Transformation Processes, Pathways, and Possible Sources of

3, Townsville MC 4810 Queensland, Australia,. Ergo Forschungsgesellschaft mbH, Geierstrasse 1,. D-22305 Hamburg, Germany, and Universität Tübingen,...
0 downloads 0 Views 169KB Size
Download PDF
Environ. Sci. Technol. 2002, 36, 3542-3549

Transformation Processes, Pathways, and Possible Sources of Distinctive Polychlorinated Dibenzo-p-dioxin Signatures in Sink Environments C A R O L I N E G A U S , * ,†,‡ GREGG J. BRUNSKILL,§ DES W. CONNELL,‡ JOELLE PRANGE,‡ JOCHEN F. MU ¨ L L E R , † O L A F P A¨ P K E , | A N D R O L A N D W E B E R * ,# National Research Centre for Environmental Toxicology, 39 Kessels Road, Coopers Plains 4108, Queensland, Australia, School of Public Health, Griffith University, Nathan 4111 Queensland, Australia, Australian Institute of Marine Science, PMB No. 3, Townsville MC 4810 Queensland, Australia, Ergo Forschungsgesellschaft mbH, Geierstrasse 1, D-22305 Hamburg, Germany, and Universita¨t Tu ¨ bingen, Institute for Organic Chemistry, Auf der Morgenstelle 18, 72076 Tu ¨ bingen, Germany

In recent years, studies on environmental samples with unusual dibenzo-p-dioxin (PCDD) congener profiles were reported from a range of countries. These profiles, characterized by a dominance of octachlorinated dibenzodioxin (OCDD) and relatively low in dibenzofuran (PCDF) concentrations, could not be attributed to known sources or formation processes. In the present study, the processes that result in these unusual profiles were assessed using the concentrations and isomer signatures of PCDDs from dated estuarine sediment cores in Queensland, Australia. Increases in relative concentrations of lower chlorinated PCDDs and a relative decrease of OCDD were correlated with time of sediment deposition. Preferred lateral, anaerobic dechlorination of OCDD represents a likely pathway for these changes. In Queensland sediments, these transformations result in a distinct dominance of isomers fully chlorinated in the 1,4,6,9positions (1,4-patterns), and similar 1,4-patterns were observed in sediments from elsewhere. Consequently, these environmental samples may not reflect the signatures of the original source, and a reevaluation of source inputs was undertaken. Natural formation of PCDDs, which has previously been suggested, is discussed; however, based on the present results and literature comparisons, we propose an alternative scenario. This scenario hypothesizes that an anthropogenic PCDD precursor input (e.g. pentachlorophenol) results in the contamination. These results and hypothesis imply further investigations * Corresponding authors phone: +61 7 3274 9147; fax: +61 7 3274 9003; e-mail: [email protected] (Gaus) and phone: +81 45 759 2164; fax: +81 45 759 2149; E-mail: roland_weber@ ihi.co.jp (Weber). † National Research Centre for Environmental Toxicology. ‡ Griffith University. § Australian Institute of Marine Science. | Ergo Forschungsgesellschaft mbH. # Universita ¨ t Tu ¨ bingen. 3542

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 16, 2002

are warranted into possible anthropogenic sources in areas where natural PCDD formation has been suggested.

Introduction Extensive research on the formation, fate, and pathways of polychlorinated dibenzo-p-dioxins (PCDDs) and dibenzofurans (PCDFs) has enabled the correlation of homologue and congener signatures found in environmental samples to those produced by specific sources (1-4). However, although key pollutant sources have been identified in many countries, local and global PCDD/F source emission estimates are not in balance with their environmental deposition (both in terms of composition and concentration), particularly concerning significantly higher levels of octachlorodibenzodioxin (OCDD) in environmental sinks (2, 5-8). Such mass and composition imbalances have been attributed to (i) inadequately understood transformation processes of PCDD/Fs along their transport pathways (2) or OCDD synthesis from precursors in the environment (5) and/or (ii) underestimated/yet unidentified contributions from anthropogenic (6) or natural PCDD/F sources (8, 9). Numerous studies have demonstrated the formation of PCDD/Fs, in particular OCDD, from anthropogenic phenolic precursors (10-13), and various abiotic and biotic transformation processes along the environmental pathways of PCDD/Fs have been shown to result in considerable alterations of original PCDD/F emission signatures (2, 14-17). On the other hand, indications for natural formation of PCDDs in the environment have been presented from several locations, including the coastal environment of Queensland, Australia (18-20). Previously, we reported on elevated PCDD levels throughout all deposition layers and as early as the mid 1600s in dated sediment cores collected from estuarine embayments in North Queensland (18). PCDD/F concentrations were similar throughout the sediment cores, ranging from 1000 to 2500 pg g-1 dry weight (dw) in Hinchinbrook Channel and from 760 to 1300 pg g-1 dw in Burdekin estuary core sediments. The PCDD/F congener distribution throughout the cores was characteristically dominated by OCDD, whereas PCDFs were generally near or below the limit of quantification. The spatial distribution of similar PCDD/F profiles in topsoil and surface sediments from rural Queensland was found to be extensive, encompassing a wide variety of environments and types of land-use within (but confined to) the entire coastal region (19, 20). The spatial and temporal extent of similar PCDD/F signatures suggested a land-based, widespread, qualitatively similar, and historically old PCDD/F input into the Queensland environment over several centuries. PCDD/F characteristics similar to these Queensland samples were also reported from a number of other studies on ball clay and kaolinite from Australia (20), the U.S.A. (21, 22), and Germany (23) as well as sediments from the State of Mississippi, U.S.A. (24-26), Hong Kong (27), Japan, China, and the Philippine Basin (28, 29). All these studies concluded that (i) the characteristic and unusual PCDD/F profiles could not be attributed to known anthropogenic source inputs and/ or (ii) PCDDs have been present in sediments deposited prior to the production of organochlorines, indicating that the specific PCDD formation process may be of nonanthropogenic nature. To date it has not been possible to identify the processes that result in these characteristic PCDD/F signatures and elevated levels. Little is known about the natural formation of PCDD/Fs and their significance in the environment. On 10.1021/es025674j CCC: $22.00

 2002 American Chemical Society Published on Web 07/09/2002

FIGURE 1. Isomer distribution of tetra- to heptachlorinated PCDDs in A. Queensland sediment cores and river samples from the Hinchinbrook area and B. Kasumigaura Lake sediments (56), Mississippi river sediments (26), German kaolinite clay (31), Osaka Bay core sediment (28), and Hong Kong core sediments. (Note that some isomers could not be obtained from the literature). White arrows indicate the pathway of lateral dechlorination; black arrows indicate the isomers resulting from lateral- (L), peri-lateral- (LP), and peri- (P) dechlorination from OCDD. Note: 1,2,3,4,6,8-HxCDD was co-eluted in Osaka Bay core sediments with 1,2,4,6,7,9-/1,2,4,6,8,9-HxCDD. the other hand, transformation processes of PCDD/Fs in sediment and soil environments are inadequately understood. A first step to understanding the possible processes involved in the formation of the characteristic OCDD dominated PCDD signatures found in some environmental samples is to determine whether these represent the actual source output pattern or are the result of postdepositional transformation processes obscuring the original patterns. To investigate this, we have assessed our previously reported spatial and temporal PCDD/F distributions from Queensland, Australia (18, 19) using additional normalized concentrationand isomer-specific analysis and compared them to the isomer distribution of samples collected elsewhere. The implications from these results on environmental processes occurring in the environment, the original source patterns, and hence possible sources are discussed, and two scenarios are provided. These source scenarios are similar with respect to the processes involved but fundamentally different in the origin of the source signatures that seem to characterize “natural PCDD signatures”.

Materials and Methods This paper presents a reassessment of samples collected for previous studies on spatial and temporal PCDD/F concentrations in the coastal zone of Queensland (18, 19). For the present study, these data have been used to perform additional, ΣPCDD normalized homologue concentrationand isomer-specific analysis for sediment, soil, and sediment core samples (Hinchinbrook Channel and Burdekin estuary dated sediment cores). Materials and methods for sample collection, locations, PCDD/F analysis, and quality control during sampling, storage, and analysis of samples are presented elsewhere (18, 19). Gamma spectrometric mea-

surements and interpretations of the sediment deposition age of core sections have been described previously (30).

Results and Discussion PCDD/F Fingerprints. The isomer distributions of tetra- to heptachlorinated PCDDs from sediments of top and bottom core sections of Hinchinbrook Channel and Burdekin estuary cores are presented in Figure 1A (see also chromatograms presented in Figure 4, Supporting Information). In general, core slices from both cores, ranging from sediments deposited recently up to approximately 350 years ago, show similar isomer patterns. These are characterized by a striking dominance of isomers fully chlorinated in the 1,4,6,9positions (the dominance of 1,4,6,9-isomers was confirmed by GC-MS analysis using DB5 and SP 2331 columns). The average contribution of the single 1,4,6,9-substituted isomers to their respective homologue groups in Burdekin estuary and Hinchinbrook Channel core sediments respectively was 80% and 84% among both HpCDDs, 77% and 87% among all 10 HxCDDs (sum of three 1,4,6,9-substituted isomers), 34% and 53% among all 14 PnCDDs, and 25% and 57% among all 22 TCDDs. (Note that the concentration of 1,4,6,9-TCDD was determined by subtraction of the coeluted isomers on SP 2331- from those coeluted on DB 5 column injections.) In the following we will refer to this isomer distribution as a “1,4-pattern” which, to our knowledge, is unlike any known direct anthropogenic source output pattern and has not been described previously. However, most data available to date on PCDD/F sources and sinks are typically biased toward reporting 2,3,7,8-substituted congeners only, hence sources producing 1,4-patterns may simply be unnoticed. Isomerspecific analysis also revealed a similar 1,4-pattern in most other sediment samples analyzed to date from Queensland, VOL. 36, NO. 16, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3543

FIGURE 2. Relative increase (contribution of Σhomologue groupto ΣPCDD concentrations) of tetra- to heptachlorinated PCDDs and relative decrease of OCDD with increasing sedimentation age in Burdekin and Hinchinbrook core sediments andswithin the shaded areasin topsoil and surface river sediments from the Hinchinbrook area. including those collected from riverine, estuarine, and marine locations along the 2000 km coastline (Figure 1A shows the 3544

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 16, 2002

1,4-pattern from river sediments in the Hinchinbrook area). Apart from these unique isomer characteristics all these samples have in common a strong dominance of OCDD and low or nondetectable concentrations of PCDFs (18, 19). Comparing the Queensland PCDD/F fingerprints to isomer patterns or concentrations published from other “natural formation samples”, interesting similarities become apparent. Figure 1B shows the isomer distributions of a range of samples reported in the literature where a natural formation of PCDDs was suggested. Isomers from sediments in a Japan estuary (Osaka Bay (Figure 1B) and Harima Nada) (28) as well as from the Yellow and East China Sea and the Philippine Basin (29) also show the distinctive 1,4-pattern (note: TCDDs and PnCDDs were reported below the limit of detection). These samples included two dated sediment cores with an estimated deposition age of several thousand to million years. The 1,4-patterns are also apparent from PCDD isomer patterns published by Rappe et al. (25, 26) from Southern Mississippi, U.S.A., in particular from river sediment (Figure 1B) and dried oxbow samples as well as rural lake sediments. Further, estuarine sediments from Mai Po Marshland near Hong Kong (27) are also dominated by 1,4,6,9-substituted isomers (Figure 1B) and kaolinite clay chromatograms from Germany (31) (Figure 1B) and Australia (20) show the distinctive 1,4-pattern. The striking similarities of these PCDD fingerprints, in combination with the overall similarities of PCDD homologue distributions and the general lack or low concentrations of PCDFs, strongly suggests that specific formation and/or transformation processes occur (or have occurred) not only in Queensland but also possibly globally. These processes result in the formation of PCDDs, in particular OCDD and the 1,4-pattern from a yet unidentified PCDD/F source. Transformation of PCDD/F Signatures. Indications for transformations of PCDD/F signatures were apparent from historical PCDD/F distributions in Queensland sediment cores. Previously we reported relatively consistent PCDD/F concentrations in Queensland sediment cores over three centuries of deposition (18). Further analysis of these data reveals that although no significant differences in ΣPCDD/F concentrations over time were apparent, the historical data showed relatively small, but significant and consistent changes of the percent contribution of tetra- to octahomologue groups to the ΣPCDD concentration over the core’s sedimentation history (Figure 2). In both core locations and particularly in sediments from the Burdekin area, the contribution of tetra- to heptachlorinated PCDDs to the ΣPCDDs significantly increases with core depth (r2 values are given in Figure 2). Corresponding with this was a significant and continuous decrease in contribution of OCDD to the ΣPCDDs with sediment age (Figure 2). Surface river samples and forest as well as agricultural soil collected in the Hinchinbrook region during earlier studies show similar PCDD/F concentrations compared to the Hinchinbrook Channel core sediments (18); however, higher contributions of OCDD and lower contributions of tetra- to heptachlorinated PCDD congeners are apparent from these samples (Figure 2). This suggests that the trends of increasing tetrato heptachlorinated PCDDs with core depth are a function of age after deposition in the marine system due to degradation processes. The processes involved resulted in a shift toward higher concentrations of tetra- to heptachlorinated homologues with increasing “age” after deposition and suggest a dechlorination of higher chlorinated PCDDs. The regression slopes obtained from Figure 2 represent a proxy for the relative rate of appearance of tetra-to heptachlorinated PCDDs and the relative rate of disappearance of OCDD with age of deposition. The time of appearance (2-fold increase in concentration) of lower chlorinated PCDDs decreased with decreasing degree of chlorination (see Figure

FIGURE 3. Schematic of dechlorination processes of higher chlorinated PCDDs and the resultant production of lower chlorinated PCDDs with time of deposition in the sediment environment to result in the homologue shift observed in the sediment cores. Scenario A postulates the deposition of mainly higher chlorinated PCDDs on land and their subsequent transport and dechlorination in the marine environment. Scenario B postulates the deposition of anthropogenic CPs (and their derivatives) during the mid 20th century, their transport, migration into deep sediment layers, and subsequent formation and dechlorination of OCDD in the marine environment. 5, Supporting Information), which corresponds to results from experimental dechlorination studies (16, 17). The estimated time for OCDD disappearance (halving in concentration, calculated using the regression slopes from Figure 2) were 300-700 years, and the times of appearance of lower chlorinated PCDDs ranged from 200 to 300 years, lowest for ΣTCDDs with 60-200 years (see Figure 5, Supporting Information). Reported half-lives for OCDD in sediment or soil environments vary and range from less than 10 years to more than 100 years (7, 32, 33). In addition, the production of lower chlorinated PCDDs from OCDD and HpCDD has been observed to occur within days (16, 17, 34), which may indicate that the processes in the Queensland environment are relatively slow. The present study suggests that dechlorination processes have altered the original PCDD/F source signature. After formation, the PCDDs are subject to progressive degradation processes, predominantly during the long-term sedimentation within the marine environment. PCDDs in sediments from the deepest core sections have been subject to longerterm degradation processes since their deposition in the marine system compared to surface sediments, i.e., the

PCDD/F profiles in the surface slices of the sediment cores, and those found in samples from river sediments and soil, represent “young” signatures and are more similar to the original source output profile. Dechlorination Pathways. Dechlorination pathways of OCDD can, due to symmetry considerations, proceed via the lateral (2,3,7,8-) and/or peri (1,4,6,9-) positions. The 1,4patterns in Queensland samples indicate the predominant lateral chlorine elimination from OCDD, resulting in higher concentrations of 1,2,3,4,6,7,9-HpCDD compared to 1,2,3,4, 6,7,8-HpCDD (Figure 1A). In subsequent lateral chlorine eliminations the 1,4,6,9-substituted HxCDDs, PnCDDs, and TCDDs are formed as the main isomers in the respective homologue groups with increasing time of transformation (Figure 1A). This results in an increase of 1,4,6,9-substituted isomer contributions to the respective homologue groups with increasing age after deposition observed in the present study (Figure 1A illustrates this for the top and bottom sediment slices of both Queensland cores). As expected in a preferred lateral dechlorination process from OCDD, the ratios of the lateral-dechlorinated (L) to peri-dechlorinated (P) HpCDDs (1,2,3,4,6,7,9- to 1,2,3,4,6,7,8VOL. 36, NO. 16, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3545

TABLE 1. Ratios of Lateral Dechlorinated (L) and Peri Dechlorinated (P) HpCDDs (i.e. 1,2,3,4,6,7,9-HpCDD: 1,2,3,4,6,7,8-HpCDD) and L:LP:P HxCDDs (i.e. 1,2,4,6,7,9-/ 1,2,4,6,8,9-HxCDD:1,2,3,6,7,9-/1,2,3,6,8,9-HxCDD:1,2,3,6,7,8-/ 1,2,3,7,8,9-/1,2,3,4,6,7-HxCDD) and Percent Contribution of L PnCDD (i.e. 1,2,4,6,9-PnCDD) and L TCDD (i.e. 1,4,6,9-TCDD) to the ΣHomologue Group in a Range of Samples from Queensland and Elsewhere L:P L:LP:P HpCDD HxCDD %L %L ratio ratio PnCDD TCDD

Marine and Estuarine Sediments Hinchinbrook (150 cm)a Hinchinbrook (0-2 cm)a Burdekin (390 cm)a Burdekin (0-2 cm)a Hong Kong (33-40) cma Yellow Sea (surface)b Osaka Bay (200 cm)c Philippine Basin (1050-1080 cm)b

8:1 4:1 5:1 4:1 5:1 4:1 ∼3:1 3:1

45:5:1 53:7:1 22:4:1 11:2:1 23:6:1 ndi ∼10:3:1 ndi

56 49 37 36 28 ndi ndi ndi

66 50 32 26 36 ndi ndi ndi

∼34:2:1 4:1:1 ∼4:2:1 ∼4:2:1

nah 19 ∼33 nah

nah 20 nah nah

5:2:1

21

15

nah 13

nah 3

Lake and River Sediments Mississippi lake (bottom)d River, Hinchinbrook areaa Mississippi river (Pascagoula)e Mississippi lake (Rogue Homa)d

nah 2:1 ∼2:1 nah

Kaolinite Clay German kaolinite claya

2:1

Lake Sediments with Known PCP Input Lake Kasumigaura (Takahama Iri)f nah Lake Shinji (25-26 cm)g 2:1

∼4:2:1 4:2:1

a Present study. b Calculated from PCDD concentrations published in ref 29. c Calculated from isomer distribution published in ref 28. d Calculated from isomer distribution published in ref 25. e Calculated from isomer distribution published in ref 26. f Calculated from isomer distribution published in ref 56. g Calculated from PCDD concentrations published in ref 55. h na, not available. i nd, not detected.

HpCDD) were found to influence the ratios of L (1,2,4,6,7,9and 1,2,4,6,8,9-) to LP (1,2,3,6,7,9- and 1,2,3,6,8,9-) to P (1,2,3,6,7,8- and 1,2,3,7,8,9-) HxCDDs, and contributions of lower chlorinated L isomers (1,2,4,6,9-PnCDD and 1,4,6,9TCDD) to the respective Σhomologues. Figure 1A,B presents the isomer patterns, Table 1 the corresponding ratios for a range of 1,4-pattern samples with different degrees of initial dechlorination preference (i.e. ratios of the two HpCDD isomers). It is apparent that among Queensland samples, the most distinct 1,4-pattern shows the highest L:P HpCDD ratio and is present in the deepest sediment layer (i.e. longest sedimentation history). Correspondingly, this sample shows the highest L:LP:P HxCDD ratio and the highest percent contribution of L PnCDD and L TCDD among Queensland samples. On the other hand, Burdekin core sediments show lower L:P HpCDD ratios and less distinct 1,4-patterns. Further, surface river sediments from the Hinchinbrook area, with the least distinct 1,4-pattern, show the lowest L:P HpCDD ratio and corresponding lowest L:LP:P HxCDD ratio and percent contribution of L isomers. These trends indicate that the formation of 1,4,6,9-isomers, and, therefore, the overall isomer distribution of tetra- to hexa-chlorinated PCDDs is already determined from the ratio of the two HpCDD isomers (i.e. the appearance of 1,4,6,9-substituted isomers is governed mainly by the dechlorination preference from OCDD and resulting PCDDs). Table 1 compares the L:P and L:LP:L ratios and percent contributions of L isomers of a range of samples with 1,4patterns. Ratios obtained from other estuarine and marine core or surface sediments (Osaka Bay, Harima Nada, Philippine Basin, Yellow Sea, Hong Kong) (28, 29) show similar ratios to those from estuarine sediments in Queensland. The bottom stratum from a Mississippi lake (Lake Mallard) (25) shows a very distinct 1,4-pattern and correlates to the highest 3546

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 16, 2002

ratios from the estuarine core sediments. Samples collected from Mississippi rivers (Pascagoula River) (26) and lakes (Lake Rogue Homa cores) (25) are comparable to ratios obtained for river samples in the Hinchinbrook area. Interestingly, German kaolinite ratios are similar to those obtained from the river samples. Considering the results of this and other studies, the differences in the extent of 1,4,6,9-substituted isomer dominance and the resulting differences in isomer distributions can probably be explained by the dechlorination preference. Laboratory studies on dechlorination of PCDD/Fs demonstrate that these processes can range from preferred perito peri/lateral and preferred lateral chlorine elimination, depending on various parameters (35, 36). The heterogeneous catalyzed, thermal dechlorination leads to a 2,3,7,8-HxCDD pattern with a dominance of 1,2,3,7,8,9-HxCDD (36, 37), which is one of the characteristics often reported for kaolinite, ball clay, Mississippi, and Queensland samples (19-21, 24, 26, 38). This characteristic of the HxCDD isomer pattern was to date supporting the theory of natural PCDD formation in these samples since it could not be attributed to a known anthropogenic source output. However, the data presented here strongly suggests that dechlorination processes are at present the only formation mechanisms that can explain the specific distribution of 2,3,7,8-substituted HxCDDs. A preferred lateral dechlorination was recently reported by Ohtsuka et al. from studies on photodegradation pathways using UV (254 nm) radiation (39). Remarkably, the Cl-H photochemical substitution pathways resulted in a 1,4,6,9dominated isomer pattern, strikingly similar compared to the Queensland samples. Therefore, patterns with L:LP:P HxCDD ratios as they are found in the 1,4-patterns are likely to result from dechlorination processes. With respect to Queensland’s and other deep core sediments, however, UV radiation is unlikely the catalyst of the observed transformation of PCDD signatures. Since the sediments analyzed during the present study are predominantly anaerobic to semianaerobic, reductive dechlorination seems a more likely mechanism for the transformation processes with respect to the sediment environment of Queensland. Dechlorination of PCDDs (by microbial and/or abiotic processes) has been demonstrated during a number of studies under the reducing conditions prevailing in sediments (16, 17, 40-42). Several pathways have been identified depending on microorganisms present, the substitution patterns of the original source profile, and a number of abiotic factors. Abiotic and biotic dechlorination was observed at room temperature in historically PCDD/F contaminated estuarine sediments (16). These processes, however, resulted mainly in the removal of a peri-chlorine (ratio of L:P HpCDD 1:4). A dichotomous microbial dechlorination pathway for highly chlorinated PCDDs was reported from Passaic River sediments (43), including a mixed peri-lateral dechlorination for non-2,3,7,8-substituted congeners. Similar to the 1,4-patterns of the present study, an increase in 1,2,3,7,8,9-HxCDD was reported in anaerobic sediment cores from Lake Ketelmeer (44). A subsequent study from Lake Ketelmeer indicated that specific microorganisms are capable of simultaneous lateraland peri-dechlorination activities (42). Recently, bacteria from Spittel Wasser sediment were found to selectively dechlorinate the lateral positions (“SP-pathway”) (41). The results indicate that a specific lateral dechlorination by microorganisms is possible and may therefore provide a potential mechanism for the transformation of PCDD source profiles into the 1,4,6,9-dominated signatures observed. However, only the 1,2,3,4-tetra-CDD to 1,2,4-tri-CDD pathway was investigated in this study, with 1,2,3,4-TCDD exhibiting different charge distributions compared to higher chlorinated PCDDs. Bunge et al. concluded that the position and congener specificity of microbial PCDD reductive dechlorination might

depend on specific bacteria. Differences in L:P and L:LP:P isomer ratios observed in the 1,4-patterns of estuarine/marine and river/lake samples may therefore depend, in addition to time frames available for transformation, on the biotic environment of the specific locations. The 1,4-patterns with lower ratios observed in lake and river sediments (L:P HpCDD ratio of about 2:1) could even be expected from the results of Buerskens et al. (42) with a high impact of peridechlorination. In addition, the presence of phenols and humic constituents has been suggested to influence the preferred pathways of dechlorination reactions (17, 45). Natural and Anthropogenic Source Implications. To date, the specific 1,4-patterns observed in samples from Queensland and elsewhere could not be correlated to any known PCDD/F source. This supported the hypothesis that natural PCDD formation is a potential cause of the contamination. The chronology of the two sediment cores in Queensland (dated from the 1640s and 1840s to present) and the relatively consistent PCDD/F levels throughout all sediment layers, in combination with the absence of significant tertiary industries near the sampling areas, previously suggested a PCDD formation prior to the production of organochlorines and industrialization (18). Scenario A, schematized in Figure 3A, postulates a historically old (at least 3 centuries) or continuous land-based source of higher chlorinated PCDDs (predominantly OCDD) to account for the PCDDs found in sediments deposited during the last 350 years. Subsequent dechlorination of OCDD, occurring predominantly in the sediment environment, resulted in the formation of lower chlorinated, 1,4,6,9-dominated PCDDs with depositional age. Hence, the PCDD/F profiles in the surface slices of the sediment cores are more similar to the original source profile. The results of the present study, however, strongly suggest that not the PCDD/F source but post-formation dechlorination processes in the environment result in the characteristic 1,4-patterns (including the specific distribution of 2,3,7,8-substituted HxCDD present in all 1,4-patterns). Consequently, the 1,4-pattern does not reflect the original PCDD/F source output signature, and any formation process that results in the production of predominantly OCDD represents a possible source input. This provides new possibilities with respect to the identification of potential PCDD/F sources and their formation processes. Various anthropogenic compounds, in particular chlorinated phenols (CPs) such as pentachlorophenol (PCP) and sodium pentachlorophenol (Na-PCP) (and their derivates e.g. phenoxyphenols (46)) have been shown in numerous studies not only to contain large amounts of OCDD but also to produce OCDD via both biological and abiotic processes (11, 13, 47, 48). These compounds are relatively water-soluble, depending on their ionization states in the environment, which significantly impacts their aqueous solubility, sorption, and transport (e.g. PCP solubility in water ranges from 0.2 g/L at pH 5 to 200 g/L at pH 10 (12)). At pH >7, for example, PCP is 99% disassociated, moving easily through soils (12, 49). PCP is a common groundwater contaminant (50) and a number of studies have observed that PCP, used extensively during the 1940s to 1990s as fungicides and/or herbicides, can be highly leachable in soils and sediments (51, 52). Estimation of the diffusion coefficient of PCP in water (DBW ) 7.29 × 10-6 cm2/s) using the Wilke-Chang equation (53) supports that disassociated PCP may have diffused to the deepest core sections of the sediment cores in Queensland, although the diffusion coefficient in water will overestimate the extent of the diffusion in sediment pore water environments. Studies on migration of PCP (aqueous and kerosene solutions) in soils, however, show that PCP column breakthrough occurs quickly indicating its deep penetration into the columns (54). Hence, we cannot ignore the potential for

anthropogenic influences as possible contamination sources, in particular CPs that had the potential to migrate through the sediment cores in the past 50-60 years of their use and represent precursors for the formation of OCDD in deep sediments. Based on the results from our work in Queensland, in combination with the possibilities of a diffusion and transformation of anthropogenic OCDD precursors in deep sediments, we propose a second scenario (Figure 3B). Scenario B includes the dechlorination processes discussed in this paper, but in contrast to the natural formation Scenario A, hypothesizes a precursor source input of commercial chlorophenols (e.g. PCPs and their derivates) during the midend 20th century and the subsequent condensation of predominantly OCDD from CPs in the environment. CPs and their predominant OCDD-impurity runoff to the marine system, together with an OCDD condensation within the sediment environment would result in the strong OCDD dominated profiles observed in the sediment layers from the mid-end 20th century. From these sediment sections however, diffusion of the deposited CPs into the deeper sediment layers is likely to have occurred. A condensation of OCDD from diffused CPs in these deeper sediments would again result in OCDD dominated PCDD profiles and may have contaminated sediment layers that have been deposited prior to the production of organochlorines. Subsequent dechlorination of OCDD results in the distinctive 1,4-pattern typical of Queensland samples. The (predominant) OCDD formation from diffused PCP could explain the lower PCDF concentrations (PCP has been shown to produce predominantly OCDD and to a lesser extent HpCDDs, whereas both PCDDs and PCDFs are typically present in PCP as impurities). Further, in this scenario an order of magnitude higher times of OCDD disappearance (approximately 30-70 years) and PCDD appearance (approximately 6-20 years for ΣTCDDs) are obtained compared to estimates using the actual 350 years sediment deposition periods (see Figure 4, Supporting Information). This is due to the considerably shorter deposition periods (since 40-60 years) in the case of Scenario B. Support for Scenario B is found in reports from Japan lake- and tributary river sediments with known PCP contamination from agricultural use. Similar to the 1,4-pattern discussed in the present study, the HxCDD distribution in Lake Shinji core sediments (55) and some Lake Kasumigaura (Takahama Iri) surface sediments (56) show a distinctive 1,4pattern (including the dominance of 1,2,3,7,8,9/(1,2,3,4,6,7)HxCDD among the toxic HxCDDs). The HxCDD isomer pattern from Lake Kasumigaura (Takahama Iri) is presented in Figure 1B. In addition, the L:P HpCDD and L:LP:P HxCDD ratios calculated from reported PCDD/F concentrations in these lakes are remarkably similar to Hinchinbrook and Mississippi river/lake ratios (Table 1). Principal component analysis from Kasumigaura Lake/river sediments revealed that one group of samples was characterized by these unusual HxCDD patterns, contributing 18% to the total variation, but could not be attributed to a known source or transformation process (56). Importantly, particularly the sediments from the lake showed the unusual HxCDD patterns, whereas the tributary river sediments correlated to principal components representing PCP and/or chloronitrophen (CNP) impurities, used in rice paddy cultivations of the area. The authors suggested that “either some input or lack of some loss/ transformation” could have resulted in these signatures. Results from the present study suggest that dechlorination of OCDD, derived from the extensive PCP input into this system, has resulted in the transformation of Lake Kasumigaura sediment PCDD signatures. Core sediments from Lake Shinji also showed high PCDD/F contributions (average 68%) from PCP impurities, VOL. 36, NO. 16, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3547

derived from its extensive use as rice paddy herbicide in the area during the 1960s and 1970s (55). Apart from characteristic 1,4-patterns present in the core samples, it is particularly interesting that the PCDD/F concentrations in the dated sediment core increased several years ahead of PCP use. Sediments deposited during the time of PCP application showed relatively high PCDF concentrations and low PCDD: PCDF (D/F) ratios (11:14), corresponding well to the D/F ratios reported from PCP formulas (57). Older sediments, deposited before the application of PCP, still showed relatively high PCDD concentrations (>2500 pg g-1 dw); however, PCDF concentrations were considerably lower, resulting in a D/F ratio of up to 82 in the deepest sediment layer. Considering the history of land-use in this area, it seems possible that this trend may be due to the formation of higher chlorinated PCDDs after migration of PCP into deeper sediment layers, whereas PCDD/F impurities originally present in PCP are relatively immobile and sorbed strongly to the organic material in the original sediment layer of deposition. A subsequent long-term dechlorination of newly formed OCDD can result in the formation of the characteristic 1,4,6,9isomers as suggested by the results of the present study. OCDD condensation from PCP may also explain the discrepancies observed in Lake Shinji sediments with respect to mass balance estimations, where the ratio of PCDD/Fs lost from soil to that of the emission input, and the ratios of PCDD/Fs deposited in the lake sediments to that lost from the basin was considerably higher for PCP- compared to CNP-calculated PCDD/Fs impurities (55). Similarly, Baker (5) suggested that the conversion of PCP to OCDD, and to a lesser extent HpCDD in the environment, may represent the major cause for the large discrepancies between atmospheric PCDD/F emission and deposition estimates observed from mass balance estimations around the world. Their estimates suggested that photochemical PCP conversion processes in the atmosphere can close the mass balance for both OCDD and HpCDD. Our enquiries regarding past CP application quantities within the coastline of Queensland or even within Australia have been unsuccessful, and reliable data have been inaccessible to date. However, PCP has been produced in Australia (until the mid 1990s), imported into Queensland, and found application for preservation of timber (58, 59). Use within agricultural areas of Queensland is unknown at this stage. It is noteworthy in this respect though that recent studies on forest soil cores in Queensland have shown that almost exclusively OCDD and HpCDD are present in deep soil layers, while the surface sections show PCDD congeners typical for Queensland profiles and PCDFs that indicate an impact of PCP contamination (60). Considering all present knowledge about the PCDD contamination in Queensland, a land-based source of primarily OCDD (and/or OCDD precursors) seems most probable. A subsequent transformation via dechlorination after OCDD formation (and for Scenario B migration and condensation) is likely to result in the accumulation of lower chlorinated PCDDs, in particular the peri-chlorinated isomers producing the characteristic 1,4-patterns observed along the Queensland coastline. With the present knowledge, however, it cannot be elucidated if the source input is of anthropogenic or nonanthropogenic origin. Preliminary results suggest that more evidence for Scenario B may be obtained from some surface sediments and soil samples, where PCDFs signatures are clearly present and comparable to those found in PCP impurities (e.g. 1,2,3,4,6,8,9-HpCDF), although concentrations are relatively low in most samples analyzed to date. More work is currently underway to determine the historical distribution and levels of likely anthropogenic compounds as well as their potential to result in the widespread contamination of coastal Queensland, including historically 3548

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 16, 2002

old sediment layers. A thorough understanding of the PCDD/F transformation processes and pathways in the environment will be imperative to the identification of the original PCDD/F source profiles and the specific formation processes involved in the formation of OCDD dominated 1,4-patterns observed in Queensland and elsewhere.

Acknowledgments We thank Mark O’Donohue for valuable discussions and helpful reviews. This research was supported by the ARCSPIRT scheme, Ergo Forschungsgesellschaft and the Great Barrier Reef Marine Park Authority. Queensland Health Scientific Services provides funding for NRCET. Gregg Brunskill’s work was partially supported by the CSIRO Coastal Zone Project and the CRC Reef.

Supporting Information Available Isomer patterns of tetra- to heptachlorinated CDDs from Hinchinbrook core surface and bottom sediments (Figure 4) and estimated half-lives of OCDD and time of appearances of lower chlorinated PCDDs in both cores (Figure 5). This material is available free of charge via the Internet at http:// pubs.acs.org.

Literature Cited (1) Hagenmaier, H.; Lindig, C.; She, J. Chemosphere 1994, 29(911), 2163. (2) Brubaker, W. W.; Hites, R. A. Environ. Sci. Technol. 1997, 31(6), 1805. (3) Eitzer, B. D. Environ. Sci. Technol. 1993, 27(8), 1632. (4) Ogura, I.; Masunaga, S.; Nakanishi, J. Chemosphere 2001, 45(2), 173. (5) Baker, J. I.; Hites, R. A. Environ. Sci. Technol. 2000, 34(14), 2879. (6) Brzuzy, L. P.; Hites, R. A. Environ. Sci. Technol. 1996, 30(6), 1797. (7) Eisenberg, J. N. S.; Bennett, D. H.; McKone, T. E. Environ. Sci. Technol. 1998, 32(1), 115. (8) Prange, J. A.; Mu ¨ ller, J. F. Organohalogen Compd., in press. (9) Gribble, G. W. Environ. Sci. Technol. 1994, 28(7), 310A. (10) Klimm, C.; Schramm, K. W.; Henkelmann, B.; Martens, D.; Kettrup, A. Chemosphere 1998, 37(9-12), 2003. (11) Wittsiepe, J.; Kullmann, Y.; Schrey, P.; Fidelis, S.; Wilhelm, M. Toxicol. Lett. 1999, 106, 191. (12) Crosby, D. G.; Beynon, K. I.; Greve, P. A.; Korte, F.; Still, G. G.; Vonk, J. W. Pure Appl. Chem. 1981, 53, 1051. (13) O ¨ berg, L. G.; Rappe, C. Chemosphere 1992, 25, 49. (14) Alcock, R. E.; Sweetman, A. J.; Jones, K. C. Chemosphere 2001, 43, 183. (15) Barkovskii, A. L.; Adriaens, P. Int. Biodeterior. Biodegred. 1996, 37(3-4), 247. (16) Albrecht, I. D.; Barkovskii, A. L.; Adriaens, P. Environ. Sci. Technol. 1999, 33(5), 737. (17) Fu, Q. S.; Barkovskii, A. L.; Adriaens, P. Environ. Sci. Technol. 1999, 33, 3837. (18) Gaus, C.; Brunskill, G., J.; Weber, R.; Pa¨pke, O.; Mu ¨ ller, J. F. Environ. Sci. Technol. 2001, 35, 4597. (19) Gaus, C.; Pa¨pke, O.; Dennison, N.; Haynes, D.; Shaw, G. R.; Connell, D. W.; Mu ¨ ller, J. F. Chemosphere 2001, 43(4-7), 549. (20) Prange, J. A.; Gaus, C.; Pa¨pke, O.; Muller, J. F. Chemosphere 2002, 46(9-10), 1335. (21) Ferrario, J. B.; Byrne, C. J.; Cleverly, D. H. Environ. Sci. Technol. 2000, 34(21), 4524. (22) Rappe, C.; Andersson, R. Organohalogen Compd. 2001, 46, 9. (23) Jobst, H.; Aldag, R. UWSF - Zeitschrift fu ¨ r Umweltchemische O ¨ kotoxikologie 2000, 12(1), 2. (24) Rappe, C.; Andersson, R.; Cooper, K.; Bopp, R.; Fiedler, H.; Howell, F.; Bonner, M. Organohalogen Compd. 2000, 46, 19. (25) Rappe, C.; Andersson, R.; Cooper, K.; Fiedler, H.; Lau, C.; Bonner, M.; Howell, F. Organohalogen Compd. 1997, 43. (26) Rappe, C.; Anderson, R.; Bonner, M.; Cooper, K.; Fiedler, H.; Howell, F.; Kulp, W. E.; Lau, C. Chemosphere 1997, 34(5-7), 1297. (27) Mu ¨ ller, J. F.; Gaus, C.; Prange, J. A.; Pa¨pke, O.; Fai Poon, K.; Lam, M. H. W.; Lam, P. K. S. Mar. Pollut. Bull. 2002, in press. (28) Hashimoto, S.; Wakimoto, T.; Tatsukawa, R. Chemosphere 1990, 21(7), 825.

(29) Hashimoto, S.; Wakimoto, T.; Tatsukawa, R. Mar. Pollut. Bull. 1995, 30(5), 341. (30) Brunskill, G. J.; Zagorskis, I.; Pfitzner, J. Estuarine, Coastal Shelf Sci., in press. (31) Pa¨pke, O. Ergo Forschungsgesellschaft mbH, Germany, unpublished results. (32) Mackay, D.; Shiu, W. Y.; Ma, K. C. Illustrated handbook of physical-chemical properties and environmental fate for organic chemicals - polynuclear aromatic hydrocarbons, polychlorinated dioxins and dibenzofurans; 1992; Vol. 2. (33) Kjeller, L.-O.; Rappe, C. Environ. Sci. Technol. 1995, 29(2), 346. (34) Segstro, M. D.; Muir, D. C. G.; Servos, M. R.; Webster, G. R. B. Environ. Toxicol. Chem. 1995, 14(10), 1799. (35) Weber, R.; Nagai, K.; Nishino, J.; Shiraishi, H.; Ishida, M.; Takasuga, T.; Konndo, K.; Hiraoka, M. Chemosphere, in press. (36) Hagenmaier, H.; Brunner, R.; Haag, H.-J.; Kunzendorf, M.; Kraft, K.; Tichaczeck; Weberruss, U. VDI Berichte 1987, 634, 61. (37) Wiesmu ¨ ller, T. Ph.D. Dissertation, University of Tu ¨ bingen: Germany, 1992. (38) Rappe, C.; Bergek, S.; Fiedler, H.; Cooper, K. R. Chemosphere 1998, 36(13), 2705. (39) Ohtsuka, N.; Minomo, K.; Hosono, S.; Kurata, Y.; Sugisaki, M. Organohalogen Compd. 2001. (40) Adriaens, P.; Fu, Q.; Grbic-Galic, D. Environ. Sci. Technol. 1995, 29, 2252. (41) Bunge, M.; Ballerstedt, H.; Lechner, U. Chemosphere 2001, 43, 675. (42) Beurskens, J. E. M.; Toussaint, M.; De Wolf, J.; Van der Steen, J. M. D.; Slot, P. C.; Commandeur, L. C. M.; Parsons, J. R. Environ. Toxicol. Chem. 1995, 14, 939. (43) Barkovskii, A. L.; Adriaens, P. Appl. Environ. Microbiol. 1996, 62(12), 4556. (44) Beurskens, J. E. M.; Mol, G. A. J.; Barreveld, H. L.; Van Munster, B.; Winkels, H. J. Environ. Toxicol. Chem. 1993, 12, 1549. (45) Barkovskii, A. L.; Adriaens, P. Environ. Toxicol. Chem. 1998, 17, 1013.

(46) Huwe, J. K.; Feil, V. J.; Zaylskie, R. G.; Tiernan, T. O. Chemosphere 2000, 40, 957-962. (47) Lamparski, L. L.; Stehl, R.; Johnson, R. L. Environ. Sci. Technol. 1980, 14(2), 196. (48) Vollmuth, S.; Zajc, A.; Niessner, R. Environ. Sci. Technol. 1994, 28(6), 1145. (49) Seip, H. M.; Alstad, J.; Carlberg, G. E.; Martinsen, K.; Skaane, R. Sci. Total Environ. 1986, 50, 87. (50) Lee, L. S.; Rao, P. S. C.; Nkedi-Kizza, P.; Delfino, J. Environ. Sci. Technol. 1990, 24(5), 654. (51) Jackson, D. R.; Bisson, D. L. J. Air Waste Manage. Assoc. 1990, 40, 1129. (52) Hagenmaier, H.; She, J.; Lindig, C. Chemosphere 1992, 25(710), 1449. (53) Mackay, D.; Shiu, W. Y.; Ma, K. C. Illustrated handbook of physical-chemical properties and environmental fate for organic chemicals; 1995; Vol. 4. (54) Wang, J.; Wang, J.-M.; Banerji, S. K.; Bajpai, R. K. Appl. Biochem. Biotechnol. 1995, 51/52, 693. (55) Masunaga, S.; Yao, Y.; Ogura, I.; Nakai, S.; Kanai, Y.; Yamamuro, M.; Nakanishi, J. Environ. Sci. Technol. 2001, 35(10), 1967. (56) Sakurai, T.; Suzuki, N.; Masunaga, S.; Nakanishi, J. Chemosphere 1998, 37(9-12), 2211. (57) Hagenmaier, H.; Brunner, R. Chemosphere 1987, 16, 1759. (58) Kookana, R. S.; Baskaran, S.; Nidu, R. Aust. J. Soil Res. 1998, 36(5), 715. (59) ABS, Australian Bureau of Statistics. Consultancy data report, 2002. (60) Prange, J.; Gaus, C.; Pa¨pke, O.; Weber, R.; Mueller, J. F. Organohalogen Compd., in press.

Received for review March 26, 2002. Revised manuscript received May 29, 2002. Accepted June 4, 2002. ES025674J

VOL. 36, NO. 16, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3549