Environ. Sci. Technol. 2001, 35, 3213-3218
Determination of Perchlorate in Tobacco Plants and Tobacco Products J . J A C K S O N E L L I N G T O N , * ,† N. LEE WOLFE,† A. WAYNE GARRISON,† JOHN J. EVANS,† JIMMY K. AVANTS,† AND QUINCY TENG‡ United States Environmental Protection Agency, National Exposure Research Laboratory, Ecosystems Research Division, 960 College Station Road, Athens, Georgia 30605, and Chemistry Department, University of Georgia, Athens, Georgia 30602
Previous field and laboratory studies with vascular plants have shown that perchlorate is transported from perchlorate fortified soils and is accumulated in the plant tissues and organs. This paper contains results of preliminary investigations on the occurrence of perchlorate in tobacco plants grown in soils amended with a fertilizer whose nitrogen content is derived from naturally occurring sodium nitrate (Chile saltpeter). Ion chromatography (IC) and capillary electrophoresis (CE) were used for quantitative analysis, while nuclear magnetic resonance (NMR) spectroscopy was used for qualitative analysis of perchlorate. Results show that perchlorate is accumulated by tobacco plants into the leaves from soils amended with fertilizers that contain perchlorate. Also, perchlorate can persist over an extended period of time and under a variety of industrial processes as shown by its presence in off-theshelf tobacco products including cigarettes, cigars, and pouch and plug chewing tobaccos in concentrations ranging from nd to 60.4 ( 0.8 mg/kg on a wet weight basis.
Introduction Perchlorate (ClO4-) is receiving increased attention as an inorganic contaminant in soils, groundwaters, surface waters, and irrigation waters used for crop production. The United States Environmental Protection Agency (EPA) is currently assessing the risks posed by perchlorate in the environment and has placed perchlorate on the Contaminant Candidate List (1) and the Unregulated Contaminant Monitoring Rule (2). Perchlorate has been extensively used as an oxidant in the propulsion systems of missiles and in fireworks (3). It is generally known that perchlorate is found in fertilizers whose source of nitrate is caliche ores that are rich in sodium nitrate. The major market for the nitrate fertilizer produced from the ore (N-P-K ratio: 16-0-0) is for use on tobacco, citrus fruit, cotton, and some vegetable crops for which the low chloride content of the fertilizer is beneficial (4). Harvey et al. (5) reported 885 mg/kg of perchlorate in kelp harvested from the Atlantic Ocean, suggesting a high degree of accumulation. They also reported perchlorate in fish meal and in bone meal that suggests dietary uptake by * Corresponding author phone: (706)355-8204; fax: (706)355-8202; e-mail:
[email protected]. † United States Environmental Protection Agency. ‡ University of Georgia. 10.1021/es0106321 CCC: $20.00 Published on Web 07/06/2001
2001 American Chemical Society
fish and livestock. In view of the increasing reports of perchlorate occurrence in the environment, there is a need for information on fate and transport of perchlorate in soils and ground/surface waters. Nzengung et al. (6) reported the remediation of perchlorate-contaminated water using willow trees. Two other recent reports on phytoremediation have addressed the uptake and persistence of perchlorate in 13 vascular plant species (7, 8). Perchlorate was depleted from the aqueous solution for all but two species. Other studies with lettuce have shown that perchlorate is accumulated through the growth cycle (9, 10). Questions remain, however, as to the generality of accumulation processes. Some plants accumulate perchlorate, while others mediate its transformation to products with fewer oxygen atoms, even to the chloride ion (8). To date, there are no apparent correlations between perchlorate reduction products (ClO3-, ClO2-, ClO-, and Cl-) and soil/plant chemistry. In this study, perchlorate in the aqueous extracts of green and flue-cured tobacco leaf that was grown in soil that was amended with perchlorate containing fertilizers was determined quantitatively by ion chromatography (IC) and capillary electrophoresis (CE). Nuclear magnetic resonance spectroscopy (NMR) was used for spectral identification of perchlorate in the aqueous extracts. Using the same methodology, the levels of perchlorate in 10 randomly selected off-the-shelf tobacco products, including seven chewing tobaccos, two brands of cigarettes, and one cigar, were determined.
Experimental Section Perchlorate analysis by ion chromatography and some of the sample preparation procedures have been reported in more detail elsewhere (11). A. Sample Preparation. Tobacco Plants and Products. Green (uncured) tobacco leaves from the upper stalk positions of Nicotiana tabacum var. K326 were obtained at the Coastal Plain Experiment Station (CPES), Tifton, GA in late July of 1999. The chewing tobaccos, cigar, and cigarettes were purchased at retail stores in Athens, GA, in NovemberDecember 1999 and represented products from seven tobacco companies. The green tobacco leaves were not washed before freeze-drying and separation into lamina and midrib. The tobacco products were also freeze-dried prior to extraction. All freeze-dried material was ground through a 30-mesh screen in a thoroughly cleaned Wiley mill and stored in airtight containers at 3 °C until analyzed. The flue-cured sample was also from the 1999 CPES crop. A composite sample of the flue-cured leaves was obtained according to a standard method (12). The leaf midribs were removed, and the lamina was ground through a 30-mesh screen. Duplicate 600 mg ground samples were weighed and placed in 45-mL screw top centrifuge tubes. Thirty milliliters of 18 megohm-cm water was added to each tube. The tubes were tightly capped and placed in a boiling bath for 0.5 h both to precipitate protein and to saturate the dried material. Then, the tubes were gently shaken several times during a total extraction time of 20 h at 3 °C; they were then centrifuged at 20 000g for 30 min. The supernatants were gently poured through a layer of Kimwipes ,and the supernatant was again centrifuged at 20 000g for 30 min. The pellets were discarded, and the supernatants were filtered through a 0.2 µm Pall Gelman Acrodisc ion membrane syringe filter (Fisher Scientific, Fairlawn, NJ, Gelman part #4483). These solutions were designated “IC extracts”. Prior to analysis by IC, 1 mL of each “IC extracts” was allowed to stand over 500 mg of VOL. 35, NO. 15, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3213
water washed (11) DD-6 alumina (Alcoa, Port Allen, LA) for 18 h, diluted 1/10, and filtered through a second 0.2 µm Acrodisc filter and an OnGuard RP cartridge (Dionex, Sunnyvale, CA, part # 39595). The first 0.75 mL from the OnGuard RP cartridge was discarded, and four 2-mL aliquots were collected for IC analysis. For CE analysis for perchlorate, the “IC extracts” of most of the tobacco samples were not concentrated enough. It was necessary to extract 600 mg of tobacco leaf or product with only 6 mL of water rather than the 30 mL used for IC. Subsequent treatment was the same as for preparation of the “IC extracts” except for a final dilution of 1/3 instead of 1/10 with deionized water. A few samples that contained lower levels of perchlorate and/or very high concentrations of interfering anions required special cleanup procedures. For these, the “IC extracts” were diluted 1/3, allowed to stand over DD-6 alumina, and filtered as described above for IC analysis. (They were not passed through the OnGuard RP cartridge.) A Maxi-Clean IC/Ag cartridge (Alltech, Deerfield, IL, part # 30266) was prepared by passage of 10 mL of deionized water. Without allowing the cartridge to become dry, 1 mL of the filtered extract was passed through this cartridge at a rate of 1 mL/min or less. The first approximately 0.1 mL of this eluent was discarded, and the remainder was then passed at 1 mL/min through a Maxi-Clean IC/H cartridge (Alltech, part #30264), which had been prepared by passage of 10 mL of deionized water. The first approximately 0.1 mL of this final eluent was discarded, and the remainder was used for CE analysis. For NMR analysis, the final extract volume of several of the samples was increased by extracting 1.8 g samples with 18 mL of deionized water. The only extract cleanup before analysis by NMR was centrifugation followed by filtration through 0.2 µm filters. Fertilizer Samples. Bags that remained from the 1999 crop application were opened, and approximately 500-g samples removed from each. Sodium nitrate (purified Chile saltpeter) was the sole source of nitrogen in the 16-0-0 fertilizer (Bulldog Soda, 50 lb bag). The source of nitrogen in the 6-6-18 fertilizer (Super Rainbow, Tobacco, 50 lb bag) was ammonium nitrate and potassium nitrate. The low chloride content 6-6-18 fertilizer was formulated for CPES by a local supplier. A 250-g sample of each fertilizer was ground through a 30-mesh screen in a thoroughly cleaned Wiley mill and mixed thoroughly. A 5-g sample of the ground material was mixed with 20 mL of 18 megohm-cm water in a 45-mL centrifuge tube. The tubes were shaken gently several times during an extraction time of 20 h at 3 °C. For analysis by IC, the centrifugation parameters and cleanup were the same as used for the tobacco plants and products except that boiling for protein precipitation was not necessary. For CE analysis, the same extracts prepared for IC were cleaned up by the method described for tobacco plants and products for samples with high concentrations of interfering anions. Soil Samples. In early December 1999, eight approximately equal random samples of the Tifton series loamy sandy soil to a depth of 10 cm were collected from the plot where the tobacco plants were grown in the summer of 1999. Care was taken to exclude plant material. The eight samples were combined and mixed, and a composite sample was taken for analysis of perchlorate. The soil was passed through a 30mesh screen, and duplicate 5-g samples were added to 10 mL of 18 megohm-cm water in a 45-mL centrifuge tube. The tubes were shaken gently several times during an extraction time of 20 h at 3 °C. The samples were centrifuged, and the supernatants were filtered through 0.2 µm filters and analyzed by IC and CE without further clean up. B. Instrumentation. Ion Chromatography (IC). Ion chromatography was performed on a Dionex DX-500 system (Sunnyvale, CA). The chromatograph was equipped with a 3214
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 15, 2001
GP50 gradient pump, ED40 electrochemical detector, LC30 chromatography oven, and AS3500 autosampler. The ED40 was equipped with a conductivity cell and a DS3 detection stabilizer maintained at 30 °C within the LC30 oven. The conductivity cell was mounted at the end of the ASRS ULTRA (2 mm) self-regenerating suppressor that was operated at 100 mA and allowed detection of anions in the 0-10 µS/cm conductivity range. Anions were separated on an IonPac AS16 analytical microbore separation column (2 × 250 mm) in tandem with an IonPac AG16 guard column (2 × 50 mm). Samples and standards were run in the isocratic mode (0.38 mL/min) using 50 mM NaOH as eluent. In all cases, a 1000 µL injection loop was used. Elution time of perchlorate on a new column varied from 11.4 to 11.7 min. A 1000 mg/L stock solution of perchlorate (anion) was prepared by weighing 1.231 g of sodium perchlorate into a 1 L volumetric flask and bringing to volume with 18 megohm-cm water. Dilutions of the stock solution were made to span the range 5-200 µg/L. Duplicate injections of these standards yielded calibration of the IC. A typical equation for the regression line was y ) 4909 ( 55 (x) - 3953 ( 5543, and the correlation coefficient for the regression line was greater than 0.9999. Capillary Electrophoresis (CE). Capillary electrophoresis was performed on a Beckman P/ACE 5000 with Beckman System Gold software. Conditions were as follows: columns fused silica, 57 cm total length, 50 cm to the detector, 300 µm o.d., and 75 µm i.d.; column temperatures23 °C; voltages 20 kV with the detector at the anode; detectorsUV/vis, 214 nm; run buffers40 mM phosphate, pH 7.0, containing 2.24 mM pyromellitic acid (Sigma, # 8544) and 0.5 mM tetradecyl trimethylammonium bromide (TTAB, Sigma, #T-4762); sample injections5 s hydrodynamic. The sample run time was 6 min. After each sample analysis, the column was washed with deionized water for 1 min, then with 0.1 M NaOH for 4 min, and then with water for 4 min more; it was washed with buffer for 4 min before the next sample was injected. The 1000 mg/L stock solution was prepared the same as for IC. Aqueous (deionized water) solutions of 0.560 mg/L of perchlorate were used to prepare a standard curve. Each solution was analyzed twice, and the resulting absorbance values were averaged; the correlation coefficient for regression of the averaged values was 0.9973. Nuclear Magnetic Resonance (NMR). All 35Cl NMR data were acquired at 20° C on a Bruker AMX400 NMR spectrometer with a resonance frequency of 39.25 MHz for 35Cl. The aqueous extracts were adjusted to 10% v/v D2O in a 10-mL sample probe by adding 9 mL of sample to 1 mL of D2O. The 35Cl chemical shift (1004.5 ppm) was referenced to saturated KCl in the same solvent blend at 20 °C as 0.0 ppm. The data were collected using 10 mm probes with 16K scans, a spectral window of 20 kHz, acquisition time of 300 ms, relaxation time of 2 s, and 90° pulse. Typical NMR experiment times were about 12 h. Data were Fourier transformed with 5 Hz line broadening. After baseline correction was applied, signal-to-noise (S/N) was calculated using Bruker software. The relationship of concentration to S/N was established using standard sodium perchlorate samples with concentrations of 25, 20, 8, and 5 mg/L. The limit of detection was estimated to be 3 mg/L in the aqueous extract and was based on a S/N of 5 for the 5 ppm standard. The method detection limit for samples on a dry weight basis was calculated to be 30 mg/kg.
Results and Discussion In response to the need for a standard method of analysis for perchlorate in water EPA promulgated an IC based method (EPA Method 314.0) with a method detection limit of 0.5 µg/L and a minimum reporting level of 4.0 µg/L in reagent water (13). Recently Method 314.0 was adapted for the analysis of perchlorate in fertilizers and related materials
TABLE 1. Perchlorate Content in Soil, Fertilizers, and Tobacco Samples sample tobacco laminac tobacco midrib flue-cured tobaccoc 6-6-18 fertilizer 16-0-0 fertilizer soil
ICa analysis (mg/kg) dry wet
CEb analysis (mg/kg) dry
96.0 ( 0.6d 14.6 ( 0.1 96.8 ( 0.5d 12.5 ( 0.4 1.1 (
