Environ. Sci. Technol. 2004, 38, 3608-3613
Kinetics of Zn Release in Soils and Prediction of Zn Concentration in Plants Using Diffusive Gradients in Thin Films H A O Z H A N G , * ,† E N Z O L O M B I , ‡,§ ERIK SMOLDERS,+ AND STEVE MCGRATH‡ Department of Environmental Sciences, IENS, Lancaster University, Lancaster LA1 4YQ, United Kingdom, Agriculture and Environment Division, Rothamsted Research, Harpenden Herts AL5 2JQ, United Kingdom, and Laboratory of Soil and Water Management, K. U. Leuven, K. Kasteelpark Arenberg 20, 3001 Leuven, Belgium
Effective concentrations (CE) of Zn measured by the technique of DGT (diffusive gradients in thin films) were compared, along with total concentrations of Zn and the concentrations of Zn in soil solutions, to Zn concentrations in plants. Soils variously contaminated with Zn were collected in the vicinity of two galvanized electrical transmission towers (pylons) and two motorway crash barriers. Lepidium sativum was grown in each soil and in corresponding control soils amended with ZnCl2 to similar total Zn concentrations. CE, concentrations in soil solution, and total Zn were measured in all soils, and total Zn was measured in the plant shoots. The CE values, soil solution Zn, and shoot Zn concentrations were all larger in ZnCl2 amended soils than in field contaminated soils at corresponding total Zn. Correlations between the concentration of Zn in the plants and the measured soil parameter followed the order CE > soil solution > total Zn. The low scatter in the plot of log plant concentration versus log CE revealed a relationship with two distinct features. Plant Zn was between 100 and 300 mg/kg up to an effective Zn concentration of about 2 mg/L, above which plant Zn increased steadily with increasing CE. Use of a dynamic model to interpret the DGT measurement suggests that the intrinsic rate of release of Zn from solid phase to solution, expressed as a rate constant, is much higher for soils that receive fresh supplies of Zn. This finding provides a mechanistic basis for reconciling laboratory experiments, where metal is freshly amended, to data obtained in the field. The potential of DGT as a surrogate for metal availability to plants is further confirmed by this work.
Introduction Increasingly, it is being realized that total metal in soil is a poor predictor of availability of metals to plants because soil * Corresponding author telephone: +44 1524 593899; fax: +44 1524 593985; e-mail:
[email protected]. † Lancaster University. ‡ Rothamsted Research. § Present address: CSIRO Land and Water, PMB2, Glen Osmond, SA5064, Australia. + Laboratory of Soil and Water Management. 3608
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 13, 2004
properties such as soil pH, redox potential, texture, and organic matter content affect bioavailability (e.g., ref 1). It is recognized that the supply of metal to plants is through the soil solution (2-4). When metal is also supplied from the solid phase, it must first be transferred to solution before it can be taken up by plant roots. Assessment of the potential supply from the solid phase has traditionally been made by various extraction techniques that attempt to quantify the size of the available solid-phase pool (5). Recently, a different approach, namely, DGT (diffusive gradients in thin films) (6), has been used to assess the contribution of metal from the solid phase. Like a plant, DGT locally lowers the concentration of metals. It then measures the metals available from the complete soil system. The measured supply to DGT devices is controlled by a combination of the concentration in soil solution, the size of the available labile pool in the solid phase, and the kinetics of exchange between the two (7-9). Highly significant correlations have been obtained between the concentration of Cu and other trace metals in plants and the DGT measurements (10, 11). DGT measurements can be related to fundamental kinetic and capacity properties of the soil (9). This technique may therefore be able to reconcile field observations and laboratory results. This is extremely important because toxicity tests are often conducted on metal salt amended soils, and the results obtained are applied with difficulty to field situations, where metal contamination generally occurs over a long period of time and where the fixation reaction can reduce metal availability. In this study, Zn concentrations measured in the soils, soil solutions, and by DGT were compared to the Zn taken up by the plant Lepidium sativum (garden cress). The field soils received inputs of Zn from two types of galvanized structures, electrical transmission towers, and motorway crash barriers. Two sites, in the U.K. and Belgium, were studied for each type of structure, and uncontaminated soils from each location were amended with Zn at similar concentrations to those observed in the field. The DGT measurements were used to estimate the kinetics of transfer from solid phase to solution.
Materials and Methods Field Soil Samples. Soil samples (0-20 cm depth) were collected in the vicinity of two galvanized electrical transmission towers (electricity pylons) and two safety barriers (galvanized crash barriers) along major roads. Pylon 1 (P1) was located approximately 20 km North of Central London (England) on arable land currently not being used (set-aside). The pylon was unpainted and is at least 30 years old. Samples were collected from a transect 0, 1, 2, 4, and 40 m from the pylon base. Pylon tower 2 (P2) was located in Putte in Belgium on sandy arable land. Samples were collected along a transect at 0, 2, 4, 10, and 60 m from one of the pylon corners. To ensure that soil properties were as consistent as possible, roadside locations were selected in flat areas. Samples were collected along transects perpendicular to the road border. Roadside 1 (RS1) was located along the A421 road between Woburn and Bedford in the U.K. Soils were collected at 0.5, 1, and 10 m from the road border. Roadside 2 (RS2) was located along the E19 motorway in Machelen (Belgium). Soils were collected at 1, 1.5, and 8 m from the road border. At road sites, contamination by Zn and other metals may also occur due to tire debris and vehicle emissions. To assess the relative importance of Zn from galvanized crash barriers, at both roadsides, sample A was collected directly under galvanized safety barriers, whereas 10.1021/es0352597 CCC: $27.50
2004 American Chemical Society Published on Web 05/25/2004
TABLE 1. Dependence of pH and Total Zn Concentrations in Soil and in Plants Grown in Soils Collected at Different Distances from Galvanized Structures in the U.K. and Belgiuma site pylon 1 (U.K.)
pylon 2 (Belgium)
roadside 1 (U.K.)
roadside 2 (Belgium)
distanceb (m)
pH
total (mg/kg)
soln (mg/L)
CE (mg/L)
plant (mg/kg)
yield (SD) (mg)
0 1 2 4 40
7.0 7.1 7.0 6.9 6.9
450 300 240 207 91
0.03
0.19
216
0.01 0.01 0.01
0.07 0.04 0.04
184 123 100
78 (11) 149 (30) 85 (14) 84 (6) 102 (18)
0 2 4 10 60
4.9 4.9 4.9 4.9 4.7
192 147 135 108 28
4.44 2.79 1.93 2.13 d
d
0.5 (A) 1 10 no barriere
6.8 6.5 6.8 6.8
1120 161 103 165
1.1 0.05 0.01 0.01
1 (A) 1.5 8 no barriere
6.9 6.4 4.9 7.1
749 186 16 29
1.55 0.37 0.09 0.01
29.6 21.2 12.8 13.6
1108 748 553 462
49 (33) 62 (27) 69 (8) 40 (15) 18 (18)
1.84 0.20 0.04 0.09
315 110 117 107
47 (6) 202 (156) 112 (68) 88 (30)
18.6 1.48 0.53 0.07
890 165 103 126
126 (29) 90 (23) 40 (6) 60 (37)
amendingc (mg/kg)
0, 50, 200
0, 50, 100, 150, 200
100, 1100
200, 700
a Concentrations of Zn added to the control soils and the yield per plant with standard deviations are also shown. b Distance from the galvanized structure. c Zn added by amending the reference soil with a solution of ZnCl2 (mg kg-1). d Not determined as no growth in some pots. e Sample collected at the same distance from the road border as sample A but with no safety barrier present.
an additional sample D was collected at the same distance from the road as sample A but beyond the end of the safety barrier. All soils collected, with the exception of the samples furthest away from the galvanized structure (reference soil), were contaminated by Zn but to different extents (Table 1). The soil samples were sieved at 4 mm and stored at 5 °C until preparation of the pot experiment. Amended Soil Samples. The uncontaminated soils at each of the four sites were preincubated at 20 °C for 5 days prior to amending with Zn. To achieve gradients of Zn concentrations similar to those observed in the respective transects at the field sites, appropriated amounts of a solution of ZnCl2 were mixed into the uncontaminated soils. ZnCl2 was added to the uncontaminated soils from the transmission tower transects to achieve four different concentration levels and to the uncontaminated roadside soils to achieve two levels (Table 1). Amended soils were left to equilibrate for two weeks before pot experiments were started. Analyses of Soil Properties. Subsamples of moist soils were air-dried at room temperature (20 to ∼25 °C), ground, and used for chemical analyses. Total metal concentrations were determined by inductively coupled plasma atomic emission spectroscopy (ICP-AES: Fisons ARL Accuris, Ecublens, Switzerland), following aqua regia digestion (12). Total nitrogen and carbon concentrations were determined using a combustion analyzer (LECO CNS 2000, St. Joseph, MI). Soil pH was measured in a suspension of 1:2.5 soil/0.01 M CaCl2 (w/v). Pot Experiment. L. sativum was sown in pots (1 kg) filled with each soil from the transects and with the amended soils (13). Three replicates per soil were prepared and placed in a glasshouse with the following conditions: 14 h day period, 20/16 °C day/night temperature, and minimum photon flux of 250 µmol m-2 s-1. Soil moisture samplers (Rhizosphere Research Products, Wageningen, Holland) specifically designed for heavy metal research were inserted in each pot to collect soil pore water, as described by McGrath et al. (14). Soil pore waters were collected at the beginning of the experiment, and the Zn concentration was determined by ICP-AES. Five weeks after germination, plants were harvested, rinsed with distilled water, and oven-dried at 80 °C for 24 h. Dry weights of the plants were recorded, and Zn concentra-
tions in plants were determined by ICP-AES following digestion with HNO3/HClO4 (15). DGT Device and Measurements. Cylindrical DGT devices (www.dgtresearch.com) (Figure 2) were prepared using polyacramide gel with 0.2% agarose derivative cross-linker and Chelex-100 ion-exchange resin according to standard procedures (16). Gels were cast in an oven at 42 °C until they were fully set. Subsequently, they were hydrated in ultrapure water MQ (MilliQ, Millipore Ltd.) for a day. The diffusive gels were transferred into a 0.01 M NaNO3 solution, and the resin gels were kept in MQ water at 4 °C until use. The DGT devices consist of a plastic base, a layer of resin embedded in gel (resin gel), a layer of diffusive gel, a protective membrane filter, and a plastic cap that holds all the layers tightly together. The diffusive coefficient of Zn in the diffusive gel was measured using a diffusion cell with the same procedure as previously published (17). The value obtained of 5.29 × 10-6 cm2 s-1 at 20 °C agreed well with previous measurements. Soils for DGT deployment were wetted to 100% field capacity by placing 50 g of soil of known moisture content and field capacity into a small plastic beaker and adding the appropriate amount of MQ water. The soil and water were mixed thoroughly using a plastic spatula until a smooth paste was formed. The soil paste was left to equilibrate for 1 day at the room temperature (typically 19-21 °C) before DGT deployment. DGT devices were carefully placed on the soil paste with gentle pressure to ensure complete contact between the filter membrane of the device and the soil. For most soils, they were deployed for 24 h at 20 ( 1 °C. For amended soils with total concentrations exceeding 500 mg kg-1, a 2 h deployment time was used to avoid saturation of the resin gel in the DGT devices. As the DGT measurement is sensitive to temperature (∼2% per degree), substantial temperature fluctuations were avoided during deployment. Upon retrieval, the surface of the DGT devices was jetwashed with MQ water to remove soil particles. Then, the excess water was removed from the devices using clean tissue papers. To elute metals, 1 mL of HNO3 (1 M) was added to each resin gel in a microvial. Soil solutions were sampled by centrifugation (10 000 rpm, 3 min, room temperature) of a subsample of the soil paste after the DGT device was retrieved. The supernatant was filtered using 0.45 µm pore size VOL. 38, NO. 13, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3609
disposable polysulfone filter assemblies (Whatman Puradisk) and then acidified using HNO3. The concentrations of Zn in the resin gel elution samples and soil solutions were analyzed by ICP-MS with Rh as internal standard. DGT Data Interpretation. DGT directly measures the amount of metal, M, accumulated on the resin gel (6, 16).
TABLE 2. Values of Particle Concentration (Pc), Porosity (φ), Tortuosity (θ2), Effective Diffusion Coefficient of Zn in Soils (Ds), and Rdiff a Used in DIFS Model Calculations for Each Soil Type soils
M ) C(Vacid + Vgel)/fe
(1)
C is the metal concentration in elution samples measured by ICP-MS. Vacid is the volume of acid used for elution (Vacid ) 1 mL), and Vgel is the volume of resin gel (Vgel ) 0.16 mL). fe is the elution factor ( fe ) 0.8) (16). From the DGT measured M, the time averaged concentration of metal at the interface of the DGT device and the soil (CDGT) can be calculated using eq 2 (11):
CDGT ) M∆g/(DAt)
(2)
where A is the surface area of the DGT sampling window, t is the deployment time, ∆g is the total thickness of the diffusive gel layer (0.8 mm) and the filter membrane (0.14 mm), and D is the diffusion coefficient of Zn in the diffusive gel. To further interpret DGT measurements, the concentration directly measured by DGT (CDGT) can be converted to an effective concentration (CE) using eq 3 (11):
CE ) CDGT/Rdiff
(3)
CE represents the concentration of Zn that is effectively available from both soil solution and solid-phase labile pool. Rdiff is the ratio of the theoretical Zn concentration at the DGT surface for the diffusion only case (i.e., no resupply from the solid phase) to the concentration in soil solution. The actual Zn concentration at the DGT surface is lower than the value given by eq 2 because desorption processes rarely buffer completely depletion of the solution phase induced by DGT. Rdiff was calculated using the DIFS dynamic numerical model of the DGT-soil system (8). Input parameters of particle concentration (Pc), soil porosity (φ), and the diffusion coefficient of Zn in the soil (Ds) were calculated using eqs 4-6 (8, 18):
Pc ) m/V
(4)
φ ) dp/(Pc + dp)
(5)
Ds ) Do/(1 - ln φ2)
(6)
where m is the total mass of all soil particles; V is the porewater volume in a given volume of total soil; Do is the diffusion coefficient in water; and dp is the density of the soil particles, which is commonly assumed to be 2.65 g cm-3 in soils (19). As DIFS is a 1-D model, a correction factor derived from a 2-D model (8) was applied to obtain values of Rdiff for 2-D diffusion. As the DGT device continuously takes metals out of the soil system during deployment, it is inevitable that for most cases some depletion of the concentration of metal in soil solution at the interface of the DGT device and soil will occur (9). The extent of the depletion is indicated using the ratio (R) of CDGT to the independently measured initial concentration of metal in the soil solution (Csoln):
R ) CDGT/Csoln
(7)
The value of R is affected by the solid-phase labile pool size and the response time of the soil to the depletion (Tc) (8). The DIFS model uses the partition coefficient for labile species 3610
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 13, 2004
Pc (g/mL)
U.K. road (R1) 0.7 ( 0.03 Be road (R2) 1.19 ( 0.11 U.K. pylon (P1) 0.98 ( 0.01 Be pylon (P2) 1.56 ( 0.05 a
φ
θ2
Ds (cm2/s)
Rdiff
0.79 ( 0.01 0.69 ( 0.02 0.73 ( 0.03 0.63 ( 0.01
1.47 1.74 1.63 1.94
4.23 3.57 3.82 3.21
0.15 0.12 0.13 0.10
See text.
(Kdl) to calculate the labile pool size in the solid phase and quantifies the relationship between R, Kdl, and Tc. As Tc is directly related to the rate constant of the supply process from solid phase to solution (k-1), it can be obtained if R and Kdl are known.
Results and Discussion Effect of Galvanized Structures. The total concentration of Zn in soils close to the pylons was much higher than in samples collected at 40 m (U.K.) and 60 m (Belgium) distance (Table 1). The concentration declined progressively with distance away from the pylons but was still significantly elevated at 4 m (U.K.) and 10 m (Belgium). Measurements of more available fractions of metal (i.e., the concentration in soil solution and effective concentration CE) were, relative to total concentrations, much lower in the clay soils of the U.K. than the sandy soils of Belgium (Table 1). CE and soil solution Zn were especially elevated in the samples collected near the pylon in Belgium. This is probably due to the low pH (
