A Nationwide Survey on Organic Solvent Components in Various Solvent Products: Part 2. Heterogeneous Products Such as Paints, Inks and Adhesives*

Transcription

1 Industrial Health, 1983, 21, A Nationwide Survey on Organic Solvent Components in Various Solvent Products: Part 2. Heterogeneous Products Such as Paints, Inks and Adhesives* M. KUMAI1), A. KOIZUMI1), K. SAITO2), H. SAKURAI3), T. INOUE4), Y. TAKEUCHI4), I. HARA5), M. OGATA6), T. MATSUSHITA7) and M. IKEDA1) 1) Department of Environmental Health, Tohoku University School of Medicine, Sendai 980, Japan 2) Department of Hygiene and Preventive Medicine, Hokkaido University School of Medicine, Sapporo 060, Japan 3) Department of Preventive Medicine and Public Health, Keio University School of Medicine, Tokyo 160, Japan 4) Department of Hygiene, Nagoya University School of Medicine, Nagoya 446, Japan 5) Department of Public Health, Kansai Medical University, Moriguchi 570, Japan 6) Department of Public Health, Okayama University Medical School, Okayama 700, Japan 7) Department of Hygiene, Kagoshima University Faculty of Medicine, Kagoshima 890, Japan (Received May 9, 1983 and in revised form June 13, 1983) Abstract : Commercial heterogeneous solvent products (e.g., paints, inks and adhesives) were collected nationwide in Japan in The vapor phase of the product containers were analyzed for volatile organic solvent constituents by means of FID-gas chromatography on two FS-WCOT (OV-101 and PEG-600) capillary columns. Of 657 products collected (358 paints, 62 inks, 165 abhesives and 72 others), 136 samples were not analyzable because 75 gave numerous peaks (presumably containing gasoline) and others had no volatile component. Among the remaining 521 samples (298 paints, 52 inks, 120 achesives and 51 others), 70 gave only one peak while others gave multiple peaks, indicating the mixture of solvents rather than single solvent was commonly used. Of the organic solvent components identified, toluene was the most popular solvent throughout paints * This work was supported in part by a Grant-in-aid for Co-operative Research for 1980 (Chief ; Prof. T. Inoue : Grant Number ) from the Ministry of Education, Science and Culture of the Government of Japan. Offprint requests to : Prof. M. Ikeda, Department of Environmental Health, Tohoku University School of Medicine, Sendai 980, Japan.

2 186 M. KUMAI, et al. (appearing in 80%), inks (62%), adhesives (51%) and others (65%), being detected in 70% of the total products analyzed. This popularity of toluene was followed by xylenes [predominantly m- (66%) and p-isomer (61%)] in the case of paints, isopropyl alcohol (35%) in inks, and n-hexane (27%) and methyl ethyl ketone (23%) in adhesives, while benzene was not detected in any samples analyzed nor even in the gasoline-containing products. The concentrations of each solvent component in the vapor phases varied depending on the products, following a log-normal distribution. When the relative exposure risk of each of eight leading solvents of various chemical structures was calculated as the geometric mean concentration multiplied by the frequency of detection, n-hexane in adhesives was highest in the risk ranking while toluene and xylenes in paints scored much less because of low volatility. Key words : Adhensives-GC analysis-inks-organic solvent-paints INTRODUCTION The survey of solvent components in commercial solvent-containing products is expected to offer one of the basic information requested for the protection of health of workers in workshops where organic solvents are used. This is also the case when heterogeneous solvent products containing nonvolatile components (such as paints, inks and adhesives) are considered. Information is, however, extremely scarce for paints and inks1-3), and the solvent constituents of adhesives were reported only occasionally in relation to the etiology of polyneuropathy4,5). In conducting a nationwide mass survey on heterogeneous products, it became self-evident that the gas chromatographic (GC) analysis for organic solvent components of these products is disturbed by the existence of nonvolatile constituents and that the direct injection of such sample into a GC apparatus is not recommendable. It was further considered impractical to distillate numerous heterogeneous products in obtaining liquid solvent fractions; distillation temperature should be variable depending on the solvent components and decomposition of nonvolatile fraction often takes place especially when plastics-containing adhesives are heated. In the present study, head space technique was applied for GC analysis of solvent components in heterogeneous materials6), to avoid troublesome process of distillation so that a large number of paints, inks, adhesives and other heterogeneous products will be analyzed for volatile constituents even if semiquantitatively. The results will be presented in this report to show the present situation of the use of various organic solvents in heterogeneous nonvolatile fraction-containing products. The cases of homogeneous solvent products like thinners will be described in a twin paper7) which will be published elsewhere.

3 SOLVENT ANALYSIS OF PAINTS, INKS, ADHESIVES, ETC. 187 MATERIALS AND METHODS Heterogeneous solvent products (e.g., paints, inks, adhesives) were collected in 1980 from seven districts of Japan to cover an entire country, as summarized in Table 1. In total, 358 paints, 62 inks and 165 adhesives were collected together with 72 other miscellaneous products (e.g., turbid degreasers, stain removers, resin hardners, paint materials, and reagents). It should be noted that those products known beforehand to contain gasoline were excluded because of analytical limitations. Each sample was kept in a 200 ml glass bottle with about 50 ml vapor phase space over the liquid phase. After standing at room temperature for at least one week, the cap was minimally opened and 2 ml of the vapor in the bottle was taken to inject immediately to a gas chromatograph (GC). The GC used was a Hitachi Model 163 (Hitachi Ltd., Tokyo, Japan), equipped with FID's and two ƒó 0.25 mm ~ 50 m FS-WCOT capillary columns [Silicone OV 101 and polyethylyne glycol (PEG) 6000] and connected with both a Hitachi Recorder Model 056 and a Takeda-Riken Automatic Integrator (Takeda-Riken Industry Co. Ltd., Tokyo, Japan). The system was assembled so that the vapor injected was transferred to the columns with a split ratio of 1/40 and that the signal voltage from the GC was reduced to one tenth for the integrator. The GC conditions employed were as follows; supply of H2 and air to FID's at 0.8 and 1.5 kg/cm2, respectively, N2 flow to the OV 101 column at 35 ml/min and to the PEG 6000 column at 20 ml/ min, and the temperatures of the oven and the injection port at 80 Ž and 180 Ž, respectively. The attenuation and the range of the GC were set at 16 and 102, respectively. Selector nobs of the integrator were set at the positions as follows; CH Selector at CH 1, Peak Width at 20S and Minimum Area at Usual durations of GC analysis were as summarized in Table 2, but the duration could be prolonged when considered necessary. The constituents were identified by the comparison of retention times on the two columns of samples with those of authentic materials. To eliminate day by day variation in the retention time, toluene was analyzed as a standard every day and the retention time of toluene was employed as a measure. Repeated deter- Table 1. Collection of heterogeneous solvent products by various districts in Japan

4 188 M. KUMAI, et al. Table 2. Duration of GC analysis minations, 173 in total during the 8 month period of analysis, resulted in }0.95 min as the mean }SD of the retention time for toluene on the OV 101 column (C.V. 9.3% ) and 6.27 }0:51 min (8.2% ) on the PEG 6000 column. A gradual shortening in retention time was observed during the 8 months. The first, middle and last 10 determinations each among the 173 determinations on the OV 101 column gave 11.9 }0.15 min (1.3% ), 9.44 }0.06 min (0.6% ), and 9.29 }0.12 min (1.3% ), respectively, while the counterpart palues on the PEG 6000 column were 7.08 }0.17 min (2.4% ), 5.83 }0.20 min (3.5% ) and 5.80 }0.03 min (0.5% ), respectively. In order to obtain the factors to convert the observed peak area values to the vapor concentrations, 2 ƒêl each of the authentic liquid materials were injected and the peak area value was determined. The conversion factor for each chemical was calculated from the peak area value after correction for the specific gravity and the molecular weight of the authentic solvent and the theoretical volume ( ) of the vapor to be occupied by one mole of the chemical at 25 Ž under 760 mm Hg. From the peak area values observed, it was also possible to calculate the "relative molar response" (the value for n-heptane set as 700)8), after correction for the specific gravity and the molecular weight. The theoretical relative molar response could be figured out based on effective carbon numbers9). Such calculations were made on 44 chemicals which were identified as the constituents of the solvent products as shown in Table 6. For carbon disulfide and diethyleneglycol monopropyl ether, however, the theoretical values were excluded due to the lack of the information. The comparison between the theoretical values and the experimental values revealed a very close correlation with a regression line of Y= X (r=0.981, p<0.001, n=44) where X is the experimental value and Y is the theoretical value. Thee slope is very close to one, and the intercept on the Y axis is negligible as the mean } SD of X is 423 }215 while that of Y is 412 }228. The applicability of the experimental values was thus validated. Sensitivity of the method employed RESULTS Trials were made to compare the sensitivity of the vapor phase analysis with

5 SOLVENT ANALYSIS OF BAINTS, INKS, ADHESIVES, ETC. 189 that of the liquid phase analysis. For this purpose, 14 degreasers with a few components and 9 thinners of multiple components both with very homogeneous appearance were subjected to the two phase analyses, i.e., 2 ml of the vapor phase and 2 Đl of the liquid phase were injected in series to GC and the results were compared. With degreasers, it was possible by the vapor phase analysis to detect all the major components of every samples, e.g., trichloroethylene, tetrachloroethylene, 1,1,1-trichloroethane, acetone, methyl ethyl ketone, and methanol. The vary minor components existing as stabilizers (e.g., 1,4-dioxane in chlorinated hydrocarbon solvents) or impurities (e.g., 2- or 3-methyl pentane in n-hexane) were not always detectable by the vapor phase method especially when the concentration in the liquid was 1% or less, while the liquid phase method could detect them. The sensitivity observed in degreaser analysis was essentially applicable also in thinner analysis. Typical cases are summarized in Table 3 to show the reliability of the vapor phase analysis on qualitative and semiquantitative basis. It is clear that all the components in the liquid could be detected by the vapor phase in Cases 1, 2 and 3. When a component is more volatile, the relative concentration in the vapor phase used to be higher than the percentage in the liquid phase. In Case 4, the summation of the percentage in the liquid is not 100%. This means that the component(s) in the liquid (about 15% ) was less volatile than 3-methyl cyclohexanone and was not detected by the vapor phase Table 3. Comparison between vapor and liquid phase analysis results

6 190 M. KUMAI, et at analysis. In fact, while 3-methyl cyclohexanone shared 16% in the liquid phase, its vapor concentration was much less than that of other minor, yet more volatile components. It should, however, also be mentioned that the vapor phase analysis of pure 3-methyl cyclohexanone could detect 6591 ppm of the compound regardless of its poor volatility. Removal of unanalyzable samples W hen the vapor phase of the sample was subjected to GC, it was found that no peak could be detected with some products (e.g., water paint). Some other samples gave multiple peaks, suggesting that the organic phase was probably gasoline with multiple component. In practice, when more than 10 peaks appeared in one chromatogram, the solute was considered to be gasoline. Such case was most frequently encountered with paints (15% of the samples collected) followed by adhesives (10% ). It should be noted that no benzene was detected in any of gasoline-containing products. This positive exclusion was achieved by successive or simultaneous analysis of the suspicious product and the sample containing authentic benzene for precise comparison of the retention time. A few organic solvent components in some products could not be identified with any of the authentic chemicals tested. As a result, about 80% of the heterogeneous porducts collected was found analyzable for organic solvent components (Table 4). Number of peaks per product Table 5 summarizes the numbers of peaks observed per product together with those identified. The distribution pattern of the numbers of peaks per product is not clear, and normal distribution is provisionally considered so that the mathematical means and SDs are given in Table 5. The large coefficient of variation, however, might suggest poor fittness of the distribution. It is apparent that most products (more than 80%) contain more than one kind of organic solvent, i.e., the solvent of these porducts is a mixture of solvents and not of a single component. When paints, inks and adhesives are compared, paints embrace most multiple solvents having about 6 peaks per product, followed by adhesives Table 4. Analyzable samples among the products collected.

7 SOLVENT ANALYSIS OF BAINTS, INKS, ADHESIVES, ETC. 191 Table 5. Number of peaks per product of paints, inks, adhesives, and others (about 4) and inks (about 3) in the decreasing order. It was not always possible to identify all the peaks observed; there used to remain one unidentifiable peak per product. It should be stated, however, that efforts were concentrated to identify major peaks (i.e., peaks with larger peak areas) first and that those unidentified were minor components with peak areas of less than one-tenth to one-handredth when compared with major ones. Solvent components in various products The solvents identified as the components of organic phase of the heterogeneous products totaled 46, and are classified depending on the chemical structures in Table 6, in which the frequencies are shown together with the vapor phase concentrations observed. A preliminary analysis with 364 toluene concentrations observed revealed that the concentrations distribute log-normally. Accordingly, the observed concentration of each solvent is expressed as a geometrical mean in ppm and a gometrical standard deviation (dimensionless). Toluene is the most commonly used solvent; the frequency was highest in paints (i.e., 80%) but almost equally high in other kinds of products with a grand mean of 70%. Three isomers of xylenes are also widely used, espicially in paints; the common use of p-xylene should be noted in addition to popular m-isomer. They were detected usually in combination with toluene, and ethylbenzene was present only together with toluene/xylenes. The use of styrene was not expected but it was present in paints and adhesives. Reversely, benzene was not found in any products. Among the well known hepatotoxins, carbon tetrachloride, 1,1,2-trichloroethane and 1,1,2,2-tetrachloroethane were not detected and chloroform was present only in one case of the adhesive. The use of trichlorothylene, tetrachloroethylene and

8 192 M. KUMAI, et at.

9 SOLVENT ANALYSIS OF BAINTS, INKS, ADHESIVES, ETC. 193

10 194 M. KUMAI, et al. 1,1,1-trichloroethane appear to be very limited among the products other than degreasers. Of alcohols, only methanol and isopropyl alcohol are frequently detected, while ethyl acetate and methyl ethyl ketone are exclusively utilized among esters and ketones, respectively. Application of ethers (including glycol derivatives), alicyclic compounds and other miscellaneous compounds are not popular except for n-hexane in adhesives. The only one case in which carbon disulfide was detected was the production material collected from a rayon factory, a site wellknown for the use of this solvent. In paints, the most popular solvents are toluene (80% ) and xylenes (61-66% ), even though their vapor phase concentrations are rather low and not higher than 10,000 ppm. Their high frequencies are followed by two ketones, methyl ethyl ketone (26% ) and methyl isobutyl ketone (26% ) with rather high vapor concentration for the former. The next frequent is methanol (17% ); while it is not met very frequently, its high vapor concentration should be taken into account in relation to the insidious toxicity to the optic system. The most popular organic components in inks are toluene (62% ), two alcohols of methanol (25% ) and isopropyl alcohol (35% ), and methyl ethyl ketone (21 % ). Application of xylenes in inks are much less (4 to 13% ) than in paints. Adhesives contain toluene (51 % ), n-hexane (27% ) and methyl ethyl ketone (23% ). Attention should be paid to the high vapor concentration of n-hexane in relation to the causative role in peripheral neuropathy. No trial was made to define typical solvent components of other miscellaneous products because they consist of turbid degreasers, production materials, hardners, and other various ones. It is yet possible to point out that toluene is the most popular solvent even in the products of this poorly defined category. DISCUSSION As no information is available apparently on the total population of solventcontaining products in Japan, no suitable samping strategy could be found to ensure that the collected samples represent the population. Efforts were made, however, to collect samples from varieties of factories in various districts of Japan as shown in Table 1, so that the results of the present study may hopefully depict the current general use pattern of organic solvents in paints, inks and adhesives. Noteworthy is the fact that benzene, a human leukemogen10), was not detected in any heterogeneous solvent products so far studied. This makes a sharp contrast with the observation that the current unleaded automobile gasoline used to contain benzene up to several percents11,12). Such contrast may owe to the difference in production process, e.g., the benzene in the unleaded gasoline could be solely attributable to the generation during reforming process of crude gasoline to increase aromatic fractions. It is also worthy of attention that n-hexane, a well known causative agent of occupational polyneuropathy13,14), still stands as the leading

11 SOLVENT ANALYSIS OF PAINTS, INKS, ADHESIVES, ETC. 195 solvent in the adhesives for industrial use, while the use of methyl butyl ketone, another potent neurotoxin15,16). is very limited in Japan. In order to obtain a typical composition of paints, ink and adhesives, one most popular solvent was selected from each of the eight categories of chemical structure in Table 6 (xylenes were selected from the aromatics in addition to toluene, while ethers, glycol derivatives and alicyclic compounds were represented by one solvent). The frequencies of the eight solvents, i.e., toluene, xylenes (the average of the three isomers), dichloromethane, isopropyl alcohol, ethyl acetate, methyl ethyl ketone, tetrahydrofuran and n-hexane are shown in the form of an octangle in Fig. 1. It is apparent that paints contain toluene and xylenes much more frequently than inks and adhesives, while isopropyl alcohol is more often observed in inks than in other two. The high popularity of n-hexane in adhesives in addition to toluene (but not xylenes) should not be ignored. The frequencies were further multiplied by the observed mean concentrations in the vapor phase to obtain an index for exposure risk via inhalation, and the results are shown in an octangular form in Fig. 2. The exposure risk of utilizing adhesives is now characterized by the three highly volatile solvents, primarily n-hexane followed by dichloromethane and methyl ethyl ketone. Paints appear to be less dangerous for handling because the risks due to the two major components, toluene and xylenes, are brought down by the low volatilities. Regarding the analytical procedures employed, it is true that the results from vapor phase analysis does not reflect the proportion of the components in the liquid phase17). It is however also true that, without by passing the troublesome procedure of distillation, it should be very difficult to carry out massive analyses Fig. 1. Frequency octangles for paints, inks and adhesives.

12 196 M. KUMAI, et al. Fig. 2. Exposure risk octangles for paints, inks and adhesives. of a large number of heterogeneous solvent products, and that, by vapor phase analysis, the most volatile (and therefore dangerous) fractions in the product can be detected at least semiquantitatively as the causative agents of inhalaion-induced occupational poisoning. It is no use to state the importance of estimation of the liquid composition. The difficulties are associated with the fact that the vapor pressure of the organic solvent in question is readily modified by the existenct of other solvent(s) in the liquid phase, especially when the physico-chemical characteristics of the two are different from each other17,18). Now that the typical composition of paints, inks and adhesives are identified, further theoretical and experimental trials are currently in progress in our laboratories to develop a system so that the liquid phase composition be estimated based on the results from the vapor phase analysis. REFERENCES 1) Shoji, H., Yamamoto, T., Nishida, K. and Ozaki, Y. (1966). The gas chromatographic analysis of laquer thinner. Jpn. J. Hyg., 20, 364 (in Japanese). 2) Tokunaga, R., Takahata, S., Onoda, M., Ishii, T., Sato, K., Hayashi, M. and Ikeda, M. (1974). Evaluation of exposure to organic solvent mixture. Mt. Arch. Arbeitsmed., 33, ) Hanninen, H., Eskelinen, L., Husman, K. and Nurminen, M. (1976). Behavioral effects

13 SOLVENT ANALYSIS OF PAINTS, INKS, ADHESIVES, ETC. 197 of long-term exposure to a mixture of organic solvents. Scand. J. Work Environ. Health ) Suzuki, T., Shimbo, S., Nishitani, H., Oga, T., Imamura, T. and Ikeda, M. (1974). Muscular atrophy due to glue shiffing. Int. Arch. Arbeitsmed, 33, ) Takenaka, S., Tawara, T., Yamada, T., Okajima, T. and Tokuomi, H. (1972). Polyneuropathv due to glue sniffing. Jpn. Med. J. (Nippon Iji Shinpo), 2515, 33. (in Japanese). 6) Drexler, H.-J. and Osterkamp, G. (1977). Head-space analysis for the quantitative determination of trichloroethylene and tetrachloroethylene in oils and liquid paraffin. J. Clin. Chem. Clin. Biochem., 15, ) Inoue, T., Takeuchi, Y., Hisanaga, N., Ono, Y., Iwata, M., Ogata, M., Saito, K., Sakurai, H., Hara, I., Matsushita, T. and Ikeda, M. (1983). A nationwide survey on organic solvent components in various solvent products : Part 1. Homogeneous products such as thinners, degreasers, reagents and others. Ind. Health, 21, ) Ettre, L.S. (1962). Relative molar response of hydrocabons on the ionization detectors. In : Gas Chromatography, (Edited by Brenner, N., Callen, J.E. and Weiss, M.D.), p Academic Press, New York and London. 9) Sternberg, J.C., Gallaway, W.S. and Johnes, D.T.L. (1962). The mechanism of response of flame ionization detectors. In : Gas Chromatography (Edited by Brenner, N., Callen, E.J. and Weiss, M.D.) p Academic press, New York and London. 10) International Agency for Research on Cancer (1982). Benzene. IARC Monographs on the Evaluation of Carcinogenic Risk of Chemicals to Humans, 29, ) Berlin, M., Gage, J. and Jonnson, E. (1974). Increased aromatics in motor fuels : A review of the environmental and health effects. Work Environ. Health, 11, 1. 12) Ikeda, M. (1977). Aromatic contents in automobile gasoline. Jpn. J. Hyg., 32, 62 (in Japanese). 13) Yamamura, Y. (1969). n-hexane polyneuropathy. Folia Psychiatr. Neurol. Jpn., 23, ) Inoue, T., Takeuchi, Y., Takeuchi, S., Yamada, S., Suzuki, H., Matsushita, T., Miyagaki, H., Maeda, K. and Matsumoto, T. (1970). A health survey on vinyl sandal manufactures with high incidence of "n-hexane" intoxication. Jpn. J. Ind. Health (Sangyo Igaku), 12, 73 (in Japanese). 15) Billmaier, D.J., Yee, H.T., Allen, N., Craft, B., Williams, N., Epstein, S. and Fontaine, R. (1974). Peripheral neuronathy in a coated fabrics plant. J. Occup. Med., 16, ) Allen, N., Mendell, J.R., Billmaier, D.J., Fontaine, R.E. and O'Neill, J. (1975). Toxic polyneuropathy due to methyl n-butyl ketone. Arch. Neurol., 32, ) Gmehling_ Onken. U. and Arlt. W. ( ). Vapor-Liquid Equilibrium Data Collection, Chemistry Data Series Vol. 1, DECHEMA. Frankfurt. 18) The Chemical Society of Japan (1975). Kagaku Binran (Chemistry handbook), 2nd revised edition, p Maruzen, Tokyo (in Japanese).