Environ. Sci. Technol. 2008, 42, 1472–1478
Assessing Dioxin Precursors in Pesticide Formulations and Environmental Samples As a Source of Octachlorodibenzo-p-dioxin in Soil and Sediment E V A H O L T , * ,†,∇ R O L A N D V O N D E R RECKE,‡ WALTER VETTER,‡ D A R R Y L H A W K E R , †,§ V I N C E N T A L B E R T S , 4 BERTRAM KUCH,⊥ ROLAND WEBER,¶ AND CAROLINE GAUS∇ School of Environment and Centre for Environmental Systems Research, Griffith University, Nathan 4111, Australia, University of Hohenheim, Institute of Food Chemistry (170b), DE-70593 Stuttgart, Germany, Centre for Environmental Systems Research, Griffith University, Nathan 4111, Australia, Queensland Health and Scientific Services, 39 Kessels Road, Coopers Plains 4108, Australia, Institut für Siedlungswasserbau, Wassergüte- and Abfallwirtschaft Abt. Hydrochemie Bandtäle 2, D-70569 Stuttgart, POPs Environmental Consulting DE-73035 Göppingen, Germany, and National Research Centre for Environmental Toxicology (EnTox), The University of Queensland, 39 Kessels Road Coopers Plains 4108, Australia
Received July 10, 2007. Revised manuscript received November 11, 2007. Accepted November 27, 2007.
An as yet unidentified origin of elevated concentrations of polychlorinated dibenzo-p-dioxins (PCDDs) in soil and sediment has repeatedly been described from different locations around the world, including Australia. Natural sources have been hypothesized to account for such contamination, which is characterized by a distinctive dioxin profile, in particular, elevated levels of octachlorodibenzo-p-dioxins (OCDD) as well as relatively low contributions of polychlorinated dibenzofurans (PCDFs). The present study investigated whether OCDD formation via anthropogenically derived precursors represents a possible source in such samples. Soil and sediment from Australia and Hawaii were screened for known pesticide derived dioxin precursors. Two pesticide formulations containing pentachlorophenol (PCP), which are well-known to contain predominantly OCDD impurities, were also analyzed. Polychlorinated phenoxyphenols (PCPPs), common byproducts of pesticide production, were detected at parts-per-billion (ppb) levels in two PCP formulations and in five environmental samples. Of particular interest was the presence of the PCPP isomer 3,4,5,6tetrachloro-2-(2,3,4,5,6-pentachlorophenoxy)phenol(nonaC2PP), often also termed predioxin, in these samples. This compound * Corresponding author: e-mail:
[email protected]. † School of Environment, Griffith University. ‡ University of Hohenheim. § Centre for Environmental Systems Research, Griffith University. 4 Queensland Health and Scientific Services. ⊥ Institut für Siedlungswasserbau, Wassergüte- and Abfallwirtschaft Abt. Hydrochemie Bandtäle 2. ¶ POPs Environmental Consulting. ∇ National Research Centre for Environmental Toxicology. 1472
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 5, 2008
readily undergoes ring closure to form OCDD under a range of conditions and environments. In addition, the pesticide PCP itself, which also represents a potent precursor to OCDD formation and is known to contain OCDD impurities, was detected in some environmental samples. The evidence from this study indicates that pesticides and their impurities play an important role in the dioxin contamination of Australian soils and sediments, as well as other locations with similar PCDD/F patterns. The results further suggest that formation of OCDD from pesticide derived precursors may be a possible past, present, and future pathway for contamination of environmental samples.
Introduction Several studies from different countries (e.g., Australia, China, Japan, Germany, and the USA) have described the presence of a unique congener and isomer pattern of polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/Fs) in soil or sediment (1–7). The source of these contaminations remains unidentified to date. This PCDD/F pattern is characterized, in particular, by the dominance of octachlorodibenzo-pdioxin (OCDD) (contributing up to 99% to ΣPCDD/F concentration) and relatively low contributions of PCDFs, resulting in high PCDD to PCDF (D/F) ratios (>10 to several thousand) (1–7). Many samples where such contamination has been observed originate from within or near areas with agricultural/ rural land-use, where organochlorine pesticides may have been or are being extensively applied (1–7). Some organochlorine manufacturing processes are well-known to result in the formation of PCDD/Fs and their precursors, which may remain in the product as impurities (8). It has been recognized that the PCDD/F profile in these environmental samples shows similarities to that described from impurities in pesticide formulations containing pentachlorophenol (PCP) and sodium pentachlorophenate (NaPCP) (4, 9). In particular, both share the dominance of OCDD and a low contribution of PCDFs to ΣPCDD/Fs. However, PCP formulations were previously considered unlikely as a source, because the D/F ratios in environmental samples were not in the same range as those observed for PCP formulations (1–10 (3, 10), and PCP usage was not suspected, unknown, or not considered in these areas (2, 3, 11, 12). However, more recent studies demonstrate that the contribution of PCDFs can be highly variable among pesticide formulations, and D/F ratios of up to 82 have been reported from the Japanese PCP formulations used in this study (13). This variability is most likely due to differences in production process parameters such as temperature, alkalinity, and production pathway. Alterations in production processes and conditions are known to account for considerable differences of product impurity levels as well as profiles (13). In addition, some of the PCDD/F isomers that are considered markers for PCP impurities (1,2,3,4,6,7,8-HpCDF, 1,2,3,4,6,8,9-HpCDF, and OCDF 10, 13) could not be detected in many samples or were present at lower relative levels than would be expected (3). Investigating PCP-containing pesticides as a possible source is further complicated by their relatively short halflife and high mobility as compared to dioxins, which facilitates their more rapid loss in soil and sediment (14). In addition to containing dioxin impurities, higher chlorinated phenols (CPs) are potent precursors to PCDD formation. OCDD in particular is formed as the product of a condensation reaction involving PCP and has been reported 10.1021/es071687r CCC: $40.75
2008 American Chemical Society
Published on Web 01/31/2008
FIGURE 1. Formation pathways of OCDD from pentachlorophenol and polychlorinated phenoxyphenols. as the main formation product during photochemical (15–17), thermal (18), and biochemical transformation (19) of PCP and NaPCP. 3,4,5,6-Tetrachloro-2-(2,3,4,5,6-pentachlorophenoxy)phenol (nonaC2PP) has also been demonstrated to form OCDD under the influence of heat (20), sunlight (21), and biochemical processes (22) (Figure 1). Polychlorinated phenoxy phenols (PCPPs), in particular their higher chlorinated homologues (hexa- to nonaCPPs), have been reported as contaminants in pesticide formulations containing NaPCP and PCP (23) with the OCDD precursor nonaC2PP as the main congener in PCP (20, 24). It was previously hypothesized that formation of OCDD from such anthropogenic dioxin precursors represented a possible pathway to soil/sediment contamination, and hence a possible source for OCDD-dominated contaminations in Australia and elsewhere (4). Similarly, Baker and Hites (25) hypothesized that formation of OCDD during photolysis of PCP in the atmosphere may represent a significant source for OCDD on a global scale. However, research on OCDD precursors is very limited, and there is a significant knowledge gap with respect to their levels, distribution, and fate in the contemporary environment. In addition, information on their occurrence in pesticide formulations is currently restricted to a few studies focused on PCP. Moreover, it is unknown whether pesticide derived precursors are present in areas contaminated with OCDD of unidentified origin, the underlying premise for the above hypothesis. The present study focused on investigating whether pesticide derived dioxin precursors are present in soil/ sediment contaminated with the characteristic OCDDdominated dioxin pattern and whether these are comparable to samples with confirmed pesticide contamination and impurities found in pesticide formulations. Archived samples from a range of areas (in Australia and Hawaii) and known to contain relatively high levels of OCDD, in addition to three pesticide formulations containing PCP and NaPCP, were analyzed by high resolution gas chromatography/low resolution mass spectrometry (HRGC/LRMS) for CPs and PCPPs. Mass spectra were scrutinized for congeners that are known OCDD precursors.
Materials and Methods Sample Details. Eight archived samples (six soils and two sediments) that have previously been analyzed for PCDD/Fs were selected for CP and polychlorinated phenoxyphenol (PCPP) analysis. Samples were chosen to represent a range of contamination levels and on the basis of sample availability. To facilitate the detection of PCPPs, samples with OCDD levels on the order of ppb were chosen (Table 1). The details of sample collection are provided in references 11, 12, and 26–28. Samples were stored in either aluminum foil or solvent washed glass jars with Teflon sealed lids.
Analyzed environmental samples included soil and sediment that could be linked to PCP based on their dioxin congener profiles (H4, E2, and PR) as well as those characterized by the typical Queensland dioxin profile, where the PCDFs considered markers for PCP were near or below the limit of detection and which contained high D/F ratios (E3, S400, CAR, CNP, and MOS) (Table 1). Soil samples originated from a discontinued pineapple farm in Southeast Queensland (E2); a small National park approximately 500 m from the same pineapple farm (E3), the North Queensland regional center of Cairns (CAR), and two sugar cane farms (Maui, Hawaii (H4) and Mossman, Queensland (MOS)). One soil sample (CNP) originated from Cooloola National park in Southeast Queensland, an area previously classified as “remote” (26). Marine sediment samples originated from Hervey Bay (S400), a Southeast Queensland region dominated by sugar cane agriculture land use, and from the Parramatta River in Sydney (PR), which has recently been demonstrated to be impacted by emissions from a pesticide manufacturing plant where 2,4,5-trichlorophenoxy acetic acid (2,4,5-T) and PCP were produced (29). Because dioxin precursors have been documented as impurities of PCP and NaPCP formulations, three technical products (M-1 (PCP), production year unknown; M-2 (NaPCP), produced in 1967; and K-9 (NaPCP), produced in 1971) were also analyzed for PCPPs. The PCP products were obtained from Japan (kindly provided by Professor Masunaga, Yokohama University) and, as typical for these formulations, contained high levels of OCDD (Table 1; full PCDD/F congener and isomer patterns are reported in ref 13). Analysis for Chlorophenols and Chlorinated Phenoxyphenols. A subsample of freeze-dried soil or sediment was collected from each archived sample, sieved to remove particles >2 mm diameter, and homogenized. From this material, 10–20 g (dry weight, dw) was taken for CP and PCPP analysis. An internal standard (13C6 pentachlorophenol (13C6 PCP); Cambridge Isotope Laboratories, Andover, Massachusetts, USA) was added pre-extraction. The samples were Soxhlet extracted for 24 h in methanol/diethylether/ hydrochloric acid (100:10:0.01). After Soxhlet extraction, the soil and sediment extracts were concentrated, transferred to n-hexane, and subjected to alkaline liquid/liquid partitioning (∼pH 12–14) with 1 M KOH. The alkaline (aqueous) phase was subsequently acidified to a pH of 1–3 and extracted with dichloromethane. The dichloromethane phase was transferred to hexane and methylated with diazomethane in ether solution and then concentrated, subjected to cleanup on a column packed with 1 g of activated silica gel, and topped with anhydrous sodium sulfate. The column was eluted with n-hexane, followed by n-hexane: dichloromethane [1:1]. Both the n-hexane and [1:1] eluants were collected, concentrated under a gentle nitrogen stream, and reconstituted in 30 µL of toluene for analysis by HRGC/LRMS. Samples H4 and PR were subject to repeated silica cleanup to improve the quality of chromatograms. Subsamples of known quantities of the agrochemical formulations were ground and added to 4 mL of 1 M KOH (pH >12). The mixtures were shaken until most of the agrochemical formulations had dissolved. An aliquot of the aqueous mixture was added to 8 mL of 1 M KOH and subjected to alkaline liquid/liquid partitioning with hexane, followed by methylation and cleanup as described above. Samples were analyzed on a HRGC/LRMS (Shimadzu GC (GC-17A), Shimadzu MS (QP5050A), splitless injection) fitted with a nonpolar column (ZB-5MS; 0.25 mm internal diameter, length: 30 m, film thickness: 0.25 µm). The column was operated under temperature programmed conditions of 130-270 at 20 °C min-1 followed by 270–340 at 10 °C min-1. The injection port temperature was maintained at 250 °C and the interface temperature at 345 °C. The MS was operated VOL. 42, NO. 5, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1473
1474
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 5, 2008
26 000
72 2.0
15 190 4.0 1.0 440 200 31 14
