Environ. Sci. Technol. 2001, 35, 1974-1981
Further Evidence for the Existence of PCDD/Fs in the Environment Prior to 1900 N I C H O L A S J . L . G R E E N , * ,† JOANNE L. JONES,† A. E. JOHNSTON,‡ AND KEVIN C. JONES† Institute of Environmental and Natural Sciences, Lancaster University, Lancaster LA1 4YQ, U.K., and Institute of Arable Crop Research, Rothamsted Experimental Station, Harpenden Hertfordshire AL5 2JQ, U.K.
PCDD/Fs and PCBs have been analyzed in a series of archived soil samples collected from various depths during the 1800s and early 1900s. PCBs were not found in soil samples collected before 1900, whereas PCDD/Fs were present in concentrations between 43 and 110 pg/g in surface soils, and between 9 and 150 pg/g in soils collected from below the surface. The PCDD/F homologue patterns of all surface soils were consistent with each other. The homologue pattern of deeper soils altered with depth to one that was dominated by highly chlorinated PCDDs. The highest Σ(4-8)PCDD/F concentration (150pg/g) was found in the deepest soil analyzed (230-250 cm below the surface). The cork from one of the storage bottles contained considerable quantities of both PCBs and PCDD/Fs. However, contamination of the soils, either by diffusion through the cork or by cork particles, was discounted on the basis that no PCBs were evident in the soil, and that the PCDD/F homologue pattern in the cork was very different to that found in the soil. Similar arguments were used to discount contamination of the soil by dust. A sample of ashed vegetation from the archive, that had no cork stopper, contained high concentrations of PCBs (78 ng/g), concentrations of mono- to tri-CDFs that were higher than in any of the soils (190 pg/g), but very low concentrations of Σ(4-8)PCDD/F (12 pg/g). This pattern of analytes was considered to be representative of contamination from store room air and was completely different from the pattern observed in the soils. Taken together these observations indicate that contamination during storage, or subsequent handling, is unlikely to have occurred in archived soil samples that were stored with cork and wax seal intact. The results imply surface soil Σ(4-8)PCDD/F concentrations of around 60 pg/g at Rothamsted (southeast England) in the late 1800s, compared with ∼300 pg/g reported for rural UK soils in the 1990s.
Fs to the modern environmentscombustion processes and their occurrence as impurities in a range of deliberately manufactured chlorinated aromatic products, such as chlorophenols and polychlorinated biphenyls. The latter category of source(s) is obviously of fairly recent origin, related to the inception of manufacturing processes for these compounds. Large-scale combustion processes, such as the burning of coal, wood, petroleum, and the smelting of metals are processes which have been occurring for much longer. Indeed the smelting of metals in Europe goes back to pre-Roman times, while the burning of wood, volcanic emissions, etc., are “natural combustion processes”, which have no meaningful chronological bounds. PCDD/F source inventories point to the burning of municipal waste, domestic burning, and metal smelting as the most important primary atmospheric emissions of PCDD/Fs to the modern environment. It may therefore seem reasonable to assume that these processes have been important, ongoing (though variable through time) sources of PCDD/Fs to the environment. The potential for purely biogenic formation of some PCDD/F congeners has been mooted by various workers (e.g., ref 1). It remains difficult to demonstrate categorically that PCDD/ Fs were present prior to the last century, because of uncertainties about postcollection contamination of samples. One of our previous papers reporting the occurrence of PCDD/Fs in former times by using archived soils (2) attracted criticism (3). It was suggested that the presence of PCDD/Fs in soil and vegetation samples collected in the mid-1800s was a result of contamination from more recent times. Since then a number of direct studies and anecdotal implications have been published which all lend credence to the hypothesis that PCDD/Fs are not entirely a recent phenomenon (for a brief review, see 4). In this paper we report the findings of a study designed to shed more light on this topic. We previously published results of PCDD/Fs found in archived soils from the 1800s (5). These soils were dried at the time of collection and stored in glass bottles with cork bungs. Some of these bungs had also been wax-sealed and lead-capped. The soils form part of the unique Rothamsted Archive. The particular soils analyzed previously had originally been sub-sampled for metals analysis in 1986 (6). The remainder of the subsample, not used for metals analysis, was then stored for 5 years in paper bags before being analyzed for PCDD/Fs. This handling procedure caused some ambiguity as to the validity of the PCDD/Fs detected in these soils. It was suggested that the results might be an artifact of postcollection contamination rather than an indication of the presence of PCDD/Fs in the 1800s (3). Some work was undertaken to address the possibility of contamination (2) but some questions remained. The purpose of the current study was, therefore, 2-fold: (i) to investigate more fully whether contamination of the soils from the Rothamsted Archive had occurred and (ii) to obtain further evidence for or against the hypothesis that PCDD/Fs existed in the U.K. environment prior to 1900.
Materials Introduction PCDD/Fs are ubiquitous contaminants present at ultratrace concentrations in the modern environment. Source inventories suggest that there are two principal sources of PCDD/ * To whom correspondence should be addressed. E-mail: n.green@ lancaster.ac.uk. † Institute of Environmental and Natural Sciences. ‡ Institute of Arable Crop Research. 1974
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 10, 2001
Soils. To test the validity of our previous results we analyzed archived soil samples with demonstrably different recent storage histories. Some had been stored for the past 12 years in glass jars with foil-lined lids and some in sealed plastic bags, and others had remained in their original, sealed glass bottles and were sampled directly from them for this study. The set of samples analyzed comprised nine surface (0-23 cm) soils, four subsurface (23-46 cm) soils, and four deep (up to 250 cm) soils collected between 1850 and 1938 from 10.1021/es0002161 CCC: $20.00
2001 American Chemical Society Published on Web 04/12/2001
TABLE 1. Samples Analyzed, with Comments on Their Storagea sample
date of collection
depth (cm)
site
storage comments
R56a R79a R81a R93a R93b R93c R93d R04a R13a W76a W76b W76c W76d W27a W27b W38a W38b DustR DustW AshR Cork (U) Cork (L)
1856 1879 1881 1893 1893 1893 1893 1904 1913 1876 1876 1876 1876 1927 1927 1938 1938
0-23 0-23 0-23 0-23 23-46 46-69 231-251 0-23 0-23 0-23 23-46 46-69 92-115 0-23 23-46 0-23 23-46
Bk G Bk Bk Bk Bk Bk BW HB S S S S S S S S R W R G G
glass jar/foil lined lid since 1986 original wax and lead seals in tact glass jar/foil lined lid since 1986 plastic bag since 1986 plastic bag since 1986 plastic bag since 1986 plastic bag since 1986 glass jar/foil lined lid since 1986 glass jar/foil lined lid since 1986 original wax seal intact original wax seal intact original wax and lead seals intact original wax seal intact original corked bottle, no seal original corked bottle, no seal original corked bottle, no seal original corked bottle, no seal brushed from several Rothamsted bottles brushed from several Woburn bottles found uncorked in store upper, exposed half (incl. wax seal) lower, unexposed half (no seal)
1856 1879 1879
a Bk ) Broadbalk, BW ) woodland adjacent to Broadbalk, HB ) Hoosfield, and G ) Dr. Gilbert’s Meadow are all from Rothamsted (R). S ) Stackyard from Woburn (W).
two farms (Rothamsted and Woburn) approximately 20 km apart. The samples came from several differently treated experimental plotssBroadbalk (Bk), woodland adjacent to Broadbalk (BW), Hoosfield (HB) and Dr. Gilbert’s Meadow (G) at Rothamsted, and Stackyard (S) at Woburn. Additionally, a sample of ashed vegetation was found in the Archive store that had no stopper. This was analyzed to indicate what PCDD/F concentrations might be expected for a contaminated sample. A “field blank” was generated by laying sodium sulfate on a dish in the archive storage hall for ca. 24 h while the archive samples were being obtained. This was stored in a glass jar and reopened at the laboratory while the archive storage jars were opened and sampled. A map locating the two farms can be found in a previous article (2). Storage Jar Corks and Dust. To test the potential for PCDD/Fs to diffuse through the sealed cork during the ∼150 years of storage, we analyzed sections of the cork from one of the bottles. This also gave information on the potential for contamination of the soils by fragments of cork falling into the soil when the bottles were opened. The corks were so dry and brittle that some of them sheared horizontally when opening the oldest bottles. The one taken for analysis sheared at the point of tightest seal, into two roughly equal parts. These were termed upper cork (outermost) and lower cork (innermost). Similarly, we analyzed samples of dust collected from the surface of the bottles to address the possibility of soils becoming contaminated by dust falling into the soil on opening. A list of all the samples and their storage history is provided in Table 1. Supplementary Analysis of PCBs. In addition to PCDD/ Fs, PCBs were analyzed in many of the samples. PCBs were first manufactured around 1930, and in the U.K. between 1954 and 1977. They are detected in modern ambient air at concentrations that are around 3 orders of magnitude higher than those of PCDD/Fs. PCBs, therefore, offer a much more sensitive measure of postcollection contamination of the archived samples. If the soils were to have been contaminated by PCDD/Fs of a more recent origin then they would also be contaminated by PCBs at very much higher concentrations. Conversely, if PCBs were not found in the soils then whatever PCDD/Fs were detected could be regarded as having been present in the soil at the time of their original bottling.
Analysis All samples were dried at the time of their original collection. Subsamples of approximately 100 g of soil were weighed directly into extraction thimbles. Cork sections were extracted whole. In anticipation of high concentrations in the dust, 1 g samples of dust were extracted and 1/1000th of the extract was spiked with surrogates, cleaned and quantified. The entire ashed vegetation sample (ca 6 g) was taken for analysis. Each sample was spiked with 13C12-labeled surrogates of all 17 2,3,7,8-substituted PCDD/F congeners and three dior trichlorinatedDD/Fs. Those samples in which PCBs were analyzed were additionally spiked with seven 13C12-labeled PCB congeners. Samples were Soxhlet extracted in toluene overnight. Clean up was by elution through a mixed column containing 44% (w/w) sulfuric acid/silica gel, 33% 1 N sodium hydroxide/silica gel and activated silica gel. Fractionation of PCDD/Fs from PCBs was achieved for all samples using Alumina B Super-1 (ICN Biomedicals). An injection standard of 37Cl4-labeled 2,3,7,8-TCDD was added to both the PCDD/F and PCB fractions, and the final extracts were concentrated to ca. 15 µL. PCDD/Fs and PCBs were quantified by HRGC/HRMS using an HP6890 gas chromatograph connected to a Micromass Autospec Ultima mass spectrometer tuned to 10 000 resolving power and running in selected ion monitoring mode. PCBs and total PCDD/F homologues were quantified using a 30 m HP5-ms capillary column, individual PCDD/F congeners were quantified using a 60 m SP2331 column. As such, 1,2,3,7,8-PeCDF is not resolved from 1,2,3,4,8-PeCDF, and 1,2,3,4,7,8-HxCDF is not resolved from 1,2,3,4,7,9HxCDF. Method Validation. Five subsamples of a modern soil were fortified at different levels with unlabeled 2,3,7,8substituted PCDD/Fs. The degree of fortification ranged from 2 to 2000% of the native congener concentrations, as identified by triplicate analyses of the unspiked soil. Standard deviations of congener concentrations for the triplicate analyses were between 2 and 9%. Measured values for the spiked soils were between 87 and 125% of the anticipated values for all congeners (mean value 98%). Recoveries of unlabeled PCDD/Fs from a spiked blank analyzed at the same time were between 85 and 103%. VOL. 35, NO. 10, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1975
TABLE 2. PCDD/F Concentrations (pg/g) in Soils from Rothamsted and Woburn Stackyard 1856
1881
Rothamsted 1904 1913
1893a
1893b
1893c
1893d
depth (cm)
0-23
0-23
0-23
0-23
0-23
23-46
46-69
231-251
congener 2,3,7,8-TCDF 1,2,3,7,8-PeCDF 2,3,4,7,8-PeCDF 1,2,3,4,7,8-HxCDF 1,2,3,6,7,8-HxCDF 1,2,3,7,8,9-HxCDF 2,3,4,6,7,8-HxCDF 1,2,3,4,6,7,8-HpCDF 1,2,3,4,7,8,9-HpCDF OCDF
0.56 0.76 0.62 0.74 0.52 0.03 0.50 1.75 0.17 1.16
0.45 0.60 0.53 0.66 0.47 0.03 0.50 1.89 0.13 1.21
0.50 0.68 0.65 0.81 0.64 0.06 0.66 2.37 0.18 1.51
1.54 1.18 0.91 1.08 0.82 0.09 0.76 3.34 0.32 3.31
0.50 0.48 0.49 0.64 0.48 0.03 0.49 1.94 0.18 1.52
0.24 0.07 0.06 0.07 0.05 0.01 0.05 0.39 0.02 0.15
0.15 0.05 0.04 0.05 0.04 0.02 0.04 0.42 0.01 0.20
0.11 0.04 0.02 0.02 0.02 0.01 0.02 0.30 0.01 0.17
2,3,7,8-TCDD 1,2,3,7,8-PeCDD 1,2,3,4,7,8-HxCDD 1,2,3,6,7,8-HxCDD 1,2,3,7,8,9-HxCDD 1,2,3,4,6,7,8-HpCDD OCDD
0.03 0.15 0.17 0.16 0.18 1.5 10.7
0.03 0.15 0.14 0.19 0.23 2.5 14.0
0.04 0.20 0.24 0.36 0.32 3.1 21.9
0.07 0.31 0.32 0.50 0.46 4.1 23.6
0.04 0.14 0.18 0.37 0.26 2.1 10.3
0.03 0.06 0.09 0.31 0.22 2.4 23.0
0.02 0.06 0.09 0.25 0.25 3.1 67.1
0.03 0.10 0.13 0.52 0.48 5.7 121
10.3 5.0 4.8 7.5 7.1 5.4 1.9 1.2
67.1 33.8 14.0 7.2 6.1 4.9 1.5 1.2
12.0 4.5 4.2 6.7 6.1 5.8 3.2 1.5
19.8 12.7 10.8 11.2 9.4 7.8 4.0 3.3
12.9 36.6 7.7 5.7 5.5 4.9 2.5 1.5
9.9 30.5 6.2 1.8 0.88 0.73 0.29 0.15
7.3 22.4 4.6 1.3 0.54 0.42 0.27 0.20
2.7 15.1 3.0 1.0 0.45 0.35 0.39 0.17
mono-dioxins di-dioxins tri-dioxins tetra-dioxins penta-dioxins hexa-dioxins hepta-dioxins OCDD
0.19 0.29 0.48 1.4 1.7 3.2 4.7 10.7
1.00 2.39 1.37 2.2 1.7 5.0 6.9 14.0
0.19 0.34 0.45 2.3 3.0 5.5 7.0 21.9
0.28 0.61 0.89 3.1 4.1 7.7 9.5 23.6
0.19 1.04 0.45 1.6 2.2 4.2 4.7 10.3
0.12 1.04 0.29 1.1 1.0 3.5 6.5 23.0
0.11 0.87 0.25 0.93 0.80 3.0 8.7 67.1
0.07 0.40 0.22 0.95 1.04 6.3 18.3 121
ΣP(4-8)CDD/F ΣP(1-8)CDD/F
45 66
homologue mono-furans di-furans tri-furans tetra-furans penta-furans hexa-furans hepta-furans OCDF
51 170
39 87
83 119
150 172
1927
1938
1938
23-46
46-69
92-115
0-23
23-46
0-23
23-46
0.58 0.51 0.39 0.51 0.40 * 0.50 1.99 0.15 1.71
0.05 0.07 0.05 0.06 0.06 0.12 0.09 0.22 * 0.11
0.03 0.05 0.03 * 0.04 0.08 0.06 0.22 0.01 0.10
0.02 0.01 0.01 * 0.02 0.04 0.02 0.05 * 0.06
0.50 0.97 0.66 0.87 0.60 0.11 0.59 2.80 0.25 3.24
0.13 0.18 0.12 0.16 0.13 0.13 0.12 0.49 0.04 0.45
0.44 0.69 0.53 0.63 0.48 * 0.44 2.19 0.18 2.37
0.07 0.13 0.09 0.12 0.10 0.18 0.10 0.46 0.03 0.30
0.02 0.13 0.28 0.36 0.51 3.4 11.4
* 0.03 0.14 0.15 0.32 1.9 7.8
0.00 0.03 0.10 0.11 0.24 1.6 8.7
0.00 0.01 0.04 0.05 0.09 0.60 6.7
0.05 0.22 0.28 0.43 0.43 3.1 9.3
0.01 0.05 0.14 0.16 0.27 1.9 7.7
0.03 0.18 0.31 0.36 0.48 3.4 11.1
0.01 0.05 0.22 0.21 0.44 2.9 12.0
29.3 12.3 3.7 5.0 4.5
8.2 2.3 0.55 0.54 0.59
8.0 2.3 0.41 0.33 0.40
4.7 2.3 0.21 0.11 0.08
4.6 1.9 0.99 1.7 1.4
6.8 3.9 3.8 6.4 6.0
1.7 0.87 0.90 1.1 1.2
depth (cm)
0-23
congeners 2,3,7,8-TCDF 1,2,3,7,8-PeCDF 2,3,4,7,8-PeCDF 1,2,3,4,7,8-HxCDF 1,2,3,6,7,8-HxCDF 1,2,3,7,8,9-HxCDF 2,3,4,6,7,8-HxCDF 1,2,3,4,6,7,8-HpCDF 1,2,3,4,7,8,9-HpCDF OCDF 2,3,7,8-TCDD 1,2,3,7,8-PeCDD 1,2,3,4,7,8-HxCDD 1,2,3,6,7,8-HxCDD 1,2,3,7,8,9-HxCDD 1,2,3,4,6,7,8-HpCDD OCDD
9
43 102
1927
1876
1976
84 129
Woburn Stackyard 1876 1876
1876
homologue mono-furans di-furans tri-furans tetra-furans penta-furans
63 85
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 10, 2001
12.2 6.9 5.9 9.7 8.1
TABLE 2 (Continued) Woburn Stackyard (cont.) 1904 1913
1856
1881
homologue (cont.) hexa-furans hepta-furans OCDF
4.0 2.3 1.7
0.54 0.26 0.11
0.42 0.12 0.10
mono-dioxins di-dioxins tri-dioxins tetra-dioxins penta-dioxins hexa-dioxins hepta-dioxins OCDD
0.63 0.46 0.44 1.4 2.2 4.4 5.8 11.4
0.13 0.09 0.06 0.21 0.37 1.5 2.9 7.8
0.14 0.05 0.05 0.17 0.39 1.3 2.8 8.7
ΣP(4-8)CDD/F ΣP(1-8)CDD/F
43 90
15 26
15 26
1893a
1893b
1893c
1893d
0.12 0.03 0.06
6.6 3.3 3.2
1.3 0.63 0.45
4.9 2.6 2.4
1.1 0.59 0.30
0.07 0.05 0.01 0.05 0.09 0.49 1.2 6.7
0.63 0.63 1.1 3.2 5.4 8.2 6.5 9.3
0.26 0.18 0.22 0.54 0.90 2.0 3.2 7.7
0.40 0.43 0.53 1.6 3.2 5.6 6.1 11.1
0.09 0.09 0.09 0.31 0.62 2.2 4.2 12.0
50 66
24 27
9 16
63 91
20 28
TABLE 3. PCDD/F and PCB Homologue Concentrations in Test Materials from the Rothamsted Archive soil 1879 (G) (pg/sample)
material
upper cork 1879 (G) (pg/sample)
lower cork 1879 (G) (pg/sample)
uncorked ashed leaves (pg/sample)
dust Rothamsted (pg/sample)
dust Woburn (pg/sample)
field blank 100 g (pg/sample)
lab blank 100 g (pg/sample)
amount analyzed
91 g
half cork
half cork
6g
1g
1g
Na2SO4
Na2SO4
PCDD/Fs mono-furans di-furans tri-furans tetra-furans penta-furans hexa-furans hepta-furans OCDF mono-dioxins di-dioxins tri-dioxins tetra-dioxins penta-dioxins hexa-dioxins hepta-dioxins OCDD ΣP(4-8)CDD/F ΣP(1-8)CDD/F
1760 960 600 1210 1750 1700 770 200 22 45 63 280 540 1010 920 1560 9940 (108/g) 13 400
1060 1040 540 305 167 102 89 43 19 55 48 79 62 69 53 72 1040 3800
508 135 27 12 6 3 9 16 4 4 3 2 1 1 4 25 79 760
817 270 61 30 11 6 3 3 2 12 5 4 0.4 2 3 10 72 (12/g) 1240
5300 10 200 3500 2800 6800 1030 31 500 21 900 * 820 430 710 650 2900 10 600 25 600 114 000 134 000
5700 19 100 12 800 9100 5400 10 600 44 100 25 600 370 2100 1300 1100 57 6000 15 900 35 100 153 000 194 000
1.2 9 0.8 0.4 0.2 0.3 * 1.5 * * 0.1 * 0.1 * 1.8 8 13 23
1.1 31 0.1 0.0 0.0 0.1 0.6 0.7 1.4 0.4 0.1 0.0 0.0 0.1 0.3 3 5 39
material
soil 1879 (G) (ng/sample)
upper cork 1879 (G) (ng/sample)
lower cork 1879 (G) (ng/sample)
uncorked ashed leaves (ng/sample)
dust Rothamsted (ng/sample)
dust Woburn (ng/sample)
field blank 100 g (ng/sample)
lab blank 100 g (ng/sample)
PCBs tri-CBs tetra-CBs penta-CBs hexa-CBs hepta-CBs octa-CBs ΣPCB
2 2 1 0.7 0.3 0.1 6
311 797 382 123 20 3 1640
34 23 7 2.4 0.4 0.04 67
250 178 35 7 0.6 0.04 470 (78/g)
2780 5840 3680 1970 464 68 14 800
3310 7110 4060 2020 680 138 17 300
4 5 2 1 0.3 0.02 12
4 5 2 1 0.2 0.02 12
A sediment sample was analyzed blind as part of an international interlaboratory calibration exercise [described by van Bavel et al. (7)]. The analytical procedure was the same as was adopted for this study except that two further cleanup steps were used to eliminate sulfur and hydrocarbons from the sediment. Values measured for all 2,3,7,8-substituted congeners and all PCDD/F homologues were within 0.4 standard deviations of the respective median values. On account of their high levels in the ambient environment, quantifiable concentrations of PCBs in laboratory blanks and field blanks were found. Detection limits were set at three times the mean blank concentration. For PCB congeners found in higher concentrations than the limit of detection, values are quoted as the quantified amount less
the level quantified in the laboratory blank associated with that sample. Blank concentrations of PCDD/Fs were generally nondetectable. The blank peak, or noise, values were used to generate limits of detection, calculated as three times the standard deviation plus the mean. PCDD/F concentrations in samples are not blank-corrected.
Results Surface Soils. ΣP(4-8)CDD/F concentrations in the nine samples analyzed ranged between 40 and 110 pg/g dry soil (Tables 2 and 3). No clear increase in concentration was observed with time at Woburn, even though the set contained soils collected in 1927 and 1938. At Rothamsted, an upward trend was visible in surface soil ΣP(4-8)CDD/F concentrations, VOL. 35, NO. 10, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1977
FIGURE 1. Surface (0-23 cm) and subsurface (23-46 cm) soil concentrations from Woburn Stackyard at three dates. excluding Dr. Gilbert’s Meadow (1879) (Table 3). These results are generally slightly higher than those reported for the original set of soils analyzed from the Rothamsted archive by Kjeller et al. (5). No apparent consistent differences in concentration or homologue pattern were found between surface soil samples from the original storage jars and those stored in plastic bags for several years. Rural U.K. soils collected in the 1990s have been reported as typically containing ∼300 pg/g ΣPCDD/F (8), roughly five times higher than we found for archived soils. The homologue patterns for the surface soils from Rothamsted were consistent with those from Woburn (Table 2). Subsurface Soils. The ΣP(4-8)CDD/F concentrations in the three subsurface soils (23-46 cm) were about one-third of those found in the respective surface soil (Table 2). The value in 1938 was closer to half the surface concentrations, mainly due to increased OCDD. The homologue patterns of the three subsurface soils were consistent with each other but were markedly different from the surface soils. The subsurface patterns were dominated by highly chlorinated PCDDs, with all PCDFs making only minor contributions (Figure 1). Deep Soils. The shift in homologue pattern in favour of Hx-OCDD that was observed between surface and subsurface soils was amplified with increasing depth in both the deep soil sets (Figures 2 and 3). At the Rothamsted site concentrations of Hx-OCDDs increased with depth, with the ΣP(4-8)CDD/F concentrations reaching a maximum of 150 pg/g at 1978
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 10, 2001
a depth of 2.3m. This was the highest ΣP(4-8)CDD/F concentration measured in all the soil samples we analyzed from the Rothamsted archive. At the Woburn site, the homologue pattern changed with depth in a similar way to the Rothamsted series, although at Woburn the concentrations decreased with increasing depth. Lower Chlorinated PCDD/Fs. Mono- to triCDD/Fs are measured as a routine part of our analysis, although these rely on just three commercially available labeled surrogates for quantitation. The trends observed for these homologues were a consistent extension of what was seen for the tetraocta homologues. In common with the tetra- and pentaCDD/ Fs, concentrations of the less chlorinated congeners decreased with depth. All data pertaining to these homologues are presented in Tables 2 and 3, but ΣP(4-8)CDD/F concentrations are used in the text to allow comparison to other studies on PCDD/Fs in soils. The mono- to triCDD/F values are included in Figures 2-4. The exclusion of mono- and diCDD/Fs from Figure 1 is for purposes of clarity. Their high concentrations relative to tri- through OCDD/Fs would stunt the scale of the graphs and so obscure the data for the higher chlorinated PCDD/Fs. PCBs in Archived Soils. Soils from the Woburn field were analyzed for PCBs at the same time as PCDD/Fs, and these are presented in Table 4. No PCBs were found above the detection limits in any of the soils from 1876. Small amounts of triCBs were found in both the surface and subsurface soils of 1927, 0.1 and 0.2 ng/g, respectively. The surface soil from 1938 contained 0.6 ng/g triCB and 0.15 ng/g tetraCB, but concentrations of PCBs in the subsurface soil did not change between 1927 and 1938. For comparison, rural surface soils in the U.K. had a mean ΣPCB concentration of ∼4 ng/g in 1993 (8) and were higher than that in the 1960s (9). Test Materials. The results for the two dust samples, from Rothamsted and Woburn were almost identical (Table 3). ΣP(4-8)CDD/F concentrations were 114 000 pg/g, around 3 orders of magnitude higher than was found in the soils. The homologue pattern was unlike any of the soil samples analyzed, with PCDF concentrations increasing with degree of chlorination. PCBs were measured in the dust samples at 14 800 000 pg/g (15 µg/g). The two sections of cork were roughly the same weight. The upper half (exposed to the atmosphere) contained 1040 pg of ΣP(4-8)CDD/F and 1 640 000 pg of ΣPCB in total, while the lower half contained 79 pg of ΣP(4-8)CDD/F and 66 600 pg of ΣPCB (Table 3). The two sections also had differences in their homologue profiles (as seen in Figure 4). The sample of ashed vegetation that was found uncorked in the archive store contained 12 pg/g of ΣP(4-8)CDD/F, less than was found in all bar one of the soils. By contrast it contained higher levels of mono- and dichlorofurans (180 pg/g) and ΣPCBs (78 000 pg/g) than the soils (Table 3).
Discussion Potential Contamination. It has been suggested that the archived soils in the Rothamsted archive could have been contaminated during storage or subsequent handling. Indeed the original samples analyzed by Kjeller et al. (5) had been kept in paper bags in a cupboard at Lancaster for six years before they were analyzed. Some of the soils analyzed as part of this study had also been stored at Lancaster, although in plastic bags. The rest of the soils were taken directly from the original storage bottles, breaking the wax (and lead) seals to do so. Increasing concentrations of OCDD with depth at Rothamsted might be considered to be due to permeation through the plastic bags from dust particles during storage at Lancaster. However, this can be dismissed because a similar increase in OCDF would then be expected from the homologue profile found in dust (Table 4 and ref 2). Furthermore, use of such an argument to account for the change in
FIGURE 2. PCDD/F homologue patterns in soil depths (Rothamsted Broadbalk).
FIGURE 3. PCDD/F homologue patterns in soil depths (Woburn Stackyard). homologue pattern with depth at Rothamsted (Figure 2) is inconsistent with the same homologue pattern change being observed at Woburn (Figure 3), where the soils never came into contact with plastic bags. The fact that these two classes of soils gave very comparable results suggests that the storage at Lancaster did not compromise the integrity of the PCDD/F record of the soils. It has also been suggested that on uncorking the bottles a minute amount of dust falling into the soil would cause severe contamination. The original bottles held several kilograms of soils. If, however, all the dust settled in the 100 g samples we took for analysis it would have taken just 40 mg of dust to account for the ΣP(4-8)CDD/F concentrations observed. The differences in homologue patterns between dust and soils (Figures 4) indicates that the dust is not responsible for the observed PCDD/Fs in the soils. Furthermore, if the soils had received 40 mg of
dust we would have measured 600 000 pg of PCBs in the soil sample whereas no PCBs were detectable in the soils collected before 1900, and only very minor amounts in the early 1900s. A similar argument could be used to refute a further suggestion that PCDD/Fs might have permeated through the cork bung and into the soil during storage. We found both PCDD/Fs and PCBs in the lower half of the cork that was analyzed, and it is possible that they migrated through the cork. It appears unlikely that they were then able to transfer from the cork to the soil, however, as no PCBs were found in the soil. Another piece of evidence that implies the soils were not contaminated in storage comes from the analysis of an uncorked sample. It is not known how long this sample had been left uncorked but it contained less ΣP(4-8)CDD/F than all bar one of the corked soils. It did, however, contain considerable quantities of PCBs. This VOL. 35, NO. 10, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1979
FIGURE 4. PCDD and PCDF homologue patterns in test materials.
TABLE 4. PCB Concentrations (pg/g) in Soils from Woburn Stackyarda
depth (cm) TriCBs TetraCBs PentaCBs HexaCBs HeptaCBs OctaCBs ΣPCB
1876
1876
1876
1876
1927
1927
1938
1938
blank equivalent
0-23 * * * * * * 0
23-46 * * * * * * 0
46-69 * * * * * * 0
92-115 * * * * * * 0
0-23 131 * * * * * 131
23-46 196 10 * * * * 206
0-23 602 149 2 * * * 752
23-46 184 5 * * * * 190
29 26 12 7 2 0 76
observation is in keeping with our findings for the contamination of soils by laboratory air (2, 10). In that experiment we left soil on a work bench in the laboratory and took increments of it for analysis over time. The soil picked up PCBs from the laboratory air very quickly, whereas PCDD/F concentrations changed little, if at all. Small amounts of triCBs were detected in both the surface and subsurface soils from 1927. Commercial PCB formulations were not manufactured at this time, and so it may be possible that these particular soils have indeed been exposed to some atmospheric PCBs in storage. Although tightly corked, the storage bottles of these soils were not wax sealed and it may be that this omission has allowed some minor contamination. Similar amounts of triCBs were found in the subsurface soil from 1927 and the same argument could apply to this sample. The higher concentrations of TriCBs, and presence of tetraCBs, in the surface soil from 1938 may be the combined effect of minor exposure in storage with genuine environmental contamination of surface soils in the U.K. by this time. Although PCBs were not manufactured in the U.K. until 1954, they were imported prior to this and also may have undergone long-range transport from source areas. Time-trends of environmental concentrations of PCBs in the U.K. have been reconstructed from their record in a dated rural lake sediment core (11), and concentrations were first seen to rise above their background levels around the 1930s. If these samples have been partially exposed to atmospheric PCBs during storage, this appears to have had no impact on their PCDD/F concentrations. The more volatile, lower chlorinated PCDD/F congeners were present in the 1927 and 1938 samples at very similar (or lower) concentrations to the equivalent samples collected in 1876, although no PCBs were quantifiable in these earliest soils. 1980
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 10, 2001
The consistency of the results for PCDD/Fs in all surface soils collected between 1856 and 1938 could be interpreted as being indicative of ubiquitous contamination of the Rothamsted Archive samples. However, the results of the subsurface and deep soils emphatically counter such an argument. The four subsurface soils consistently display a homologue pattern (as well as absolute concentrations) that is distinct from that of respective surface soils (see Figure 1). Furthermore, the homologue profiles of the deeper soils are also consistently distinct from that of either the surface or subsurface soils. Historical Record. The soil samples in the Rothamsted Archive can be considered to be free from post-collection PCDD/F contamination and therefore provide a reasonable indication of the amount of PCDD/Fs in the environment at the time of their collection. Half-lives of PCDD/Fs in soils are not known precisely (12) but are considered to be of the order of 10 years, maybe longer. PCDD/Fs have few elimination pathways in soils. They are very strongly bound and it is currently thought that volatilisation back to the atmosphere is minimal at the present time (13). They do not percolate through the soil into the aquifer, and microbial degradation under aerobic conditions is minor. Unlike the atmosphere or human tissues, there is as yet no indication that PCDD/F concentrations in background soils have declined over the past 10-20 years. Following exposure to the PCDD/F emissions of the last (twentieth) century, typical ΣP(4-8)CDD/F concentrations in rural U.K. soils in the 1990s were around 300 pg/g dry weight (8). A direct comparison with the data from the archived surface soils indicates that PCDD/F concentrations in the second half of the nineteenth century were ∼20% of modern soils. This implies PCDD/F environmental contamination
was considerable before the introduction of many of the major “known” PCDD/F source activities, including largescale municipal waste incineration, electricity generation and the chlorine industry (14). Results from the deep soils are particularly intriguing. Concentrations of all PCDFs decreased with depth at the Rothamsted site, and the higher the degree of chlorination the more quickly the concentration decreased. The monoto tetraCDDs were also seen to decrease in concentration with depth. However, the hepta- and octaCDDs increased dramatically with depth. Penta- and hexaCDDs first decreased, then increased with depth. The net effect on ΣPCDD/F is an initial decline with depth followed by a subsequent increase, such that the deepest soil was the most contaminated of all the soils analyzed in this study. Although the same changes to the homologue pattern were seen in the Woburn set, the absolute concentrations did not increase with depth. This might imply that the same processes were operating at both sites but to different degrees. It is difficult to account for the observations in terms of all the PCDD/Fs having derived from surface contamination. If surface PCDD/Fs were to be leached through the soil in drainage water, it would be expected that the lower chlorinated congeners would have increased most and that these would have a greater influence on the homologue pattern at depth. The converse is seen for the PCDDs. If surface PCDD/ Fs moved through the soil associated with very fine particles or colloids, concentrations would be expected to drop off exponentially with depth, even if the movement was in water moving through the soil profile by preferential flow. Movement of PCDD/Fs by association with colloids or dissolved organic matter would not bring about the homologue pattern changes observed with depth. Certainly no kind of mechanical mixing or worm action could account for the observations. If PCDD/Fs present at depth are not a result of migration of surface contamination, they may constitute a natural part of this soil. The soil in this part of the U.K. is thought to have been laid down as alluvial deposits as the glaciers of the last ice-age retreated. It is therefore possible that the OCDD measured in the deep soil is a remnant of PCDD/Fs present over 10 000 years ago or may have been formed by poorly understood processes in the subsurface soil environment. The possibility of PCDD/Fs being present in very old material has been raised by other researchers. For example, Ferrario et al. (15) and Jobst and Aldag (16) found PCDD/Fs in clay deposits in the U.S. and Germany, respectively. Rappe et al. (17) and Gaus et al. (18) have reported PCDD/Fs
(principally OCDD) in old sediments from the U.S. and Australia, respectively, and Hashimoto et al. (19) reported PCDD/Fs in Pacific Ocean sediment that was deposited around 8000 years ago.
Acknowledgments We thank the Trustee Directors of the Lawes Agricultural Trust for the samples used in these analyses, and Eurochlor for providing funding for this project.
Literature Cited (1) Hoekstra, E. J.; de Weerd, H.; de Leer, E. W. B.; Brinkman, U. A. T. Environ. Sci. Technol. 1999, 33, 2543-2549. (2) Alcock, R. E.; McLachlan, M. S.; Johnston A. E.; Jones, K. C. Environ. Sci. Technol. 1998, 32, 1580-1587. (3) Baker, J. L.; Hites, R. A. Environ. Sci. Technol. 1999, 33, 205. (4) Green, N. J. L.; Alcock, R. E.; Johnston, A. E.; Jones, K. C. Chemosphere 2001 (submitted for publication). (5) Kjeller, L.-O.; Jones, K. C.; Johnston, A. E.; Rappe, C. Environ. Sci. Technol. 1991, 25, 1619-1627. (6) Jones, K. C.; Symon C. J.; Johnston, A. E. Sci. Total Environ. 1987, 61, 131-144. (7) van Bavel, B.; Rappe, C.; Tysklind, M.; Takeda, N. Organohalogen Compd. 1999, 40, 297-300. (8) Her Majesty’s Inspectorate of Pollution (HMIP). Determination of PCDDs and PCDFs in UK soil; HMSO: London, 1995. (9) Lead, W. A.; Steinnes, E.; Bacon, J. R.; Jones, K. C. Sci. Total Environ. 1997, 193, 229-236. (10) Alcock, R. E.; McLachlan, M. S.; Johnston A. E.; Jones, K. C. Environ. Sci. Technol. 1994, 28, 1838-1842. (11) Gevao, B.; Hamilton-Taylor, J.; Murdoch, C.; Jones, K. C.; Kelly, M.; Tabner, B. J. Environ. Sci. Technol. 1997, 31, 3274-3280. (12) Mackay, D., Shiu, W. Y., Ma, K.-C. Illustrated Handbook of Physical-Chemical Properties and Environmental Fate for Organic Chemicals; Lewis Publishers: Michigan, Vol. 2, 1992. (13) Harner, T.; Green, N. J. L.; Jones, K. C. Environ. Sci. Technol. 2000, 34, 3109-3114. (14) Duarte-Davidson, R.; Sewart, A.; Alcock, R. E.; Cousins, I. T.; Jones, K. C. Environ. Sci. Technol. 1997, 31, 1-11. (15) Ferrario, J.; McDaniel, D.; Byrne, C. Organohalogen Compd. 1999, 40, 95-99. (16) Jobst, H.; Aldag, R. Z. Umweltchem. O ¨ kotox. 2000, 12, 2-4. (17) Rappe, C.; Bergek, S.; Andersson, R.; Cooper, K.; Fiedler, H.; Bopp, R.; Howell, F.; Bonner, M. Organohalogen Compd. 1999, 43, 111-116. (18) Gaus, C.; Organohalogen Compd. 2000, 46, 15-18. (19) Hashimoto, S.; Wakimoto, T.; Tatsukawa, R. Chemosphere 1990, 21, 825-835.
Received for review September 15, 2000. Revised manuscript received February 13, 2001. Accepted February 20, 2001. ES0002161
VOL. 35, NO. 10, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1981
