Science of the Total Environment 416 (2012) 490–500
Contents lists available at SciVerse ScienceDirect
Science of the Total Environment
j ourna l homepage: www.e lsev ie r .com/ locate /sc i totenv
Cadmium and Zn availability as affected by pH manipulation and its assessment bysoil extraction, DGT and indicator plants
Iqbal Muhammad, Markus Puschenreiter ⁎, Walter W. WenzelBOKU, University of Natural Resources and Life Sciences, Vienna, Department of Forest and Soil Sciences, Konrad Lorenz Straße 24, A-3430 Tulln, Austria
Manipulation of soil pH by soil additives and / or rhizosphere processes may enhance the efficiency of metalphytoextraction. Here we report on the effect of nitric acid additions to four polluted soils on Cd and Zn con-centrations in soil solution (Csoln) and 0.005 M Ca(NO3)2 extracts, and related changes in the diffusive fluxesand resupply of the metals as assessed by diffusive gradients in thin films (DGT). The responses of thesechemical indicators of bioavailability were compared to metal uptake in two indicator plant species, commondandelion (Taraxacum officinale F.H. Wigg) and narrow leaf plantain (Plantago lanceolata L.) grown for75 days in a pot experiment.Lowering soil pH increased Csoln, the 0.005 M Ca(NO3)2-soluble fractions and the DGT-measured Cd and Znconcentrations (CDGT) in the experimental soils. This was associated with enhanced uptake of Cd and Zn onsoils acidified to pH 4.5 whereas plants did not survive at pH 3.5. Toxicity along with decreased kinetics ofmetal resupply (calculated by the 2D DIFS model) in the strong acidification treatment suggests that moder-ate acidification is more appropriate to enhance the phytoextraction process.Each of the chemical indicators of bioavailability predicted well (R2>0.70) the Cd and Zn concentrations inplantain shoots but due to metal toxicity not for dandelion. Concentration factors, i.e. the ratio betweenmetal concentrations in shoots and in soil solution (CF) indicate that Cd and Zn uptake in plantain was notlimited by diffusion which may explain that DGT did not perform better than Csoln. However, DGT is expectedto predict plant uptake better in diffusion-limited conditions such as in the rhizosphere of metal-accumulating phytoextraction crops.
Cadmium and Zn share a similar geochemical behavior and oftenoccur together as environmental pollutants in soils, sediments, plantsand other endpoints (Adriano, 2001). Main pollution sources of bothelements in soil include mining and smelter activities, urban and in-dustrial wastes, and for Cd long-term fertilization of rock phosphateto agricultural soils (Adriano, 2001; McLaughlin et al., 1999).
Normal (background) concentrations in world soils are reportedto range between 0.02 and 6.2 mg kg−1 for Cd, and 1 and900 mg kg−1 for Zn (Adriano, 2001). Soils containing 5–20 mg Cdkg−1 or 600–3000 mg Zn kg−1 typically require remedial action asthey are likely to pose a risk for the environment (Eikmann andKloke, 1991).
Currently a number of technologies are available to remediate Cd/Zn-polluted soils. Engineering-based technologies are typically inva-sive, expensive, and may generate secondary wastes and additionalrisks to the environment (Wenzel et al., 1999). Phytoremediationhas been considered as less expensive, virtually non-invasive and
. Puschenreiter).
rights reserved.
environmentally-friendly alternative with high public acceptance.Among the various phytoremediation technologies, phytoextraction,i.e. the removal of soil pollutants by cropping and harvesting metal-accumulating plants is the method of choice for relatively mobilemetals such as Cd and Zn.
Phytoextraction largely depends on (i) the plant's ability to accu-mulate the pollutants in the shoots, and (ii) the bioavailability ofsoil pollutants. The bioavailability of metals in soil is controlled bysoil factors such as pH, cation exchange capacity, drainage status ofsoil, organic matter and soil clay types. Soil pH is a master variableof metal solubility, speciation in soils and plant uptake. The solubilityof cationic metals such as Cd and Zn is known to increase with de-creasing pH (Adriano, 2001).
Lowering soil pH by rhizosphere processes, e.g. by co-cropping ofroot zone acidifiers such as red alder and metal accumulators suchas willows, or soil amendments such as elemental sulfur has beenproposed as effective means to increase metal cation uptake by phy-toextraction crops (e.g. Wenzel et al., 1999; Wenzel, 2009) and hasbeen demonstrated in pot and field experiments (Kayser et al.,2000; Kukier et al., 2004; Wang et al., 2006).
Regarding their uptake behavior, plants may be categorized intothree types, i.e. indicators, accumulators and excluders (Adriano,
491I. Muhammad et al. / Science of the Total Environment 416 (2012) 490–500
2001). Indicator plants are widely used for biological monitoring ofseveral environmental indicators, e.g., pollutant level in soil, waterlogging, nitrogen deficiency and fertility. Many authors have previ-ously discussed the advantages of biological monitoring (e.g., Wittig,1993; Bargagli, 1998). Taraxacum officinale is characterized by wide-spread ecological distribution, which makes this plant particularly in-teresting for biological monitoring of pollutants (Djingova and Kuleff,1993; Simon et al., 2006). Plantago lanceolata is also a common peren-nial plant species in native and agricultural ecosystems of temperateclimatic zones and has been suggested as an indicator plant for bio-available forms of Cd and Zn in soil (Leštan et al., 2003; Zupan et al.,1997).
The potential metal bioavailability in soil is traditionally assessedby chemical extraction (e.g. by 1 M NH4NO3, 0.005 M Ca(NO3)2 or0.05 M Na2-EDTA) or by measuring soil solution concentrations inwater extracts (Prüeß, 1997; Gray et al., 1999). However, theseequilibrium-based approaches do not account for the depletion atthe root–soil interface and depletion-induced resupply from thesolid phase. A promising tool to study the dynamics of trace elementsin soil solution and to mimic the processes in the rhizosphere is thediffusive gradients in thin films technique, DGT (Davison and Zhang,1994). Deployment of DGT lowers the concentration of metals locallyin the soil solution at the DGT–soil interface. The mass of the metalaccumulated in the DGT device mainly depends on the soil solutionconcentration and the fluxes, i.e. the diffusional transport and the ki-netics of metal resupply from labile pools in the solid phase (Zhang etal., 2001; Harper et al., 1998). The ratio of CDGT (interfacial metal con-centration) to the concentration in the soil solution is termed R anddescribes the metal resupply characteristics in a given soil as it is re-lated to the solid phase labile pool size (Kdl, partition coefficient forthe labile species) and the response time (Tc) of the soil to depletion.The latter is directly related to the rate constant of the resupply pro-cess (Zhang et al., 2001; Harper et al., 1998).
DGT has been used for more than a decade to study the bioavail-ability and resupply of essential and toxic elements in soil (Zhang etal., 2001; Nolan et al., 2005). Several studies have confirmed goodcorrelation betweenmetal concentration in plants and their measure-ment by DGT (Zhang et al., 2001, 2004; Song et al., 2004; Fischerovaet al., 2005; Koster et al., 2005).
Efficient phytoextraction within acceptable time is often hinderedby limited availability of the target pollutants (Wenzel, 2009). Selec-tion of appropriate soil amendments or the design of rhizosphere ma-nipulation to adjust soil pH requires detailed information on therelated changes in metal solubility, extractability, metal resupplyfrom the solid phase and the resulting phytoavailability to plants. Assoils are highly diverse in their properties, it is difficult to predictthe response of metal bioavailability to soil pH manipulation.
We selected four polluted soils representing a range of soil solu-tion concentrations of Cd and Zn and related Kdl values to determinethe effects of lowering the pH on (1) Cd and Zn extractability and sol-ubility; (2) metal fluxes and resupply determined by the DGT tech-nique; (3) metal accumulation in indicator plants, including theassessment of potential toxicity effects. Using this dataset we evaluat-ed the predictive power of the chemical and biological indicators ofbioavailability and their application to design proton-aided phytoex-traction technologies.
2. Materials and methods
2.1. Experimental soils and acidification treatments
Four soils, ARNB (Arnoldstein, Austria), PR2 (Příbram, Czech Re-public), GÖ (Gyöngyösoroszi, Hungary) and SK (Banská Štiavnica,Slovakia) were used in this experiment (Table 1). These soils are clas-sified as A horizons of Cambisols (IUSS Working Group WRB, 2006)and had been polluted with Zn and Cd over several hundred years
due to atmospheric deposition derived from metal smelters and pro-cessing. Initial soil pH values and other characteristics are shown inTable 1.
Prior to use in the pot experiment, all four soils were air dried,passed through a 2-mm sieve to remove pieces of stones and homog-enized. Water holding capacity (WHC) was determined by placingthe soils on filter paper on a shallow pan of water until the soil wassaturated. Then the soils were allowed to drain in water-saturated at-mosphere until the drainage was complete and the water content wasdetermined.
The soils were acidified with HNO3 to obtain the targeted pHlevels, further referred as A1 (pH 4.5) and A2 (pH 3.5); the originalpH level is termed A0. For calculation of the amounts of H+ neededto obtain the targeted pH levels in each soil, a preliminary incubationexperiment was conducted in which a known amount of H+ wasadded in the form of HNO3 to the soils prior to 5 days incubation at25 °C and 65% of the WHC. The amounts of the H+ and the corre-sponding volumes of 14.44 M nitric acid required to obtain the tar-geted pH values in the main experiment are shown in Table 2.Three batches for each soil (A0, A1 and A2 level) were adjusted to
492 I. Muhammad et al. / Science of the Total Environment 416 (2012) 490–500
65% WHC and subsequently the appropriate amount of HNO3 wasadded. The soils were incubated in plastic bags at room temperaturein the dark for equilibration. After 7 days of incubation, soils wereair dried and kept for the plant experiment and DGT deployment.
2.2. Pot experiment
We purchased seeds of two indicator plant species, common dan-delion (T. officinale F.H. Wigg) and narrow leaf plantain (P. lanceolataL.) from Rühlemann's Kräuter & Duftpflanzen, Horstedt, Germany.Eight seeds of the respective plant species were sown in drainedpots (7-cm depth and 9-cm diameter) containing 100 g untreated(A0) or acidified (A1, A2), air-dry soil. Each treatment was replicatedthree times. Additionally 24 unplanted blank pots containing only soilwere set up containing only soil in two replicates for each pH treat-ment. The purpose of including these blank pots in the experimentwas to observe the time course of pH and metal extractability in thesoils in the absence of roots and rhizosphere processes.
Seeds germinated after four (T. officinale) and seven (P. lanceolata)days. The number of plants was subsequently reduced to three plantsper pot. As the plants could not germinate in the A2 treatment, bothspecies were pre-germinated in a peat potting medium and the seed-lings were grown until three to four leaves had developed. Theseseedlings were then transplanted into the A2 pots. However, alsothe transplanted seedlings did not survive in this treatment.
The plants were allowed to grow for a period of 75 days in agreenhouse with automatically controlled light and temperature con-ditions (16 ⁄ 25 °C day ⁄ night temperature; 60% relative humidity,16 h light per day). During the plant experiment, all pots includingblank pots (without plants) were watered every two or three daysto maintain soil moisture at approximately 65% WHC.
Soil samples were taken 10, 20, 35 and 55 days after starting theexperiment from control pots for analysis of pH and 0.005 MCa(NO3)2-extractable Cd and Zn. Soil of each blank pot was properlymixed before taking samples with a clean spatula. Upon terminationof the experiment after 75 days plants were harvested and the soilssieved tob2 mm using a stainless steel mesh to remove very fineroots. Before that, coarser roots were removed manually.
2.3. Soil and plant analysis
The air-dried experimental soils were analyzed for soil texture(sand 2000-63 μm, silt 63-2 μm, clayb2 μm), total carbon using thecombustion methods and infrared detection, carbonate equivalentsby the Scheibler method, and pH (H2O) at a soil:water ratio of 1:2.5(w/v). Organic carbon (OC) was calculated from total carbon contentsand the carbon equivalent of the carbonates. Total metal concentra-tions were determined after digestion by aqua regia. Potentially labilefractions of Cd and Zn were measured following extraction with0.05 M Na2-EDTA and 1 M NH4NO3. All these measurements followedstandard procedures described in Blum et al. (1996).
The soils collected at different times from the non-planted blankpots were analyzed for pH (H2O) and metals in 0.005 M Ca(NO3)2 ex-tracts according to McLaren et al. (2007) but employing a modifiedsoil : solution ratio (1:5 instead of 1:6; m/v). Prior to measurementthe Ca(NO3)2 extracts were filtered using Munktell Filter PaperGrade (14/N, 2.7 μm and 150 mm diameter) and acidified.
Plants were harvested by clipping the stem just above the soil sur-face using scissors. Shoots were washed several times with deionizedwater and were oven dried at 65 °C until constant weight. All plantsharvested from the same pot were mixed to obtain a composite sam-ple. The plant material was ground in a stainless steel plant mill. Sub-samples of 0.2 g were digested in 4 mL concentrated HNO3 (reagentgrade., p.a., Sigma-Aldrich Handels GmbH, Vienna, Austria) and1 mL concentrated HClO4 (reagent grade, p.a., Sigma-Aldrich HandelsGmbH, Vienna, Austria) at 225 °C using an automated heating block
(Digester DK 42/26, Velp Scientifica, Milano, Italy). After digestionZn and Cd were measured by using ICP-MS (Elan 9000 DRCe, PerkinElmer). Blank samples and certified reference materials CTA-OTL-1and DCI 7004 for plant and soil samples, respectively) were used forquality control. Recoveries for Zn and Cd from certified reference ma-terial were in the range of 90 to 95%. All devices used for metal extrac-tion and digestion were soaked in diluted HNO3 and rinsed withdeionized water.
2.4. DGT and soil solution measurements and calculations
Diffusive fluxes of Cd and Zn were measured in soil samplesaccording to Zhang et al. (2001) using diffusive gradients in thinfilms (DGT) fitted with Chelex-100-resin-impragnated gels. Accord-ing to Zhang and Davison (1995), the efficiency of Cd and Zn adsorp-tion by the chelex resin is reduced at pHb4, but still in an acceptablerange of ~80–100% at pH 3.5–4. The remaining portions of experi-mental soils after setting up the pot experiment were used to deployDGT. Soil slurries were prepared by the addition of deionized water(0.066 μS cm−1) to adjust to 100% WHC. The slurries were placed inan incubator (Cooled Incubators IPP with Peltier-technology, IPP500, Memmert) at 23 °C for 24 h for equilibration. The DGT deviceswere prepared according to the procedure described by Fischerovaet al. (2005). Approximately 3 grams of soil slurry were placed onthe top of the DGT devices and kept in tightly closed plastic boxescontaining a water-saturated atmosphere. These plastic boxes wereplaced in an incubator at 23 °C for a period of 5 and 36 h, respectively.The purpose of measuring two time points was to observe the tempo-ral changes in Zn and Cd resupply from labile pools of the solid phase.After incubation, DGT devices were stream-washed with a jet of deio-nized water to remove the soil. Subsequently, devices were opened toremove the filter and the diffusive membrane. The chelex resin gelswere eluted by 1 mL of 1 M HNO3. The eluate was diluted 5 timesand analyzed for Zn and Cd using ICP-MS. Total mass of metals (M),time averaged interfacial concentrations (CDGT) and R (i.e., the ratiobetween CDGT and concentration of metals in soil solution, Csoln)were calculated according to Ernstberger et al. (2002). The responsetime (Tc) was calculated using the 2D DIFS model employing themode "R fitting" according to Sochaczewski et al. (2007).
To determine Csoln, the slurries remaining after DGT measurementwere stirred and 5–10 mL portions were transferred into a 25-mLPTFE tube and centrifuged at 614×g for 10 min. The supernatantwas collected with a plastic syringe and filtered into micro vialsusing disposable polysulfone filter assemblies with 13-mm diameterand 0.45-μm pore size (Whatman Puradisk). After filtration, Cd andZn concentrations in soil solution were measured using ICP-MS.
2.5. Statistical analysis
The pot experiment was conducted in a completely randomizeddesign. Treatment effects on plant growth were evaluated using oneway analysis of variance (ANOVA). Duncan's multiple range test(DMRT) (Pb0.05) was used for comparison between treatmentmeans. Statistical analyses were performed using PASW statistics 18(SPSS, Inc., Somers, NY, USA). The regression/correlation analysiswas calculated by Microsoft Excel 2003 professional edition.
3. Results
3.1. Correlation between metal concentrations in soil solution and 0.05 MCa(NO3)2
The 0.005 M Ca(NO3)2-extractable metal concentrations at the be-ginning of the experiment (day 2, non-planted pots) were linearlycorrelated with the soil solution concentrations (Csoln) obtained bycentrifugation of the slurries (100% WHC) prepared for the DGT
493I. Muhammad et al. / Science of the Total Environment 416 (2012) 490–500
measurements (Cd: R2=0.93; Zn: R2=0.83; Fig. 1). Compared tothose in soil solution, the concentrations of Cd and Zn in theCa(NO3)2 extract were larger by approximately ten and five times, re-spectively (Fig. 1).
3.2. Time course of pH in non-planted pots
The pH values in the A0 treatments (no nitric acid addition)remained virtually unchanged throughout the duration of the pot ex-periment. However, pH generally increased in the acidification treat-ments, with steeper increases observed at stronger acidification (A2).The pH increase was most pronounced (>1 unit) in soil ARNB, fol-lowed by the soils PR2, GÖ and SK. Soil PR2 showed only marginalchanges of pH in the A1 treatment and increased only by 0.43 unitsin the A1 treatment (Fig. 2). The pH for GÖ soil started from pH>5which indicates very fast buffering of protons in GÖ soil alreadyprior to the start of the experiment.
3.3. Time course of 0.005 M Ca(NO3)2-extractable Cd and Zn
In the A0 treatments Ca(NO3)2-extractable Cd and Zn remained atthe initial concentrations throughout the experiment (Fig. 3).
In the A1 treatments, changes of Cd and Zn concentrations weremarginal in soil PR2 whereas initial moderate decreases were ob-served in soil ARNB until day 20, with no further changes thereafter.The soils GÖ and SK showed initial increases of Cd and Zn extractabil-ity, followed by decreases after day 10. These changes were more pro-nounced in soil SK (Fig. 3). In the A1 treatments, the maximalconcentrations of Cd decreased in the order SK>ARNB>GÖ>>PR2,and for Zn according to SK>>GÖ>ARNB>>PR2.
Znsoln (µg cm-3)
Znsoln (µg cm-3)
0 50 100 150 200 250Ca(
NO
3)2
extr
acta
ble
Zn
(mg-1
kg
soil)
on 2
nd d
ay in
non
-pla
nted
pot
s
0
50
100
150
200
250
c
0 50 100 150 200 250
Zn D
GT (
µg c
m-3
)
0
2
4
6
8
10
a
e0 20 40 60 80 100120
0
2
4
6
8
10
R2 = 0.83
R2 = 0.44
R2 = 0.98
Fig. 1. Correlations between metal concentrations in soil solution and in 0.005 M Ca(NO3)2-iment (day 2, non-planted pots). Insets (e) and (f) show the plots after excluding the A2 levethe blank pots data having standard errors of the mean (n=2).
Pronounced temporal changes of metal concentration were foundin all A2 treatments. In soil GÖ, the metal concentrations steeply de-creased throughout the experiment. In all other soils, a pronounceddecrease of Cd and Zn concentrations was observed after an initial in-crease until day 10 (Fig. 3). In the A2 treatments, the maximal con-centrations of Cd decreased in the order SK>ARNB>>GÖ>PR2,and for Zn according to GÖ>SK>ARNB>>PR2.
3.4. Effect of pH on 0.005 M Ca(NO3)2-extractable Zn and Cd
The correlations between soil pH and the metal extractability areshown in Fig. 4. For each soil, Ca(NO3)2-extractable Cd and Zn wereinversely correlated with soil pH (R2>0.70).
3.5. Diffusive metal fluxes and their resupply potential determined byDGT
The results of the DGT measurements (CDGT and R values) after 5and 36 h deployment are shown in Fig. 5. The masses (M) of Cd andZn accumulated in the chelex gel increased in all soils with the dura-tion of deployment. Similarly, the diffusive flux (F) of both metals was(with few exceptions) subsequently decreasing with time (data notshown). The ratio of the DGT-measured concentration to the solutionconcentration (Csoln), R, showed a decrease with increase of deploy-ment time but for some soils (e.g. GÖ) no clear temporal trend wasfound.
For both Cd and Zn, CDGT (36 h) was linearly but weakly related toCsoln (Cd: R2=0.35; Zn: R2=0.44) if all soils and acidity levels wereincluded (Fig. 1). Excluding the A2 levels of soils ARNB and SK im-proved the correlations significantly (Cd: R2=0.85; Zn: R2=0.98).
Ca(
NO
3)2
extr
acta
ble
Cd
(mg
kg-1
soi
l)
on 2
nd d
ay in
non
-pla
nted
pot
s b
0 1000 2000 3000 4000 5000 60000.0
0.5
1.0
1.5
2.0
2.5
3.0
R2 = 0.93
Cd D
GT (
ng c
m-3
)
d
0 1000 2000 3000 4000 5000 60000
20
40
60
80
f0 500 1000 1500 2000
0
20
40
60
80
Cdsoln (ng cm-3)
Cdsoln (ng cm-3)
R2 = 0.85
R2 = 0.35
extracts (a–b) or the DGT-measured concentration (c–d) at the beginning of the exper-ls of soils ARNB and SK. Error bars represent standard errors of the mean (n=3) except
Time (days)0 20 40 60 80
pH (
H2O
)
3
4
5
6
7
8PR2
GÖ
Time (days)
0 20 40 60 80
pH (
H2O
)
3
4
5
6
7
8
A0 (orignal soil pH) A1 (pH 4.5)A2 (pH 3.5)
Time (days)
0 20 40 60 80
pH (
H2O
)
3
4
5
6
7
8SK
A0 (orignal soil pH) A1 (pH 4.5)A2 (pH 3.5)
A0 (orignal soil pH) A1 (pH 4.5)A2 (pH 3.5)
A0 (orignal soil pH) A1 (pH 4.5)A2 (pH 3.5)
ARNB
Time (days)0 20 40 60 80
pH (
H2O
)
3
4
5
6
7
8
Fig. 2. Temporal changes of pH (H2O) in unplanted pots. Error bars represent standard errors of the mean (n=2).
494 I. Muhammad et al. / Science of the Total Environment 416 (2012) 490–500
3.6. Effect of soil pH manipulation on shoot biomass
As mentioned in the methods section, none of the plants was ableto germinate in the A2 treatment. The shoot biomass production ofplantain and dandelion is shown for the A0 and A1 treatments inTable 3.
Plantain produced the largest shoot biomass in the A1 treatmenton soil PR2 while significant differences were found between theother soils. Shoot biomass production of plantain on soil PR2 in theA1 treatments decreased according to PR2>ARNB>>SK. The plantswere not able to grow on soil GÖ.
In the A0 treatments, the shoot biomass production of dandelionwas not statistically significant among the soils. In the A1 treatmentthe plants only survived on soils PR2 and ARNB, with significantlyhigher shoot biomass produced on PR2 compared to ARNB.
3.7. Cadmium and Zn concentrations in plant shoots and their correlationwith chemical indicators of metal bioavailability
Cadmium concentrations in shoots ranged from 0.50 (PR2, A0) to30.3 (SK, A1) mg kg−1 in plantain, and between 0.62 (PR2, A0) to(ARNB, A1) 66.1 mg kg−1 in dandelion (Table 3). The Zn concentrationsvaried between 56.6 (PR2, A0) to 648 (SK, A1) mg kg−1 in plantain and67.8 (PR2, A0) to 1000 (ARNB, A1) in dandelion shoots (Table 3). Thelargermetal concentrations in dandelion correspond to smaller biomassproduction of this species compared to plantain (Table 3).
Cadmium and Zn concentrations in plantain shoots correlated well(R2>0.80) with the metal concentrations in soil solution (Csoln) andthe 0.005 M Ca(NO3)2 extract, respectively (R2≥0.80) (Fig. 6). Incase of dandelion, both Cd and Zn concentrations in shoots werestrongly correlated (R2>0.80) with the metal concentrations in the0.005 M Ca(NO3)2 extract but in contrast only weakly (R2b0.50)
correlated with the corresponding metal concentration in soil solu-tion (Csoln) (Fig. 6).
Similarly, strong (R2≥0.70) correlations between metal concen-trations in shoots and CDGT were obtained for plantain but not fordandelion plants (Fig. 6).
4. Discussion
4.1. Effect of nitric acid addition on soil pH and metal solubility
Addition of appropriate amounts of nitric acid efficiently de-creased soil pH with the exception of soil GÖ (Fig. 2). In this soil, pHalready increased significantly during the short incubation period be-fore the experiment started. The rapid buffering of protons in soil GÖis related to the presence of carbonates and larger amounts of ex-changeable Ca2+ as indicated by the higher CEC (Table 1). Exchangeof Ca2+ for protons is a rapid process. Also CaCO3 is known to reactquite quickly with protons by releasing Ca2+ and bicarbonate ionsthat can further react with protons to form CO2 and water(Schwertmann et al., 1987). In the absence of carbonates, the kineticsof proton buffering slow down as—apart from exchange reactions—silicates including clay minerals and metal oxides become the mainbuffer substances (Schwertmann et al., 1987). Such slower buffer re-actions are probably reflected by the steady increase of pH during theexperiment (Fig. 2).
Application of nitric acid effectively increased Cd and Zn concen-trations in the Ca(NO3)2 extracts (Fig. 4) and in soil solution. Thetime course of metal concentrations shows high initial release of Cdand Zn even in the well-buffered soil GÖ but rapid subsequent de-clines of the metal concentrations especially in the strongly acidifiedA2 treatments (Fig. 4). The effective solubilisation of metals in theGÖ soil may be related to co-dissolution from soil carbonates.
PR2 (Zn)
Time (days)
0 20 40 60 800
10
20
30
40
50
GÖ (Cd)
Time (days)0 20 40 60 80
0.0
0.2
0.4
0.6
0.8
1.0
SK (Cd)
GÖ (Zn)
Time (days)0 20 40 60 80
0
20
40
60
80
100
120
140
160
180
Time (days)0 20 40 60 80
0
50
100
150
200
250SK (Cd)
Time (days)0 20 40 60 80
0
1
2
3
4
5
ARNB (Cd)
Time (days)
0 20 40 60 800
1
2
3
4
PR2 (Cd)
Time (days)
0 20 40 60 800.0
0.1
0.2
0.3
0.4
0.5
0.6
SK (Zn)
ARNB (Zn)
Time (days)
0 20 40 60 80
Zn
(mg
kg-1
soil)
Zn
(mg
kg-1
soil)
0
20
40
60
80
100
120
140
160
Cd
(mg
kg-1
soil)
Cd
(mg
kg-1
soil)
Zn
(mg
kg-1
soil)
Cd
(mg
kg-1
soil)
Zn
(mg
kg-1
soil)
Cd
(mg
kg-1
soil)
A0 (orignal soil pH) A1 (pH 4.5)A2 (pH 3.5)
A0 (orignal soil pH) A1 (pH 4.5)A2 (pH 3.5)
A0 (orignal soil pH) A1 (pH 4.5)A2 (pH 3.5)
A0 (orignal soil pH) A1 (pH 4.5)A2 (pH 3.5)
A0 (orignal soil pH) A1 (pH 4.5)A2 (pH 3.5)
A0 (orignal soil pH) A1 (pH 4.5)A2 (pH 3.5)
A0 (orignal soil pH) A1 (pH 4.5)A2 (pH 3.5)
A0 (orignal soil pH) A1 (pH 4.5)A2 (pH 3.5)
Fig. 3. Temporal changes of Cd and Zn extractability by 0.005 M Ca(NO3)2 in unplanted pots. Error bars represent standard errors of the mean (n=2).
495I. Muhammad et al. / Science of the Total Environment 416 (2012) 490–500
With progressing acidification we found a steep decrease of thedistribution coefficients, i.e. Kdl values (the ratio between Csoln andthe labile metal concentration in the solid phase, obtained by 0.05 MNa2-EDTA or 1 M NH4NO3 extraction). The observed decline in Kdl isclearly reflecting the solubilisation of metals by the added nitric acidand indicates a concomitant strong decrease of the buffer power.
DGT-measured fluxes and concentrations (CDGT) (Fig. 5) of Cd andZn generally increased upon acidification, indicating enhanced solu-bility. The ratio between metal concentration in bulk soil solution(Csoln) and at the DGT interface (CDGT), i.e. R, was increased for Cdin soils ARNB, PR2 and SK and for Zn in soils ARNB and GÖ when acid-ified to pH 4.5 (A1 treatment) but decreased again upon further
3.0 3.5 4.0 4.5 5.0 5.5 6.00
1
2
3
4
R2 = 0.91
R2 = 0.91
ARNB (Cd)
PR2 (Cd)
3.5 4.0 4.5 5.0 5.5 6.00.0
0.1
0.2
0.3
0.4
0.5
0.6
GÖ (Cd)
5.5 6.0 6.5 7.0 7.50.0
0.2
0.4
0.6
0.8
1.0
R2 = 0.87
SK (Cd)
3.5 4.0 4.5 5.0 5.5 6.0 6.50
1
2
3
4
5
R2 = 0.71
ARNB (Zn)
3.0 3.5 4.0 4.5 5.0 5.5 6.0Ca(
NO
3)2
extr
acta
ble
Zn
(mg
kg-1
soi
l)
0
20
40
60
80
100
120
140
160
R2 = 0.90
PR2 (Zn)
3.5 4.0 4.5 5.0 5.5 6.00
10
20
30
40
50
R2 = 0.87
5.5 6.0 6.5 7.0 7.50
20
40
60
80
100
120
140
160
180
R2 = 0.89
SK (Zn)
3.5 4.0 4.5 5.0 5.5 6.0 6.50
50
100
150
200
250
R2 = 0.79
GÖ (Zn)
Ca(
NO
3)2
extr
acta
ble
Cd
(mg
kg-1
soi
l)
Ca(
NO
3)2
extr
acta
ble
Zn
(mg
kg-1
soi
l)
Ca(
NO
3)2
extr
acta
ble
Cd
(mg
kg-1
soi
l)
Ca(
NO
3)2
extr
acta
ble
Zn
(mg
kg-1
soi
l)
Ca(
NO
3)2
extr
acta
ble
Cd
(mg
kg-1
soi
l)
Ca(
NO
3)2
extr
acta
ble
Zn
(mg
kg-1
soi
l)
Ca(
NO
3)2
extr
acta
ble
Cd
(mg
kg-1
soi
l)
pH (H2O)
pH (H2O)
pH (H2O)
pH (H2O)
pH (H2O)
pH (H2O)
pH (H2O)
pH (H2O)
Fig. 4. Correlation between soil pH and the 0.005 M Ca(NO3)2 extractable fractions of Zn and Cd in the unplanted pots. Error bars represent standard errors of the mean (n=2).
496 I. Muhammad et al. / Science of the Total Environment 416 (2012) 490–500
acidification (Fig. 5). This finding can be related to a stronger increaseof Cd and Zn concentrations in soil solution compared to their DGT-measured concentrations, indicating that in these cases metal
resupply from the solid phase was enhanced upon moderate acidifi-cation (pH 4.5) but slowed down in the highly acidic treatments(pH 3.5).
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
5 36 5 36 5 36
A0 A1 A2
R (
Zn)
ARNB
PR2
GÖ
SK
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
5 36 5 36 5 36
A0 A1 A2
R (
Cd)
ARNB
PR2
GÖ
SK
0
5
10
15
20
25
5 36 5 36 5 36
A0 A1 A2
CD
GT
for
Zn
(µg
cm-3
) ARNB
PR2
GÖ
SK
0
50
100
150
200
250
300
5 36 5 36 5 36
A0 A1 A2
CD
GT
for
Cd
(ng
cm-3
) ARNB
PR2
GÖ
SK
Fig. 5. DGT parameters CDGT (interfacial concentrations of metals) and R (the ratio between CDGT and Csoln) at A0 (original pH level), A1 (pH 4.5) and A2 (pH 3.5) after 5 and 36 h ofdeployment.
497I. Muhammad et al. / Science of the Total Environment 416 (2012) 490–500
The distribution coefficients, Kdl, and R are both indicators of po-tential resupply if the metal concentration in soil solution is loweredby plant uptake. As Kdl is strictly describing the metal distribution atequilibrium, it will reflect metal resupply in the depletion zonearound plant roots only if there is no rate limitation of metal desorp-tion at the time scale of uptake by the plant root. In contrast, the DGT-based R value is also related to the kinetics of desorption processeswith a lower R value indicating slower desorption (Degryse et al.,2009a). Taking together our Kdl and R value data it appears that acid-ification decreased the instantaneous soil buffer power for Cd and Znbut increased the contribution of slower (kinetically controlled) de-sorption of the metals from soil in some of the moderate acidificationtreatments. Indeed, DIFS modeling revealed that the above men-tioned increases of R (36 h) were related to pronounced increases ofthe response time (Tc), indicating enhanced kinetics of resupply inthese treatments as compared to the controls (A0) and highly acidic(A2) treatments.
Our findings suggest that moderate acidification to pH 4.5 is supe-rior to stronger acidification (pH 3.5) as the metal concentrationsmay be kept below phytotoxic levels while increasing both themetal concentration in soil solution and the resupply rate from thesolid phase. Further acidification is unlikely to improve the phytoex-traction process as the plants may suffer frommetal toxicity thus pro-ducing smaller biomass while the resupply rates are often lower thanat moderate acidification. This is in accordance with the findings ofWang et al. (2006) who found the largest uptake of Zn in Thlaspi caer-ulescens shoots at moderate reduction of soil pH.
4.2. Evaluation of bioavailability indicators
The success of phytoremediation of metal-polluted soils depends onplant as well as soil properties. Phytoextraction may be limited by thechemical availability of the metal of concern in soil as soon as themore labile metal fraction becomes depleted due to high uptake rates
by the accumulator plant. It is therefore crucial to assess the availabilityof the metal in soil before applying phytoextraction at field scale. Com-mon phytoavailability testing includes biological methods such asgrowing indicator plants and chemical extraction of soils. The advan-tage of indicator plants relates to the direct assessment of phytoavail-ability but when applied to polluted soils this approach is also limitedto the range of tolerance of the test plant. In contrast, chemical soil ex-traction is independent of plant properties but is using an equilibriumapproach that does not reflect the dynamic uptake process.
More recently, the DGT technique has become a valuable tool formimicking metal uptake by plants (Zhang et al., 2004; Song et al.,2004; Koster et al., 2005). As pointed out by Degryse et al. (2009a),DGT is expected to mimic the process of uptake only if uptake is lim-ited by diffusion, i.e. if the uptake rate exceeds the rate of supply. TheDGT technique accounts for processes such as slow desorption anddissociation of labile complexes that are not assessed by traditionalsoil extraction. However, strong correlations between plant uptakeand DGT measurements may be also found in other cases as CDGTand soil solution concentrations may co-vary across a wide range ofsoil properties (Degryse et al., 2009a).
Metal hyperaccumulator plants available for phytoextraction ofmetal-polluted soils typically grow slow and therefore may not bethe first choice for rapid testing of phytoavailability of metals in pol-luted soils.
In this work we compare chemical indicators, i.e. metal concentra-tions in soil solution (Csoln) and in a 0.05 M Ca(NO3)2 extract as wellas DGT-measured metal concentrations with the uptake of Cd andZn in two indicator plant species.
Among the two extraction methods, Cd and Zn concentrationsobtained by centrifugation of the soil slurries at 100% WHC (Csoln)are considered to reflect the real metal concentrations in soil solutionbetter than those measured in 0.005 M Ca(NO3)2 extracts at soil : so-lution ratios of 1:5. Employing wider solution:soil ratios may result inincomplete desorption during the relatively short extraction periods
Table 3The effect of pH treatments on shoot biomass (mg pot−1) and heavy metals concentrations (mg kg−1 plant dry matter) in T. officinale (dandelion) and P. lanceolata (plantain). Sig-nificant difference between the treatment means at pb0.05 is indicated by different letters.
ARNB Shoot biomass mg pot−1 498 ab 1060 b 35 a 90 aPR2 Shoot biomass mg pot−1 2530 c 2840 c 21 a 620 bGÖ Shoot biomass mg pot−1 150 a – 90 a –
SK Shoot biomass mg pot−1 145 a 40 a 14 a –
ARNB Zn concA in shoots mg kg−1 plant dry matter 295 b 456 c 897 b 1000 bPR2 Zn concA in shoots mg kg−1 plant dry matter 56.6 a 118 a 67.8 a 105 aGÖ Zn concA in shoots mg kg−1 plant dry matter 82 a * 222 a *SK Zn concA in shoots mg kg−1 plant dry matter 314 b 648 d 295 a *ARNB Cd concA in shoots mg kg−1 plant dry matter 12.7 d 13.2 d 57.7 b 66.1 bPR2 Cd concA in shoots mg kg−1 plant dry matter 0.5 a 2.4 b 0.62 a 1.8 aGÖ Cd concA in shoots mg kg−1 plant dry matter 1.1 ab * 14 a *SK Cd concA in shoots mg kg−1 plant dry matter 9.3 c 30.3 e 6.9 a *
– No plant growth.⁎ Not determined.
A Concentration.
R2 = 0.82
R2 = 0.22
0
200
400
600
800
1000
1200
0 100 200 300Zn
conc
entr
atio
n in
sho
ots
(mg
kg-1
)
R2 = 0.85
R2 = 0.47
0
10
20
30
40
50
60
70
80
0 2000 4000 6000
Cdsoln (ng cm-3)
Cd
conc
entr
atio
n in
sho
ots
(mg
kg-1
)
Zn
conc
entr
atio
n in
sho
ots
(mg
kg-1
)
Cd
conc
entr
atio
n in
sho
ots
(mg
kg-1
)
Zn
conc
entr
atio
n in
sho
ots
(mg
kg-1
)
Cd
conc
entr
atio
n in
sho
ots
(mg
kg-1
)
R2 = 0.70
R2 = 0.18
0
200
400
600
800
1000
1200
0 2 4 6 8 10
ZnDGT (µg cm-3)
R2 = 0.84
R2 = 0.22
0
10
20
30
40
50
60
70
80
0 20 40 60 80 100
CdDGT (ng cm-3)
R2 = 0.85
R2 = 0.97
0
10
20
30
40
50
60
70
80
0 0.2 0.4 0.6 0.8 1
Cd in Ca(NO3)2 (mg kg-1 soil)
R2 = 0.80
R2 = 0.83
0
200
400
600
800
1000
1200
0 20 40 60
Zn in Ca(NO3)2 (mg kg-1 soil)
Znsoln (µg cm-3)
Fig. 6. Correlations between chemical predictors of metal bioavailability and metal concentrations in plant shoots. Filled cycles and the continuous line represent the data for plan-tain, whereas the open triangles and the dashed line represent the data for dandelion. For CDGT, the values of the 36 h deployment time were used. Error bars represent standarderrors of the mean (n=3).
498 I. Muhammad et al. / Science of the Total Environment 416 (2012) 490–500
499I. Muhammad et al. / Science of the Total Environment 416 (2012) 490–500
(2 hours) and cause dilution of complexing ligands thus loweringmetal concentrations (Degryse et al., 2009b; Molina Millán et al.,2006). This dilution effect is clearly reflected by the measured lowerconcentrations of Cd and Zn in the 0.05 M Ca(NO3)2 extract (Fig. 1).However, soil extraction methods are for several reasons easierto handle than obtaining soil solutions at narrow solution: soilratios by centrifugation (Degryse et al., 2009b). Our results showthat Cd and Zn concentrations obtained by extraction correlatedwell with those in the solution centrifuged from the soil slurry(Fig. 1).
DGT-measured concentrations (CDGT) of Cd and Zn were also cor-related with that in soil solution (Csoln), but in contrast to otherstudies (Degryse et al., 2009a), the degree of determination wasrather weak (Fig. 1) unless the A2 levels of soils ARNB and SK wereexcluded.
As both bioindicator plants were not able to grow in the stronglyacidified A2 treatments (Table 3), our comparison of chemical and bi-ological indicators of metal bioavailability is limited to the control andthe moderate acidification treatment (A1). Metal concentrations gen-erally increased in the acidification treatments relative to the control(Table 3). Cadmium and Zn concentrations in dandelion were gener-ally larger than in plantain shoots but the higher uptake was associat-ed with lower tolerance as indicated by more frequent total loss ofgrowth (Table 3).
Probably as a consequence of toxicity, Cd and Zn concentrations indandelion shoots were only loosely related to the metal concentra-tions in soil solution (Csoln) and CDGT (Fig. 6). This is well in linewith Degryse et al. (2009a) who suggested that DGT does not workwell in the toxic range. In contrast, we found fairly good correlationsbetween metal concentrations in plantain shoots and CDGT or Csoln(Fig. 6). As the Ca(NO3)2-extractable metal concentrations correlatedwell with Csoln, this extract was also a good predictor of Cd and Zn up-take in plantain (Fig. 6).
To further investigate whether the close relations between CDGTand metal concentrations in plantain shoots are an artifact of covari-ance between CDGT and Csoln (Degryse et al., 2009a) or if they reflectdiffusion control of metal uptake, we calculated the concentrationfactors (CF), i.e. the ratio between metal concentrations in plantainshoots and in soil solution (Csoln). With few exceptions, the CF valuesare clearly below 100, indicating that mass flowwould be sufficient toexplain the metal concentrations in shoots (Degryse et al., 2009a).However, metal supply may become limited by diffusion if metal(hyper-)accumulator plants as used in phytoextraction would begrown on the same soils. Recalculating the CF values by employingthe hyperaccumulation thresholds of Cd (100 mg kg−1 d.m.) and Zn(10,000 mg kg−1 d.m.) strongly suggests that in response to thehigh plant demand uptake might be controlled by diffusion. This issupported by the findings of Fitz et al. (2003) who showed that therate of resupply in the rhizosphere of the As hyperaccumulator Pterisvittata was decreased to one third of that in the bulk soil after onlyone cropping period. Therefore we argue that DGT is likely to outper-form the prediction of metal uptake in hyperaccumulator plants fromsoil solution or extraction data. Moreover, DGT is not limited to therange of tolerance of a given bioindicator plant.
5. Conclusions
Manipulation of pH may be a useful approach in enhancing theprocess of phytoextraction. However, it is critical to not exceedtoxic levels of metals in soil solution for the selected phytoextractioncrop. Our data suggest that moderate acidification in most cases keepsmetal concentrations below toxic levels while enhancing metalresupply as indicated by the measured R values and DIFS modeling.Chemical assessment of bioavailability using soil solution and DGTmeasurements may provide the information required to determine
the optimal acidification level before acidity-aided phytoextractionis applied.
Acknowledgements
The stay of the first author in Austria was funded by Higher Educa-tion Commission (HEC), Pakistan, under the project title “OverseasScholarship for Ms/Mphil Leading to PhD in Selected Fields Phase-II,Batch-I” and supported by the Austrian Exchange Service (ÖAD).
We also acknowledge funding of the experimental work throughthe project NUTZRAUM by the Austrian Federal Ministry of Agricul-ture, Forestry, Environment and Water Management handled byKommunalkredit Austria.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.scitotenv.2011.11.029.
References
Adriano DC. Trace elements in the terrestrial environment. 2nd ed. New York: SpringerVerlag; 2001.
Bargagli R. Trace elements in terrestrial plants: an ecophysiological approach to biomo-nitoring and biorecovery. Berlin: Springer Verlag; 1998 [324].
BlumWEH, Spregel H,WenzelWW.Bodenzustandsinventur, konzeption,Duruchführung,Bewertung. 2nd ed. Vienna: Bundesministerium für Land- und Forstwirtschaft; 1996.p. 102.
Davison W, Zhang H. In situ speciation measurements of trace components in naturalwaters using thin-film gels. Nature 1994;367:546–8.
Degryse F, Smolders E, Zhang H, DavisonW. Predicting availability of mineral elements toplants with the DGT technique: a review of experimental data and interpretation bymodelling. Environ Chem 2009a;6:198–218.
Degryse F, Smolders E, Parker DR. Partitioning of metals (Cd, Co, Cu, Ni, Pb, Zn) in soils:concepts, methodologies, prediction and applications. Eur J Soil Sci 2009b;60:590–612.
Djingova R, Kuleff I.Monitoring of heavymetal pollution by Taraxacumofficinale. In:MarkertB, editor. Plants as biomonitors. Weinheim, Germany: VCH Verlagsgesellschaft; 1993.p. 433–60.
Eikmann T, Kloke A. Nutzungs-und schutzgutbezogene Orientierungswerte für(Schad-) Stoffe in Böden. In: Rosenkranz, Einsele, Harreß, editors. Bodenschutz,Ergänzbares Handbuch, Erich Schmidt Verlag; 1991.
Ernstberger H, DavisonW, Zhang H, Tye A, Young S.Measurement and dynamicmodelingof trace metal mobilization in soils using DGT and DIFS. Environ Sci Technol 2002;36:349–54.
Fischerova Z, Szakova J, Pavlíková D, Tlustos P. The application of diffusive gradienttechnique (DGT) for assessment of changes in Cd, Pb, and Zn mobility in rhizo-sphere. Plant Soil Environ 2005;51:532–8.
Fitz WJ, Wenzel WW, Zhang H, Nurmi J, Stipek K, Fischerova Z, et al. Rhizospherecharacteristics of the arsenic hyperaccumulator Pteris vittata L. and monitoringof phytoremoval efficiency. Environ Sci Technol 2003;37:5008–14.
Gray CW, McLaren RG, Roberts AHC, Condron LM. Effect of soil pH on cadmium phytoa-vailability in some New Zealand soils. New Zeal J Crop Hort 1999;27:169–79.
Harper MP, Davison W, Zhang H, Tych W. Kinetics of metal exchange between solidsand solutions in sediments and soils interpreted from DGT measured fluxes. Geo-chim Cosmochim Acta 1998;62:2757–70.
IUSS Working Group WRB. World reference base for soil resources. World Soil Re-sources Reports No. 103. Rome: FAO; 2006.
Kayser A, Wenger K, Keller A, Attinger W, Felix HR, Gupta SK, et al. Enhancement ofphytoextraction of Zn, Cd, and Cu from calcareous soil: the use of NTA and sulfuramendments. Environ Sci Technol 2000;34:1778–83.
Koster M, Reijnders L, van Oost NR, Peijnenburg WJ. Comparison of the method of dif-fusive gels in thin films with conventional extraction techniques for evaluatingzinc accumulation in plants and isopods. Environ Pollut 2005;133:103–16.
Kukier U, Peters CA, Chaney RL, Angle JS, Roseberg RJ. The effects of pH on metal accu-mulation in two Alyssum species. J Environ Qual 2004;33:2090–102.
Leštan D, Grcman H, Zupan M, Bačac N. Relationship of soil properties to fractionationof Pb and Zn in soil and their uptake into Plantago lanceolata. Soil Sediment Contam2003;12:507–22.
McLaren MG, Clucas LM, Speir TW, van Schaik AP. Distribution and movement of nutri-ents and metals in a Pinus radiata forest soil following applications of biosolids. En-viron Pollut 2007;147:32–40.
McLaughlin MJ, Parker DR, Clarke JM. Metals and micronutrients–food safety issues.Field Crops Res 1999;60:143–63.
Molina Millán I, Fitz WJ, Unterbrunner R, Wenzel WW. Comparison of methods formeasuring metal desorption from soils for parameterizing rhizosphere models.Eur J Soil Sci 2006;57:38–46.
Nolan AL, Zhang H, McLaughlin MJ. Prediction of zinc, cadmium, lead, and copper avail-ability to wheat in contaminated soils using chemical speciation, diffusive
500 I. Muhammad et al. / Science of the Total Environment 416 (2012) 490–500
gradients in thin films, extraction, and isotopic dilution techniques. J Environ Qual2005;34:496–507.
Prüeß A. Action values for mobile (NH4NO3-extractable) trace elements in soils basedon the German national standard DIN 19730. In: Prost R, editor. Contaminatedsoils3rd International Conference on the Biogeochemistry of Trace Elements,INRA Editions. Paris, France; 1997. p. 415–23.
Schwertmann U, Susser P, Natsher L. Porton buffer compounds in soils. J Plant Nutr SoilSci 1987;150:174–8.
Simon L, Tamás J, Kovács E, Kovács B, Biró B. Stabilisation of metals in mine spoil withamendments and growth of red fescue in symbiosis with mycorrhizal fungi. PlantSoil Environ 2006;52:385–91.
Sochaczewski L, Tych W, Davison B, Zhang H. 2D DGT induced fluxes in sediments andsoils (2D DIFS). Environ Modell Softw 2007;22:14–23.
Song J, Zhao FJ, Luo YM, McGrath SP, Zhang H. Copper uptake by Elsholtzia splendensand Silene vulgaris and assessment of copper phytoavailability in contaminatedsoils. Environ Pollut 2004;3:307–15.
Wang AS, Angle JS, Chaney RL, Delorme TA, Reeves RD. Soil pH effects on uptake of Cdand Zn by Thlaspi caerulescens. Plant and Soil 2006;281:325–37.
Wenzel WW. Rhizosphere processes and management in plant-assisted bioremediation(phytoremediation) of soils. Plant and Soil 2009;321:385–408.
Wenzel WW, Adriano DC, Salt D, Smith R. Phytoremediation: a plant-microbe-basedremediation system. In: Adriano DC, Bollag JM, Frankenberger WT, Sims RC,
editors. Bioremediation of Contaminated Soils. Madison Wisconsin: American So-ciety of Agronomy, Crop Science Society of America, Soil Science Society of Amer-ica; 1999. p. 457–508.
Wittig R. General aspects of biomonitoring trace elements by plants. In: Markert B, editor.Plants as biomonitors. Weinheim, Germany: VCH Verlagsgesellschaft; 1993. p. 3-27.
Zhang H, DavisonW. Performance characteristics of diffusion gradients in thin films forthe in situ measurement of trace metals in aqueous solution. Anal Chem 1995;67:3391–400.
Zhang H, Zhao FJ, Sun B, DavisonW, McGrath SP. A newmethod to measure effective soilsolution concentration predicts copper availability to plants. Environ Sci Technol2001;35:2602–7.
Zhang H, Lombi E, Smolders E, McGrath S. Kinetics of Zn release in soils and predictionof Zn concentration in plants using diffusive gradients in thin films. Environ SciTechnol 2004;38:3608–13.
Zupan M, Hudnik V, Lobnik F, Kadunc V. Accumulation of Pb, Cd and Zn from contam-inated soil to various plants and evaluation of soil remediation with indicator plant(Plantago lanceolata L.). In: Prost R, editor. Contaminated soils3rd InternationalConference on the Biogeochemistry of Trace Elements, INRA Editions, Paris, France;1997. p. 415–23.
