Ann. occup. Hyg., Vol. 49, No. 8, pp. 719–725, 2005 # 2005 British Occupational Hygiene Society Published by Oxford University Press doi:10.1093/annhyg/mei040
Exposure to Low Molecular Weight Isocyanates and Formaldehyde in Foundries Using Hot Box Core Binders ˚ 1 ˚ ¨ 1,2 ´ 1 HAKAN WESTBERG *, HAKAN LOFSTEDT , ANDERS SELDEN , Downloaded from https://academic.oup.com/annweh/article/49/8/719/127584 by guest on 24 September 2021 BENGT-GUNNAR LILJA3 and PETER NAYSTRO¨ M4
1Department of Occupational and Environmental Medicine, O¨ rebro University Hospital, SE-701 85 O¨ rebro, Sweden; 2Department of Public Health Sciences, Division of Occupational Medicine, Karolinska Institutet, SE-171 77 Stockholm, Sweden; 3TMV-Environmental Consultant, Box 506 SE-541 28 Sko¨vde, Sweden; 4Swedish Foundry Association, Box 2033 SE-550 02 Jo¨nko¨ping, Sweden
Received 7 April 2005; in final form 5 July 2005; published online 26 August 2005
Emissions from a chemical core binder system (Hot Box) based on a formaldehyde–carbamide resin have been investigated. The binder is used in some Swedish die-casting foundries. During core-making and casting, low molecular monoisocyanates, in particular methyl isocyanate (MIC) and isocyanic acid (ICA), were identified. Exposure to air concentrations of MIC, ICA and formaldehyde were subsequently determined in all Swedish foundries using the Hot Box binder, and involved three brass and one grey iron foundry. The survey was carried out in the winter period of 2001, and involved core-makers, casters and fettlers in the brass foundries, whereas only core-makers were included in the grey iron foundry. For each worker, four to five short-term samples of isocyanates (n = 298) and one 8 h sample of formaldehyde (n = 64) were collected during one shift for 15 die-casters, 39 core-makers and 10 other workers in the foundry. The air concentrations of the MIC short-term samples varied between <4 and 68 mgm 3, with corresponding ICA levels between <4 and 280 mgm 3. Calculated 8 h time weighted average air concentrations of MIC, based on short-term samples for each indi- vidual, varied between <4 and 31 mgm 3; for ICA the corresponding levels varied from <4to 190 mgm 3. The formaldehyde time weighted average concentration levels ranged from 14 to 1600 mgm 3, and the Swedish occupational exposure limit (600 mgm 3) was exceeded only in 3% of the samples. In general, the core-makers were exposed to higher average formaldehyde levels compared to the casters, the latter being more exposed to monoisocyanates. During core- making and die-casting, low molecular monoisocyanates, in particular MIC and ICA, were identified. Compared to the American Conference of Governmental Industrial Hygienists (ACGIH) threshold limit value-time weighted average (TLV-TWA) for MIC, the exposures were low. The lack of toxicological and human data for ICA and the relatively high air concentrations call for medical examination and preventive measures in production, ventilation and the use of personal safety equipment in the investigated foundries.
Keywords: casting; core-making; exposure; foundries; isocyanic acid; methyl isocyanate
INTRODUCTION bisphenyl diisocyanate (MDI; Archibald and Smith, 1988). During the thermal degradation of this binder, Exposure to isocyanates in iron, aluminium and metal monoisocyanates and diisocyanates could be formed, foundries has historically been associated with the use such as phenyl isocyanate and MDI (Renman et al., of isocyanate based chemical binders for core pro- 1986). Lately this has become true also for carbamide duction, the most common known as the Cold Box containing binders like the Hot Box system, and here process, using core binders containing methylene the monoisocyanates methyl isocyanate (MIC; CAS no 624-83-9) and isocyanic acid (ICA; CAS no *Author to whom correspondence should be addressed. Tel: +46-19-6022493; fax: +46-19-120404; 75-13-8) are emitted at high levels. In addition, e-mail: [email protected] monoisocyanates have been determined in a large
719 720 H. Westberg et al. number of new exposure situations in foundries, 40 workers) and in the grey iron foundry 24 core- including the use of Hot Box core-binders in static makers. Exposure measurements covering the whole die-casting foundries (Lilja et al., 1999). shifts were performed for all included workers, To specifically determine low air concentrations mostly (>90%) as 4–5 short-term samples of MIC of monoisocyanates, new analytical methods for and ICA. Formaldehyde was determined with diffu- the determination of low molecular isocyanates sive samplers (GMD) as full-shift samples. like ICA and MIC have been developed (Karlsson, 1998b; Spanne, 1999). Sampling and analysis Isocyanates are a group of chemically very reactive Sampling of isocyanates was performed by liquid agents; di-, poly- and prepolymerized isocyanates are chemosorption using impinger bottles, containing used to form polyurethanes. The industrial use of 0.01 M dibutylamine (DBA) dissolved in toluene Downloaded from https://academic.oup.com/annweh/article/49/8/719/127584 by guest on 24 September 2021 polyurethanes has increased drastically during the (Tinnerberg et al., 1997; Karlsson et al., 1998b). last decades (IARC, 1999a,b) and the products To adsorb fine particulate aerosol normally passing cover a wide range of industrial applications. Adverse through the adsorbing liquid, the impinger bottles health effects have been studied almost exclusively were connected to a cellulose ester filter with a for diisocyanates, the main effects are respiratory pore size of 0.3 mm. After the sampling, the filter disorders and irritative effects on the mucous mem- was introduced into the impinger bottle. The sam- branes (Baur et al., 1994). Little, if any knowledge pling flow was 1 l min 1 and the sampling was per- exists on exposure to monoisocyanates (MIC and formed with personal high-flow sampling pumps with ICA) and adverse health effects in industrial settings flow-rates between 1 and 5 l min 1. We used SKC (NIWL, 2002). 224-PCXR-9, SKC 224-PCXR-3, GILIAN HFS-513 In this study the exposures to MIC and ICA in and -113, MSA FLOW-LITE 34RI sampling pumps, Swedish foundries using Hot Box binders are and the air flow was controlled with calibrated described. The study includes core-makers, die- rotameters. Sampling of formaldehyde was carried casters and other workers in the foundry area. In out with diffusive samplers GMD, based on a reaction parallel to the exposure investigation, a study of res- between aldehydes and dinitrophenylhydrazine piratory symptoms and lung function was conducted (Levin et al., 1988). (reported elsewhere), and formaldehyde was, there- The analysis of formaldehyde was performed with fore, included in the monitoring programme of poten- high performance liquid chromatography techniques, tially harmful agents. the corresponding analysis of monoisocyanates and diisocyanates with liquid chromatography mass MATERIAL AND METHODS spectrometry techniques (Karlsson et al., 1998a). The detection concentration level for formaldehyde dur- 3 Study design ing an 8 h sampling is 20 mgm , and for methyl isocyanate and ICA 4 mgm 3 for 15 min short- The exposures to MIC, ICA and formaldehyde term sampling. All analyses were performed by were investigated in all four Swedish foundries the laboratory at the Department of Occupational using Hot Box binders, three small brass and one and Environmental Medicine, O¨ rebro University large grey iron foundry. The binder in use at all found- Hospital. The laboratory is accredited by the Swedish ries was based on a carbamide–formaldehyde resin Board for Accreditation and Conformity Assessment (<1% formaldehyde) and a curing agent containing (SBACA). ammonium nitrate (10–15%), urea and sodium hydroxide or water. The brass foundries were produc- ing armatures for households, and the grey iron foun- Statistical methods dry spare parts for the automobile industry. During The air concentrations for the individual ICA, MIC core production, both manual and enclosed automatic and formaldehyde samples were determined and 8 h core machines were used, and during die-casting time-weighted averages calculated for the total num- manually operated as well as enclosed robots were ber of measurements, different jobs and foundries. used. In the brass foundries, the alloy in use contained The data included a number of measurements (n), 62% copper, 32% zinc and 1–2% lead, in addition to and due to approximate log normal distributions of smaller amounts of aluminium and tin. The corres- air concentrations, the parameters of the air concen- ponding grey iron alloy contained 94% iron, 3% car- tration distributions were presented as the geometric bon, 2% silicone, 0.7% manganese, 0.3% chromium, mean (GM), and the corresponding geometric and 0.1% copper. The grey iron foundry was large, standard deviation (GSD). However, depending on producing some 75 000 ton per year, in contrast to the the further use of air concentration data, the arith- smaller brass foundries producing between 500 and metic mean and standard deviation were also 1000 ton per year. In the brass foundries, all core- presented (Seixas et al., 1988). Concentration values makers, die-casters and fettlers were included (in total less than the detection limits were estimated by Chemical pollutants in foundries 721
100 MIC 90
80
70
60 ICA
% 50
40
30 Downloaded from https://academic.oup.com/annweh/article/49/8/719/127584 by guest on 24 September 2021
20
10
0 –24 25–48 49– –24 25–48 49– µg/m3
Fig. 1. Distribution of total short-term samples of methyl isocyanate (MIC) and isocyanic acid (ICA) air concentrations (mgm 3). pffiffiffi multiplying the detection limit by 1= 2 (Hornung Table 1. Exposure measurements, short-term samples of and Reed, 1990). Univariate ANOVA with a post hoc methyl isocyanate (MIC) and isocyanic acid (ICA), different test by Tukey (SPSS 12.0) was applied for ICA and foundries and job titles MIC with the different foundries and job titles as the Chemical agent Air concentration (mgm 3) explaining variables. The determination of within- n GM GSD AM SD Range and between worker variability was carried out Total using variance component estimation (SPSS 12.0). MIC 298 4.9 2.2 7.4 9.2 <4–68 ICA 297 24 2.7 38 40 <4–280 Core-maker RESULTS MIC 179 4 1.8 5.1 6.2 <4–60 For a majority of the workers (>90%), 4–5 short- ICA 178 22 2.3 30 23 <4–130 term samples of isocyanates, in all 298 samples, were Die-caster taken, representing the exposures of 15 casters, 39 core-makers and 10 others. In addition, one full shift MIC 71 10 2.6 16 13 <4–68 sample of formaldehyde was carried out. ICA 71 48 2.8 73 61 4–280 Time-weighted averages (8 h TWA) of the air Others concentrations of MIC were calculated for each MIC 48 3.4 1.5 3.8 2.2 <4–12 individual, based on the short-term samples, and ICA 48 12 2.3 16 13 <4–66 varied between <4 and 31 mgm 3 (Table 2). No TWA exceeded the ACGIH TLV-TWA (ACGIH, n, number of samples; AM, arithmetic mean; SD, standard 2004) for MIC (48 mgm 3). For ICA the correspond- deviation; GM, geometric mean; GSD, geometric standard deviation. ing levels varied from <4 to 190 mgm 3, and 27% exceeded the TLV for MIC. For the short-term samples of MIC, the air concen- between foundries revealed air concentrations of trations varied between <4 and 68 mgm 3, and ICA and MIC for both core-makers and die-casters the corresponding ICA levels ranged from <4to at 2–3 times higher in one of the foundries. Notably, 280 mgm 3. For MIC, only 1% of the short-term high exposures to ICA, ranging up to 66 mgm 3, were samples exceeded 48 mgm 3, and 8% exceeded determined for workers with secondary exposure to 24 mgm 3 (Table 1; Fig. 1). The corresponding the Hot Box process emissions working in the figures for ICA were 26 and 47%, respectively. foundry premises (Table 1). The GM of the total ICA concentrations was higher The formaldehyde levels ranged from 14 to than that of the corresponding MIC concentrations, 1600 mgm 3, but the ACGIH-TLV (600 mgm 3) 24 versus 4.9 mgm 3, and this was also true for was exceeded in only 3% of the samples (Table 2). the different jobs such as core-makers (22 versus These high levels were attributed to one particular 4 mgm 3) and die-casters (48 versus 10 mgm 3; foundry. The air concentrations of formaldehyde Table 1) and for other exposed workers present in were higher for core-makers (GM = 200 mgm 3) the foundry area (Figs 2 and 3). A comparison than for die-casters (GM = 63 mgm 3). 722 H. Westberg et al.
100
90 MIC - die caster MIC - core maker MIC - others 80
70
60
% 50
40
30 Downloaded from https://academic.oup.com/annweh/article/49/8/719/127584 by guest on 24 September 2021
20
10
0 –24 25–4849– –24 25–48 49– –24 µg/m3
Fig. 2. Distribution of total short-term samples of MIC air concentrations (mgm 3) for different job titles (core-makers, die-casters and others).
100
90 ICA - die caster ICA - core maker ICA - others
80
70
60
% 50
40
30
20
10
0 −24 25−48 49−−24 25−48 49−−24 25−48 49−µg/m3
Fig. 3. Distribution of total short-term samples of ICA air concentrations (mgm 3) for different job titles (core-makers, die-casters and others).
An ANOVA regarding determinant factors influ- ratio (l) for the different variability measures encing variability of the air concentration levels range from 0.77 to 1.1. The ratio between 5 and showed for both ICA and MIC that different foundries 95% confidence limits for the GSDB variability *** *** *** (F = 7.50 and 17.11 ) and job titles (F = 41.40 (R0.95B) for all workers was higher or equal to and 54.14***) had a significant effect (P < 0.001) on those of the job titles (Table 4). the variability of the air concentration, implying that When comparing core making for the grey iron and no sole foundry or job title alone could represent the three brass foundries, the average air concentra- the distribution of air concentrations for the whole tions of MIC (GM = 3.4 and 5.9 mgm 3, respectively) group when used for compliance or epidemiology and ICA (GM = 24 mgm 3 for both types) were purposes. However, job titles explained most of the similar (Table 5). variability (Table 3). The overall and job title between-worker (GSDB) DISCUSSION variability and the within-worker (GSDW) variability varied from 1.5 to 2.3, and the analysis also revealed Comprehensive exposure measurements in slightly higher GSDB variability than variability Swedish brass and grey iron foundries using the GSDW for ICA, but not for MIC. The variance Hot Box core binder system were performed. MIC, Chemical pollutants in foundries 723
Table 2. Exposure measurements, means, SDs and ranges Table 4. Within- and between worker variation expressed as of individual 8 h TWA for isocyanates and formaldehyde, GSD l and R0.95 for log-transformed MIC, ICA and different foundries and job titles formaldehyde air concentrations by agent and job title
3 Chemical agent Air concentration (mgm ) n GSDW GSDB l R0.95B n GM GSD AM SD Range MIC Total All workers 298 1.7 1.9 0.89 12 MIC 64 5.3 2.1 7.3 7.1 <4–31 Core-makers 179 1.6 1.5 1.1 5 ICA 64 27 2.3 38 34 <4–190 Die-casters 71 2.0 1.9 1.0 12 Formaldehyde 64 120 2.8 190 250 14–1600 ICA Core-maker All workers 297 1.8 2.3 0.77 26 MIC 39 4.2 1.8 5.1 3.9 <4–20 Core-makers 178 1.6 2.0 0.83 15 Downloaded from https://academic.oup.com/annweh/article/49/8/719/127584 by guest on 24 September 2021 ICA 39 24 2 30 19 5.3–84 Die-casters 71 2.0 2.2 0.89 22 Formaldehyde 39 200 2.1 280 290 50–1600 n, number of samples; GSDW, within worker standard deviation; GSDB, between worker standard deviation; Die-caster l, variance ratio of within- and between worker variance; MIC 15 12 2.1 15 9.6 <4–31 R0.95B, ratio of 5 and 95% confidence limits of ICA 15 55 2.3 72 49 6–190 between worker variability. Formaldehyde 15 63 2.1 80 58 21–220 Table 5. Exposure measurements, means, standard deviations Others and ranges of individual short-term ICA, MIC samples and 8 h TWA of formaldehyde in the grey iron and brass foundries MIC 10 3.5 1.5 3.8 1.9 <4–8.2 3 ICA 10 14 2 16 9.4 <4–32 Chemical agent Air concentration (mgm ) Formaldehyde 10 32 2.2 46 50 14–170 n GM GSD AM SD Range Grey iron foundry n, number of samples; GM, geometric mean; GSD, geometric standard deviation; AM, arithmetic mean, SD, standard Total deviation. Core-maker MIC 108 3.4 1.6 4.2 6.3 <4–60 Table 3. ANOVA of log-transformed air concentrations ICA 107 24.0 2.0 30.5 21.4 4–96 for MIC ICA (n = 298) and formaldehyde (n = 64) Formaldehyde 24 260 2.1 343 339 100–1600 Factor df MS FP-value MIC Brass foundries Foundry 3 440.3 7.50 <0.001 Total Job title 2 2429.2 41.40 <0.001 MIC 190 5.9 2.4 9.1 10 <4–68 Foundry * Job title 4 115.3 1.96 0.100 ICA 190 24 2.0 31 21 <4–280 Formaldehyde 40 71 2.5 105 9.6 14–440 ICA Core-maker Foundry 3 17 147.3 17.11 <0.001 MIC 71 4.8 2.0 6.4 5.8 <4–28 Job title 2 54 274.1 54.14 <0.001 ICA 71 19 2.7 2.8 2.5 <4–126 Foundry * Job title 4 4373.6 4.36 0.002 Formaldehyde 15 136 2.0 169 110 50–440 Formaldehyde Die-caster Foundry 3 71 208.6 1.38 0.260 MIC 71 10 2.6 16 13 <4–68 Job title 2 86 007.5 1.66 0.199 ICA 71 47 2.8 73 61 <4–280 Foundry * Job title 4 15 695.7 0.30 0.874 Formaldehyde 15 63 2.1 80 58 21–220 df, degrees of freedom; MS, mean squares; F, F-value; P = level of significance. Others MIC 48 3.4 1.5 3.8 2.2 <4–12 ICA 48 12 2.3 16 13 <4–66 ICA and formaldehyde were analysed. The exposures to MIC were all below the TLV-TWA, however the Formaldehyde 10 32 2.2 46 50 14–170 ICA concentration levels were substantially higher. n, number of samples; GM, geometric mean; GSD, geometric In general, the die-casters were more exposed to standard deviation; AM, arithmetic mean; SD, standard isocyanates than the core-makers. The formaldehyde deviation. concentrations were higher for the core-makers, although still below the ACGIH-TLV for most of included in the study for the brass foundries consisted the exposed workers. of die-casters, core-makers and other workers, as for All Swedish foundries using the Hot Box core the grey iron foundry only the core-makers were binder system were included. The total work force included, whereas the casting area was left out. 724 H. Westberg et al.
The grey iron foundry was actually using a blend of Our measurements performed during the winter core binders like the Cold Box cores, Epoxy-SO2 season, were likely to provide higher air concentra- cores and Hot Box cores as well as green sand mould- tions than average thus representing a worst-case sea- ing. Measurements of the casting process would most son sampling scenario. All Swedish foundries using certainly include the emissions of thermal degrada- the Hot Box core method were taking part in our tion products like isocyanates emanating from other survey, and the problem of external validity therefore binders than the binder under study, and therefore did not exist. However, all potential exposures were casters were excluded. The core-making and casting not determined, but historical data exist. Dust and processes cover old and new techniques regarding the phenol measurements were performed by the com- core machines, as well as different sizes and shapes of pany health services and their safety engineers at the cores, both manual and automatic die-casting the three brass foundries. The total dust concentra- techniques were used in each of the brass foundries. tions varied between 0.1 and 3 mg m 3 for the Downloaded from https://academic.oup.com/annweh/article/49/8/719/127584 by guest on 24 September 2021 Notably, much lower ICA and MIC air concentrations die-casters and between 0.1 and 1.5 mg m 3 for were determined when the old and new techniques the core-makers. In the grey iron foundry, the cor- were compared, for ICA 49 versus 21 mgm 3 and for responding respirable dust concentrations varied MIC 8.2 versus 5.2 mgm 3. The same pattern was between 0.2 and 0.6 mg m 3. seen for die-casting, the average ICA concentrations The air concentrations of phenol were historically were reduced from 93 to 39 mgm 3, the correspond- equally low, ranging from 0.09 to 0.17 mg m 3. The ing MIC levels were reduced from 19 to 9.4 mgm 3. earlier measurements of formaldehyde in the grey Between 80 and 90% of the core- and the die- iron foundry showed air concentrations ranging casting machines were used during the time of our from 0.2 to 8 mg m 3; the same exposure pattern investigations. was seen in this study. Our survey was preceded Our sampling programme was initially based on by measurements of emissions during the thermal measurements of different monoisocyanates and degradation of different nitrogen containing binders diisocyanates, in particular MIC and ICA. However, for cores and moulds carried out by the Swedish for practical and economical reasons, sampling of Foundry Association (Lilja et al., 1999). In this total dust (containing potential respirable quartz), survey, the MIC levels for core-makers ranged phenol and mineral oil mist was left out. The meas- from 2 to 8 mgm 3, and the ICA levels varied urement programme, originally designed to deter- between 5 and 72 mgm 3. The corresponding MIC mine compliance with occupational exposure limits, air concentrations for die-casters ranged from 2 to was also intended to provide information in a parallel 29 mgm 3, the ICA levels from 13 to 190 mgm 3 medical study on respiratory symptoms and lung (Lilja et al., 2000). These figures were in the function impairments, and phenol could then be same order of magnitude as in our study and the ruled out. Due to its potential effect on the respiratory same patterns of exposures were seen. The overall tract, formaldehyde was included. The sampling strat- within- and between-worker variabilities expressed egies for the short- and long-term samples as well as as GSD were 2.2 for MIC and 2.7 for ICA, well the number of samples for each job title followed reflecting normal variability in industrial settings developed theory and practice (Leidel et al., 1977). (Rappaport, 1991). For the sampling of isocyanates (in particular MIC The ACGIH TLV-TWA for MIC is 48 mgm 3 and ICA) we used the latest developed analytical (ACGIH, 2004), the corresponding German occupa- methods, an impinger sampling technique based on tional exposure limit (OEL) is 24 mgm 3 (DFG, chemosorption with dibutylamine and liquid 2004). In Sweden, the OELs for isocyanates are chomatography, and mass spectrometry analysis derived from the TLV for MDI (0.005 p.p.m.) and enabling the determination of a blend of isocyanates recalculated by molecular weight, giving TWA-TLVs in each sample. No other isocyanates than the MIC for MIC and ICA of 12 and 9 mgm 3, respectively and ICA were determined (Levin et al., 1988; (NBOSH, 2000). Comparison with the Swedish OELs Karlsson et al., 1998b) and during the course of implies a large number of ICA air concentrations this project, several interlaboratory controls regard- exceeding the threshold value-ceiling and ing the DBA-method were performed. The sampling TLV-TWA. and analysis of formaldehyde was performed with an accredited method (Levin et al., 1988) at our laboratory, accredited by the SBACA. CONCLUSIONS When comparing core-making for different types of foundries using basically the same core-making Emissions from a chemical core binder system (Hot techniques, the average air concentrations of MIC Box) based on a formaldehyde–carbamide resin have (GM = 3.4 and 5.9 mgm 3, respectively) and ICA been investigated. During core making and casting, (GM = 24 mgm 3 for both types) were similar, low molecular monoisocyanates, in particular ICA implying good internal validity. and MIC were identified. Compared to ACGIH Chemical pollutants in foundries 725
TLV-TWA for MIC and formaldehyde, the exposures Part 4. Determination of aliphatic isocyanates as dibuty- were low. The lack of toxicological and human data lamine derivatives using liquid chromatography and mass for ICA and the relatively high air concentrations spectrometry. Analyst; 123: 117–23. Karlsson D, Dalene M, Skarping G. (1998b) Determination calls for further medical examination and preventive of complex mixtures of airborne isocyanates and amines. measures in production, ventilation and personal Part 5. Determination of low molecular weight aliphatic safety equipment. isocyanates as dibutylamine derivatives. Analyst; 123: 1507–12. Leidel N, Busch K, Lynch J. (1977) Occupational sampling Acknowledgements—The authors thank the participating com- strategy manual, NIOSH. Cincinnatti, OH: US Department panies, employers and employees as well as company health of Health, Education and Welfare. services who helped us in our practical field work. We also Levin JO, Lindahl R, Andersson K. (1988) High-performance thank Krister Berg and Lisbeth Viklund for the field measure- liquid chromatographic determination of formaldehyde in ments and Ing-Liss Bryngelsson and Cecilia Fedeli for the air in the p.p.b. and p.p.m. range using diffusive sampling Downloaded from https://academic.oup.com/annweh/article/49/8/719/127584 by guest on 24 September 2021 statistical analyses and data management. The study was and hydrazone formation. Environ Technol Lett; 9: financed by VINNOVA research fund, grant nos. 2001- 1423–30. 03954 and 2001-03393. Lilja BG, Westberg H, Naystro¨m P. (1999) A survey of isocyanate exposure in Swedish foundries. Part I. Emission measurements. Jo¨nko¨ping, Sweden: Swedish Foundry Association. p. 31. REFERENCES Lilja BG, Westberg H, Naystro¨m P. (2000) A survey of isocyanate exposure in Swedish foundries. Part II Exposure ACGIH. (2004) Threshold limit values for chemical and physi- measurements. Jo¨nko¨ping, Sweden: Swedish Foundry cal agents and biological exposure indices. Cincinnati, Association. p. 83. OH: American Conference of Governmental Industrial NBOSH. (2000) Occupational exposure limits 2000:3. Hygienists. Stockholm, Sweden: National Board of Occupational Safety Archibald JJ, Smith RL. (1988) Resin binder processes. In: and Health. Stefanescu DM, Davis JR, editors. Metals handbook. NIWL. (2002) Scientific basis for Swedish Occupational Vol. 15, Casting, 9th edn. Metals Park, OH: ASM Standards XXIII. Criteria group for occupational standards. International. pp. 214–21. Arbete och Ha¨lsa, 2002:19. Stockholm, Sweden: National Baur X, Marek W, Ammon J et al. (1994) Respiratory and other Institute of Working Life. pp. 1–63. hazards of isocyanates. Int Arch Occup Environ Health; 66: Rappaport SM. (1991) Assessment of long-term exposures to 141–52. toxic substances in air. Ann Occup Hyg; 35: 61–121. DFG. (2004) List of MAK and BAT values. Deutsche Renman L, Sango¨ C, Skarping G. (1986) Determination of Forschungsgemeinschaft, Bonn Germany. isocyanate and aromatic amine emissions from thermally Hornung RW, Reed LD. (1990) Estimation of an average degraded polyurethanes in foundries. Am Ind Hyg Assoc concentration in the presence of nondetectable values. Appl J; 47: 621–8. Occup Environ Hyg; 6: 458–64. Seixas NS, Robins TG, Moulton LH. (1988) The use of IARC. (1999a) Toluene diisocyanates. IARC monographs on geometric and arithmetic mean exposures in occupational the evaluation of carcinogenic risks to humans. Vol. 71, part epidemiology. Am Ind J Med; 14: 465–77. 2. Lyon: International Agency for Research on Cancer. Spanne M, Grzybowski G, Boghard M. (1999) Collection pp. 865–79. efficiency for submicron particles of a commonly used IARC. (1999b) 4,40-methylenediphenyl diisocyanate and impinger. Am Ind Hyg Assoc J; 60: 540–4. polymeric 4,4–methylene diphenyl diisocyanate. IARC Tinnerberg H, Spanne M, Dalene M et al. (1997) Determination monographs on the evaluation of carcinogenic risks to of complex mixtures of airborne isocyanates and amines. humans. Vol. 71, part 3. Lyon: International Agency for Part 3. Methylenediphenyl diisocyanate, methylendiphenyla- Research on Cancer. pp. 1049–58. mino isocyanate and methylendiphenyldiamine and struc- Karlsson D, Spanne M, Dalene M et al. (1998a) Determination tural analogues after thermal degradation of polyurethane. of complex mixtures of airborne isocyanates and amines. Analyst; 122: 275–8.