Evaluation of in Situ DGT Measurements for ... - ACS Publications

Jun 20, 2012 - (13) Lux, A.; Martinka, M.; Vaculi, M.; White, P. J. Root responses to cadmium in the rhizosphere: A review. J. Exp. Bot. 2011, 62 (1),...
0 downloads 0 Views 965KB Size
Download PDF
Article pubs.acs.org/est

Evaluation of in Situ DGT Measurements for Predicting the Concentration of Cd in Chinese Field-Cultivated Rice: Impact of Soil Cd:Zn Ratios Paul N. Williams,†,* Hao Zhang,† William Davison,† Shizhen Zhao,† Ying Lu,‡,* Fei Dong,‡ Lin Zhang,‡ and Qi Pan‡ †

Lancaster Environment Centre, Lancaster University, Lancaster, LA1 4YQ, England South China Agricultural University, College of Natural Resources & Environment, Guangzhou, 510642, Guangdong, People's Republic of China

‡

S Supporting Information *

ABSTRACT: DGT (diffusive gradients in thin-films) has been proposed as a tool for predicting Cd concentrations in rice grain, but there is a lack of authenticating data. To further explore the relationship between DGT measured Cd and concentrations in rice cultivated in challenging, metal degraded, field locations with different heavy metal pollutant sources, 77 paired soil and grain samples were collected in Southern China from industrial zones, a “cancer village” impacted by mining waste and an organic farm. In situ deployments of DGT in flooded paddy rice rhizospheres were compared with a laboratory DGT assay on dried and rewetted soil. Total soil concentrations were a very poor predictor of plant uptake. Laboratory and field deployed DGT assays and porewater measurements were linearly related to grain concentrations in all but the most contaminated samples where plant toxicity occurred. The laboratory DGT assay was the best predictor of grain Cd concentrations, accommodating differences in soil Cd, pollutant source, and Cd:Zn ratios. Field DGT measurements showed that Zn availability in the flooded rice rhizospheres was greatly diminished compared to that of Cd, resulting in very high Cd:Zn ratios (0.1) compared to commonly observed values (0.005). These results demonstrate the potential of the DGT technique to predict Cd concentrations in field cultivated rice and demonstrate its robustness in a range of environments. Although, field deployments provided important details about in situ element stoichiometry, due to the inherent heterogeneity of the rice rhizosphere soils, deployment of DGT in dried and homogenized soils offers the best possibility of a soil screening tool.

■

are better indicators of Cd bioavailability in paddy soil.6 When plant uptake rather than supply flux from the soil governs uptake, free ion activity can also explain metal accumulation in many plants. However, these methods are not holistic, failing to encompass important soil properties such as pH and cation exchange capacity (CEC).6,7 Models of plant uptake based on these measurements can be improved by the inclusion of secondary parameters (pH, CEC, soil organic carbon (SOC), dissolved organic carbon (DOC), clay content, Zn, and field order away from the irrigation source).6,8−10 Unlike other techniques, which are reliant on additional factors to improve predictive capabilities, a single DGT measurement automatically takes several important soil factors into account6 thus acting as a better proxy of a plant root. A particular merit is that the technique fully considers the

INTRODUCTION Food is the dominant source of Cd, a group 1 human carcinogen, for the general population1 with rice subsistence based regimes constituting the greatest exposure threat.2 Soil to plant transfer of Cd is high in rice. Compounding the problem further, the polished grain is of a poor nutritional quality with respect to Zn and Fe; enhancing gastrointestinal absorption of Cd even in marginally jejune diets.2 Natural soils are rarely enriched in Cd so it is anthropogenic sources that are the dominant contamination risk: examples of the most prolific are mining waste, sewage sludge, industrial effluent, and phosphate fertilizers.3,4 Safety thresholds for Cd in paddy soil are commonly set in accordance with concentrations determined by digesting soil samples at high temperatures in strong mineral acids.5 In broad terms, this method can differentiate between uncontaminated and seriously metal impacted soils. However, as a measurement of phytoavailability it is weak and thereby an ineffective surrogate of rice grain concentrations. Simple chemical treatments and soil solution measurements of available Cd © 2012 American Chemical Society

Received: Revised: Accepted: Published: 8009

March 27, 2012 June 20, 2012 June 20, 2012 June 20, 2012 dx.doi.org/10.1021/es301195h | Environ. Sci. Technol. 2012, 46, 8009−8016

Environmental Science & Technology

Article

resupply dynamics of metals from the solid phase to solution. Although DGT has been identified as a promising tool for assessing the safety of rice paddy soils,6 further evaluation of its capabilities are required. It may sometimes fail to accommodate all the factors that affect the uptake of metals in rice, including the impact of competitive/protective ions. Cd enters roots, opportunistically via carriers for other divalent cations, such as Zn, Cu, or Fe, because the chemical properties of those divalent cations are analogous.11 Therefore, when the activity of protecting ions in solution is high, the relationship between DGT measured Cd (CdCDGT) and plant uptake may deviate. He et al.12 demonstrated that Cd uptake by rice roots was inhibited by 30−40% when Zn was added to nutrient solutions at equal molar concentrations with Cd. Other studies have reported similar findings for other plants.13 Wide variations in Cd:Zn ratios usually reflect diverse origins of the Cd. Mine wastes and metal smelting emissions are major global sources of Cd, yet, due to the Zn content of the ore from which they are derived, the Cd:Zn ratios are typically comparable to background signatures (c. 0.005).14 Processing of Cd for industrial uses and products, such as Ni−Cd batteries, plastics, colored pigments, metal alloys, and anticorrosion agents alters the Cd:Zn ratios in their associated wastes. The consequence of this is far higher Cd:Zn ratios, and thereby a potentially greater human Cd risk.14 A further confounding issue is that redox-related processes are important modifiers of Cd and Zn mobility. Under flooded conditions, soil pH will commonly rise, favoring Cd and Zn sulphide formation, which reduces phytoavailability.15 However, Cd behavior upon reoxidation (which can be enacted by soil draining or localized aeration of the rhizosphere by rice roots) is distinct from Zn,14 either due to differential oxidation of sulphide minerals,16 or competitive sorption of Zn over Cd with precipitated Fe and manganese (Mn) (oxyhydr)oxides.17,18 The success of any single measurement approach for predicting Cd in rice depends, therefore, on it being representative of the redox conditions in the field and for variation in Zn concentrations in the soil porewater not impinging on Cd uptake by rice roots. In this investigation, rice paddies from four contrasting sites with different Cd pollutant sources, input intensities and Cd:Zn ratios were selected to test if grain Cd concentrations could be predicted reliably from soil chemical characteristics. Specifically, four questions were posed. (i) Can DGT be used to predict successfully Cd uptake in rice from challenging metal degraded field locations with different heavy metal pollutant sources? (ii) Does in situ sampling using DGT provide a more holistic measurement of Cd availability? (iii) What is the importance of Cd:Zn ratios and Cd/Zn soil concentrations as factors in predicting uptake by rice? (iv) How does the Cd:Zn stoichiometry in rice paddy soil differ between total soil concentrations and those measured by DGT?

■

that spans from Hong Kong to Guangzhou (the capital city of Guangdong). Plant and Soil Sampling. Seventy-seven late season rice plants from 27 paddy fields with associated soils were sampled as grain reached maturity (November 1−5, 2009). Individual plants were uprooted enabling approximately 1 kg wet weight of soil, encompassing rhizosphere and neighboring bulk soil to be collected. Rice was cultivated under field conditions, reflecting a range of commonly grown rice cultivars. Normal field management practices were used, e.g., submerged in the growth stage, with sampling timed just prior to fields being fully drained for harvest. Therefore, the soil environment was either water saturated or fully submerged. To account for Cd soil loading patterns due to irrigation water inputs and in-field irrigation flows,8,21 rice/soil samples were collected along a transect originating at the water entry point of each field at 10 m intervals. Four different case study sites were targeted during the sampling campaign: (i) Cd- Zn- Control. Dong Dong village (n = 29). Location = 60 km northeast of Guangzhou. Organic rice farmland in an area with little industrial activity. (ii) Cd+ Zn- Light Industry Zone. Gaoming city (n = 23). Location = 50 km southeast of Guangzhou. Urban area with extensive industrial development with small but localized zones of peri-urban agriculture. (iii) Cd+ Zn+ Mining Impacted. Shangba village. (n = 12). Location = 150 km northeast of Guangzhou and c.18 km downstream of one of the largest opencast mines in South China. From 1984 to 2001. One hundred fiftythree people in Shangba died from tumors in the alimentary tract, which accounted for fifty-one percent of the recorded deaths.22 (iv) Cd+ Zn++ Mining Impacted. Shangba village (n = 13). Paddy fields in this group were located near to the river. In addition to having elevated soil Cd concentrations, high levels of Zn and other heavy metals were a characteristic of these soils. Grain Preparation and Analysis. Grain was rinsed with MQ water, dehusked, subjected to a secondary MQ wash and oven-dried at 100 °C until a constant weight was reached,23 then samples were pulverized into a flour (MM2 ball mill, Retsch, Germany). Powdered rice, 0.2 g, was weighed into Teflon digestion bombs and steeped overnight in 2 mL of conc. HNO3 before being heated to 95 °C for 30 min (MARSExpress. CEM).24 Measuring Potential Cd Bioavailability Total Soil Cd (Cdt). 0.2 g subsamples of air-dried and 2 mm-sieved soil samples were weighed into Teflon digestion bombs and steeped overnight in 10 mL of conc. HNO3 before being heated to 175 °C (MARS-Express. CEM).25 DGT and Porewater (Laboratory Assay). This study used cylindrical DGT piston devices with an exposure window of 2.54 cm2. The most widely used agarose derived cross-linked polyacrylamide diffusive gel (0.8 mm) overlaid a 0.4 mm chelex gel.26 Air-dried and 2 mm-sieved soil samples (c. 50 g, n = 39) were wetted to 60% maximum water holding capacity (MWHC) and incubated for two days, then raised to 90−100% MWHC for 24 h, at which point the slurry was transferred into small cylindrical lidded dishes (50 × 20.3 mm).6 DGT devices were deployed for 24 h and retrieved following standard

MATERIALS AND METHODS

Sample Site Background. Guangdong province in southern China was selected as the most relevant case study areas to trial the deployments because rice in this region is one of the most threatened by Cd contamination due to its high bioavailability (a consequence of the acidic lateritic red earths that typify the agricultural soils in this part of China19) and the variety and intensity of Cd pollutant sources. Furthermore, the rice grown in this region is a crucial component of the food supply chain for one of the world’s largest urban conurbations20 8010

dx.doi.org/10.1021/es301195h | Environ. Sci. Technol. 2012, 46, 8009−8016

Environmental Science & Technology

Article

Table 1. Grain, Soil, Porewater, and DGT Characteristics

Cd (ngg−1) Zn (ngg−1) Cd (ngg−1) Zn (ngg−1) Cd:Zn CDGT (μgL−1) CDGT (μgL−1)

Cd Zn

CDGT (μgL−1) CDGT (μgL−1) Cd Pore water (μgL−1) Zn Pore water (μgL−1) Cd R Cd CE (μgL−1) Cd Zn

organic farmland

industry impacted

Cd- Zn- (Dong Dong)

Cd+ Zn- (Gaoming)

mining impacted Cd- Zn+ (Shangba)

GRAIN mean ± standard deviation 175 ± 88 519 ± 374 1070 ± 429 19400 ± 2600 19500 ± 1320 18400 ± 2210 SOIL mean ± standard deviation 0.16 ± 0.04 0.63 ± 0.65 0.85 ± 0.31 50 ± 17 46 ± 29 168 ± 36 0.004 ± 0.002 0.011 ± 0.005 0.005 ± 0.001 FIELD (in situ) mean ± standard deviation 0.2 ± 0.17 0.4 ± 0.25 1.5 ± 1.02 2.8 ± 1.76 3.6 ± 1.98 12.9 ± 8.00 LABORATORY mean ± standard deviation 0.34 ± 0.11 0.80 ± 0.66 1.80 ± 0.90 21 ± 25 21 ± 22 77 ± 35 1.7 ± 0.83 2.9 ± 1.71 5.3 ± 2.6 76 ± 82 69 ± 82 190 ± 113 0.21 ± 0.06 0.25 ± 0.06 0.34 ± 0.04 5 ± 0.4 13 ± 2 31 ± 4

methodologies.6,27 After DGT retrieval, soil slurries were transferred into 50 mL centrifuge tubes, then spun at 5000 ×g for 15 min to extract the porewaters. The resulting soil solutions were then filtered through 13 mm diameter, 0.45 μm polysulfone filters and acidified to 0.1 M HNO3 for metal analysis. DGT (in Situ Deployment). Commonly, DGT devices are placed on the soil surface.28 However, in rice paddies, surface layers can be a poor surrogate for the underlying soil due to fresh deposition of sediments and macro/micro algae and because the soil/water interface is a specific biogeochemical niche in its own right.29 To target the rooting zones of the rice plants, DGT devices were located c. 5 cm from the soil surface, at the end of burrows, formed by excavating a soil core, that extended into the rhizosphere/rooting zone. After the devices were deployed, the burrows were immediately backfilled with the excavated material. Oxygen penetration either from air or oxygenated water into the burrow was unavoidable, yet the process of installing the DGT device, which involves the unit being gently driven into the wet matrix, ensures that fresh and unexposed soil made contact with the measurement window. As the DGT device only depletes the soil solution concentrations a few millimeters adjacent to the device’s filter, the depleted zone only affects at most 1−2 cm3 of soil.28 This measurement localization helps minimize redox change artifacts incurred during sampling. After inserting the DGT devices into the soil, starting times and soil temperatures were recorded. At each point, a marker peg was used for locating the submerged DGT units. A detailed record of the diurnal temperatures was achieved with a series of data loggers (StowAway Tidbit) that were also deployed concurrently with the DGT. After c. 24 h, DGT devices were collected, jet-washed with MQ water to remove soil particles and placed in a cool box prior to transfer back to the laboratory for disassembling.30 Resin gels, once removed from the DGT device, were placed for 24 h in 1.5 mL PVC tubes containing 1 M HNO3, before dilution and analysis. Chemical Analysis. Trace Metal Characterization. ICPMS (Thermo Elemental X7) was used to determine Cd, Zn, Pb, Cu, Ni, Co, Cr, Se, and As in soils, rice grain, DGT, and

Cd+ Zn++ (Shangba) 618 ± 512 19400 ± 2900 0.77 ± 0.27 254 ± 38 0.003 ± 0.001 6.7 ± 4.0 195 ± 122 11.7 1190 62.3 8011 0.21 198

± ± ± ± ± ±

5.9 561 44 3500 0.04 28

porewater samples (see Supporting Information (SI) Tables S1−S5). Major nutrients, Ca, Fe, K, Mg, Mn, P, and Si, were measured using an ICP-OES iCAP 6000 (Thermo Elemental). Soil pH was measured both in extracted porewaters and in a soil:water ratio of 1: 2.5, after letting the solution equilibrate for sixty minutes. SOC was determined by wet digestion with K2Cr2O7/H2SO4.9 Calculating DGT Labile Cd. DGT labile metals were calculated using eq 1. C DGT = M × Δg /(D × A × t )

(1)

where M is the mass accumulated on the binding gel, Δg 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 (2.54 cm2), t is the deployment time, and D is the diffusion coefficient of the metal in the gel. D was calculated using element specific temperature corrected diffusion coefficients. Calculating Effective Cd Concentrations from DGT. The effective concentration (CE) is the hypothetical porewater concentration that would be needed to accumulate the observed amount of metal on the resin if there was only diffusional supply. The effective concentration differs from CDGT with a factor that depends on the geometry of the device, deployment time and soil tortuosity.27,31The DGT measured concentration in a soil, CDGT can be converted to CE using eq 2. C E = C DGT/R diff

(2)

where Rdiff is the ratio of CdCDGT to the soil solution concentration when supply to the DGT device is only by diffusion. It was calculated using the numerical model of the DGT-soil system 2D DIFS (DGT induced fluxes in soils).32 Ds (the diffusion coefficient in soil) was calculated using eqs 3, 4, and 5:

8011

Pc = m /V

(3)

ϕ = d p/(Pc + d p)

(4)

Ds = D0 /(1 − ln ϕ2)

(5)

dx.doi.org/10.1021/es301195h | Environ. Sci. Technol. 2012, 46, 8009−8016

Environmental Science & Technology

Article

Figure 1. Cd:Zn and soil Ni trends with increasing soil Cd across the four sample sites. (a) Cd:Zn ratios. Mean ± s.e. (b) Soil [Ni] vs [Cd]. △Organic Farmland, Dong Dong(Cd‑Zn‑), ● Light Industry Gaoming(Cd+Zn‑), □ Mining Impacted, Shangba(Cd+Zn+), and ◊ Mining Impacted, Shangba(Cd+Zn++).

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 in this model was assumed to be 2.65 g cm−3. The term D0 refers to the diffusion coefficient in water. The ratio (CdR) of the DGT measured concentration (CdCDGT) to the bulk concentration in soil solution (CdCsol) reflects the extent to which there is dynamic resupply from the soil solid phase (eq 6): R =Cd C DGT/Cd Csol

Cd

sites ranged from 5 to 6, which is within the optimum range for DGT measurements.33 Soils located closer to irrigation inlets recorded higher Cdt’s, when concentrations exceeded 0.8 μg g−1 (SI Figure S4), suggesting irrigation waters were an important Cd source in both the mining and industry impacted sites. Similar spatial trends governed by proximity to irrigation water source were also reported for mining waste contaminated paddies in Thailand34 and industry impacted sites in Taiwan.8 Linear regressions for SOC and Cdt for both the industry and mining impacted sites were highly significant (p < 0.001) (SI Figure S5). SOC was not related to distance from the irrigation inlet, suggesting that the mechanism of the SOC-Cd relationship here was not one of input, but rather of sequestration, with soils with higher SOC levels acting to immobilize Cd more effectively. This is further supported by clear associations with SOC-Cu (GaomingCd+Zn‑ p < 0.001) and SOC-Pb (GaomingCd+Zn‑ p < 0.001; ShangbaCd+Zn+, p < 0.05) (SI Figures S6−7), both of which exhibit a high affinity for binding to SOC.35 The Cd:Zn ratios from GaomingCd+Zn‑ soils were significantly higher (ANOVA p < 0.001) than from the other case study areas (Table 1, Figure 1). Despite the high loading of Cd in the soils a co-contaminant signature from Zn, Pb, or Cu was not observed (SI Table S5). It is unlikely therefore that the Cd source originated directly from mineral ores. Moreover, Ni concentrations in the GaomingCd+Zn‑ soils were characteristically distinct from the other sites, with both total Ni and Cd:Ni ratios found to be significantly different from those of Dong DongCd‑Zn‑ and both Shangba sitesCd+Zn+, Cd+Zn++ (ANOVA p < 0.001). The relationship between Cd and Ni was supported further, by plots of soil Ni concentrations against Cd content (Figure 1) (Linear regression R2= 0.97, p < 0.001). Those data suggest that the Cd source in the industrial zones is distinct from that of the mining areas and not derived from mineral ores directly. It is more likely given the close association with Cd and Ni that the main source is from wastes arising from Ni−Cd battery manufacturing/disposal. We are unaware of such clear localized Ni and Cd elevation trends, in the absence of other heavy metal contaminants, being presented previously

(6)

Cd

A R value of >0.95 demonstrates that soil buffering is sufficient to maintain porewater Cd concentrations, whereas values equal or close to Rdiff represent a diffusion only supply to the DGT. Quality Assurance. All glassware was soaked for 12 h in 10% (v/v) HNO3 and rinsed thoroughly with high purity water 18 MΩ cm water Milli-Q (MQ). Reagent blanks, duplicate samples, and reference materials were inserted with each batch of samples. The reference material recoveries (River Clay, no.921) for Cd, Zn, and Ni were 103.2 ± 1.8, 81.0 ± 7.8, 103.9 ± 7.8% (n = 6), while the average duplicate sample relative standard deviations (R.S.D’s) for the three elements were all below 10%. Cadmium certified reference material recoveries (Rice flour GBW10010) for the grain analysis, were 100 ± 1.8% (n = 4) and average duplicate sample R.S.D’s were 2.3 ± 1.8% (n = 10). On the basis of the sample weights of 0.2 g, the limits of detection for the grain Cd analysis ranged from 3 to 20 ng g−1and the limits of quantification from 6 to 32 ng g−1 (Further quality control data are provided in the SI Tables S1−3).

■

RESULTS AND DISCUSSION Soil. The total Cd concentration in soils (Cdt) differed significantly between the four sites (ANOVA p < 0.0001), with ShangbaCd+Zn+ having the highest average Cdt, at 0.85 mg kg−1 (Table 1) followed by ShangbaCd+Zn++ GaomingCd+Zn‑ and Dong DongCd‑Zn‑. Only the Dong Dong site (0.16 mg kg−1) was below the average natural background of Cd in Chinese soils (0.2 mg kg−1 5). GaomingCd+Zn‑ exceeded soil safety thresholds (0.3 mg kg−1 5) by c. 2-fold, while levels in ShangbaCd+Zn+, Cd+Zn++ were nearly three times over this. Average soil pH across the four 8012

dx.doi.org/10.1021/es301195h | Environ. Sci. Technol. 2012, 46, 8009−8016

Environmental Science & Technology

Article

Figure 2. Predicting grain Cd concentrations. (a) Comparison between grain-[Cd] and soil-Cd availability: Grain-[Cd] vs in situ-CdCDGT (light grey circles) Grain-[Cd] vs laboratory in situ-CdCDGT (dark grey circles). (b) Grain-[Cd] vs laboratory-CdCDGT (c) Grain-[Cd] vs in situ-CdCDGT. (d) Grain [Cd] vs laboratory CdPW. △Organic Farmland, Dong Dong(Cd‑Zn‑), ● Industry ZoneGaoming(Cd+Zn‑), □ Mining ImpactedShangba(Cd+Zn+), ◊ Mining Impacted Shangba(Cd+Zn++).

Chelex is 0.5 mequiv mL−1 wet resin corresponding to 15 μequiv for a 0.4 mm resin gel.39 The accumulation of eight metals (Mn, Fe, Co, Ni, Cu, Zn, Cd, and Pb) into the resin layers of the laboratory deployed DGT devices was measured. All samples were found to be within the binding capacity (maximum 3.3%). Although the sampler also collects major cations such as Na, K, Ca, and Mg, these have much lower selectivity for the Chelex resin and are normally exchanged during exposure with other divalent/trivalent trace metal ions.33 Predicting Grain Cd Concentrations DGT. Deployment on Air-Dried Rewetted Soil under Laboratory Conditions. The order of sites ranked by mean laboratory CdCDGT (ShangbaCd+Zn++ > ShangbaCd+Zn+ > GaomingCd+Zn > Dong DongCd‑Zn‑) differed slightly from the order for Cdt trends (Table 1). Cd was shown to be most DGT-available in ShangbaCd+Zn++ despite lower Cdt’s than ShangbaCd+Zn+. The range in CdCDGT between sites was considerable; average values for ShangbaCd+Zn++ were 35-fold higher than their equivalents from Dong DongCd‑Zn‑. When only Dong DongCd‑Zn‑, GaomingCd+Zn‑, and ShangbaCd+Zn+ were considered, the regression explained 70% of the variation. A best subset analysis featuring 25 soil multielement parameters, including both laboratory and in situ-CDGT measurements was run. The findings revealed that the correlation between loggrain-Cd and loglaboratory-CdCDGT concentration was the strongest. When individual sites were considered, the regression between loggrain-Cd and loglaboratory-CdCDGT was consistently higher than for any of the other bioavailability measurements. Using CdCE instead of CdCDGT had a negligible impact on the model R2 values.

in paddy soils and suggest it is a useful contamination signature for tracking manufacturing pollutant sources. Grain. The Cd in the grain of all seventy-seven samples was above the limit of detection (LOD) (SI Table S2). After adjusting the measured wholegrain Cd concentrations to reflect losses associated with milling, using a conversion factor of 0.884, all four case study areas contained grain that exceeded the Chinese maximum contaminant level (MCL).37 Around 75% of the samples from both GaomingCd+Zn‑ and ShangbaCd+Zn+, Cd+Zn++ exceeded 200 ng g−1,36 whereas only 30% of the grain from Dong DongCd‑Zn‑ breached this. None of the rice from Dong DongCd‑Zn‑, however, was >400 ng g−1 which is the World Health Organisation’s (WHO) safety threshold.37 In contrast, 60% and 30% of samples from Gaoming Cd+Zn‑ and ShangbaCd+Zn+, Cd+Zn++ exceed this level, respectively. The most elevated Cd grain came from ShangbaCd+Zn+, Cd+Zn++ (1471 ng g−1), with nine samples containing over 1000 ng Cd g−1. However, there was no significance difference in grain Cd concentration between ShangbaCd+Zn+, Cd+Zn++ and GaomingCd+Zn‑ (Mann−Whitney p = 0.073). Evaluating Impacts of Excess Ion Loading into DGT. Saturation of the binding resin layer by high concentrations of Cd and other competing ions is a consideration for deployments in metal-enriched matrixes.38 To evaluate possible impacts of excess ion loading on the chelating resins three validation criteria were assessed: (i) DGT measurements were plotted against companion porewater metal concentrations. A disconnect in the correlation of these two parameters at high metal concentrations demonstrated by a plateauing of the DGT measurement can indicate saturation. In this study, measured Cd concentrations in the laboratory porewater (CdPW) followed similar trends to that of the laboratory-CdCDGT (Linear regression R2 = 0.95, p < 0.001) and in situ-CdCDGT (Linear regression R2 = 0.60, p < 0.001). (ii) Chelex 100 resin gels have a capacity of 1260 μg Cd disk−1 based on a gel formulation of 4 g of resin to 10 mL of gel solution.33 When Cd uptake per DGT device was calculated as a percentage of total capacity, all samples in both the laboratory (maximum 0.02%) and in situ (maximum 0.01%) assays were well within the saturation threshold. (iii) To consider the impact of competing ions, multimetal accumulation was assessed. The binding capacity of

log

Grain[Cd] = 2.738 + (0.917 ×log laboratory CdC DGT) (7)

In Situ DGT Measurements. The principal difference between conditions in soils subject to prolonged flooding and rewetted soils after air-drying would be the lower pH of the latter soils.42 The observation of this study, of a higher Cd uptake flux in laboratory DGT assays compared with field deployed devices, was consistent with a lower pH causing less adsorption of Cd to clay, humic polymer, or oxide phases.42 However, differences in the Cd availability did not cause a 8013

dx.doi.org/10.1021/es301195h | Environ. Sci. Technol. 2012, 46, 8009−8016

Environmental Science & Technology

Article

disconnect between the two measurements, as evidenced by their highly significant linear relationship (R2 = 0.6; p < 0.001). Regression of loggrain-Cd against login situ-CdCDGT was also highly significant (R2 = 0.63, p < 0.001), but was not as strong as the laboratory assay. Increasing the number of devices deployed in the rooting system of each plant may improve the predictive capacity of the in situ measurements. Alternatively, readily available DGT sediment probes offer another sampling option with larger exposure windows of 1.8 × 15 cm. Porewater Sampling on Air-Dried Rewetted Soil under Laboratory Conditions. Porewater measurements proved successful in predicting grain Cd concentrations. Similar to the DGT models, ShangbaCd+Zn++ was identified as an outlier group and removed from the analysis. Regression of loggrain-Cd against log CdPW was highly significant (R2 = 0.62, p < 0.001) but not as strong as for the CdCDGT measured in the laboratory (R2 = 0.70) (Figure 2). One explanation for the DGT technique providing better predictions of Cd uptake in rice is that the measurement encompasses resupply of metal from the solid phase. Average CdR for the four sites ranged from 0.21 to 0.34, indicating that there was significant resupply of Cd from the solid phase for all samples, but the rate of supply was low and insufficient to sustain fully porewater concentrations. The soil metal reservoirs appeared to be releasing Cd even in the highly contaminated samples. Total Soil Cd. The coefficient of determination in the regression of loggrain-Cd against logCdt concentrations was low (R2 = 0.48). When the four sites were considered separately, only GaomingCd+Zn was found to have a significant loggrain−logCdt relationship (R2 = 0.66, p < 0.001). Cd:Zn Ratios. In the control and mine impacted sites (Dong DongCd‑Zn‑, ShangbaCd+Zn+, Cd+Zn++) total Zn concentrations were c. 250 times higher than equivalent Cd concentrations. However, in the light industry zone (GaomingCd+Zn) there was only c. 50 times more Zn (Figure 3). This difference in the Cd:Zn ratio in total soil concentrations between the sample sites (Table 1; Figures 1 and 3), did not appear to affect correlations between porewater and DGT measured Cd and grain Cd concentrations. Furthermore, predictions of grain Cd concentration (eqs 6 and 7) were unaffected by either sample site or Cd:Zn ratio. This is understandable when in situ-DGT measurements of Cd and Zn are considered (Figure 3). When the metals were measured using rhizosphere located DGT under field redox conditions, the Cd:Zn ratios between the different sites were unified and the excess of available Zn compared to Cd was greatly diminished. As available Zn was only around 10-fold higher than available Cd, Zn would offer less protection against plant Cd uptake. This is important as increasing the phytoavailability of Zn in paddy soils could mitigate against Cd uptake in rice.18 Laboratory assays, both porewater and DGT, also revealed high Cd:Zn ratios of c. 0.1. Although the regressions were not as strong as for the in situ measurements, they showed a similar indication of data. There are a number of other elements besides Zn, that have been demonstrated to reduce Cd uptake in plants, namely Cu, Fe, Mn, and Ca.13 Comparisons (Mann−Whitney) of the Cd:metal ratios for total soil concentrations for Fe (p = 0.710), Mn (p = 0.066) and Ca (p = 0.563) revealed no significant difference between the industrial and mining impacted sites. Cd: Cu ratios were significantly different (Mann−Whitney p < 0.001), with the ratio being higher in GaomingCd+Zn‑. However, when the in situ DGT measured Cd and Cu were considered, then just like the Cd:Zn ratio, the sample sites were unified

Figure 3. Relationship between (a) total soil Cd and Zn and (b) in situ DGT measurements. (a) Cdt vs Znt . High Cd:Zn (Industry. GaomingCd+Zn‑) ● ( − − − R2 0.90). Low Cd:Zn (Cd-Zn-, Cd+Zn+, Cd+Zn++) ○ ( R2 0.73). (b) In situ CdCDGT vs In situ ZnCDGT ( − − − R2 0.79).

with there being proportionally more Cd released compared to Cu, resulting in a higher Cd:Cu ratio than observed from total soil metal measurements (SI Figure S9). Implications. Grain surveys predict that up to ten percent of the rice supply chain in China could exceed the national MCL for Cd, with the problem being worse in southern compared with northern regions; an issue that has been well publicised (see SI for a summary of media reports). The ability to assess the degree of Cd contamination in rice paddy soil using a single predictor measurement would be an advantage for a rapid and straightforward soil screening protocol. Furthermore a test that functions effectively on air-dried soils confers benefits for sample storage, subsampling and chemical analysis.9 Current soil safety testing in Guangdong favors total digestible Cd, yet we demonstrate that it is ineffective in explaining Cd uptake in rice. The alternative approach offered by the DGT technique, as presented by Tian et al.6 was further demonstrated here to provide a strong relationship with grain concentration, and has the potential to be used to set soil screening thresholds, pending further optimization and evaluation. The findings from this study demonstrate the robust functioning of the DGT-rice assays, even when deployed in severely metal degraded soils with different sources of Cd. Furthermore, it gave a better prediction of the grain concentration in rice than porewater measurements. We recognize that the assessment of Cd availability in Guangdong paddy soils from centrifuged porewater measurements collected from rewetted and conditioned soils might have some practical 8014

dx.doi.org/10.1021/es301195h | Environ. Sci. Technol. 2012, 46, 8009−8016

Environmental Science & Technology

Article

(3) McLaughlin, M. J.; Parker, D. R.; Clarke, J. M. Metals and micronutrientsFood safety issues. Field Crops Res. 1999, 60 (1−2), 143−163. (4) Williams, P. N.; Lei, M.; Sun, G. X.; Huang, Q.; Lu, Y.; Deacon, C.; Meharg, A. A.; Zhu, Y. -G. Occurrence and partitioning of cadmium, arsenic and lead in mine impacted paddy rice: Hunan, China. Environ. Sci. Technol. 2009, 43 (3), 637−642. (5) Soil Environmental Quality Standards. GB 15618−1995. P. R. China. http://english.mep.gov.cn/pv_obj_cache/pv_obj_id_ 10869096178A5ECCBC09D11A006442ECE35C0C00/filename/ W020070313485587994018.pdf (6) Tian, Y.; Wang, X. -R.; Luo, J.; Yu, H. -X.; Zhang, H. Evaluation of holistic approaches to predicting the concentrations of metals in field-cultivated rice. Environ. Sci. Technol. 2008, 42 (20), 7649−7654. (7) Zhao, F. J.; Rooney, C. P.; Zhang, H.; McGrath, S. P. Comparison of soil solution speciation and diffusive gradients in thin-films measurement as an indicator of copper bioavailability to plants. Environ. Toxicol. Chem. 2006, 25 (3), 733−742. (8) Romkens, P. F. A. M.; Guo, H. Y.; Chu, C. L.; Liu, T. S; Chaing, C. F.; Koopman, G. F. Prediction of Cadmium uptake by brown rice and derivation of soil-plant transfer models to improve soil protection guidelines. Environ. Pollut. 2009, 157 (8−9), 2435−2444. (9) Brus, D. J.; Li, Z. B.; Song, J.; Koopman, G. F.; Temminghoff, E. J. M.; Yin, X. B.; Yao, C. X.; Zhang, H. B.; Luo, Y. M.; Japenga, J. Predictions of spatially averaged Cadmium contents in rice grains in the Fuyang Valley, P. R. China. J. Environ. Qual. 2009, 38 (3), 1126− 1136. (10) Simmons, R. W.; Noble, A. D.; Pongsakul, P.; Sukreeyapongse, O.; Chinabut, N. Cadmium-hazard mapping using a general linear regression model (Irr-Cad) for rapid risk assessment. Environ. Geochem. Health. 2009, 31 (1), 71−79. (11) Welch, R. M.; Norvell, W. A. Mechanisms of cadmium uptake, translocation and deposition in plants. In Cadmium in soils and plants; McLaughlin, M. J., Singh, B. R., Eds.; Kluwer Academic: Dordrecht, 1999; pp 125−150. (12) He, J. Y.; Ren, Y. F.; Wang, F. J.; Pan, X. B.; Zhu, C.; Jiang, D. A. Characterization of cadmium uptake and translocation in a cadmiumsensitive mutant of rice (Oryza sativa L. ssp. Japonica). Arch. Environ. Contam. Toxicol. 2009, 57 (2), 299−306. (13) Lux, A.; Martinka, M.; Vaculi, M.; White, P. J. Root responses to cadmium in the rhizosphere: A review. J. Exp. Bot. 2011, 62 (1), 21− 37. (14) Chaney, R. L. Cadmium and zinc. In Trace Elements in Soils; Hooda, P. S., Ed.; Antony Rowe Ltd.: Wiltshire, UK, 2010; p 596. (15) Khaokaew, S.; Landrott, G.; Chaney, R. L.; Pandya, K.; Sparks, D. Speciation and release kinetics of zinc in contaminated paddy soils. Environ. Sci. Technol. 2012, 46 (7), 3957−3963. (16) Chaney, R. L.; Ryan, J. A.; Li, Y. −M.; Welch, R. M.; Reeves, P. G.; Brown, S. L. Phyto-availability and bio-availability in risk assessment for cadmium in agricultural environments. Source of Cadmium in the Environment; OECD: Paris, France, 1996; p 49−78. (17) Davranche, M.; Bollinger, J. -C. Release of metals from iron oxyhydroxides under reductive conditions: Effect of metal/solid interactions. J. Colloid Interface Sci. 2000, 232, 165−173. (18) De Livera, J.; McLaughlin, M. J.; Hettiarachchi, G. M.; Kirby, J. K.; Beak, D. G. Cadmium solubility in paddy soils: Effects of soil oxidation, metal sufides and competitive ions. Sci. Total Environ. 2011, 409 (8), 1489−1497. (19) Hseug, Y.; Li, C. Soils of China; Science Press: Beijing, China, 1990. (20) The principal agglomerations of the world; http://www. citypopulation.de/world/Agglomerations.html. (accessed Jan 2011) (21) Simmons, R. W.; Pongsakul, P.; Saiyasitpanich, D.; Klinphoklap, S. Elevated levels of cadmium and zinc in paddy soils and elevated levels of cadmium in rice grain downstream of a zinc mineralized area in Thailand: Implications for public health. Environ. Geochem. Health 2005, 27 (5−6), 501−511.

application. However, a degree of caution would be needed when using porewater as criteria for developing soil screenings thresholds. The findings from this work demonstrate that even in metal enriched sites, the contribution of the solid phase to potential bioavailability is significant. Given the large potential supply reservoir of clayed soils and that the majority of paddy soil would not be as characteristically acidic as those in this study, the role of soils in buffering Cd concentrations in porewater is likely to be even greater than found here. It would be necessary to accounting for this solid phase resupply in any bioavailability threshold. This study indicates that further focused investigations of the relationship between DGT measured Cd and grain concentrations will increase understanding. The effectiveness of the laboratory method to predict final grain concentrations from soils collected prior to planting and inundation needs confirmation.6 There is merit in testing DGT in alkaline paddy soils with elevated Cd.21 Furthermore, the use of DGT devices with values of diffusive gel layer thicknesses optimized specifically to the uptake flux of rice would provide an optimized tool, which, allied to interpretation using the DIFS dynamic model, would allow investigation of the influence of solid phase reservoir size and the kinetics of (de)sorption on Cd uptake in rice .43 Perhaps unavoidably, the relationship between CdCDGT and grain Cd concentration can falter when heavy metal exposure induces severe plant toxicity. However, screening with DGT would still ensure that the soil is not classified as agronomically suited for rice cultivation. Due to challenges with soil inhomogeneity, in situ deployments did not improve the relationship between CdCDGT and grain. However, deployments of DGT in the field were useful in demonstrating that the DGT laboratory assay reflected well the Cd and Zn mobilization trends in situ.

■

ASSOCIATED CONTENT

S Supporting Information *

Eight additional tables and nine figures as noted in the text, including additional quality control data. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Tel: +44 (0)1524 510422; fax: +44 (0)1524 510217; e-mail: [email protected] (P.N.W.). Tel: +86 20 85285853; fax: +86 20 85280292; e-mail: [email protected] (Y.L.). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We gratefully acknowledge the UK-China Science Bridge project funded by the Research Councils UK (EP/G042683/ 1).

■

REFERENCES

(1) World Health Organisation International Agency for Research on Cancer (IARC). IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. Beryllium, Cadmium, Mercury and Exposures in the Glass Manufacturing Industry; IARC Press: Lyon, 1993; Vol. 58. http:// monographs.iarc.fr/ENG/Monographs/vol58/index.php. (2) Reeves, P. G.; Chaney, R. L. Bioavailability as an issue in risk assessment and management of food cadmium: A review. Sci. Total Environ. 2008, 398 (1−3), 13−19. 8015

dx.doi.org/10.1021/es301195h | Environ. Sci. Technol. 2012, 46, 8009−8016

Environmental Science & Technology

Article

(41) Williams, P. N.; Islam, S.; Islam, R.; Jahiruddin, M.; Adomako, E.; Soliaman, A. R. M.; Rahman, G. K. M. M.; Lu, Y.; Deacon, C.; Zhu, Y.-G.; Meharg, A. A. Arsenic limits trace mineral nutrition (selenium, zinc, and nickel) in Bangladesh Rice Grain. Environ. Sci. Technol. 2009, 43 (21), 8430−8436. (42) Simmon, R. W.; Noble, A. D.; Pongsakul, P.; Sukreeyapongse, O.; Chinabut, N. Analysis of field-moist Cd contaminated paddy soils during rice grain fill allows reliable prediction of grain Cd levels. Plant Soil. 2008, 302 (1−2), 125−137. (43) Lehto, N. J.; Davison, W.; Zhang, H.; Tych, W. Theoretical comparison of how soil processes affect uptake of metals by diffusive gradients in thinfilms and plants. J. Environ. Qual. 2006, 35 (5), 1903− 1913.

(22) Liu, Y. S.; Gao, Y.; Wang, K. W.; Mai, X. H.; Chen, G. D.; Xu, T. W. Etiologic study on alimentary tract malignant tumours in villages of high occurrence. China Tropic. Med. 2005, 5 (5), 1139−1141. (23) Williams, P. N.; Price, A. H.; Raab, A.; Hossain, S. A.; Feldmann, J.; Meharg, A. A. Variation in arsenic speciation and concentration in paddy rice related to dietary exposure. Environ. Sci. Technol. 2005, 39 (15), 5531−5540. (24) Williams, P. N.; Villada, A.; Raab, A; Figuerola, J.; Green, A. J.; Feldmann, J.; Meharg, A. A. Greatly enhanced arsenic shoot assimilation in rice leads to elevated grain export; comparison with wheat & barley. Environ. Sci. Technol. 2007, 41 (19), 6854−6859. (25) Wastes-Hazardous Waste-Test methods; http://www.epa.gov/ osw/hazard/testmethods/sw846/online/3_series.htm (accessed Jan 2011) (26) Warnken, K. W.; Davison, W.; Zhang, H.; Galceran, J.; Puy, J. In situ measurements of metal complex exchange kinetics in freshwater. Environ. Sci. Technol. 2007, 41 (9), 3179−3185. (27) Zhang, H.; Zhao, F. J.; Sun, B.; Davison, W.; McGrath, S. P. A new method to measure effective soil solution concentration predicts copper availability to plants. Environ. Sci. Technol. 2001, 35 (12), 2602−2607. (28) Ernstberger, H.; Davison, W.; Zhang, H.; Tye, A.; Young, S. Measurement and dynamic modeling of trace metal mobilization in soils using DGT and DIFS. Environ. Sci. Technol. 2002, 36 (3), 349− 354. (29) Elberling, B.; Askaer, L.; Jorgensen, C. J.; Joensen, H. P.; Kuhl, M.; Glud, R. N.; Lauritsen, F. R. Linking soil O(2), CO(2), and CH(4) concentrations in a wetland soil: Implications for CO(2) and CH(4) fluxes. Environ. Sci. Technol. 2011, 45 (8), 3393−3399. (30) Nolan, A. L.; Zhang, H.; McLaughlin, M. 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. J. Environ. Qual. 2005, 34 (2), 496− 507. (31) Degryse, F.; Smolders, E.; Zhang, H.; Davison, W. Predicting availability of mineral elements to plants with the DGT technique: A review of experimental data and interpretation by modelling. Environ. Chem. 2009, 6 (3), 198−218. (32) Sochaczewski, L.; Tych, W.; Davison, B.; Zhang, H. 2D DGT induced fluxes in sediments and soils (2D DIFS). Environ. Modell. Softw. 2007, 22 (1), 14−23. (33) Zhang, H.; Davison, W. Performance characteristics of diffusion gradients in thin films for the in situ measurement of trace metals in aqueous solution. Anal. Chem. 1995, 67 (19), 3391−3400. (34) Simmons, R. W.; Noble, A. D.; Pongsakul, P.; Sukreeyapongse, O.; Chinabut, N. Cadmium-hazard mapping using a general linear regression model (Irr-Cad) for rapid risk assessment. Environ. Geochem. Health. 2009, 31 (1), 71−79. (35) Weng, L.; Temminghoff, E. J. M.; Van Riemsdijk, W. H. Contribution of individual sorbents to the control of heavy metal activity in sandy soil. Environ. Sci. Technol. 2001, 35 (22), 4436−4443. (36) Maximum Levels of Contaminants in Food, GB 2762−2005; Zhejiang Entry-Exit Inspection and Quarantine Bureau; http://www. ziq.gov.cn/portal/webfiles/web/file/12275082059288029.pdf (37) Codex General Standards for contaminants and toxins in food and feed. Codex Stan 193−1995. http://www.codexalimentarius.org/ download/standards/17/CXS_193e.pdf (38) Conesa, H. M.; Schulin, R.; Nowack, B. Suitability of using diffusive gradients in thin films (DGT) to study metal bioavailability in mine tailings: possibilities and constraints. Environ Sci. Pollut. Res. 2010, 17 (3), 657−664. (39) Garmo, O. A.; Royset, O.; Steinnes, E.; Flaten, T. P. Performance study of diffusive gradients in thin films for 55 elements. Anal. Chem. 2003, 75 (14), 3573−3580. (40) Almas, A. R.; Lombnaes, P.; Sogn, T. A.; Mulder, J. Speciation of Cd and Zn in contaminated soils assessed by DGT-DIFS, and WHAM/Model VI in relation to uptake by spinach and ryegrass. Chemosphere 2006, 62 (10), 1647−1655. 8016

dx.doi.org/10.1021/es301195h | Environ. Sci. Technol. 2012, 46, 8009−8016