twice as high in the T. caerulescens grown on the calcareous soil (p < 0.01). They were also very similar to concentrations measured in another T. caerulescens population (Prayon) grown on the same calcareous soil (Frey et al., 2000). As for Cu, also Ca, Mg, and P concentrations in S. viminalis leaves were larger for S. viminalis grown on the acidic soil. In contrast, the concentrations of those elements in the leaves of T. caerulescens were either indifferent or higher in the calcareous soil than on the acidic soil, which makes T. caerulescens a plant less sensitive than S. viminalis to soil conditions such as pH.

Comparing total metal uptake (Fig. 1) showed that in both soils, T. caerulescens took up more Cd and Zn than S. viminalis after 90 d. On the calcareous soil the difference between the two plants in metal uptake was smaller than in the acidic soil, as a result of the larger biomass of S. viminalis on the calcareous soil. This agrees well with Kayser et al. (2000), who found that on Dornach soil (calcareous soil) T. caerulescens was not the best option for phytoextraction, as its biomass production was fairly low compared with other plants.



View larger version (44K):
[in this window]
[in a new window]
  		Fig. 1. Uptake of heavy metals (HM) by Salix viminalis and Thlaspi caerulescens from the calcareous (Soil 1) and acidic (Soil 2) soils.

 
Changes in pH, Organic Carbon, and Nitrogen in the Rhizospheric Soil
As can be seen from Table 3, only slight changes in pH were observed between the control soil and the different plant treatments. All treatments with plants showed an increase in pH compared with the controls, except on the calcareous soil with S. viminalis. A maximal increase of 0.4 pH units was observed for T. caerulescens on the acidic soil. This agrees with Knight et al. (1997), who found a significant increase in soil solution pH following growth of T. caerulescens. However, McGrath et al. (1997) found a decrease of soil pH in three soils (soil to water extraction = 1:2.5) following growth of T. caerulescens and T. ochroleucum. Nevertheless, both authors concluded that a pH change in the rhizosphere was not a mechanism induced by the plants for HM solubilization. No obvious changes were observed in total carbon, organic carbon, and nitrogen after plant growth.


View this table:
[in this window]
[in a new window]
  		Table 3. The pH, carbon, and nitrogen in control and rhizospheric soil of the calcareous and acidic soil after the pot experiment.

 
Bioavailability as Measured by Single Extractants
Availability of metals according to the NaNO3 extraction depended highly on the metal and soil concerned (Table 4). While NaNO3–extractable Cu concentrations were similar for both soils, concentrations of Cd and Zn differed by a factor of 10 and 300, respectively. This pattern can be explained partly by the differences of pH between the two soils and also by their characteristics (organic matter and clay contents), as their total Cu, Cd, and Zn concentrations were in the same range. Hornburg and Bruemmer (1993) also found a linear correlation between Cd and Zn extractability and soil pH, whereas no linear pH dependency could be determined for Cu.


View this table:
[in this window]
[in a new window]
  		Table 4. Concentrations of selected elements in the control and rhizospheric soil of the calcareous and acidic soil after the 90-d pot experiment. DTPA, diethylenetriaminepentaacetic acid; EDTA, ethylenediaminetetraacetic acid.

 
When calculated in percent of total HM amount in the soils, more HM were extracted in the acidic soil by NaNO3 and DTPA but less by EDTA compared with the calcareous soil. Sodium nitrate is an unbuffered extractant and as expected, results were related to the initial soil pH. The EDTA was expected to extract more HM than DTPA and NaNO3 irrespective the soil because of its low pH. However, as mentioned above, the DTPA and EDTA extractions were defined for calcareous and acidic soils, respectively. Thus, the reason why DTPA shows a better extracting power on the acidic soil may lie in the fact that the DTPA extractant contains (i) a high concentration of CaCl2, of which Ca may exchange rapidly with bivalent cations, especially Cd and Zn in the case of the acidic soil, (ii) a high chloride concentration, which complexes these two elements, and (iii) triethanolamine (TEA), which is protonated at pH 7.3 and could exchange H+ with cations from the exchange sites as suggested by Lindsay and Norvell (1978). On the other hand, the low pH of the EDTA extractant may solubilize metals from carbonates and also from some amorphous manganese oxides in the calcareous soil as evidenced by 27% of total Ca and 14% of total Mn extracted by EDTA. In addition, EDTA was the most efficient extractant for Fe and Mn in both soils, but only 1% of total Fe was actually extracted. A solubilization of iron oxides is therefore unlikely as a liberation mechanism of HM by the EDTA extractant. Owing to these remarks we conclude that DTPA and EDTA results can only be interpreted when plant treatments of the same soil are compared.

Effect of the Plants on Heavy Metal Bioavailable Pools
A significant decrease (p < 0.05) in Cd concentrations in the rhizospheric soil was found in NaNO3 extractions of soil samples after both S. viminalis and T. caerulescens on both calcareous and acidic soils. In the DTPA extraction, significant decreases were found after T. caerulescens in both soils and after S. viminalis only in the acidic soil, whereas EDTA-extracted Cd decreased significantly only after T. caerulescens in the acidic soil. For Cu, almost no changes were detected except in the DTPA extraction. The NaNO3–extractable Zn decreased after S. viminalis on both soils and after T. caerulescens only on the acidic soil. In this soil a decrease in DTPA-extractable Zn was also observed after T. caerulescens, whereas S. viminalis induced a significant increase in DTPA-extractable Zn. No significant changes in concentration were measured for EDTA-extractable Zn. The EDTA- and most obviously NaNO3–extractable manganese pools in the acidic soil were depleted after both T. caerulescens and S. viminalis (although not reaching any deficiency levels according to the values obtained in DTPA extraction; Lindsay and Norvell, 1978). Except for Cu in the acidic soil, decreases observed in NaNO3–extractable Cd, Cu, and Zn were always lower than expected from quantities taken up by plants. Assuming that the plants were able to take up the total amount of the metals extracted by NaNO3, then NaNO3–extractable pools were rapidly replenished from less available (extractable) ones. As a consequence, both plants seemed to take up HM—actively or passively through soil reequilibration—from pools other than solely the NaNO3–extractable one. The extent of this replenishment is evidenced by the decrease observed in DTPA- and EDTA-extractable Cd and Zn in the acidic soil, which are in the same range as the quantities taken up by T. caerulescens. This means also that EDTA might be more suitable than NaNO3 to assess availability of Cd and Zn to T. caerulescens in acidic soils. The effect of S. viminalis in reducing rhizospheric metal concentrations was primarily only visible in the NaNO3–extractable pool. Indeed, the decrease in concentrations was minimal and thus could not be detected in DTPA and EDTA extractions, which solubilized larger quantities of Zn. The difference in heavy metal uptake between S. viminalis and T. caerulescens could be seen either as a result of their different demands or of their different abilities to deplete the easily exchangeable pool (NaNO3). The consequence would be a metal concentration decrease in less available pools to return to the original metal distribution between those pools. The rate of replenishment of the NaNO3 pool seems to be in the same order as the speed of uptake, as the concentration in the available pool was not zero after 90 d as could have been expected from plant uptake. Therefore, it is unlikely that the NaNO3–extractable pool remains as low as measured just after plant growth.

Availability as Measured by the Sequential Extraction
The sequential extraction split the total HM content of the soils into a six-pool system for the calcareous and a five-pool system for the acidic soil. In control pots (Fig. 2) , most Cd was associated with the HOAc-extractable pool (Step 2) for the calcareous soil and with the pyrophosphate-extractable pool (Step 3) in the acidic one. Assuming a specific selectivity of the extractants, this would mean that almost 75% of Cd was bound to carbonates in the calcareous soil and about 60% of Cd was bound to the organic matter in the acidic soil. Papadopoulos and Rowell (1988) have mentioned the similarity of the ionic radius of Ca and Cd as a possible reason for chemisorption of Cd on calcite, a mechanism that tends to be irreversible (Zachara et al., 1991). In the case of Dornach (calcareous soil), this could also be the sign of the (partly) pedo–geochemical background content of Cd, which has been mentioned for some soils of the Swiss Jura by Liebig and Dubois (1997), and more specifically for Dornach by Baize and Sterckeman (2001).



View larger version (51K):
[in this window]
[in a new window]
  		Fig. 2. Partitioning of Cd, Cu, and Zn in the calcareous and acidic soils before and after Salix viminalis and Thlaspi caerulescens growth. Pool 2 was not extracted in the acidic soil because carbonates were not found in this soil. Pool 1, exchangeable (NaNO3); Pool 2, carbonate (HOAc); Pool 3, organic matter (Na4P2O7); Pool 4, manganese oxides (NH2OH x HCl); Pool 5, iron oxides (NH2OH x HCl and HOAc); Pool 6, residual (HNO3, HF, and HClO4). The symbol * indicates a difference from the control at the 0.05 probability level.

 
In both soils most Cu was associated with the pyrophosphate-extractable pool, which is in agreement with the common view that in soil Cu is preferably associated with organic matter (McLaren and Crawford, 1973).

In the calcareous soil, Zn was equally distributed between Pools 2, 3, 4, and 5, while in the acidic soil Pools 3 and 5 were the richest in Zn, making together up to 78% of total Zn. Manganese was most abundant in Pool 4 (manganese-oxide pool) of both soils (36 and 40% in the calcareous and acidic soil, respectively). The second-largest Mn pool was Pool 2 in the calcareous soil and Pool 6 in the acidic soil. These findings confirm the suitability of the extractant used to solubilize manganese oxides in the sequential extraction. On the contrary, the extractant for Pool 5 (iron-oxide pool) was insufficiently aggressive to extract a significant proportion of the crystalline iron oxides (data not shown for Mn and Fe). In the calcareous soil, Fe in Pool 6 was about 1.2 times the amount of Fe found in Pool 5 and in the acidic soil the two pools had a similar concentration. This is in agreement with Benitez and Dubois (1998), who found about 1.1 to 2.0 times the amount of Fe in Pool 6 compared with Pool 5 in their soils. The authors showed also in the same work that the extractant dissolved goethite but failed to dissolve hematite. Although hematite might not exist in significant amounts in Soils 1 and 2, the hydroxylamine and acetic acid extractant might fail to dissolve other well-structured crystalline oxides (Keller and Védy, 1994).

Differences in the methodology of sequential extraction, historical background, and extent of the contamination of soils render it very difficult to further compare HM distribution of soils from the literature with these two soils. Nevertheless, Zeien and Bruemmer (1991) found similar results in two soils with comparable pH. Most Cd was found in the more easily extractable fractions (Pools 1 and 2) and most Cu was bound to organics and iron oxides. In the residual fraction HM concentrations were quite low, even though 20 yr had passed since the two soils we studied had been exposed to HM inputs. We observed limited concentrations of Cu and Cd in the less accessible pools, as also found by Riise et al. (1994) for Cd with tracer studies. In contrast, Zn was found distributed between the five first pools in both soils. The reason for this remains unclear. This could be the result of either a faster redistribution of Zn into the less-accessible pools or the reflection of the form of the Zn input to the soil. Indeed, with another sequential extraction procedure, Kayser (2000) also found in the Dornach soil (calcareous soil) a high amount of Zn in the well-crystallized iron oxides and residual fractions and argued that it could be related to the particulate input of Zn emitted by the nearby smelting plant. However, this can be ruled out for the acidic soil, which was contaminated by liquid waste from septic tanks.

Effect of Plants on Heavy Metal Pools Defined by the Sequential Extraction
As already shown by the single extractants, the influence of plant growth on depletion of the HM pool system in these two soils was mostly evident for Cd in the acidic soil. The main reason is a smaller soil to plant concentration ratio for Cd than for Zn.

As assessed by adding the different pools of the sequential extraction, T. caerulescens reduced total Cd in the acidic soil by 36%, assuming that the plant roots had access to all of the soil (1.54 kg dry soil). This is more than the expected 28% calculated from total Cd uptake by T. caerulescens. Therefore, the assumption that the samples taken from the rhizosphere represented the whole soil is not entirely correct. There are two reasons for that: (i) root uptake could not be taken into account and (ii) we overestimated the effect on the soil pools because calculations were made assuming the whole pot was rhizospheric soil.

In this soil, Cd in Pools 1, 3, and 4 was significantly affected by T. caerulescens, whereas S. viminalis affected Pool 3 only. This suggests that part of the organic matter–bound fraction (Pool 3) was depleted by these plants and accordingly has to be considered as bioavailable. On the other hand, no decrease of the Cd concentration in Pool 3 was observed in the calcareous soil.

In the case of Zn, comparison between plant uptake and the soil mass balance was less obvious because of the above-mentioned higher soil to plant ratio of Zn. In the acidic soil Zn uptake of T. caerulescens was about 3% of the total soil Zn, which matches well with the 3.8% calculated from the Zn mass balance of the sequential extraction. However, no such accordance between Zn measured in the plant (0.2%) and the Zn difference in the sequential extraction (5%) was found for the treatment with S. viminalis on the acidic soil.

On the calcareous soil, S. viminalis did not seem able to attack the "carbonate-bound" pool. However, although no decrease in pH was observed, T. caerulescens seemed able to extract Cd, Cu, and Zn from carbonates. Also, the carbonates appeared to have a protective effect on the other pools (especially on Pool 3), as none of the other pools showed significant change between the control and the T. caerulescens pots.

In general, no major changes were detected for Cu apart from a mobilization in Pool 1 and a slight decrease in Pool 2 for the treatment with T. caerulescens. As this element was not taken up by any plant in significant amounts, no changes were to be expected. However, the mechanisms responsible for the dissolution of some of the soil phases may have contributed to the increase of Cu in Pool 1.

Increase or decrease of Fe in the different pools between control and treatments were all below 1.5% and not significant (data not shown). But the ratio of total iron in soil to plant demand is high, and changes in the distribution may not have been detectable.

On the contrary, in the calcareous soil a significant 4% increase in Pool 2 of Mn and an 8% decrease in Pool 4 were observed for the S. viminalis treatment. A similar 3.5% decrease in Pool 4 was observed on the acidic soil (all p < 0.001, Student t test, data not shown). This confirms the observation made with the single extractants that Mn was depleted after the growth of T. caerulescens.

Unlike the extractants, plants are potentially able to access all pools (until Pool 4) at the same time. This was observed in the acidic soil for the EDTA and DTPA single extractions, which are not selective of one specific soil component. However, Pools 1 and 3 (exchangeable and organic matter) were more depleted than Pool 4 (manganese oxides), suggesting that plants might not be as efficient in extracting elements from all pools. The sequential extraction procedure may be seen as a set of successive extractions of increasing strength (progressive decreasing pH, complexing agents of increasing strength, etc.) and the results may thus reflect the greater difficulty for plants to attack the more resistant pools. As in the single-extraction procedure, the NaNO3–extractable Cd and Zn pools (Pool 1) were not totally depleted, which means that part of the depleted Pools 3 and 4 may have contributed to its replenishment. Salix viminalis seemed to be less able to access Pool 4 than T. caerulescens. However, this might be related to the limited amount of Cd and Zn that S. viminalis was able to accumulate.


   SUMMARY AND CONCLUSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 SUMMARY AND CONCLUSION
REFERENCES
 
We used the sequential extraction to determine the chemical association of Cd, Cu, and Zn with the soil components, whereas we used the single extractants to estimate the bioavailable pools of the HM in the soil. Although it was not possible to relate the "component" pools with the amounts extracted by EDTA and DTPA, results from both approaches agree well and have shown that T. caerulescens and, to a lesser extent S. viminalis, were able to alter so-called "bioavailable" Cd and Zn pools. The effect was more pronounced in the acidic soil than in the calcareous one and was directly related to the total metal uptake of the plant. In terms of percent total metal, Cd pools were always more depleted than Zn ones due to higher initial Zn concentrations in the soils. No depletion was observed for Cu. Additionally, specific soil components were attacked by the plants as evidenced by the sequential extraction. In the most favorable case (Cd, T. caerulescens, acidic soil), all pools were depleted except Pool 5 and the residue. Whatever the mechanisms involved—direct solubilization, metal extraction from the pool, or reequilibration between the different soil phases—this implied deep modifications of the soil conditions. However, as the extent of depletion of the pools is related to the total amount extracted by the plant, the question remains whether a larger uptake (e.g., by repeated planting or extending the growth period) might not result in a decrease in more pools, both for T. caerulescens and S. viminalis.

The other observed phenomenon was a partial replenishment of the easily exchangeable pool. Therefore, the use of phytoextraction to reduce the bioavailable fraction of HM in the soil requires further investigation. In particular, more research is needed to determine the extent and the stability of the modifications of the heavy metal status in the soil with time. A technique that would be efficient in the long term would need to reduce total metals in the soil to such an extent that metal fractionation is permanently altered.


   ACKNOWLEDGMENTS
 
This research is part of the Project ENV4-CT97-0598/OFES97.0362 and was mainly conducted in D. Genske's laboratory. Salix viminalis cuttings were provided by M. Greger, University of Stockholm, Sweden. We thank J.-D. Teuscher for help in the soil and plant analyses and R. Hamon for corrections of the text.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 SUMMARY AND CONCLUSION
 REFERENCES
 




This article has been cited by other articles:


Home page
 J. Environ. Qual.Home page
P. Vervaeke, F. M. G. Tack, N. Lust, and M. Verloo
Short- and Longer-Term Effects of the Willow Root System on Metal Extractability in Contaminated Dredged Sediment
J. Environ. Qual., May 1, 2004; 33(3): 976 - 983.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (45)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Hammer, D.
Right arrow Articles by Keller, C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Hammer, D.
Right arrow Articles by Keller, C.
Agricola
Right arrow Articles by Hammer, D.
Right arrow Articles by Keller, C.
Related Collections
Right arrow Heavy Metals
Right arrow Soil Pollution
Right arrow Soil Analysis
Right arrow Soil Chemistry

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Soil Science Society of America Journal Journal of Plant Registrations The Plant Genome