- Email: [email protected]
Direct measurement of octanol–water partition coefficients of some environmentally relevant brominated diphenyl ether congeners
Chemosphere 51 (2003) 563–567 www.elsevier.com/locate/chemosphere
Direct measurement of octanol–water partition coefficients of some environmentally relevant brominated diphenyl ether congeners Eric Braekevelt a, Sheryl A. Tittlemier b, Gregg T. Tomy a
a,*
Department of Fisheries and Oceans, Freshwater Institute, 501 University Cresent, Winnipeg, Manitoba, Canada R3T 2N6 b Food Research Division, Health Canada, 2203D Frederick G. Banting Building, Ottawa, Ont., Canada K1A 0L2 Received 5 July 2002; received in revised form 27 November 2002; accepted 3 December 2002
Abstract Octanol–water partition coefficients ðKOW Þ of nine environmentally relevant brominated diphenyl ether (BDE) congeners present in two technical mixtures were directly measured using a slow-stir technique. Log KOW values of tri- to heptabrominated BDE congeners ranged from 5.74 to 8.27, and were related to bromine content by the equation log KOW ¼ 0:621ð#BrÞ þ 4:12 ðR2 ¼ 0:970Þ. The directly determined KOW values were generally lower than those calculated using fragment constant methods, particularly at higher levels of bromine substitution. The quasi-experimental approach of using fragment constants to modify a ‘‘backbone’’ compound of known KOW was much more successful than using the fragment constants to ‘‘build’’ the entire molecule. The tri- and tetrabrominated congeners are in the range of optimum bioaccumulation potential. Crown Copyright Ó 2003 Published by Elsevier Science Ltd. All rights reserved. Keywords: Brominated flame retardants; KOW ; PBDEs; Physico-chemical properties; Slow stir
1. Introduction Polybrominated diphenyl ethers (PBDEs) are a class of organic compounds that have been added to a variety of consumer products to reduce their flammability. PBDEs have become a concern because concentrations in humans and their food are rapidly increasing (Luross et al., 2000; Noren and Meironyte, 2000). However, the ability to predict their environmental fate is fairly limited, because the physico-chemical data required for predictive models are largely unavailable. The octanol–water partition coefficient ðKOW Þ is an important property in determining the environmental
*
Corresponding author. Tel.: +1-204-983-5167; fax: +1-204984-2403. E-mail address: [email protected] (G.T. Tomy).
fate of hydrophobic organic chemicals, particularly in biota. Octanol acts as a lipid surrogate: chemicals of high KOW tend to be highly lipophilic and bioaccumulate to a great extent. To our knowledge, there are no published KOW data for individual BDE congeners, hindering their incorporation into quantitative environmental models. Values have been reported as a range for each homologue group (Watanabe and Tatsukawa, 1990), or they have been calculated using fragment constant methods (Palm et al., 2002), where KOW is calculated by adding together empirically derived fragment values. The main purpose of this study was to experimentally determine KOW for individual congeners present in technical PBDE mixtures, allowing their quantitative prediction in environmental models. Another goal of the study was to examine the accuracy of fragment constant calculations of KOW for extremely hydrophobic chemicals such as PBDEs.
0045-6535/03/$ - see front matter Crown Copyright Ó 2003 Published by Elsevier Science Ltd. All rights reserved. doi:10.1016/S0045-6535(02)00841-X
564
E. Braekevelt et al. / Chemosphere 51 (2003) 563–567
Several techniques have been used to experimentally measure KOW . Older methods involved the determination of the concentration of a test chemical in octanol and water phases after extensive shaking. However, the applicability of this technique is limited to chemicals of low hydrophobicity, because the formation of octanol emulsions in the aqueous phase depresses KOW . Other methods relate hydrophobicity to liquid chromatographic retention time (Veith et al., 1979), but these are correlative techniques rather than direct measurements, and are dependent on the accurate determination of reference compound KOW by direct methods (the PBDE KOW values of Watanabe and Tatsukawa (1990) were derived using this method). We chose to directly measure KOW by the slow-stir method, in which test chemical equilibrates between octanol and water phases in a slowly stirred, temperature-controlled vessel (de Bruijn et al., 1989). This technique is intuitive and simple, and less susceptible to analytical artifacts than the shakeflask technique. It is also less labour-intensive than generator column methods of direct KOW measurement (DeVoe et al., 1981; Woodburn et al., 1984). KOW was also calculated using various fragment constant methods to compare with the directly measured data.
2. Materials and methods The slow-stir experiment was conducted using five specially made Erlenmeyer flasks (4 l) equipped with spigots at the bottom for collection of the aqueous phase. These flasks were housed in a controlled environment chamber held at 25 0:5 °C. Four liters of HPLC-grade water (Fisher Scientific, Ottawa, ON, Canada) and a small (1 cm diameter) stir bar were added to each flask. Several polychlorinated biphenyls (PCBs) were included as reference chemicals in addition to the PBDE test chemicals. Individual PCB congeners (CBs 180, 202, 206 and 209, AccuStandard, New Haven CT, USA) and two technical PBDE mixtures (DE-71 and DE-79, Great Lakes Chemical Co., West Lafayette, IN, USA) were accurately weighed (10–60 mg) into individual scintillation vials. Approximately 10 ml octanol (Fisher Scientific, Ottawa, ON, Canada) was added and the vials were sonicated to dissolve the chemicals. The dissolved compounds were then combined into a single flask. Fifteen milliliters of the octanol containing the study compounds were then carefully pipetted onto the surface of the water in each flask. Stir plates were adjusted so that a vortex no greater than 1 cm was formed. We did not equilibrate octanol and water phases before the addition of test chemicals. Water and octanol phases appear to equilibrate within hours, a much shorter time scale than the test chemicals, which reach equilibrium in several days (de Bruijn et al., 1989; Fisk
et al., 1999). We recommend that phases not be preequilibrated, because even slight agitation of octanolsaturated water can create emulsions. Equilibration of phases is only necessary to avoid errors associated with changes in phase volumes when slightly soluble phases are mixed, i.e., when a constant phase volume is assumed (Dearden and Bresnen, 1988). Because volumetric samples were collected after phase equilibration, pre-equilibration was not considered necessary. After 70 days, 1.0 ml octanol was collected from each flask with a volumetric pipette, and diluted a thousandfold with hexane. Solid-phase extraction (SPE) disks (47 mm ENVI-18 Disks, Supelco, Bellefonte, PA, USA) were rinsed with dichloromethane and conditioned with methanol before use. Water was collected from the spigot at the bottom of each flask directly into a SPE vessel, and spiked with 10 ll octachloronaphthalene (OCN) in methanol to measure extraction efficiency. After 1.1 l of water was extracted, SPE disks were placed in a dessicator for several days, then extracted twice with 5 ml toluene and evaporated to 100 ll. Water extracts and diluted octanol phases were analyzed by gas chromatography–mass spectrometry (GC– MS). Chromatographic separation was performed on a Hewlett-Packard 5890 Series II gas chromatograph, fitted with a 30 m 0:25 mm ID (0.1 lm film thickness) DB-5MS capillary column (J&W Scientific, CA, USA). Helium was used as the carrier gas. Samples were run using splitless injection with the injector temperature set at 260 °C. The initial column temperature was 80 °C; after 1 min the oven was ramped at 30 °C min1 to 200 °C, then at 5 °C min1 to a final temperature of 310 °C and held for 10 min. Electronic pressure programming was used to maintain a constant flow of 1 ml min1 during the run. Sample injections of 2 ll were made by a CTC A200SE autosampler. The gas chromatograph was connected by a heated transfer line maintained at 250 °C to a Kratos Concept high-resolution mass spectrometer (EBE geometry) controlled by a Mach 3X data system. Mass spectrometry was performed in the electron ionization mode at an ion source temperature of 250 °C and an ion accelerating voltage of 8 kV. Selected ion monitoring was performed at a resolving power of 10 000, with perfluorokerosene (PFK) used as the mass calibrant. The initial electron beam energy was adjusted during tuning for maximum sensitivity of the PFK ions that were used as lock masses. The cycle time for each window was 1 s, with equal dwell time for each monitored ion. The two most abundant ions in the Mþ cluster (for PCB reference compounds, tri- and tetrabrominated BDEs) or [M–2Br]þ cluster (for penta- to heptabrominated BDEs) were monitored. The most abundant ion was used for quantitation and the next most abundant as a confirmation ion. BDE congeners in each homologue group were identified by retention time from a standard mix-
E. Braekevelt et al. / Chemosphere 51 (2003) 563–567
ture (Cambridge Isotope Laboratories, Andover, MA, USA). The following congeners in the two technical BDE mixtures were examined: 2,20 ,4-tribromodiphenyl ether (BDE-17); 2,4,40 -tribromodiphenyl ether (BDE-28); 2,20 , 4,40 -tetrabromodiphenyl ether (BDE-47); 2,20 ,3,4,40 pentabromodiphenyl ether (BDE-85); 2,20 ,4,40 ,5-pentabromodiphenyl ether (BDE-99); 2,20 ,4,40 ,6-pentabromodiphenyl ether (BDE-100); 2,20 ,4,40 ,5,50 -hexabromodiphenyl ether (BDE-153); 2,20 ,4,40 ,5,60 -hexabromodiphenyl ether (BDE-154); 2,20 ,3,4,40 ,50 ,6-heptabromodiphenyl ether (BDE-183). The KOW of each congener was determined as the ratio of electronically integrated peak area in the octanol phase to that in the aqueous phase, and a standard deviation was calculated from the five replicates. Various fragment constant methods were also used to calculate log KOW of the PCB reference chemicals and PBDEs (Leo et al., 1975; Leo, 1983; Broto et al., 1984; Ghose and Crippen, 1986, 1987, 1988; Viswanadhan et al., 1989; Meylan and Howard, 1995).
3. Results and discussion Recoveries of the OCN surrogate standard were 90– 95%. Concentrations in water were recovery-corrected but not blank corrected, as PBDE concentrations in blanks were extremely low (less than 1% of the sample concentrations). The KOW values of the PCB reference compounds are slightly higher than other direct measurements of KOW , with differences ranging from 0.36 log units for PCB 180 to 0.86 log units for PCB 202 (Table 1). This may be partly due to differences in methodology, because the greatest differences from the results reported here are values that have been determined using the generator column method. However, considerable variability has been reported for some PCB congeners even when using
Table 1 Comparison of PCB log KOW values PCB congener #
This study
180 202
7.72 0.20 8.10 0.24
206 209
8.47 0.25 8.82 0.35
a
Literature Experimental
Calculateda
7.36b 7.24b 7.73c 8.09b 8.18b 8.27c
7.36–8.27 7.88–8.91 8.39–9.55 8.91–10.2
Various fragment constant methods (Leo, 1983; Broto et al., 1984; Ghose and Crippen, 1986, 1987; Viswanadhan et al., 1989; Meylan and Howard, 1995). b Hawker and Connell (1988, generator column). c de Bruijn et al. (1989, slow stir).
565
the same method (Miller et al., 1984; Woodburn et al., 1984; Hawker and Connell, 1988). Although the lower end of the range of calculated values are fairly close to experimental values, the higher end of the range of calculated values are considerably higher, often by over an order of magnitude (Table 1). The variability associated with the reference compounds was higher than that of the PBDE test chemicals. With the exception of the hexabrominated congeners, the KOW values reported here are slightly higher than the homologue ranges reported by Watanabe and Tatsukawa (1990). Like the PCB reference compounds, the lower end of the range of calculated values are close to experimental values. However, at higher degrees of bromination, even the lower end of the range of calculated values are considerably higher than experimental values, and the higher end of the range of calculated values diverge significantly from the experimentally determined PBDE KOW values, often by over an order of magnitude (Table 2). Because fragment values are determined empirically from small molecules, minor differences can result in large errors when multiple fragments are added. Alternatively, fragment constants can be used to modify a ‘‘backbone’’ compound of similar structure and known KOW . Polychlorinated diphenyl ether (PCDE) KOW values from Kurz and Ballschmiter (1999) were modified by adding the difference between Br and Cl fragment constants from the various fragment constant methods. The PBDE log KOW values from this study are within the range of values calculated from the PCDE backbone, which are presented in Table 2 as ‘‘quasiexperimental.’’ As with almost all halogenated organic compounds, KOW increased with increasing halogen content. The addition of a bromine atom to a PBDE molecule increased KOW to a greater extent than the addition of a chlorine atom to PCBs and PCDEs (Fig. 1). The high KOW values of these compounds suggest that they would be present mainly in relatively immobile environmental media such as soil and sediment. The triand tetrabrominated BDEs are in the KOW range of optimum bioaccumulation potential (Fisk et al., 1998; Meylan et al., 1999). The more heavily brominated congeners may be too nonpolar or too large to bioaccumulate, but they may debrominate to congeners of higher bioaccumulation potential (Tysklind et al., 2001). Octanol–water partition coefficients continue to be used in environmental models because octanol is considered an appropriate surrogate for environmental lipid and organic matter. Variation in KOW is assumed to be dominated by the activity of the chemical in the aqueous phase. The activity of the chemical in octanol is assumed to be near unity, suggesting that the choice of organic phase is relatively unimportant (Schwarzenbach et al., 1993). Whereas this appears to be true for chemicals of
566
E. Braekevelt et al. / Chemosphere 51 (2003) 563–567
Table 2 Comparison of PBDE log KOW values Congener # BDE-17 BDE-28 BDE-47 BDE-85 BDE-99 BDE-100 BDE-153 BDE-154 BDE-183
This study 5.74 0.22 5.94 0.15 6.81 0.08 7.37 0.12 7.32 0.14 7.24 0.16 7.90 0.14 7.82 0.16 8.27 0.26
Literature Experimentala
Calculatedb
Quasi-experimentalc
5.47–5.58 5.47–5.58 5.87–6.16 6.46–6.97 6.46–6.97 6.46–6.97 6.86–7.92 6.86–7.92 nd
5.88–6.60 5.88–6.60 6.77–7.49 7.66–8.38 7.66–8.38 7.66–8.38 8.54–9.27 8.53–9.27 8.99–10.2
5.41–5.77 5.98–6.34 6.55–7.03 7.03–7.63 7.13–7.73 6.86–7.46 7.62–8.34 7.39–8.11 nd
nd: not determined. a Range reported in Watanabe and Tatsukawa (1990). b Various fragment constant methods (Leo et al., 1975; Leo, 1983; Broto et al., 1984; Ghose and Crippen, 1986, 1987, 1988; Viswanadhan et al., 1989; Meylan and Howard, 1995). c PCDE KOW values from Kurz and Ballschmiter (1999), modified by adding the difference between Br and Cl fragment constants from various fragment constant methods (Leo et al., 1975; Leo, 1983; Broto et al., 1984; Ghose and Crippen, 1986, 1987, 1988; Viswanadhan et al., 1989; Meylan and Howard, 1995).
Fig. 1. Relationships of PBDE (this study), PCDE (Kurz and Ballschmiter, 1999), and PCB (Hawker and Connell, 1988) experimental log KOW values with degree of halogen substitution.
moderate to high hydrophobicity (log KOW < 7), log KOW appears to asymptotically approach an upper limit of about nine, regardless of hydrophobicity (de Bruijn and Hermens, 1990). This may be due to the solubilizing effect of the small amount (103 M) of octanol in the aqueous phase, or to decreased chemical solubility in water-saturated octanol, both of which become more important with increasing hydrophobicity (Chiou et al., 1982). The sensitivity of extremely hydrophobic compounds to the presence of water in the octanol phase suggests that they do not behave ideally in (water-saturated) octanol. KOW then becomes at least partly dependent on the nature of the organic phase. At this point, the suitability of octanol as a lipid and organic matter surrogate may no longer be appropriate. The
environmental relevance of KOW values for such hydrophobic chemicals as PBDEs will not be known until models that incorporate these data successfully predict environmental fate. Octanol–water partitioning behavior of PBDEs is generally not accurately predicted by fragment constant methods. These methods were designed to evaluate the environmental behavior of newly developed drugs, which generally have a variety of bioactive sites and are considerably less hydrophobic than PBDEs. The fragment constants were determined empirically using small molecules: suitable empirical relationships using larger and more hydrophobic chemicals have not been determined. The quasi-experimental approach of using fragment constants to modify ‘‘backbone’’ chemicals of similar structure and known KOW was much more successful than using the fragment constants to ‘‘build’’ the entire molecule. The inability of fragment constant methods to predict changes in chemical activity in the octanol phase may be the reason why they fare so poorly when it comes to predicting KOW of the extremely hydrophobic PBDEs.
References Broto, P., Moreau, G., VanDycke, C., 1984. Molecular structures: perception, autocorrelation descriptor and sar studies. System of atomic contributions for the calculation of the n-octanol/water coefficients. Eur. J. Med. Chem. 19, 71–78. Chiou, C.T., Schmedding, D.W., Manes, M., 1982. Partitioning of organic compounds in octanol–water systems. Environ. Sci. Technol. 16, 4–10. de Bruijn, J., Hermens, J., 1990. Relationships between octanol/ water partition coefficients and total molecular surface area
E. Braekevelt et al. / Chemosphere 51 (2003) 563–567 and total molecular volume of hydrophobic organic chemicals. Quant. Struct.––Act. Relat. 9, 11–21. de Bruijn, J., Busser, F., Seinen, W., Hermens, J., 1989. Determination of octanol/water partition coefficients for hydrophobic organic chemicals with the ‘‘slow-stirring’’ method. Environ. Toxicol. Chem. 8, 499–512. Dearden, J.C., Bresnen, G.M., 1988. The measurement of partition coefficients. Quant. Struct.––Act. Relat. 7, 133– 144. DeVoe, H., Miller, M.M., Wasik, S.P., 1981. Generator columns and high pressure liquid chromatography for determining aqueous solubilities and octanol–water partition coefficients of hydrophobic substances. J. Res. Nat. Bur. Stand. 86, 361–366. Fisk, A.T., Norstrom, R.J., Cymbalisty, C.D., Muir, D.C.G., 1998. Dietary accumulation and depuration of hydrophobic organochlorines: bioaccumulation parameters and their relationship with the octanol/water partition coefficient. Environ. Toxicol. Chem. 17, 951–961. Fisk, A.T., Rosenberg, B., Cymbalisty, C.D., Stern, G.A., Muir, D.C.G., 1999. Octanol/water partition coefficients of toxaphene congeners determined by the ‘‘slow-stirring’’ method. Chemosphere 39, 2549–2562. Ghose, A.K., Crippen, G.M., 1986. Atomic physicochemical parameters for three-dimensional structure-directed quantitative structure–activity relationships. I. Partition coefficients as a measure of hydrophobicity. J. Comput. Chem. 7, 565–577. Ghose, A.K., Crippen, G.M., 1987. Atomic physicochemical parameters for three-dimensional-structure-directed quantitative structure–activity relationships. 2. Modeling dispersive and hydrophobic interactions. J. Chem. Inf. Comput. Sci. 27, 21–35. Ghose, A.K., Crippen, G.M., 1988. Atomic physicochemical parameters for three-dimensional structure-directed quantitative structure–activity relationships. III. Modeling hydrophobic interactions. J. Comput. Chem. 9, 80–90. Hawker, D.W., Connell, D.W., 1988. Octanol–water partition coefficients of polychlorinated biphenyl congeners. Environ. Sci. Technol. 22, 382–387. Kurz, J., Ballschmiter, K., 1999. Vapour pressures, aqueous solubilities, HenryÕs law constants, partition coefficients between gas/water ðKgw Þ, n-octanol/water ðKow Þ and gas/octanol ðKgo Þ of 106 polychlorinated diphenyl ethers (PCDE). Chemosphere 38, 573–586. Leo, A., 1983. The octanol–water partition coefficient of aromatic solutes: the effect of electronic interactions, alkyl chains, hydrogen bonds, and ortho-substitution. J. Chem. Soc., Perkin Trans. II 1983, 825–838. Leo, A., Jow, P.Y.C., Silipo, C., Hansch, C., 1975. Calculation of hydrophobic constant ðlog P Þ from p and f constants. J. Med. Chem. 18, 865–868. Luross, J.M., Alaee, M., Sergeant, D.B., Whittle, D.M., Solomon, K.R., 2000. Spatial and temporal distribution of
567
polybrominated diphenyl ethers in lake trout from the Great Lakes. Organohalogen Compds. 47, 73–76. Meylan, W.M., Howard, P.H., 1995. Atom/fragment contribution method for estimating octanol–water partition coefficients. J. Pharm. Sci. 84, 83–92. Meylan, W.M., Howard, P.H., Boethling, R.S., Aronson, D., Printup, H., Gouchie, S., 1999. Improved method for estimating bioconcentration/bioaccumulation factor from octanol/water partition coefficient. Environ. Toxicol. Chem. 18, 664–672. Miller, M.M., Ghodbane, S., Wasik, S.P., Tewari, Y.B., Martire, D.E., 1984. Aqueous solubilities, octanol/water partition coefficients, and entropies of melting of chlorinated benzenes and biphenyls. J. Chem. Eng. Data 29, 184– 190. Noren, K., Meironyte, D., 2000. Certain organochlorine and organobromine contaminants in Swedish human milk in perspective of past 20–30 years. Chemosphere 40, 1111– 1123. Palm, A., Cousins, I.T., Mackay, D., Tysklind, M., Metcalfe, C., Alaee, M., 2002. Assessing the environmental fate of chemicals of emerging concern: a case study of the polybrominated diphenyl ethers. Environ. Pollut. 117, 195– 213. Schwarzenbach, R.P., Gschwend, P.M., Imboden, D.M., 1993. Environmental Organic Chemistry. John Wiley & Sons Inc., New York. Tysklind, M., Sellstr€ om, U., S€ oderstr€ om, G., de Wit, C., 2001. Abiotic transformation of polybrominated diphenylethers (PBDEs): photolytic debromination of decabromo diphenyl ether. In: Proceedings, Second International Workshop on Brominated Flame Retardants, Stockholm, Sweden, 14–16 May, 2001, pp. 51–54. Veith, G.D., Austin, N.M., Morris, R.T., 1979. A rapid method for estimating log P for organic chemicals. Water Res. 13, 43–47. Viswanadhan, V.N., Ghose, A.K., Revankar, G.R., Robins, R.K., 1989. Atomic physicochemical parameters for three dimensional structure directed quantitative structure–activity relationships. 4. Additional parameters for hydrophobic and dispersive interactions and their application for an automated superposition of certain naturally occurring nucleoside antibiotics. J. Chem. Inf. Comput. Sci. 29, 163– 172. Watanabe, I., Tatsukawa, R., 1990. Anthropogenic brominated aromatics in the Japanese environment. In: Freij, L. (Ed.). Proceedings, Workshop on Brominated Aromatic Flame Retardants, Skokloster, Sweden, 24–26 October, 1989. Swedish National Chemicals Inspectorate (KemI), Solna, Sweden, pp. 63–70. Woodburn, K.B., Doucette, W.J., Andren, A.W., 1984. Generator column determination of octanol/water partition coefficients for selected polychlorinated biphenyl congeners. Environ. Sci. Technol. 18, 457–459.
Direct measurement of octanol–water partition coefficients of some environmentally relevant brominated diphenyl ether congeners Eric Braekevelt a, Sheryl A. Tittlemier b, Gregg T. Tomy a
a,*
Department of Fisheries and Oceans, Freshwater Institute, 501 University Cresent, Winnipeg, Manitoba, Canada R3T 2N6 b Food Research Division, Health Canada, 2203D Frederick G. Banting Building, Ottawa, Ont., Canada K1A 0L2 Received 5 July 2002; received in revised form 27 November 2002; accepted 3 December 2002
Abstract Octanol–water partition coefficients ðKOW Þ of nine environmentally relevant brominated diphenyl ether (BDE) congeners present in two technical mixtures were directly measured using a slow-stir technique. Log KOW values of tri- to heptabrominated BDE congeners ranged from 5.74 to 8.27, and were related to bromine content by the equation log KOW ¼ 0:621ð#BrÞ þ 4:12 ðR2 ¼ 0:970Þ. The directly determined KOW values were generally lower than those calculated using fragment constant methods, particularly at higher levels of bromine substitution. The quasi-experimental approach of using fragment constants to modify a ‘‘backbone’’ compound of known KOW was much more successful than using the fragment constants to ‘‘build’’ the entire molecule. The tri- and tetrabrominated congeners are in the range of optimum bioaccumulation potential. Crown Copyright Ó 2003 Published by Elsevier Science Ltd. All rights reserved. Keywords: Brominated flame retardants; KOW ; PBDEs; Physico-chemical properties; Slow stir
1. Introduction Polybrominated diphenyl ethers (PBDEs) are a class of organic compounds that have been added to a variety of consumer products to reduce their flammability. PBDEs have become a concern because concentrations in humans and their food are rapidly increasing (Luross et al., 2000; Noren and Meironyte, 2000). However, the ability to predict their environmental fate is fairly limited, because the physico-chemical data required for predictive models are largely unavailable. The octanol–water partition coefficient ðKOW Þ is an important property in determining the environmental
*
Corresponding author. Tel.: +1-204-983-5167; fax: +1-204984-2403. E-mail address: [email protected] (G.T. Tomy).
fate of hydrophobic organic chemicals, particularly in biota. Octanol acts as a lipid surrogate: chemicals of high KOW tend to be highly lipophilic and bioaccumulate to a great extent. To our knowledge, there are no published KOW data for individual BDE congeners, hindering their incorporation into quantitative environmental models. Values have been reported as a range for each homologue group (Watanabe and Tatsukawa, 1990), or they have been calculated using fragment constant methods (Palm et al., 2002), where KOW is calculated by adding together empirically derived fragment values. The main purpose of this study was to experimentally determine KOW for individual congeners present in technical PBDE mixtures, allowing their quantitative prediction in environmental models. Another goal of the study was to examine the accuracy of fragment constant calculations of KOW for extremely hydrophobic chemicals such as PBDEs.
0045-6535/03/$ - see front matter Crown Copyright Ó 2003 Published by Elsevier Science Ltd. All rights reserved. doi:10.1016/S0045-6535(02)00841-X
564
E. Braekevelt et al. / Chemosphere 51 (2003) 563–567
Several techniques have been used to experimentally measure KOW . Older methods involved the determination of the concentration of a test chemical in octanol and water phases after extensive shaking. However, the applicability of this technique is limited to chemicals of low hydrophobicity, because the formation of octanol emulsions in the aqueous phase depresses KOW . Other methods relate hydrophobicity to liquid chromatographic retention time (Veith et al., 1979), but these are correlative techniques rather than direct measurements, and are dependent on the accurate determination of reference compound KOW by direct methods (the PBDE KOW values of Watanabe and Tatsukawa (1990) were derived using this method). We chose to directly measure KOW by the slow-stir method, in which test chemical equilibrates between octanol and water phases in a slowly stirred, temperature-controlled vessel (de Bruijn et al., 1989). This technique is intuitive and simple, and less susceptible to analytical artifacts than the shakeflask technique. It is also less labour-intensive than generator column methods of direct KOW measurement (DeVoe et al., 1981; Woodburn et al., 1984). KOW was also calculated using various fragment constant methods to compare with the directly measured data.
2. Materials and methods The slow-stir experiment was conducted using five specially made Erlenmeyer flasks (4 l) equipped with spigots at the bottom for collection of the aqueous phase. These flasks were housed in a controlled environment chamber held at 25 0:5 °C. Four liters of HPLC-grade water (Fisher Scientific, Ottawa, ON, Canada) and a small (1 cm diameter) stir bar were added to each flask. Several polychlorinated biphenyls (PCBs) were included as reference chemicals in addition to the PBDE test chemicals. Individual PCB congeners (CBs 180, 202, 206 and 209, AccuStandard, New Haven CT, USA) and two technical PBDE mixtures (DE-71 and DE-79, Great Lakes Chemical Co., West Lafayette, IN, USA) were accurately weighed (10–60 mg) into individual scintillation vials. Approximately 10 ml octanol (Fisher Scientific, Ottawa, ON, Canada) was added and the vials were sonicated to dissolve the chemicals. The dissolved compounds were then combined into a single flask. Fifteen milliliters of the octanol containing the study compounds were then carefully pipetted onto the surface of the water in each flask. Stir plates were adjusted so that a vortex no greater than 1 cm was formed. We did not equilibrate octanol and water phases before the addition of test chemicals. Water and octanol phases appear to equilibrate within hours, a much shorter time scale than the test chemicals, which reach equilibrium in several days (de Bruijn et al., 1989; Fisk
et al., 1999). We recommend that phases not be preequilibrated, because even slight agitation of octanolsaturated water can create emulsions. Equilibration of phases is only necessary to avoid errors associated with changes in phase volumes when slightly soluble phases are mixed, i.e., when a constant phase volume is assumed (Dearden and Bresnen, 1988). Because volumetric samples were collected after phase equilibration, pre-equilibration was not considered necessary. After 70 days, 1.0 ml octanol was collected from each flask with a volumetric pipette, and diluted a thousandfold with hexane. Solid-phase extraction (SPE) disks (47 mm ENVI-18 Disks, Supelco, Bellefonte, PA, USA) were rinsed with dichloromethane and conditioned with methanol before use. Water was collected from the spigot at the bottom of each flask directly into a SPE vessel, and spiked with 10 ll octachloronaphthalene (OCN) in methanol to measure extraction efficiency. After 1.1 l of water was extracted, SPE disks were placed in a dessicator for several days, then extracted twice with 5 ml toluene and evaporated to 100 ll. Water extracts and diluted octanol phases were analyzed by gas chromatography–mass spectrometry (GC– MS). Chromatographic separation was performed on a Hewlett-Packard 5890 Series II gas chromatograph, fitted with a 30 m 0:25 mm ID (0.1 lm film thickness) DB-5MS capillary column (J&W Scientific, CA, USA). Helium was used as the carrier gas. Samples were run using splitless injection with the injector temperature set at 260 °C. The initial column temperature was 80 °C; after 1 min the oven was ramped at 30 °C min1 to 200 °C, then at 5 °C min1 to a final temperature of 310 °C and held for 10 min. Electronic pressure programming was used to maintain a constant flow of 1 ml min1 during the run. Sample injections of 2 ll were made by a CTC A200SE autosampler. The gas chromatograph was connected by a heated transfer line maintained at 250 °C to a Kratos Concept high-resolution mass spectrometer (EBE geometry) controlled by a Mach 3X data system. Mass spectrometry was performed in the electron ionization mode at an ion source temperature of 250 °C and an ion accelerating voltage of 8 kV. Selected ion monitoring was performed at a resolving power of 10 000, with perfluorokerosene (PFK) used as the mass calibrant. The initial electron beam energy was adjusted during tuning for maximum sensitivity of the PFK ions that were used as lock masses. The cycle time for each window was 1 s, with equal dwell time for each monitored ion. The two most abundant ions in the Mþ cluster (for PCB reference compounds, tri- and tetrabrominated BDEs) or [M–2Br]þ cluster (for penta- to heptabrominated BDEs) were monitored. The most abundant ion was used for quantitation and the next most abundant as a confirmation ion. BDE congeners in each homologue group were identified by retention time from a standard mix-
E. Braekevelt et al. / Chemosphere 51 (2003) 563–567
ture (Cambridge Isotope Laboratories, Andover, MA, USA). The following congeners in the two technical BDE mixtures were examined: 2,20 ,4-tribromodiphenyl ether (BDE-17); 2,4,40 -tribromodiphenyl ether (BDE-28); 2,20 , 4,40 -tetrabromodiphenyl ether (BDE-47); 2,20 ,3,4,40 pentabromodiphenyl ether (BDE-85); 2,20 ,4,40 ,5-pentabromodiphenyl ether (BDE-99); 2,20 ,4,40 ,6-pentabromodiphenyl ether (BDE-100); 2,20 ,4,40 ,5,50 -hexabromodiphenyl ether (BDE-153); 2,20 ,4,40 ,5,60 -hexabromodiphenyl ether (BDE-154); 2,20 ,3,4,40 ,50 ,6-heptabromodiphenyl ether (BDE-183). The KOW of each congener was determined as the ratio of electronically integrated peak area in the octanol phase to that in the aqueous phase, and a standard deviation was calculated from the five replicates. Various fragment constant methods were also used to calculate log KOW of the PCB reference chemicals and PBDEs (Leo et al., 1975; Leo, 1983; Broto et al., 1984; Ghose and Crippen, 1986, 1987, 1988; Viswanadhan et al., 1989; Meylan and Howard, 1995).
3. Results and discussion Recoveries of the OCN surrogate standard were 90– 95%. Concentrations in water were recovery-corrected but not blank corrected, as PBDE concentrations in blanks were extremely low (less than 1% of the sample concentrations). The KOW values of the PCB reference compounds are slightly higher than other direct measurements of KOW , with differences ranging from 0.36 log units for PCB 180 to 0.86 log units for PCB 202 (Table 1). This may be partly due to differences in methodology, because the greatest differences from the results reported here are values that have been determined using the generator column method. However, considerable variability has been reported for some PCB congeners even when using
Table 1 Comparison of PCB log KOW values PCB congener #
This study
180 202
7.72 0.20 8.10 0.24
206 209
8.47 0.25 8.82 0.35
a
Literature Experimental
Calculateda
7.36b 7.24b 7.73c 8.09b 8.18b 8.27c
7.36–8.27 7.88–8.91 8.39–9.55 8.91–10.2
Various fragment constant methods (Leo, 1983; Broto et al., 1984; Ghose and Crippen, 1986, 1987; Viswanadhan et al., 1989; Meylan and Howard, 1995). b Hawker and Connell (1988, generator column). c de Bruijn et al. (1989, slow stir).
565
the same method (Miller et al., 1984; Woodburn et al., 1984; Hawker and Connell, 1988). Although the lower end of the range of calculated values are fairly close to experimental values, the higher end of the range of calculated values are considerably higher, often by over an order of magnitude (Table 1). The variability associated with the reference compounds was higher than that of the PBDE test chemicals. With the exception of the hexabrominated congeners, the KOW values reported here are slightly higher than the homologue ranges reported by Watanabe and Tatsukawa (1990). Like the PCB reference compounds, the lower end of the range of calculated values are close to experimental values. However, at higher degrees of bromination, even the lower end of the range of calculated values are considerably higher than experimental values, and the higher end of the range of calculated values diverge significantly from the experimentally determined PBDE KOW values, often by over an order of magnitude (Table 2). Because fragment values are determined empirically from small molecules, minor differences can result in large errors when multiple fragments are added. Alternatively, fragment constants can be used to modify a ‘‘backbone’’ compound of similar structure and known KOW . Polychlorinated diphenyl ether (PCDE) KOW values from Kurz and Ballschmiter (1999) were modified by adding the difference between Br and Cl fragment constants from the various fragment constant methods. The PBDE log KOW values from this study are within the range of values calculated from the PCDE backbone, which are presented in Table 2 as ‘‘quasiexperimental.’’ As with almost all halogenated organic compounds, KOW increased with increasing halogen content. The addition of a bromine atom to a PBDE molecule increased KOW to a greater extent than the addition of a chlorine atom to PCBs and PCDEs (Fig. 1). The high KOW values of these compounds suggest that they would be present mainly in relatively immobile environmental media such as soil and sediment. The triand tetrabrominated BDEs are in the KOW range of optimum bioaccumulation potential (Fisk et al., 1998; Meylan et al., 1999). The more heavily brominated congeners may be too nonpolar or too large to bioaccumulate, but they may debrominate to congeners of higher bioaccumulation potential (Tysklind et al., 2001). Octanol–water partition coefficients continue to be used in environmental models because octanol is considered an appropriate surrogate for environmental lipid and organic matter. Variation in KOW is assumed to be dominated by the activity of the chemical in the aqueous phase. The activity of the chemical in octanol is assumed to be near unity, suggesting that the choice of organic phase is relatively unimportant (Schwarzenbach et al., 1993). Whereas this appears to be true for chemicals of
566
E. Braekevelt et al. / Chemosphere 51 (2003) 563–567
Table 2 Comparison of PBDE log KOW values Congener # BDE-17 BDE-28 BDE-47 BDE-85 BDE-99 BDE-100 BDE-153 BDE-154 BDE-183
This study 5.74 0.22 5.94 0.15 6.81 0.08 7.37 0.12 7.32 0.14 7.24 0.16 7.90 0.14 7.82 0.16 8.27 0.26
Literature Experimentala
Calculatedb
Quasi-experimentalc
5.47–5.58 5.47–5.58 5.87–6.16 6.46–6.97 6.46–6.97 6.46–6.97 6.86–7.92 6.86–7.92 nd
5.88–6.60 5.88–6.60 6.77–7.49 7.66–8.38 7.66–8.38 7.66–8.38 8.54–9.27 8.53–9.27 8.99–10.2
5.41–5.77 5.98–6.34 6.55–7.03 7.03–7.63 7.13–7.73 6.86–7.46 7.62–8.34 7.39–8.11 nd
nd: not determined. a Range reported in Watanabe and Tatsukawa (1990). b Various fragment constant methods (Leo et al., 1975; Leo, 1983; Broto et al., 1984; Ghose and Crippen, 1986, 1987, 1988; Viswanadhan et al., 1989; Meylan and Howard, 1995). c PCDE KOW values from Kurz and Ballschmiter (1999), modified by adding the difference between Br and Cl fragment constants from various fragment constant methods (Leo et al., 1975; Leo, 1983; Broto et al., 1984; Ghose and Crippen, 1986, 1987, 1988; Viswanadhan et al., 1989; Meylan and Howard, 1995).
Fig. 1. Relationships of PBDE (this study), PCDE (Kurz and Ballschmiter, 1999), and PCB (Hawker and Connell, 1988) experimental log KOW values with degree of halogen substitution.
moderate to high hydrophobicity (log KOW < 7), log KOW appears to asymptotically approach an upper limit of about nine, regardless of hydrophobicity (de Bruijn and Hermens, 1990). This may be due to the solubilizing effect of the small amount (103 M) of octanol in the aqueous phase, or to decreased chemical solubility in water-saturated octanol, both of which become more important with increasing hydrophobicity (Chiou et al., 1982). The sensitivity of extremely hydrophobic compounds to the presence of water in the octanol phase suggests that they do not behave ideally in (water-saturated) octanol. KOW then becomes at least partly dependent on the nature of the organic phase. At this point, the suitability of octanol as a lipid and organic matter surrogate may no longer be appropriate. The
environmental relevance of KOW values for such hydrophobic chemicals as PBDEs will not be known until models that incorporate these data successfully predict environmental fate. Octanol–water partitioning behavior of PBDEs is generally not accurately predicted by fragment constant methods. These methods were designed to evaluate the environmental behavior of newly developed drugs, which generally have a variety of bioactive sites and are considerably less hydrophobic than PBDEs. The fragment constants were determined empirically using small molecules: suitable empirical relationships using larger and more hydrophobic chemicals have not been determined. The quasi-experimental approach of using fragment constants to modify ‘‘backbone’’ chemicals of similar structure and known KOW was much more successful than using the fragment constants to ‘‘build’’ the entire molecule. The inability of fragment constant methods to predict changes in chemical activity in the octanol phase may be the reason why they fare so poorly when it comes to predicting KOW of the extremely hydrophobic PBDEs.
References Broto, P., Moreau, G., VanDycke, C., 1984. Molecular structures: perception, autocorrelation descriptor and sar studies. System of atomic contributions for the calculation of the n-octanol/water coefficients. Eur. J. Med. Chem. 19, 71–78. Chiou, C.T., Schmedding, D.W., Manes, M., 1982. Partitioning of organic compounds in octanol–water systems. Environ. Sci. Technol. 16, 4–10. de Bruijn, J., Hermens, J., 1990. Relationships between octanol/ water partition coefficients and total molecular surface area
E. Braekevelt et al. / Chemosphere 51 (2003) 563–567 and total molecular volume of hydrophobic organic chemicals. Quant. Struct.––Act. Relat. 9, 11–21. de Bruijn, J., Busser, F., Seinen, W., Hermens, J., 1989. Determination of octanol/water partition coefficients for hydrophobic organic chemicals with the ‘‘slow-stirring’’ method. Environ. Toxicol. Chem. 8, 499–512. Dearden, J.C., Bresnen, G.M., 1988. The measurement of partition coefficients. Quant. Struct.––Act. Relat. 7, 133– 144. DeVoe, H., Miller, M.M., Wasik, S.P., 1981. Generator columns and high pressure liquid chromatography for determining aqueous solubilities and octanol–water partition coefficients of hydrophobic substances. J. Res. Nat. Bur. Stand. 86, 361–366. Fisk, A.T., Norstrom, R.J., Cymbalisty, C.D., Muir, D.C.G., 1998. Dietary accumulation and depuration of hydrophobic organochlorines: bioaccumulation parameters and their relationship with the octanol/water partition coefficient. Environ. Toxicol. Chem. 17, 951–961. Fisk, A.T., Rosenberg, B., Cymbalisty, C.D., Stern, G.A., Muir, D.C.G., 1999. Octanol/water partition coefficients of toxaphene congeners determined by the ‘‘slow-stirring’’ method. Chemosphere 39, 2549–2562. Ghose, A.K., Crippen, G.M., 1986. Atomic physicochemical parameters for three-dimensional structure-directed quantitative structure–activity relationships. I. Partition coefficients as a measure of hydrophobicity. J. Comput. Chem. 7, 565–577. Ghose, A.K., Crippen, G.M., 1987. Atomic physicochemical parameters for three-dimensional-structure-directed quantitative structure–activity relationships. 2. Modeling dispersive and hydrophobic interactions. J. Chem. Inf. Comput. Sci. 27, 21–35. Ghose, A.K., Crippen, G.M., 1988. Atomic physicochemical parameters for three-dimensional structure-directed quantitative structure–activity relationships. III. Modeling hydrophobic interactions. J. Comput. Chem. 9, 80–90. Hawker, D.W., Connell, D.W., 1988. Octanol–water partition coefficients of polychlorinated biphenyl congeners. Environ. Sci. Technol. 22, 382–387. Kurz, J., Ballschmiter, K., 1999. Vapour pressures, aqueous solubilities, HenryÕs law constants, partition coefficients between gas/water ðKgw Þ, n-octanol/water ðKow Þ and gas/octanol ðKgo Þ of 106 polychlorinated diphenyl ethers (PCDE). Chemosphere 38, 573–586. Leo, A., 1983. The octanol–water partition coefficient of aromatic solutes: the effect of electronic interactions, alkyl chains, hydrogen bonds, and ortho-substitution. J. Chem. Soc., Perkin Trans. II 1983, 825–838. Leo, A., Jow, P.Y.C., Silipo, C., Hansch, C., 1975. Calculation of hydrophobic constant ðlog P Þ from p and f constants. J. Med. Chem. 18, 865–868. Luross, J.M., Alaee, M., Sergeant, D.B., Whittle, D.M., Solomon, K.R., 2000. Spatial and temporal distribution of
567
polybrominated diphenyl ethers in lake trout from the Great Lakes. Organohalogen Compds. 47, 73–76. Meylan, W.M., Howard, P.H., 1995. Atom/fragment contribution method for estimating octanol–water partition coefficients. J. Pharm. Sci. 84, 83–92. Meylan, W.M., Howard, P.H., Boethling, R.S., Aronson, D., Printup, H., Gouchie, S., 1999. Improved method for estimating bioconcentration/bioaccumulation factor from octanol/water partition coefficient. Environ. Toxicol. Chem. 18, 664–672. Miller, M.M., Ghodbane, S., Wasik, S.P., Tewari, Y.B., Martire, D.E., 1984. Aqueous solubilities, octanol/water partition coefficients, and entropies of melting of chlorinated benzenes and biphenyls. J. Chem. Eng. Data 29, 184– 190. Noren, K., Meironyte, D., 2000. Certain organochlorine and organobromine contaminants in Swedish human milk in perspective of past 20–30 years. Chemosphere 40, 1111– 1123. Palm, A., Cousins, I.T., Mackay, D., Tysklind, M., Metcalfe, C., Alaee, M., 2002. Assessing the environmental fate of chemicals of emerging concern: a case study of the polybrominated diphenyl ethers. Environ. Pollut. 117, 195– 213. Schwarzenbach, R.P., Gschwend, P.M., Imboden, D.M., 1993. Environmental Organic Chemistry. John Wiley & Sons Inc., New York. Tysklind, M., Sellstr€ om, U., S€ oderstr€ om, G., de Wit, C., 2001. Abiotic transformation of polybrominated diphenylethers (PBDEs): photolytic debromination of decabromo diphenyl ether. In: Proceedings, Second International Workshop on Brominated Flame Retardants, Stockholm, Sweden, 14–16 May, 2001, pp. 51–54. Veith, G.D., Austin, N.M., Morris, R.T., 1979. A rapid method for estimating log P for organic chemicals. Water Res. 13, 43–47. Viswanadhan, V.N., Ghose, A.K., Revankar, G.R., Robins, R.K., 1989. Atomic physicochemical parameters for three dimensional structure directed quantitative structure–activity relationships. 4. Additional parameters for hydrophobic and dispersive interactions and their application for an automated superposition of certain naturally occurring nucleoside antibiotics. J. Chem. Inf. Comput. Sci. 29, 163– 172. Watanabe, I., Tatsukawa, R., 1990. Anthropogenic brominated aromatics in the Japanese environment. In: Freij, L. (Ed.). Proceedings, Workshop on Brominated Aromatic Flame Retardants, Skokloster, Sweden, 24–26 October, 1989. Swedish National Chemicals Inspectorate (KemI), Solna, Sweden, pp. 63–70. Woodburn, K.B., Doucette, W.J., Andren, A.W., 1984. Generator column determination of octanol/water partition coefficients for selected polychlorinated biphenyl congeners. Environ. Sci. Technol. 18, 457–459.