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 Related articles in JEQ
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 ISI Web of Science (13)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Nolan, A. L.
Right arrow Articles by McLaughlin, M. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Nolan, A. L.
Right arrow Articles by McLaughlin, M. J.
Agricola
Right arrow Articles by Nolan, A. L.
Right arrow Articles by McLaughlin, M. J.
Related Collections
Right arrow Heavy Metals
Right arrow Plant and Soil Interactions
Right arrow Soil Pollution
Published in J. Environ. Qual. 34:496-507 (2005).
© ASA, CSSA, SSSA
677 S. Segoe Rd., Madison, WI 53711 USA

TECHNICAL REPORTS

Heavy Metals in the Environment

Prediction of Zinc, Cadmium, Lead, and Copper Availability to Wheat in Contaminated Soils Using Chemical Speciation, Diffusive Gradients in Thin Films, Extraction, and Isotopic Dilution Techniques

Annette L. Nolana,c,*, Hao Zhangb and Mike J. McLaughlina,b

a CSIRO Land and Water, PMB 2, Glen Osmond, SA 5064, Australia
b Soil and Land Systems, School of Earth and Environmental Sciences, University of Adelaide, PMB 1, Glen Osmond, SA 5064, Australia
c Current address: National Measurement Institute, PO Box 385, Pymble, NSW 2073, Australia

* Corresponding author (Annette.Nolan{at}measurement.gov.au)

Received for publication June 3, 2004.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
To predict the availability of metals to plants, it is important to understand both solution- and solid-phase processes in the soil, including the kinetics of metal release from its binding agent (ligand and/or particle). The present study examined the speciation and availability of Zn, Cd, Pb, and Cu in a range of well-equilibrated metal-contaminated soils from diverse sources using several techniques as a basis for predicting metal uptake by plants. Wheat (Triticum aestivum L.) was grown in 13 metal-contaminated soils and metal tissue concentrations (Zn, Cd, Pb, and Cu) in plant shoots were compared with total soil metal concentrations, total soluble metal, and free metal activities (pM2+) in soil pore waters, 0.01 M CaCl2–extractable metal concentrations, E values measured by isotope dilution, and effective metal concentrations, CE, measured by diffusive gradients in thin films (DGT). In the DGT technique, ions are dynamically removed by their diffusion through a gel to a binding resin, while E values represent the isotopically exchangeable (labile) metal pools. Free metal activities (Zn2+, Cd2+, and Pb2+) in soil pore waters were determined using a Donnan dialysis technique. Plant Zn and Cd concentrations were highly related to CE, while relationships for Zn and Cd with respect to the other measures of metals in the soils were generally lower, except for CaCl2–extractable Cd. These results suggest that the kinetically labile solid-phase pool of metal, which is included in the DGT measurement, played an important role in Zn and Cd uptake by wheat along with the labile metal in soil solution. Plant Pb concentrations were highly related to both soil pore water concentrations and CE, indicating that supply from the solid phase may not be so important for Pb. Predictions of Cu uptake by wheat from these soils by the various measures of Cu were generally poor, except surprisingly for total Cu.

Abbreviations: CE, effective concentration • DGT, diffusive gradients in thin films • E, isotopically exchangeable (labile) metal pool • EC, effect concentration • GF-AAS, graphite furnace atomic absorption spectroscopy • ICP, inductively coupled plasma • MS, mass spectrometry • MWHC, maximum water holding capacity • OES, optical emission spectroscopy


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
AS SOILS PROGRESSIVELY become contaminated from loadings of metalliferous waste materials, the bioavailability of toxic trace metals in soils is an increasingly important issue. Total concentrations of metals in soils are generally a poor indicator of metal toxicity because metals exist in different solution- and solid-phase forms that can vary greatly in terms of their bioavailability. Risks associated with metal contamination in soils are difficult to assess (Rieuwerts et al., 1998; Nolan et al., 2003a); however, regulators are moving toward a bioavailability-based rationale as an indicator of the risk. For example, regulators in some European countries have adopted operationally defined extractions using weak salt solutions in an attempt to improve the assessment of metal bioavailability in soils (Prüeß, 1997). The bioavailability of trace metals, their biological uptake, and ecotoxicological effects on the soil biota can be better understood in terms of their chemical speciation. Extraction of pore water from soils isolates the aqueous phase to which plant roots and microorganisms are exposed. In aquatic systems, consideration of free metal ion activities in solution has improved predictions of bioaccumulation and toxicity (Campbell, 1995), and there is some preliminary evidence that solution speciation is also important in soil systems (Sauvé et al., 1996, 1998; McGrath et al., 1999). However, the existence of an explicit relationship between free ion activity in soil pore water and bioaccumulation and/or toxicity in soil-exposed organisms (i.e., based on the free ion activity model) remains an issue for debate, particularly for uptake of metals by plants (Zhang et al., 2001). Free metal ions may be the dominant form available to biota, but this does not necessitate a correlation between free ion activity and bioaccumulation, given that numerous other important processes affect plant uptake and toxicity (Nolan et al., 2003a; McLaughlin et al., 1998). To predict the bioavailability of metals to plants, it is necessary to understand both solution- and solid-phase supply processes in soils. Contributions of an element to the plant-available pool from sources other than soil solution have long been recognized and led to the concepts of intensity, quantity, and capacity in the early soil science literature (Beckett, 1964; Barber, 1995). Processes that affect the supply of solutes to plants include diffusional and convective transport to the root, the root encountering fresh surfaces as it grows through the soil, and the influence of root microenvironments and exudates (Barber, 1995; Marschner, 1995; McLaughlin et al., 1998). Therefore unless one process is dominant, it is unlikely that any single chemical measurement could adequately assess plant uptake. Zhang et al. (2001) introduced the new concept of effective concentration, CE, which includes both the soil solution concentration and an additional term, expressed as a concentration, which represents metal supplied from the solid phase. Effective concentration was measured using the technique of diffusive gradients in thin films (DGT), which can locally lower soil solution concentrations, inducing metal resupply from the solid phase. Zhang et al. (2001) grew pepperwort (Lepidium heterophyllum Benth.) plants in 29 soil samples covering a wide range of total Cu concentrations and found that plant Cu concentrations were linearly related and highly correlated with CE but were more scattered and nonlinear with respect to other Cu measurements (free Cu2+ activity, EDTA extraction, and soil solution concentration). The results from that study indicated that overall the dominant supply processes to pepperwort plants in those soils were solution pools, coupled with diffusion and labile metal release from the solid phase induced by fast biological uptake.

The present study examined the speciation and availability of Zn, Cd, Pb, and Cu in a range of well-equilibrated metal-contaminated soils from diverse sources using several techniques as a basis for predicting metal uptake by plants. Wheat was grown in 13 contaminated soils, sampled from Australia and the United States, and metal tissue concentrations (Zn, Cd, Pb, and Cu) in plant shoots were compared with total soil metal concentrations, 0.01 M CaCl2–extractable metal concentrations, total soluble metal and free metal activities (pM2+) in soil pore waters, DGT CE concentrations, and labile metal pools (E values). Free metal activities (Zn2+, Cd2+, and Pb2+) in soil pore waters were determined using a Donnan dialysis technique that employs a semipermeable cation exchange membrane and relies on the principle of Donnan equilibrium (Fitch and Helmke, 1989; Nolan et al., 2003b; Salam and Helmke, 1998). The Donnan dialysis technique has the advantages that it is free from interference from neutral and anionic species, it minimizes perturbation to the equilibrium composition of the sample solution, a range of metals can be determined simultaneously, and it is relatively sensitive. An isotopic dilution technique was used to determine the isotopically exchangeable (labile) pool (E value) of Cd, Zn, and Cu. This technique partitions metal into equilibrium (in solution or adsorbed) and non-equilibrium (fixed) pools. These techniques were chosen as they represent the most commonly used chemical methods for assessing metal availability to plants. While there have been many studies that have examined the capabilities of these techniques in predicting metal uptake by plants, a comprehensive assessment of all the techniques in one study using soils contaminated with more than one metal has not been reported to date. The objective of this study was to assess the capabilities of the different techniques in predicting metal uptake by wheat, using well-equilibrated soils contaminated with more than one heavy metal.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Soil Samples
Eleven contaminated soils from Australia and two contaminated soils from the United States were used in this study. The soils vary widely in their origin, texture (0–29% clay), pH (3.6–8.1), total carbon content (0.1–11.4% w/w), and total metal concentrations (28.3–21289 mg Zn kg–1, 0.11–85.6 mg Cd kg–1, 19.7–5312 mg Pb kg–1, and 6.9–602.1 mg Cu kg–1). Selected soil solid-phase properties are shown in Table 1. The soils Fredrick, Florence, Horner, Lakeview, Third, and Montgomery were sampled from backyards of houses near smelters in Australia. The soil Backyard was sampled from a property in South Australia that was known to contain elevated levels of Cd because of local plating and soldering industries. Joplin and Brukunga were sampled from mine tailings areas in the United States and South Australia, respectively. The soil Zn Tower was sampled from an area affected by runoff from a galvanized electricity transmission tower in the United States. The soil Kapunda was sampled from an uncontaminated agricultural site in South Australia. Biosolids from two different wastewater treatment plants in South Australia (Murray Bridge and Bolivar) were applied to the soil 8 yr before this study at a rate of 500 Mg ha–1 equivalent. The control soil received no biosolid application. All soils in this study were collected from the A horizon (0–10 cm), dried at 40°C, and sieved to <2 mm for chemical analyses and pot experiments.


View this table:
[in this window]
[in a new window]
  		Table 1. Soil solid-phase properties.

 
Pot Experiments
Wheat (cv. frame) was grown in pots containing 200 g each of soil (dry mass). Ten wheat seedlings per pot were grown. Three replicates of each soil were used. Soil moisture was maintained at 50% maximum water holding capacity (MWHC), with additions of deionized water for 7 d before planting, which was then increased to 90% MWHC 48 h before planting for DGT deployment. Soil moisture was maintained at 50% MWHC during plant growth. To facilitate extraction of soil pore water (Knight et al., 1998), a Rhizon soil moisture sampler (Rhizon Research Products, Wageningen, the Netherlands) was placed in each pot as the soil was packed. These samplers are porous tubes (pore size ≤ 0.2 µm) that allow extraction of pore waters with a syringe and are suitable for trace metal work as they do not adsorb metals, are readily cleaned with 1% HNO3 followed by deionized water, and are considered to better represent the solution extracted by plants (Nolan et al., 2003a). Plants were grown in a growth cabinet with the following conditions: 14 h day and 10 h night, 20°C and 15°C day and night temperatures with light maintained at a minimum photon flux of 450 µmol m–2 s–1. A one-off application of macronutrient solution was added to the pots seven days after germination. Sixteen days after germination the aboveground plant material was harvested from 12 of the soils (plants failed to grow in Brukunga soil as a result of metal toxicity), rinsed with deionized water, and dried at 60°C for 72 h.

Devices and Deployment for Diffusive Gradients in Thin Films
The DGT devices consist of a plastic assembly containing a layer of resin embedded in gel, overlain by a layer of diffusive gel and a protective filter through which ions can freely diffuse. Details of the diffusive and resin gel preparation and treatment have been given previously (Zhang and Davison, 1995). A polyacrylamide hydrogel, 15% by volume acrylamide and 0.3% by volume agarose cross-linker (DGT Research Ltd., Lancaster, UK), was used as the gel solution in this work.

Before sowing, soils were maintained at 90% MWHC for 48 h. One standard DGT device with a sampling surface area of 3.14 cm2 was deployed in each pot for 2 to 24 h, depending on the concentration of the metals in the soils, at an average temperature of 18°C. It was gently pressed onto the soil surface, ensuring that good contact between the device and the soil was attained (particularly for the sandy soils). The DGT devices were also deployed after harvest, and this time the soils were maintained at 100% MWHC for 24 h before DGT deployment. DGT deployment in soils with different moisture contents has been shown to comply with theory when used at or above 80% field capacity (Hooda et al., 1999). The mass of accumulated metal in the resin layer, M, was measured by eluting with 1 M HNO3 solution and then analyzing by inductively coupled plasma–mass spectrometry (ICP–MS) (Ultramass; Varian, Palo Alto, CA).

Measurements of Zinc, Cadmium, Lead, and Copper
Total metal concentrations in the plant samples were determined using inductively coupled plasma–optical emission spectroscopy (ICP–OES) (SpectroFlame Modula; Spectro, Kleve, Germany), following digestion with HNO3 (Zarcinas et al., 1983). Total metal concentrations in the soils were determined by ICP–OES following digestion with aqua regia (Zarcinas et al., 1996). Soil pH was determined in a mixture of 1:5 soil to water (w/v). Total concentrations of carbon were determined using a combustion analyzer (CNS 2000; LECO, St. Joseph, MI). Soil solution was extracted twice from each pot, at the beginning and the end of the experiment. Graphite furnace atomic absorption spectroscopy (GF-AAS) (AAnalyst 600; PerkinElmer, Wellesley, MA) was used to determine the total soluble concentrations of Cd and, where necessary, the concentrations of Cu, Pb, and Zn. For all other elemental determinations, ICP–OES was used. Limits of reporting (5x detection limit) for Cd, Zn, Cu, and Pb by GF-AAS were approximately 4.5, 7.5, 16.0, and 5.0 nM, respectively. The pH and electrical conductivity of the extracts were measured on subsamples immediately after extraction. Pore water ionic strength (IS) was calculated from electrical conductivity using an empirical relationship (Eq. [1]) developed previously for Australian soils (McLaughlin et al., 1997):

[1]
where units for IS and electrical conductivity are mol L–1 and mS cm–1, respectively. Dissolved organic C in the samples was determined using a Formacs TOC/TN analyzer (Skalar, Breda, the Netherlands).

Free Zn2+, Cd2+, and Pb2+ activities in the soil solutions extracted before plant growth were determined using a Donnan dialysis technique (Fitch and Helmke, 1989; Nolan et al., 2003b; Salam and Helmke, 1998). Free Cu2+ activities could not be determined because the measurements were below the detection limit for GF-AAS. A total volume of 30 to 50 mL was required for Donnan dialysis, with soil solution from the replicates bulked to give a single sample. It was not possible to analyze replicate samples by Donnan dialysis due to limited pore water volumes and time constraints; however, a previous study showed good agreement between replicate samples across a wide range of soils (Nolan et al., 2003b). The Donnan dialysis technique utilizes a Teflon cell that holds a strong-acid cation exchange membrane (Nafion-117; DuPont, Wilmington, DE) separating the sample solution (donor, 30–50 mL) from an initially pure solution of Sr(NO3)2 (acceptor, 200 µL) of the same ionic strength. Nafion-117 is a copolymer of tetrafluoroethylene and sulfonyl fluoride vinyl ethers. For these experiments, membranes were soaked for 24 h in a mixture containing 10% methanol, 10% HNO3, and 80% Milli-Q water (Millipore, Billerica, MA) to remove adsorbed cations. They were then rinsed and soaked in Milli-Q water for 4 h and then in fresh acceptor solutions. The acceptor solution was replaced twice within 24 h before the membranes were used. The sample solution was continuously circulated past the bottom of the membrane by a Teflon pump at a rate of 200 mL min–1 for 2 h. An equilibration time of 2 h has been shown to be sufficient for the metals Zn, Cd, Pb, and Cu (Nolan et al., 2003b). The acceptor solution rests on the top surface of the membrane, and is contained by an annular ring in the top part of the exchange cell. The attainment of equilibrium is dependent on the ionic strengths of both solutions. To monitor and correct potential errors in matching the ionic strengths of the donor (D) and acceptor (A) solutions, the donor solution was spiked with a small volume of solution containing several kilobecquerels of carrier-free 22Na before Donnan equilibration and Eq. [2] was used to correct any observed differences (Salam and Helmke, 1998). However, this correction was generally unnecessary because Eq. [1] allowed close matching of the ionic strengths of the acceptor and donor solutions given that the 22Na activities in the donor and acceptor solutions usually varied by less than 5 to 10%:

[2]

The donor and acceptor solutions were removed after equilibrium, and 22Na activity was determined by ß scintillation. Free Zn2+, Cd2+, and Pb2+ concentrations in the acceptor solutions were determined by GF-AAS. The activities of the analytes in the acceptor solutions were determined using the activity coefficients calculated from the Davies equation.

Soil from each pot was sampled and extracted with 0.01 M CaCl2 using a standard procedure (Houba et al., 2000). Metals in the extracts were determined either by ICP–OES (Zn) or GF-AAS (Cd, Pb, Cu). The isotopically exchangeable (labile) pools of Zn, Cd, and Cu were determined using an isotopic dilution technique. The labile pool for Pb was not determined because of the lack of a suitable source of 210Pb radioisotope. Subsamples (4 g) of the soils were placed in centrifuge tubes to which was added 40 mL of 0.1 M Ca(NO3)2. This electrolyte was chosen to achieve better flocculation of metal-rich colloidal particles, as described by Young et al. (2000). The soil suspensions were equilibrated for 24 h in an end-over-end shaker. The samples were spiked with 50 µL of solution containing 109Cd (400 kBq mL–1) and 65Zn (1000 kBq mL–1). A separate batch of samples, to which was added 40 mL of 0.01 M CaCl2, was spiked with 50 µL of solution containing 64Cu (60 MBq mL–1). In the case of 64Cu the activity added to the samples was large due to the short half-life (12 h) and low efficiency of gamma counting for this isotope. The samples were then returned to the shaker and left equilibrating for 24 h. At the end of the equilibration period the samples were centrifuged at 3000 x g for 20 min and filtered through 0.2-µm cellulose acetate filters (Sartorius, Goettingen, Germany). Unlabeled Zn, Cd, and Cu in the filtrates were measured using either ICP–OES (Zn) or GF-AAS (Cd, Cu). Activities of radioactive Zn, Cd, and Cu in the filtrates were assessed using gamma spectrometry (1480 Wizard; Wallac, Turku, Finland), with care taken to correct for the interference of 65Zn on 109Cd. All analyses were performed in triplicate and included blanks as well as solutions spiked with radioisotope so that the total amount of radioisotope added could be determined. The labile pool (E) of Zn, Cd, and Cu was determined as:

[3]
where Csol is the concentration of cold metal in solution (µg mL–1), C*sol is the concentration of radioactive isotope remaining in solution after equilibration (Bq mL–1), R is the total concentration of radioisotope that was added to each sample (Bq mL–1), and V/W is the ratio of solution to sample, which in this case was 10 mL g–1.

Calculation of Effective Concentration from Diffusive Gradients in Thin Films Measurements
The DGT measured concentration in a soil, CDGT, can be converted to an effective concentration, CE, using Eq. [4] (Zhang et al., 2001):

[4]
where Rdiff expresses the extent of concentration depletion at the interface of the device and the soil for the diffusion-only case. The term CDGT was calculated using Eq. [5]:

[5]
where M is the mass accumulated on the resin gel and measured by ICP–MS, {Delta}g is the thickness of the diffusive gel layer (0.8 mm) plus the thickness of the filter membrane (0.13 mm), A is the surface area (3.14 cm2), t is the deployment time, and D is the diffusion coefficient of each metal in the diffusive gel. Diffusion coefficients were measured previously using a diffusion cell (Zhang and Davison, 1999).

The term Rdiff was calculated using the numerical model of the DGT–soil system DIFS (DGT induced fluxes in soils) (Harper et al., 1998). Input parameters of particle concentration, Pc, soil porosity, {phi}, and diffusion coefficient in soil, Ds, were calculated using Eq. [6], [7], and [8] (Boudreau, 1996):

[6]

[7]

[8]
where m is the total mass of soil particles, V is the pore water volume in a given volume of soil, and dp is the density of the soil particles, which is commonly assumed to be 2.65 g cm–3 (Bielders et al., 1990). The term Do is the diffusion coefficient in water. Diffusion coefficients for Cu, Cd, and Zn are very similar, but are quite different from that for Pb. Two values of Rdiff were calculated for each soil at each water content (90 and 100% MWHC), one representing Cu, Cd, and Zn and one representing Pb. As DIFS is a one-dimensional model, a correction factor derived from a two-dimensional model was applied to obtain a value of Rdiff for two-dimensional diffusion.

Data Analysis
The mathematical model used to resolve dose–response curves for plant yield against CE Zn or free Zn2+ activity was the effect concentration (ECx) template calculation (Barnes et al., 2003). This model uses a logistic curve, which has the following form:

[9]
where P is the amount of growth in the uncontaminated situation (upper asymptote), X is log (metal measurement + quantifiable limit), and b is a measure of the sensitivity to the toxicant near the ECx. The input parameters were CE Zn (in mmol L–1) or free Zn2+ activity (in µmol L–1), and plant yield (g). Bootstrapping was used to estimate confidence intervals around the parameter estimates (ECx). As recommended by Scholze et al. (2001), 1000 bootstrap iterations were performed to estimate bootstrap-based confidence intervals.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Log–Log Relationships
The log–log relationships between metal concentrations in plant tissue and the different measurements of metal in the soil are given in Fig. 1 to 4 . Log transformations improved the data distribution and therefore log–log relationships were used. Nonlinear (quadratic) regressions were used, as some of the relationships were nonlinear when plotted on a linear scale. The log–log relationships between metal uptake per pot and the different measurements of metal in the soil were also examined (data not shown); however, the regression coefficients were very similar to those in Fig. 1 to 4 for plant tissue metal concentrations, and therefore only the latter data will be discussed. The relationship between plant metal concentrations and metal supply on a linear scale may be either linear or curved/quadratic (reduced rate of metal accumulation per unit increase in metal supply) and this may be disguised in log–log transformed data. For example, in this study linear scale plots of plant Zn concentration against soil solution concentration or free ion activity were curved, while plant Zn and CE (Zn) were more linearly related (data not shown). A number of soil or plant factors can influence the relationship between soil solution free ion activities and metal accumulation in plants so that a linear relationship is not necessarily expected. For example, metal translocation from root to shoot is often limited in plants or root uptake from solution is saturable following quadratic or Michaelis–Menten type relationships (Mullins and Sommers, 1986; Hamon et al., 1999). Nevertheless, nonlinear regressions of log–log transformed data allowed a common frame of reference to be used for assessment of all plant–soil measurement relationships.



View larger version (30K):
[in this window]
[in a new window]
  		Fig. 1. Dependence of log concentration of Zn in plant tissue on the log of (a) effective Zn concentration, CE, before growth, (b) effective Zn concentration, CE, after growth, (c) soil solution Zn, before growth, (d) soil solution Zn, after growth, (e) free Zn2+ activity, (f) CaCl2–extracted Zn, (g) Zn E value, and (h) total soil Zn. The quadratic regression equations and regression coefficients are shown.

 


View larger version (26K):
[in this window]
[in a new window]
  		Fig. 4. Dependence of log concentration of Cu in plant tissue on the log of (a) effective Cu concentration, CE, before growth, (b) effective Cu concentration, CE, after growth, (c) soil solution Cu, before growth, (d) soil solution Cu, after growth, (e) CaCl2–extracted Cu, (f) Cu E value, and (g) total soil Cu. The quadratic regression equations and regression coefficients are shown. Free Cu2+ activities were not determined (below method detection limit).

 


View larger version (30K):
[in this window]
[in a new window]
  		Fig. 2. Dependence of log concentration of Cd in plant tissue on the log of (a) effective Cd concentration, CE, before growth, (b) effective Cd concentration, CE, after growth, (c) soil solution Cd, before growth, (d) soil solution Cd, after growth, (e) free Cd2+ activity, (f) CaCl2–extracted Cd, (g) Cd E value, and (h) total soil Cd. The quadratic regression equations and regression coefficients are shown.

 


View larger version (26K):
[in this window]
[in a new window]
  		Fig. 3. Dependence of log concentration of Pb in plant tissue on the log of (a) effective Pb concentration, CE, before growth, (b) effective Pb concentration, CE, after growth, (c) soil solution Pb, before growth, (d) soil solution Pb, after growth, (e) free Pb2+ activity, (f) CaCl2–extracted Pb, and (g) total soil Pb. The quadratic regression equations and regression coefficients are shown. Only nine soils are included as Kapunda soils were below Pb detection limit for some measurements. The E values were not determined for Pb (no suitable radioisotope available).

 
Comparison of Different Techniques
From a predictive perspective, the best measure of Zn concentrations in the wheat plants across all soils was CE, measured after plant growth (R2 = 0.96, Fig. 1). Measurement of soil solution Zn was inferior, and consideration of free Zn2+ activities improved the relationship little. Other predictors of plant Zn concentration were poor. The best measures of plant-available Cd were CaCl2–extractable Cd and CE, measured after plant growth (R2 = 0.90, Fig. 2). The best measures of plant-available Pb were soil solution Pb (R2 = 0.95) and CE (R2 = 0.94; Fig. 3). The E values for Pb were not determined because of the lack of a suitable source of 210Pb radioisotope. The best measure of plant-available Cu was total soil Cu (R2 = 0.87; Fig. 4), which is unexpected given previous results for Cu (Zhang et al., 2001) and that metals exist in different solution- and solid-phase forms in soil that can vary greatly in terms of their bioavailability. Regression coefficients for all other measures of Cu (log–log plots) were considerably lower (R2 ≤ 0.69). Determinations of free Cu2+ activities in soil pore water were not possible as they were below the Donnan dialysis method detection limit. Uptake of Cu is highly regulated by most plants, owing to homeostatic mechanisms that maintain correct concentrations of essential elements in different cellular compartments (Clemens, 2001). Therefore plant regulation may have influenced the capability of these various measurements of soil parameters to predict Cu uptake by wheat. Furthermore, phytotoxic effects of high available Zn concentrations in some soils may also have affected Cu uptake in the low concentration range. Notably, regressions of plant Cu versus the different measures of Cu (log–log data) excluding the three soils with the highest soil solution Zn concentrations (Lakeview, 12 mg L–1; Zn Tower, 30 mg L–1; and Joplin, 300 mg L–1) showed significantly improved relationships for CE (R2 = 0.86, before growth), soil solution concentrations (R2 = 0.75, before growth), CaCl2–extractable concentrations (R2 = 0.91), total soil concentrations (R2 = 0.99), and E values (R2 = 0.87). Exclusion of these same three soils from other log–log regressions of plant uptake and the different measures of metal did not produce a similar improvement in the regression coefficients (data not shown). Hence it appears that exceptionally high concentrations of Zn in soil solution affected uptake of Cu by wheat so that it is no longer predictable by the measurements used in this study.

Assessment of Processes
The results in this study indicate that for Zn, Cd, and Pb uptake by wheat from these particular soils, measurements of free metal ion activities in soil solution before plant growth did not provide the best prediction of metal supply to the plant root, in contrast to findings in other studies (Minnich et al., 1987; Sauvé et al., 1996). However, free metal ion activities in conjunction with the resupply kinetics from the solid phase may be important. Considerable analytical complexity is required to measure free metal ion activities in soil solution using the Donnan cell, which proved to be unwarranted in our experiments. Further measurements are needed on a wider range of soils, and for a wider range of environmental end points other than plant metal uptake, to determine whether free metal ion determinations are generally required to improve environmental risk assessments.

In this study, DGT measurements in soil were effective in predicting plant Zn, Cd, and Pb accumulation in wheat from contaminated soils but were unsuccessful for Cu. In contrast, Zhang et al. (2001) found that DGT measurements were able to predict Cu uptake by pepperwort from Cu-contaminated soils that covered a much wider range of Cu concentrations than those studied here. The soils used in the previous study were mainly contaminated by Cu alone and the total concentrations were up to 8645 mg kg–1, about 14 times higher than the maximum Cu concentration in this study. The maximum Cu concentration in pepperwort was about 40 times higher than the maximum concentration in wheat. Pepperwort, in contrast to most other plants, is an effective accumulator of Cu. These two studies show that the predictive capabilities of DGT may depend on the combination of metals, their concentration ranges, and the plant species. Notably, the total soil solution and CE concentrations for Zn and Cd measured after plant growth were generally lower than those before growth; however, Pb and Cu concentrations after plant growth tended to be higher than those before growth. It is therefore important to consider when these (and other) measurements are conducted.

In very highly contaminated soils it is less likely that there is a depletion of metal surrounding plant roots, as transpirational fluxes of water are likely to bring more metal to the plant root surface than is required physiologically and taken up by the plant (for essential metals such as Cu and Zn) (McLaughlin, 2001). When supply by diffusion becomes unimportant for metals in highly contaminated soils, the soil solution free ion activities may be equally effective as CE at predicting plant uptake, but determination of free ion activities in soil solution (by Donnan dialysis) is logistically more difficult than determining CE. It is worth noting that in highly contaminated soils the proportion of metals supplied to the DGT device from the solid phase will probably be lower, so that the DGT measurement approximates more closely to soil solution free ion activity. Metal complexes that are not fully labile are not completely measured by DGT (Scally et al., 2003). For some of the grossly contaminated soils in this work and in the study by Zhang et al. (2001) it is possible that supply by diffusion was not limiting due to the presence of very high concentrations of metal. However, in most soils diffusion limitations are likely to influence uptake of metal, apart from those that are grossly metal contaminated. This was reflected in the better linear relationship observed by Zhang et al. (2001) for plant Cu concentrations versus CE in comparison with the corresponding relationship with soil solution free Cu2+ activities. Similarly, it is reflected in the enhanced log–log relationships observed in this study for plant Cd, Pb, and Zn concentrations versus CE in comparison with those for the corresponding relationships with soil solution free ion activities.

The capacity of CaCl2–extractable metal concentrations to predict plant availability was metal dependant. The log–log plot of Cd plant availability versus CaCl2–extractable Cd was highly correlated over the complete concentration range; however, the corresponding relationships for the other metals were lower. Previous studies have also observed a good relationship between Cd uptake by plants and CaCl2–extractable concentrations (McLaughlin et al., 1997, 1999; Sauerbeck and Styperek, 1985), which suggests that this technique may be a useful substitute for determining plant-available Cd. The poor relationships of CaCl2–extractable concentrations with Zn, Pb, and Cu availability indicate that chloro-complexation and Ca competition are too weak for these metals and that plant-available pools of Zn, Pb, and Cu in the soils are not just the easily exchangeable fraction. In addition, compared with Zn, Pb, and Cu in soils, Cd is relatively weakly bound to soil solid phases, so that less aggressive extractants are likely to produce a better relationship with plant uptake than more aggressive methods. An alternative extraction for determining labile Cd, using a higher salt concentration of 1 M CaCl2, has been suggested by Young et al. (2000), to completely dissolve the labile Cd pool and to provide an alternative to E values obtained by isotopic dilution procedures. While this may provide a better substitute for the radio-labile fraction of Cd in soils, it may overestimate the concentration of immediately available Cd.

The log–log relationships of plant-available metal (Zn, Cd, Cu) with the isotopically exchangeable (labile) metal pool (E value) were generally low (R2 = approximately 0.6). The E value partitions the metal into pseudo-equilibrium (in solution or adsorbed) and non-equilibrium (fixed) pools by determining the distribution of the added radioisotope between the solid and solution phases of the soil suspension at a predetermined time. Most of the labile metal exchanges within 1 to 3 d, but a continuing slower exchange may take place over time. However, the labile metal pool, as operationally defined in this way, is not necessarily the pool that is immediately available for uptake by plants. The E values can be considered as an indication of the total potentially available pool, which is significantly larger than the amount of metal taken up by the plant, or the amount desorbed by weak salt solutions. For example, in this study the percentage of 0.01 M CaCl2–extractable Cd, as compared with the Cd E value, varied from 0.8 to 91.4% with an average value of 18.6%. The poor relationships of plant Zn, Cd, and Pb concentrations with total soil metal concentrations are not surprising, given that metals exist in different solution- and solid-phase forms that can vary greatly in terms of their bioavailability.

A note of caution should be expressed in this overall interpretation. Although each method has been applied carefully there are inevitably different errors associated with the different determinations, in terms of both sampling heterogeneity and analytical accuracy. For example Rhizon samplers collect soil solution from a relatively small volume of soil, similar to DGT devices deployed onto soil. However, DGT measurement has the advantage that it combines a measure of soil solution metal, as well as a proportion of the metal bound to the solid phase, hence extracting a greater amount of metal for analytical determination. Soil extractions using CaCl2 are usually performed on a homogenized sample of soil, and hence minimize the spatial heterogeneity issues inherent in DGT and Rhizon techniques. Analytically, the concentrations of metals measured in DGT eluents were 10 times greater than those measured in soil solution. Extraction with CaCl2 is the simplest and therefore probably the least error-prone method, while the Donnan dialysis method is probably the most complicated method producing solutions with the lowest concentrations of metals for analysis. Poor relationships for free ion activity may therefore be partly associated with analytical and methodological limitations.

Phytotoxicity
As Zn was the only metal present in the soils at high enough (plant-available) concentrations to cause phytotoxic effects, plant biomass yields were compared with CE Zn measured by DGT and with free Zn2+ activities in soil solution (Fig. 5) . The relationships were not as systematic as those for tissue concentrations with the different measures of soil Zn. In the nontoxic range of available Zn, absolute biomass yield was variable owing to inherent differences in the fertility of the different soils. At CE > 2 mmol Zn L–1, wheat yield decreased sharply, indicating Zn phytotoxicity. The DGT-measured Zn CE values able to inhibit plant biomass yield by 25 and 50% (EC25 and EC50) were 4.62 and 7.18 mmol L–1, respectively. The 5th and 95th percentiles were 4.50 to 4.78 and 6.98 to 7.40 mmol L–1, respectively. The effect activities of free Zn2+ able to inhibit plant biomass yield by 25 and 50% (EA25 and EA50) were 117 and 374 µmol L–1, respectively. The 5th and 95th percentiles were 106 to 124 and 366 to 387 µmol L–1, respectively. These solution activities can be compared with thresholds for Zn toxicity determined in solution culture experiments in the literature. For example, Taylor et al. (1991) studied the root inhibition of wheat seedlings in solution culture and found that the toxicity threshold for Zn was 45 µmol L–1, and the maximum unit toxicity of Zn was 0.5% growth reduction/µmol Zn L–1 giving approximate EC25 and EC50 values of 50 and 100 µmol Zn L–1, respectively. These values are somewhat lower than those calculated from our data. It is generally difficult to compare ECx values with most results reported in the literature because of the many different variables that may affect toxicity results. The growing periods are highly variable and it is well known that plant tissue response varies with the age of the plant. Furthermore, element concentrations are often not expressed as free ion activities, but as total concentration in solution, and insufficient experimental detail is given to compute ion activities.



View larger version (15K):
[in this window]
[in a new window]
  		Fig. 5. Plots of plant yields as dry weight versus (a) Zn CE before growth, and (b) free Zn2+ activities on logarithmic scales. The solid line represents the calculated dose–response curves.

 

   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
Important considerations of metal behavior in soils lend weight to the argument that single status measurements, such as free metal ion activity, may only predict toxicity under well-defined conditions. As metal bioavailability in soil is a dynamic entity, there are additional models or predictors that need to be considered for this concept to be useful for regulatory purposes. From an operational sense, measurements of free ion activities of multiple elements in soil pore water are typically neither trivial nor robust, and the reliability, reproducibility, and practicality of these measurements by any technique must be considered. Soil extraction with CaCl2 appears to be a highly effective, inexpensive, and simple technique to predict Cd uptake by plants. For Pb and Zn, assessment of soils using DGT provides a convenient and relatively simple technique (compared with soil solution extraction and speciation) to predict metal concentrations in plants, and may be useful in predicting phytotoxicity thresholds.

The distinguishing feature of the DGT measurement is that it incorporates the kinetics of metal supply from solid phase to solution, and our results indicate that this supply may be important for Zn and Cd in these soils. These results do not give support to the idea that determinations of free ion activities of Zn, Cd, and Pb are the best predictors of metal concentrations in plant tissue. Although the free ion might be the form of metal that the plant transports across the membrane, other factors, such as its dynamic supply, also appear to be important.


   ACKNOWLEDGMENTS
 
The authors wish to thank Ian Oliver and Enzo Lombi for their assistance with the isotope dilution experiments, and Gill Cozens and Daryl Stevens for their assistance with the plant experiments. Hao Zhang received a CSIRO Sir Frederick McMaster Visiting Fellowship.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 


Related articles in JEQ:

This Issue in Journal of Environmental Quality

JEQ 2005 34: 403-407. [Full Text]  




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 Related articles in JEQ
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 ISI Web of Science (13)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Nolan, A. L.
Right arrow Articles by McLaughlin, M. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Nolan, A. L.
Right arrow Articles by McLaughlin, M. J.
Agricola
Right arrow Articles by Nolan, A. L.
Right arrow Articles by McLaughlin, M. J.
Related Collections
Right arrow Heavy Metals
Right arrow Plant and Soil Interactions
Right arrow Soil Pollution

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