Screening assessment report for chlorinated naphthalenes: chapter 3

Identity, uses, sources and releases to the environment

Substance identity

The generic structure of chlorinated naphthalene (CN) is shown in Figure 1.

Figure 1. Generic structure of a chlorinated naphthalene molecule showing the naphthalene ring structure, carbon atom numbering system and potential for chlorine substitution

CNs are formed by substituting chlorine for hydrogen atoms that are attached to the numbered carbon atoms, as shown in Figure 1. CNs have the molecular formula C₁₀H_8-nCl_n (n = 1-8). There are 75 possible chlorinated naphthalenes. The system of nomenclature for CNs is similar to that for polychlorinated biphenyls and uses the numbering system shown in Figure 1. CNs are divided into eight homologue groups, based on the number of chlorine atoms in the molecule. These homologue groups are referred to using the prefixes mono- to octa- (for example: mono-CNs, di-CNs). CNs are commonly referred to in the scientific literature as polychlorinated naphthalenes (PCNs). The term CNs is used in this report because it more accurately describes the entire class of chlorinated naphthalenes, including the mono-CN congeners. The number and, to a lesser extent, the positions of the chlorine atoms within the CN molecule are the key determinants of the physical and chemical properties of the CN congeners.

CNs were produced for commercial use as mixtures. In North America, Halowax was a common brand that was manufactured by Koppers Inc. in the U.S. until 1977 (Kirk-Othmer 1980). The Halowax mixtures have commonly been used in toxicity testing of CNs; however, researchers have reported varying proportions of CN homologue groups in these mixtures.

The key physical and chemical properties of CNs and of the Halowax mixtures that are useful in predicting their environmental fate are listed in tables 1 and table2, respectively. Water solubility estimated using ECOSAR (version 0.99h) is consistently higher than measured values, typically by a factor of about 10. For chemicals with low water solubility (that is less than 1000 µg/L), there can be appreciable error in measured solubility values (Mackay et al. 1999), so the water solubility values obtained from EPIWIN can be used as conservative upper estimates.

Data source: IPCS (2001), unless otherwise noted.
^a Values outside of brackets were experimentally determined by aqueous saturation method (Opperhuizen et al. 1985) for the solid congeners; values in brackets are predicted using ECOSAR (version 0.99, physical state not given).
^b Source: Lei et al. (1999).
^c Values obtained from Puzyn and Falandysz (2007).
^d Measured Kow sources: Opperhuizen (1987), Opperhuizen et al. (1985) (shake flask method, Bruggeman et al. (1982)), Lei et al. (2000) (reversed phase HPLC method).
* Estimated value, using methodologies laid out in Lyman et al. (1982).

Data source: IPCS (2001).
^a Brinkman & De Kok (1980).
^b Estimated value.

¹ The percentage compositions reported represent the average and range of values reported in Falandysz et al. (2006a and 2006). The percentage homologue composition of the Halowax mixtures may vary, however, and different percentage compositions are given in Wiedmann and Ballschmiter (1993), Imagawa and Yamashita (1994), Harner and Bidelman (1997), Falandysz et al. (2006a), Falandysz et al. (2006b), Helm and Bidleman (2003), Espadaler et al. (1997), and Noma et al. (2004).
² The values inside brackets represent the range for each congener.

As can be seen from Tables 3a and 3b, the reported proportions of the homologue groups present in the Halowax mixtures vary.

Natural sources

The presence of CNs in the environment is thought to result mainly from human activity. A possible non-anthropogenic (natural) source of CNs is combustion of wood during, for example, forest fires. Although CNs have been reported to be released into the atmosphere from domestic combustion of wood (see, for example, Lee et al. 2005), no studies documenting release from natural combustion were identified.

Manufacture and import

Mono- and di-CNs, and "naphthalene, chloro derivatives" (a variable chemical mixture that covers the chemical class of CNs, CAS RN 70776-03-3) are on the Canadian Domestic Substances List, which indicates that these substances were in commercial use in Canada between January 1, 1984, and December 31, 1986. Domestic Substances List data indicate that an estimated 700-3500 kg/year of these substances were imported into Canada during this period. Reported uses were as, among other things, organic chemicals, abrasives, polymers, components of plastic and synthetic resins. It does not appear that CNs were ever manufactured in Canada (Holliday et al. 1982), but they were likely imported from manufacturers in the U.S.

According to responses to a voluntary industry survey carried out by Environment Canada in 2003, there was no manufacturing of CNs in Canada from 2000 to 2002. Only one company reported importing CNs (tri- and tetra-CNs) from 2000 to 2002. These imports have since been discontinued.

It is known that in the 1990s Sumitomo 3M Ltd. in Japan imported from Canada 54 tonnes of an adhesive material containing Neoprene FB, a product that contains 3% tri-CNs and 1% tetra-CNs (Yamashita et al. 2003). Neoprene is a registered trademark of DuPont.

Two Canadian industry associations indicated that mono-CNs were used in manometer fluid in very small quantities in a laboratory setting. These associations stated that the CN-containing material in the manometers was replaced with non-CN material in 2004.

Wellington Laboratories of Guelph, Ontario, is a supplier of CN standard materials for analytical purposes (for example, the series of Halowax mixtures and single congeners), and these materials are purchased by laboratories around the world, including in Japan (Wellington Laboratories 2005; Takasuga et al. 2004). It is not known whether any other laboratory supply companies in Canada sell CN standard materials for analytical purposes.

Use

CNs are not currently in commercial use in Canada, the U.S. and many other countries that belong to the Organisation for Economic Co-operation and Development. Production of CNs in the U.S. had ceased completely by 1980 (IPCS 2001). Around 15 tonnes a year of CNs were still being imported into the U.S. as of 1981, mainly for use in refractive index testing oils and capacitor dielectrics (IPCS 2001).

Beginning around 1910, CNs were produced commercially as mixtures of several homologue groups for a variety of uses. Mono-CNs and mixtures of mono- and di-CNs have been used for chemical-resistant gauge fluids and instrument seals, as heat exchange fluids, as high-boiling-point specialty solvents, for colour dispersions, as engine crank case additives and as ingredients in motor tune-up compounds. Mono-CNs have also been used as raw materials for dyes and as wood preservatives with fungicidal and insecticidal properties (Crookes and Howe 1993).

Products made with tri- and greater CNs have been used as impregnants for condensers, capacitors and dipping-encapsulating compounds in electronic and automotive applications, as temporary binders in the manufacture of ceramic components, in paper coating and impregnation, in precision casting of alloys, in electroplating stop-off compounds, as additives in gear oils and cutting compounds, in flame-proofing and insulation for electrical cable and conductors, as moisture-proof sealants, as separators in batteries, in refractive index testing oils, as masking compounds for electroplating, and in grinding wheel lubricants (Kirk Othmer 1980, Fed. Reg. 1983).

Until the early 1970s, General Motors of Canada used automotive capacitors containing CNs, which were imported from the U.S. (Holliday et al. 1982). Dow Chemical Canada reported that it did not use CNs at its Canadian chlorine manufacturing facility (Holliday et al. 1982).

Unintentional production and releases to the environment

CNs are produced unintentionally as a by-product of many industrial processes involving chlorine, especially in the presence of heat, such as waste incineration, cement and magnesium production (Falandysz 1998; SFT 2001; Takasuga et al. 2004), refining of metals such as aluminium with chlorine (Vogelgesang 1986; Aittola et al. 1994) and drinking water chlorination (Shiraishi et al. 1985). It is also thought that CNs may be produced as a by-product in the chloralkali process (Falandysz 1998; Kannan et al. 1998) and in pulp and paper production (Rayne et al. 2004). The amount of CNs released into the environment from these sources has not been well characterized.

CNs are also a contaminant found in commercial polychlorinated biphenyl (PCB) formulations. The mean concentration of CNs in technical Aroclor formulations is reported as 67 µg/g by Falandysz (1998); a later paper gives a range of 5.2-67 µg/g (Yamashita et al. 2000). Based on the amount of PCBs in use or stored as waste in Canada (Environment Canada 2003), and assuming a mean value of 67 µg CNs/g of PCBs, it is estimated that these PCBs contain less than one tonne of CNs.

Domestic burning of coal and wood in the U.K. was estimated to emit 2 kg of CNs per year, including a wide range of tri- through octa-CN congeners; mono- and di-CNs were not monitored as part of the study (Lee et al. 2005). The tri- and tetra-CN homologue groups were the dominant ones detected. In comparison, CN emissions from PCBs were estimated to produce 300 kg of CN emissions in the U.K. each year (Lee et al. 2005).

Emissions from both municipal and special waste incinerators are likely a significant source of CNs in the atmosphere, both in Canada and around the world.

• Helm et al. (2000) analyzed ashes from a medical waste incinerator, a cement kiln, a municipal solid waste incinerator and an iron sintering plant, all in Canada. The samples from the municipal solid wast incinerator, cement kiln and iron sintering plant contained 1.3-2.0 ng SCN/g of ash, while the medical waste incinerator sample contained 3600 ng SCN/g of ash.
• Helm and Bidelman (2003) found the average SCNs (tri- to octa-) in Canadian samples of fly ash to be 1.82 ng/g (municipal solid waste incinerator), 2.09 ng/g (cement kiln), 2.66 ng/g (iron sintering plant) and 5439 ng/g (medical waste incinerator).
• Helm et al. (2003) found CN combustion congener markers in Great Lakes air and around industrial sites in Toronto.

Elsewhere, Schneider et al. (1998) found that all CN homologue groups were found in fly ash from a municipal waste incinerator in Germany. Di- through penta-CNs were found in the largest amounts, and mono- through tri-CNs were also found high amounts during the gas phase of incineration. 1,2,3,5,6,7-Hexa-CN was also formed in large amounts, together with other hexa-CN isomers (Schneider et al. 1998). Takasuga et al. (2004) examined unintentional CN production in municipal waste incinerators in Japan and found that all CN homologue groups are produced in flue gas, with mono- through tetra-CNs being the predominant by-products. The formation of CNs from waste incineration and its occurrence in ambient air during start-up, steady operation and shutdown of municipal solid waste incineration machinery were examined (Takasuga et al. 2004). In 1992, mono- through octa-CNs were measured in ambient air samples from three incineration sites in western Japan. Total CN concentrations in flue gas were 15 000, 4300 and 13 000 ng/m³ during start-up, steady operation and shutdown, respectively. Concentrations of CNs in air samples collected in winter were slightly higher than in those collected summer (Takasuga et al. 2004). It should be noted that incinerators in Japan operate at lower temperatures than those required in Canada, which can effect CN production.

CNs in the environment also result from disposal of products containing CNs in landfill sites and from old industrial sites where CNs were used. The releases from these sources to the Canadian environment have not been characterized.

Environmental fate

The fate of the individual CN homologue groups has been evaluated separately because, as was noted previously, the number of chlorine atoms in the CN structure is a key determinant of a CN's physical and chemical properties (see Table 1). Air-only release is the most likely based on information on current use patterns, although direct release into water could also occur.

Partitioning

Level III fugacity modelling (Table 4) was used to predict into which environmental compartments CNs would partition. CNs tend to partition mainly into air and soil when released only into the air. CNs tend to partition mainly into water and sediment when released only into the water.

* Numbers have been rounded to two decimal places so row totals don't necessarily add to 100 percent.

Persistence in air and long-range atmospheric transport

All CNs, with the exception of mono-CNs, have estimated atmospheric half-lives of longer than two days, based on the reaction with hydroxyl radicals modelled, using rate constants predicted with the Syracuse Research Corporation AOPWIN computer program (version 1.75) model and a daily (24-hour) hydroxyl radical concentration of 5 × 10⁵ molecules/cm³, a typical value for the northern hemisphere (Table 5). There exists a measured hydroxyl radical reaction rate constant for 1,4-di-CN only, which would also give an atmospheric half-life of longer than two days (Klöpffer et al. 1988). The AOPWIN-estimated atmospheric half-life does not consider the chlorine position on the naphthalene rings; therefore, individual isomers may have somewhat higher or lower reaction rates than predicted for their congener group.

The TaPL3 model was used to estimate characteristic travel distance, defined as the maximum distance travelled by 63 percent of the substance after being released into the environment. Beyer et al. (2000) have proposed characteristic travel distances of more than 2000 km as representing high long-range transport potential (LRTP), 700-2000 km as moderate and less than 700 km as low. The LRTP for CNs after release into the air was estimated to be low for mono-CNs, moderate for di-CNs and high for tri- through octa-CNs. Based on these model predictions, mono-CNs are expected to remain primarily in the areas close to their emission sources, while tri- through octa-CNs may be transported in the atmosphere to remote regions such as the Arctic.

Tri- and greater CNs have been detected in air and biota in the Arctic, Antarctic and other regions lacking significant local sources of CNs, which suggests that long-range atmospheric transport is occurring (Helm et al. 2004). The tri- and tetra-CN congeners accounted for 90% to 95% of the total mass of CNs in the Canadian Arctic air samples collected by Harner et al. (1998), with the penta- and hexa-CNs comprising the remainder. Given that tetra-, penta- and hexa-CNs are the homologue groups most dominant in wildlife in polar regions (Corsolini et al. 2002, Helm et al. 2002), penta-and hexa-CNs may also be transported in the air to those regions in significant quantities.

Helm and Bidelman (2005) report the particle/gas distribution in Arctic air and show that the distribution is consistent with that predicted by the octanol-air partition coefficient (K_OA). Distributions on particles in the winter were 5% or less for tri- and tetra-CNs, 20% 35% for penta-CNs and 80 to 100% for hexa-, hepta- and octa-CNs. Herbert et al. (2005) reported the sum of all CN congeners detected (sum of all chlorinated naphthalene congeners detected [SCNs]) at Ny Ålesund and Tromsø, Norway, was equal to 27-48 pg/m³ and 9-47 pg/m³, respectively. Lee et al. (2007) measured SNCs in the Arctic air at levels of 1-8 pg/m³. Tri- and/or tetra-CNs dominated.

Wania (2003) used the octanol-air partition coefficients (K_OA) and the air-water partition coefficients (K_AW, the dimensionless Henry's law constant) to derive the Arctic contamination potential (ACP) of persistent organic chemicals. The two sets of partitioning characteristics with elevated ACP overlap in the range 6.5 < log K_OA < 10 and -0.5 > log K_AW > -3, which also corresponds to a log K_OW range of 5 to 8 (i.e. substances that also have a potential for bioaccumulation) (Wania 2003). The log K_AW ranges at 25°C for each homologue group (except the di- and octa-CNs) were estimated based on the Henry's law constant ranges presented in Table 1, using the following formula:

The log K_OA values were measured by Harner and Bidleman (1998) or estimated for mono-CN based on the following formula:

This information is presented in Table 6. The values that fall in the ranges specified by Wania (2003) are shaded. Based on these criteria, at least some of the congeners from the di-, tri-, tetra-, and penta-CN homologue groups appear to have elevated Arctic contamination potential (ACP).

¹ Measured values from Harner and Bidleman (1998)
* These values calculated based on KOA = KOW/KAW using a median KOW value, and the estimated KAW from the above-mentioned formula

Persistence in water, sediments and soils

There are limited data on abiotic and biotic degradation of CNs. Tetra- through hexa-CNs did not degrade during a 28-day aerobic biodegradation test using spiked natural lake sediment (Järnberg et al. 1999). Some aerobic biodegradation of mono- and di-CNs does occur: after five days in liquid culture medium, 100% of 2-CN, 95% of 1,4-di-CN and 50% of 2,7-di-CN were converted to several hydroxylated products (Kitano et al. 2003).

As Gevao et al. (2000) have noted, anaerobic dechlorination of CNs (similar to that of PCBs) cannot be ruled out, although it has not yet been reported. It is therefore possible that some lesser chlorinated CNs found in anaerobic sediment are degradation products of more highly chlorinated congeners.

Using the BIOWIN model (version 3.52, Syracuse Research Corporation) and the Boethling/SRC extrapolation method, the ultimate aerobic biodegradation half-life in water can be estimated to be 37.5 days for mono- and di-CNs, 60 days for tri- and tetra-CNs, and 182 days for penta- to octa-CNs. The BIOWIN predictions are considered more relevant measures of persistence than those using the Boethling/SRC method, since they estimate times for complete mineralization (i.e. transformation to CO2, H2O, etc.) of the compounds.

Tri- to hepta-CNs have been detected in lake sediments and soils dating back over 18 years (Gevao et al. 2000, Meijer et al. 2001), in quantities suggesting that the half-lives of CNs in sediments and soils are greater than one year, and six months, respectively (mono-, di- and octa-CNs were not analyzed in these studies). The calculations supporting this conclusion are included in Appendix A. Using the data from Meijer et al. (2001), it is possible to estimate the following degradation half-lives for soils that had been amended with CN-containing sewage sludge, based on analysis of an archived sample from 1972 and subsequent sampling in 1990: 7.4 years for tri-CNs, 13.1 years for tetra-CNs and 35.3 years for penta-CNs. Since concentrations of hexa- and hepta-CNs did not decrease significantly, Meijer et al. (2001) could not calculate half-lives.

Based on the weight of evidence, including the predicted persistence of di- to octa-CNs in the air, the high characteristic travel distances for tri- to octa-CNs, the high predicted Arctic contamination potential of di- to penta-CNs, the empirical evidence of the long-range transport of several CNs in Arctic air, the empirical evidence for persistence of tri- to hepta-CNs in sediments and soils, and the predicted persistence of penta- to octa-CNs in water, it is concluded that di- to octa-CNs are persistent, as defined in the Persistence and Bioaccumulation Regulations of CEPA (Canada 2000).

Bioaccumulation

Lipophilic organic substances with log K_ow values equal to or greater than 5, and/or bioconcentration factor (BCF) and bioaccumulation factor (BAF) values equal to or greater than 5000, such as polychlorinated biphenyls (PCBs), dichlorodiphenyltrichloroethane (DDT) and chlordane, have been shown in the field to bioaccumulate and to biomagnify into the tissues of wildlife in the upper trophic levels (Canada 1995). Therefore, if a substance has a biomagnification factor (BMF) significantly greater than 1, this may be an important supporting line of evidence indicating that the substance has the potential to bioaccumulate, especially when relevant bioconcentration factor (BCF) and bioaccumulation factor (BAF) data are limited. Information on biomagnification of CNs in food chains will therefore be considered as part of the weight of evidence that CNs are bioaccumulative substances, as understood in the context of the Persistence and Bioaccumulation Regulations of CEPA (Canada 2000).

Tri- to octa-CNs are neutral organic substances, and log K_ow values equal to or greater than 5 have been measured for these substances using either the shake flask method (Opperhuizen 1987; Opperhuizen et al. 1985) or the reversed phase-high performance liquid chromatograph (HPLC) method (Lei et al. 2000), as summarized in Table 1.

BCFs greater than 5000 have been measured for di- to penta-CNs in aquatic biota, as summarized in Table 7. No measured BCF values were available for hexa-CNs. The BCF studies by Oliver and Niimi (1984, 1985) are deemed to be acceptable, even though a solubilizer (methanol) was used in one of the studies, because the exposure concentrations (low ng/L range) were within the reported range of water solubilities of the CNs being studied (see Table 1). The BCF values reported by Matsuo (1981) are included in Table 7; however, this is a secondary reference, and the primary reference that Matsuo (1981) cites for this BCF data (Kawasaki 1980) does not appear to report these values or the methodology employed. However, no other reports of measured BCF values were found for penta-CNs.

Based on this BCF information, di-, tri-, tetra- and penta-CNs appear to bioconcentrate to a high degree in fish. Octa- and hepta-CNs do not appear to bioconcentrate significantly (Opperhuizen et al. 1985); however, the time of exposure to octa- and hepta-CNs in the experiments was short (seven days). No studies about the bioconcentration of hexa-CNs are available. However, because the log K_ow values for the higher chlorinated CNs are greater than 5, it is predicted that dietary uptake is likely to be much more significant than uptake from water (Arnot and Gobas 2003). Hexa- and hepta-CNs have been found in relatively large amounts in harbour porpoises (Falandysz and Rappe 1996) and in white-tailed sea eagles (Falandysz et al. 1996). Hexa-CNs congeners 66/67 were also found to biomagnify in many aquatic and bird predator species (Falandysz 1997), as is discussed below.

^a Exposure concentrations are estimated ranges from a graphical presentation of results.
^b Equilibrium was not reached within the duration of the experiment.

Tysklind et al. (1998) and Åkerblom et al. (2000) studied the bioaccumulation of CNs in salmon fry (Salmo salar). The fish were fed food to which Halowax 1001, 1014 and 1051 were added at concentrations of 0.1-10 µg Halowax/g of food. The study lasted 41 weeks. Biomagnification factors (BMFs) after 17 weeks that ranged from 0.73 to 2.5 were reported for single-eluting congeners, calculated for salmon fed 2 µg CNs/g of food (Tysklind et al. 1998). The congeners tetra-CN 42, penta-CNs 58 and 61 and hexa-CNs 65 and 69 were observed to have BMFs ranging from 1.0 to 2.5. A QSAR model based on the results of this study predicted that hexa-CNs 66, 67 and 68 would have BMFs ranging from 1.0 to 1.5, while hexa-CNs 64, 70 and 71 would have BMFs ranging from 0.90 to 0.97. Log BMFs for total tri- to octa-CNs after 41 weeks ranged from 3.34 to 3.54 (Åkerblom et al. 2000).

Tetra- to hepta-CNs were found to have BMFs greater than 1 in several aquatic food chains. Hanari et al. (2004) found CN levels in the St. Clair River near Marine City, Michigan, suggesting a BMF of approximately 3 for benthic algae, zebra mussels and round gobies. BMFs for hexa-CNs 66/67 and 69 in a benthic marine food chain in the Baltic Sea were found to be 1.1 and 1.4, respectively (Lundgren et al. 2002). The BMFs for the hexa-CN 66/67 congeners in a Baltic Sea pelagic food chain were found to be 10.9 for plankton/herring and 1.2 for harbour porpoise/herring (Falandysz and Rappe 1996).

Falandysz (1997) calculated BMFs for tetra- through hepta-CN congeners in many types of aquatic biota predator/prey from the Baltic Sea. All exposure was considered to be through food (prey). The predator/prey pairings included herring/plankton, stickleback/plankton, sand eel/plankton, flounder/mussel, black cormorant/fish, white-tailed sea eagle/fish, white-tailed sea eagle/cormorant and harbour porpoise/herring. Almost all of these were found to have BMFs greater than 1 for at least some of the tetra-CNs, including a maximum of 95 for the white-tailed sea eagle/fish. Almost all of the predators were found to biomagnify at least some penta-CN congeners from their food, with the exception of harbour porpoise/herring, the maximum observed BMF being 150 for white-tailed sea eagle/fish. All of the predators were found to biomagnify hexa-CN congeners 66/67 from their food, while some of the predators also biomagnifed other hexa-CN congeners. The white-tailed sea eagles were found to have the highest biomagnification from fish, with BMFs greater than 30 for some tetra-, penta- and hexa-CN congeners. The pairings that biomagnified one or both of the hepta-CN congeners included herring/plankton, flounder/mussel, eagle/cormorant, with the BMF for the white-tailed eagle/fish the largest at 5.7.

Evidence for the biomagnification of CNs can be shown by two other means (see Table 8). First, the relative level of CNs/PCBs is comparatively stable throughout a food chain (krill, icefish and south polar skua). Since PCBs are known biomagnifiers, CNs would be expected to be biomagnifiers in order to maintain this stable relationship (Corsolini et al. 2002). Second, the concentration of CNs in each trophic level increases.

Data source: Corsolini et al. 2002.

The dietary uptake efficiencies of hexa-, hepta- and octa-CNs in northern pike were found to be 35% for octa-CNs, 66% for hepta-CNs and 63% and 78% for hexa-CNs depending on the congener. This is much higher than predicted by the empirical model of Gobas et al. (1988) for molecules of this size (Burreau et al. 1997). The half-lives of certain hexa-CN congeners in rats and humans were found to be similar to those of recalcitrant compounds such as Polychlorinated dibenzo-para-dioxins (PCDDs) and polychlorinated dibenzofurans (PCDFs) that are known to accumulate in organisms and magnify in food chains (Asplund et al. 1994a, and b; Ryan et al. 1993).

Asplund et al. (1994a) studied the retention of a mixture of hexa-CN congeners 66/67 and one unidentified hexa-CN congener in rats. The half-lives were calculated to be 41 days in the adipose tissue and 36 days in the liver, which are comparable with those reported for the most slowly eliminated 2,3,7,8-substituted polychlorinated dibenzofuran (PCDF) congeners (Asplund et al. 1994a). Human blood samples from three individuals exposed in 1979 to CN-contaminated rice oil in Taiwan were monitored for the 1,2,3,4,6,7/1,2,3,5,6,7-hexa-CNs over a period of about 10 years. The measured concentrations resulted in calculated half-lives of 1.5-2.4 years (Ryan and Masuda 1994). These long half-lives in humans are similar to those reported for selected PCDF (Ryan et al. 1993).

Based on the weight of evidence, including in particular measured log K_ow values for tri- to octa-CNs (Table 1) and measured BCF values for di-to penta-CNs in fish (Table 7), and taking into account supporting information on measured BMFs for tetra- to hepta-CNs, the high dietary uptake efficiencies of hexa- to octa-CNs in northern pike, and the very slow elimination of hexa-CNs from the bodies of rats and humans, it is concluded that di to octa-CNs are bioaccumulative, as defined in the Persistence and Bioaccumulation Regulations of CEPA (Canada 2000).

Environmental effects

Aquatic organisms

There are only limited measured, reliable aquatic toxicity data for most CN homologue groups, so the toxicity of these groups was modelled using ECOSAR (version 0.99h). The K_ow and water solubility values were estimated using ECOSAR. The estimated K_ow values were generally similar to available measured K_ow values, but the estimated water solubility values were generally higher than available measured values (see Table 2). ECOSAR has log K_ow cut-offs of 5.0 and 8.0 for acute effect and chronic effect estimations, respectively, so toxicity estimations for homologues with K_ow values greater than these cut-off values were not included in Table 9. As well, any predicted toxicity values greater than the predicted water solubility values were not included in Table 9. This was the case for all of the modelled 14-day median lethal concentration (LC₅₀) values for the earthworm.

¹ LC₅₀ – The concentration of a substance that is estimated to be lethal to 50% of the test organisms.
² EC₅₀ – The concentration of a substance that is estimated to cause some toxic sublethal effect on 50% of the test organisms.

The toxicity of CNs, mostly tested as commercial Halowax mixtures, has been studied in a variety of aquatic species, including aquatic plants, invertebrates, fish and frogs. A summary of most of these studies is given in IPCS (2001). The more recent studies not included in IPCS (2001) were reviewed and key studies are noted below. Some of the tests involved using solubilizers such as acetone, with resulting test concentrations greater than the water solubility of some CN homologues or mixtures. Consequently, the validity of any given study should be considered and interpreted appropriately. It should be noted that the solubilities of the Halowax mixtures have not been well characterized: all Halowax products are described as "insoluble" in the IPCS (2001) report, rather than actual solubilities being reported, as is done for some of the individual CN isomers (refer to Table 1 for solubility estimates for homologue groups). Therefore, the solubilities of the Halowax mixtures were roughly estimated based on the solubilities listed (see Table 1) for the dominant congener groups found in the mixtures (refer to Table 3b), in order to compare the solubilities to the concentrations tested in the toxicity studies. The most sensitive aquatic toxicity values for each homologue group of CNs are summarized below.

Mono-CNs

The most sensitive invertebrate test with mono-CNs was a 96-hour LC₅₀ of 370 µg/L with mysid shrimp (U.S. EPA 1980). No chronic aquatic studies were identified. This value is higher than the ECOSAR 96-hour LC₅₀ for mysid shrimp of 190 µg/L, and is very similar to the chronic ECOSAR 16-day EC₅₀ for daphnids. The most sensitive vertebrate test was a 96-hour LC₅₀ of 690 µg/L with sheepshead minnows (Ward et al. 1981).

The most sensitive chronic toxicity result with Halowax 1000 was 100 µg/L, which caused an 11% reduction in the growth of the marine alga Dunaliella tertiolecta in a seven-day study (Walsh et al. 1977). This effect value is lower than the ECOSAR-predicted chronic value for algae of 575 µg/L, but is in the same range as other ECOSAR acute values for invertebrates. These toxicity values are all below the measured and estimated solubility values of mono-CNs (Table 1). Halowax 1000 is reported in IPCS (2001) to be comprised of 60% mono-CNs and 40% di-CNs (see Table 3a), although these values do not agree with the mean (and ranges) of the experimental percentages (see Table 3b): 15 (6.7-69), 76 (28-76.5) and 6.5 (1.2-44.1) for mono-, di- and tri-CNs, respectively.

Di-CNs

No toxicity studies of di-CNs alone were identified. The most sensitive acute toxicity result with Halowax 1000--comprised of 76% di-CNs and 15% mono-CNs, based on averages reported by Falandysz et al. (2006b) (see Table 3b)--was 100 µg/L. This caused an 11% reduction in the growth of the marine alga Dunaliella tertiolecta in a seven-day study (Walsh et al. 1977). This effect value is lower than the ECOSAR-predicted chronic value for algae of 270 µg/L but is in the same range as other ECOSAR acute and chronic values for invertebrates and fish. These toxicity values are generally below the measured and estimated solubility values of di- and mono-CNs (Table 1).

Tri-CNs

No toxicity studies of tri-CNs alone were identified. However, based on averages reported by Falandysz et al. (2006b), tri-CNs comprise approximately 38.7% of Halowax 1099 (see Table 3b). In a study with Halowax 1099 (which also contains about 48% tetra-CNs), the 96-hour LC₅₀ with the grass shrimp Palaemonetes pugio was 69 µg/L (Green and Neff 1977). Neff and Giam (1977) exposed juvenile horseshoe crabs (Limulus polyphemus) to Halowax 1099 for 96 days. They found that 50% mortality of the late T₁ stage crabs occurred at 27 days at the highest exposure concentration of 80 µg/L (nominal). Significant effects on the length of the intermoult period were observed at 80 µg/L. Acetone was used as a solubilizer in both studies. These toxicity values are within approximately 10 times the measured solubility limit for tri- and tetra-CN but are below the solubility values predicted by ECOSAR. They are also within the range of acute and chronic values predicted for invertebrates and fish using ECOSAR.

Tetra-CNs

No toxicity studies of tetra-CNs alone were identified. However, based on averages reported by Falandysz et al. (2006b), tetra-CNs comprise approximately 48% of Halowax 1099 and 53.3% of Halowax 1013 (see Table 3b). Studies conducted with Halowax 1099 are described above (see tri-CNs). Halowax 1013 (which also contains about 30% penta-CNs) reduced the growth of the marine alga Nitzschia sp. at 500 µg/L (Walsh et al. 1977). A 96-hour LC₅₀ of 74 µg/L was obtained with grass shrimp with Halowax 1013 (Green and Neff 1977). These values are generally within about 10 times the measured solubility limit for tetra-CNs and below the solubility predicted by ECOSAR. Acetone was used as solubilizer in this study. The measured grass shrimp toxicity value is higher than, but within the same range as, the chronic values for invertebrates and fish predicted with ECOSAR. The measured marine algae toxicity value is an order of magnitude higher than the ECOSAR-predicted algae values.

Penta-CNs

No toxicity studies of penta-CNs alone were identified. However, based on averages reported by Falandysz et al. (2006b), penta-CNs comprise approximately 47.7% of Halowax 1014 (see Table 3b). Halowax 1014 also contains 25.3% hexa-CNs, some of which have been found to be quite toxic to mammals, since they have a dioxin-like mode of action, as is discussed below. Penta-CNs comprise approximately 30% of Halowax 1013 (which also contains 53.3% tetra-CNs and 13.1% tri-CNs). A 96-hour LC₅₀ of 74 µg/L was obtained with grass shrimp(Palaemonetes pugio) and Halowax 1013 (Green and Neff 1977). This toxicity value is within about 10 times the measured solubility value and is slightly more than the ECOSAR-predicted solubility of penta-CNs and below the ECOSAR-predicted values for tri- and tetra-CNs. The ECOSAR-predicted chronic toxicities of penta-CNs in fish and daphnids, and green algae are less than10 µg/L and 25 µg/L, respectively.

The U.S. EPA (1980) conducted acute studies with Halowax 1014 (tetra-, penta-, hexa-CNs = 16:47.7:25.3) and brown shrimp (Penaeus aztecus), grass shrimp (Palaemonetes pugio), sheepshead minnows (Cyprinodon variegatus) and striped mullets (Mugil cephalus). The 96-hour LC₅₀ values for these studies (in µg/L) were 7.5, 248, greater than 343 and greater than 263, respectively; however, the experimental details have not been published. Halowax 1014 was not found to have a significant effect on the growth of any of four species of marine algae at concentrations of up to 1000 µg/L during seven-day studies (Walsh et al. 1977). A concentration of 100 µg/L of Halowax 1014 was fatal to 52% of frog larvae (Rana agilis) after 18 hours, while the rest of the treated tadpoles delayed their metamorphosis for three weeks (Buggiani 1980). Only the toxicity value of 7.5 µg/L LC₅₀ for brown shrimp is comparable to or below the measured and ECOSAR-predicted solubility values for penta-CNs.

Talykina et al. (2003) studied the effects of Halowax 1014 on the erythrocytes (red blood cells) of adult medaka fish (Oryzias latipes) exposed in ovo. Fragmented nuclei (micronucleated erythrocytes) were induced by Halowax 1014 at the lowest concentration tested, 300 ng of Halowax 1014/g of egg (equivalent to 300 ppb).

Villalobos et al. (2000) conducted partial lifecycle toxicity tests in which medaka fish eggs (Oryzias latipes, early gastrula) were injected with Halowax 1013, 1014 or 1051 dissolved in triolein. Following exposure, embryos developed and fry were reared to sexual maturity (four months), at which time they were euthanized. Halowax 1013 caused 64% mortality of the eggs after 16 days at a dose of 10 ng/egg (approximately 10 ppm), with half of those deaths occurring on day 3, with the greatest dose (30 ng/egg) causing 28% mortality late in development (Villalobos et al. 2000). Halowax 1013 also caused premature hatching of embryos at all doses (0.3-30 ng/egg). An LD₅₀ is the Lethal Dosage (LD) that results in mortality in half the test population. Median lethal dose (LD₅₀) measurements could not be calculated for the Halowax 1013 tests since they did not elicit monotonic dose-response relationships. Therefore, the threshold dose of Halowax 1013 in this study appears to be 0.3 ng/egg (approximately 0.3 ppm).

Hexa-CNs

No toxicity studies of hexa-CNs alone were identified. The only Halowax mixture containing a significant proportion of hexa-CNs is Halowax 1014, which contains approximately 25.3% (21-35.3%) hexa-CNs based on data reported by Falandysz et al. (2006b) (see Table 3b). Acute studies with Halowax 1014 are described above, with LC₅₀ values ranging from 7.5 µg/L to more than 343 µg/L for invertebrates, fish and frogs. The lowest of the toxicity values--7.5 µg/L LC₅₀ for brown shrimp--is within a factor of 100 of the measured solubility and below the ECOSAR predicted solubility of hexa-CNs. ECOSAR predicted chronic toxicity values for hexa-CNs with fish/daphnid are 1-3 µg/L.^{Footnote 1} 

Hepta-CNs

No toxicity studies of hepta-CNs alone were identified. The ECOSAR-predicted chronic toxicity of hepta-CNs to fish and daphnids is in the low µg/L range (see Table 9). The only Halowax mixture containing a significant proportion of hepta-CNs is Halowax 1051, which, based on data reported by Falandysz et al. (2006b), contains only about 7.6% hepta-CNs and 91.4% octa-CNs (see Table 3b). Toxicity studies with Halowax 1051 are described below.

Villalobos et al. (2000) injected medaka embryos with Halowax 1051, as described earlier. Halowax 1051 was the least toxic of the three Halowax mixtures tested; embryo mortality did not exceed 20% for all doses after 8-16 days of exposure. However, Halowax 1051 caused significantly decreased gonadosomatic indices for females after 122 days at all four doses (0.3-10 ng/egg), with no evident dose-response trend.

Talykina et al. (2003) studied the effects of Halowax 1051 on the erythrocytes (red blood cells) of adult medaka exposed in ovo. Fragmented nuclei (micronucleated erythrocytes) were induced by Halowax 1051 at the lowest concentration tested, 300 ng of Halowax 1051/g of egg (0.3 ng/embryo) (equivalent to 300 ppb).

Octa-CNs

Three acute studies of octa-CN were identified: with water fleas (Daphnia magna) (LeBlanc 1980), sheepshead minnows (Cyprinodon variegatus) (Heitmuller et al. 1981), and mysid shrimp (Mysidopsis bahia) (U.S. EPA 1980). In the first two of these tests, no effects were seen at the highest concentrations tested (530 mg/L and 560 mg/L, respectively). In the third study, the 96-hour LC₅₀ was greater than 500 mg/L (no further experimental results were available). This indicates that octa-CNs are of relatively low acute toxicity to aquatic organisms. These effect concentrations are far above both the measured and predicted solubilities of octa-CNs. No ECOSAR toxicity predictions for octa-CN were available, since the K_ow value of octa-CNs is outside of the acceptable range for this model.

Toxicity studies with Halowax 1051, which is comprised of approximately 91.4% octa-CNs, are described for hepta-CNs, above.

Terrestrial wildlife

This section focuses on the toxicity of hexa- to octa-CNs, with particular emphasis on exposure via food, since food ingestion is expected to be the principal exposure pathway for these high log K_ow (7.5-8.5) homologue groups.

Mammalian studies with CNs are summarized in IPCS (2001), and some additional studies were also reviewed. No long-term mammalian studies were identified. A study lasting up to 135 days was conducted with ewes (Brock et al. 1957). Most studies were short-term, and many were conducted with laboratory animals such as rats, rabbits and guinea pigs. Cattle, however, appear to be more sensitive to CNs than laboratory rodents and are also more sensitive to CNs than sheep (IPCS 2001). No studies with mono- or di-CNs have been conducted with cattle.

Mixtures of penta- and hexa-CNs

Only two studies of the toxicity of CNs to birds have been identified. These studies involved turkey poults and chickens a diet containing Halowax 1014 (25.3% hexa-, 47.7% penta- and 16% tetra-CNs; see Table 3b) for 40 days (Pudelkiewicz et al. 1958, 1959). A dose of 20 mg/kg feed caused 50% mortality of the turkeys but had little effect on the chickens. At 5 mg/kg feed, the CNs caused 6.5% turkey mortality, and their weight gain was reduced by 33%.

In a rabbit study, Flinn and Jarvik (1936) administered two CNs mixtures subcutaneously. After daily administration of the mixtures, which consisted mainly of tetra- and penta-CNs, and of penta and hexa-CNs at doses of 30 mg (in paraffin oil) per day per rabbit, all the rabbits died (5/5 per group) by days 12-26. Autopsies found that the livers had many yellow areas and a wide zone of necrosis (Flinn and Jarvik 1936).

Rats fed a mixture of penta- and hexa-CNs (125 mg/rat on alternate days) for 26 days showed moderate liver changes (swollen and vacuolated liver cells, as well as necrosis and degeneration of scattered cells). All other organs (no specifics given) were found to be normal after microscopic examination (Bennett et al. 1938). Groups of 10 rats fed penta- and hexa-CNs in feed, dosed at either 100 mg/rat per day for 55 days or 300 mg/rat per day for 33 days, all died or were moribund (Drinker et al. 1937, Bennett et al. 1938).

Gastrointestinal lesions and severe liver damage or death were reported in sheep orally administered 1.1 mg/kg body weight per day of Halowax 1014 for 90-135 days (Brock et al. 1957).

Decreased sperm production occurred in bulls fed penta- and hexa-CNs (50 to 200 mg daily for approximately seven weeks) (Vlahos et al. 1955).

Hexa-CNs

Oral exposure of pigs to hexa-CNs at 19-22 mg/kg body weight per day for up to 10 days caused degeneration of the liver and kidneys, and mortality, while 17.1-17.6 mg/kg body weight per day caused depressed vitamin A levels (Link et al. 1958, Huber and Link 1962). Oral exposure of rats to hexa-CNs at doses of 0.3-2.3 mg/rat per day over a period of 56-84 days caused dose-dependent increases in relative liver weights, and doses of 20 and 60 mg/rat/day caused liver damage (Weil and Goldberg 1962). Oral doses of 1.1 mg/kg body weight per day of hexa-CN congeners administered to cattle via gelatine capsules for 5-10 days resulted in severe systemic disease (hyperkeratosis) (Bell 1953).

Hepta-CNs

Oral doses of 0.69-2.4 mg/kg body weight per day of hepta-CNs administered to cattle via gelatine capsules for seven to nine days resulted in severe systemic disease (hyperkeratosis) (Bell 1953). A single oral dose of 500 mg/kg body weight of hepta-CNs resulted in the death of all three rabbits during a seven-day period (Cornish and Block 1958).

Octa-CNs

A single oral dose of 500 mg/kg body weight of octa-CNs resulted in the death of all three rabbits during a seven-day period (Cornish and Block 1958). An oral dose of 2.4 mg/kg body weight per day of octa-CNs administered to cattle via gelatine capsules for nine days resulted in severe systemic disease (hyperkeratosis), while an oral dose of 1.0 mg/kg body weight per day for 11 days resulted in mild hyperkeratosis (Bell 1953).

Toxic mode of action

As with other halogenated aryl hydrocarbons, such as PCBs and polychlorinated dibenzo-para-dioxins (PCDDs) and PCDFs, at least some of the biochemical (for example, drug-metabolizing enzyme induction, hormonal changes) and toxic (for example, skin disorders, weight loss, hepatotoxicity, reproductive toxicity) responses of CNs are believed to be mediated via the cytosolic aryl hydrocarbon (Ah) receptor (Goldstein and Safe 1989), which has been intensively studied for the model compound tetrachlorinated dibenzo-para-dioxins (TCDD) (IPCS 2001). A 3-methylcholanthrene (MC)-type induction (Ah receptor-dependent induction of CYP1A1, mostly measured as ethoxyresorufin-O-deethylase [EROD] and/or arylhydrocarbon hydroxylase [AHH] activity) is typical for TCDD-like compounds.

The strength of the binding of a compound to the Ah receptor can be measured through several bioassays, such as the ethoxyresorufin-O-deethylase (EROD), arylhydrocarbon hydroxylase (AHH) and luciferase assays.

CNs are able to induce the cytochrome P-450-dependent microsomal monooxygenases, as shown in studies with rats in vivo (Wagstaff 1973, Ahotupa and Aitio 1980, Campbell et al. 1981, 1983, Cockerline et al. 1981, Safe et al. 1981, Mäntylä and Ahotupa 1993) and in vitro (rat hepatoma cells; Hanberg et al. 1990, 1991), with chick embryos in ovo (Engwall et al. 1993, 1994) and in vitro (Brunström et al. 1995), with eider duck embryos in ovo (Engwall et al. 1993, 1994) and with fish in vivo (Holm et al. 1993, 1994, Norrgren et al. 1993).

As observed for other halogenated aryl hydrocarbons, the relative induction potencies of CNs appear to depend on the degree and position of chlorine substitution of the naphthalene ring, as described below. In several test systems, the most potent EROD and AHH inducers were 1,2,3,4,6,7/1,2,3,5,6,7-hexa-CN, 1,2,3,4,5,6,7-hepta-CN and some unidentified congeners present in Halowax 1014 (Campbell et al. 1983; Hanberg et al. 1990, 1991; Engwall et al. 1993, 1994; Norrgren et al. 1993; Brunström et al. 1995). Octa-CNs were also found to cause a dose-dependent AHH increase in microsomes of rats (Campbell et al. 1981). Halowax 1014 (tetra-, penta-, hexa-CNs = 20:40:40) and the two hexa-isomers maximally induced EROD to a level of about 15-20% of the maximal activity induced by TCDD in cultured chick embryo liver (Brunström et al. 1995). Relative potencies, compared with TCDD, determined for EROD induction in rat hepatoma H4IIE cells in vitro were given as 0.002 for the 1,2,3,4,6,7/1,2,3,5,6,7-hexa-CN and two other hexa-isomers, and as 0.003 for a hepta-CN congener (Hanberg et al. 1990, 1991). The lowest observed effective dose for EROD induction in chick embryos by the 1,2,3,4,6,7/1,2,3,5,6,7-hexa-CN mixture administered via the air sac was 0.1 mg/kg of egg (Engwall et al. 1993, 1994). An ED₅₀ is the Effective Dosage (ED) that produces an effect in half the test population, The ED₅₀ values for EROD induction in chick embryos by this hexa-mix and Halowax 1014 were estimated to be 0.06 mg/kg of egg and 0.2 mg/kg of egg, respectively (Engwall et al. 1993, 1994).