Annals of Occupational Hygiene Advance Access originally published online on May 5, 2008
Annals of Occupational Hygiene 2008 52(5):361-374; doi:10.1093/annhyg/men017
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Size-Separated Sampling and Analysis of Isocyanates in Workplace Aerosols. Part I. Denuder—Cascade Impactor Sampler
Work Environment Chemistry, Stockholm University, PO Box 460, S-281 24 Hässleholm, Sweden
* Author to whom correspondence should be addressed. Tel: +46-451-385251; fax: +46-451-385260; e-mail: gunnar.skarping{at}anchem.su.se
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Isocyanates in the workplace atmosphere are typically present both in gas and particle phase. The health effects of exposure to isocyanates in gas phase and different particle size fractions are likely to be different due to their ability to reach different parts in the respiratory system. To reveal more details regarding the exposure to isocyanate aerosols, a denuder–impactor (DI) sampler for airborne isocyanates was designed. The sampler consists of a channel-plate denuder for collection of gaseous isocyanates, in series with three-cascade impactor stages with cut-off diameters (d50) of 2.5, 1.0 and 0.5 µm. An end filter was connected in series after the impactor for collection of particles smaller than 0.5 µm. The denuder, impactor plates and the end filter were impregnated with a mixture of di-n-butylamine (DBA) and acetic acid for derivatization of the isocyanates. During sampling, the reagent on the impactor plates and the end filter is continuously refreshed, due to the DBA release from the impregnated denuder plates. This secures efficient derivatization of all isocyanate particles. The airflow through the sampler was 5 l min–1. After sampling, the samples containing the different size fractions were analyzed using liquid chromatography-mass spectrometry (LC-MS)/MS. The DBA impregnation was stable in the sampler for at least 1 week. After sampling, the DBA derivatives were stable for at least 3 weeks. Air sampling was performed in a test chamber (300 l). Isocyanate aerosols studied were thermal degradation products of different polyurethane polymers, spraying of isocyanate coating compounds and pure gas-phase isocyanates. Sampling with impinger flasks, containing DBA in toluene, with a glass fiber filter in series was used as a reference method. The DI sampler showed good compliance with the reference method, regarding total air levels. For the different aerosols studied, vast differences were revealed in the distribution of isocyanate in gas and different particle size fractions. The opportunity to obtain detailed information regarding the distribution of isocyanates in aerosols in addition to the total air levels make the DI sampler a valuable tool for studies of possible health effects in the different parts of the airways.
Keywords: aerosol • air sampling • DBA • denuder • impactor • isocyanate • LC-MS • polyurethane • thermal degradation
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Isocyanates are harmful to the human airways, and may elicit acute irritation, sensitization, bronchial hyperactivity and hypersensitivity pneumonitis (Karol, 1986; Baur et al., 1994; Ott et al., 2003). Isocyanate exposure is one of the most common causes of occupational asthma (Bernstein, 1996; Redlich and Karol, 2002). Isocyanates in the workplace atmosphere may be present both in the gas phase and in association with particles (Streicher et al., 2000). Gas-phase isocyanates are evidently deposited in the upper human airways by diffusion. Large particles (>10 µm) are also deposited here due to inertial impaction. Particles in the range 0.01–2.5 µm are able to reach the lower airways and may deposit in the lungs (Vincent, 1995; Heyder and Svartengren, 2001).
Depending on the work task performed, the distribution of isocyanates between gas phase and particle phase varies. Heating of isocyanate mixtures produces isocyanate aerosols containing monomers (Marand et al., 2004) and welding in polyurethane (PUR)-coated metal generates small particles able to carry isocyanates (Karlsson et al., 2000). In addition to isocyanates, aminoisocyanates and amines are formed (Karlsson et al., 2002). Spraying of PUR paint produces relatively large particles (>2 µm) (LaPuma et al., 2002).
The high reactivity of isocyanates demands instantaneous derivatization to minimize side reactions (Streicher et al., 1994). A number of different reagents for isocyanates have been presented. Most of them contain an amine group that reacts with the isocyanate group (Hardy and Walker, 1979; Sangö and Zimerson, 1980; Warwick et al., 1981; Spanne et al., 1996; Streicher et al., 1996). During thermal degradation of PUR, there are many other compounds present which may react with isocyanates. An efficient derivatization reagent is crucial.
Sampling of isocyanates has been described for wet sampling using impinger flasks in series with a filter and dry sampling using unimpregnated filters, impregnated filters or denuders. The impinger–filter (IF) sampler has an advantage in that particles will be dissolved in the sampling solution, making the derivatization reaction more efficient, with less sample losses. The impinger flask has low collection efficiency for particles in the range of 0.01–1.5 µm (Spanne et al., 1999), but this fraction can be collected if a filter is placed in series after the impinger. IF has been used to give a rough estimation of the distribution of isocyanates in an aerosol by analyzing the impinger and filter part separately (Karlsson et al., 2000). For most of the other suggested reagents, dry sampling techniques have been described. The drawback with dry samplers is that the samples must be field desorbed in order to dissolve isocyanates particles to make the derivatization efficient and to reduce side reactions with other compounds present.
A few other sampling systems capable of separation of gas and particles have been presented. Rando and Poovey have designed a denuder-filter system impregnated with either 9-(N-methylaminomethyl)anthracene (MAMA) or N-4-nitrobenzyl-N-n-propylamine (nitro-reagent) and an impactor using MAMA (Rando and Poovey, 1994, 1999). Lesage et al. (1992) have presented a dual-filter system with a Teflon filter for collection of particles in series with an impregnated glass fiber filter for collection of gas-phase isocyanate. A triple-filter system, with two Teflon filters in series with an 1-(2-pyridyl)piperazine (1-(2PP))-impregnated glass–fiber filter has also been presented (Tsai et al., 2003).
To characterize the exposure to airborne isocyanates, the sampler must be able to collect both gas and particles in a representative way (Streicher et al., 2000). Ideally, the sampler should also be capable to separate gas and particles in separate fractions. Moreover, derivatization reactions need to be efficient and robust against interfering reactions.
When a laminar airflow passes through a denuder tube, gas molecules in the airstream reach the walls of the denuder by diffusion. The denuder walls are covered with absorbing material or a reagent to retain the gas molecules at the walls. To increase the efficiency of denuders, they are often built as annular denuders or with narrow square channels, called parallel plate denuders. A review of different denuder applications has been presented by Namie
nik and co-workers (Kloskowski et al., 2002). An expression for the efficiency of a parallel plate denuder has been given by De Santis (1994). To determine isocyanates in gas phase, denuders have been used (Nordqvist et al., 2001). Di-n-butylamine (DBA)-impregnated denuder in series with an impregnated filter has been used for sampling of isocyanate aerosols. DBA was mixed with an equimolar amount of acetic acid to reduce the volatility of the DBA (Marand et al., 2005).
Fractionation of particles in an aerosol can be achieved by air sampling using a cascade impactor. In an impactor, airstream is passed through a small nozzle towards a substrate plate. Large particles are unable to follow the airstream as it deviates close to the plate, and are deposited on the substrate plate by inertial impaction. Small particles follow the airstream around the plate. By using several impaction stages with decreasing nozzle diameters in series, the velocity of the air is increased, and smaller particles are deposited on the successive stages. An impactor is usually defined by the cut-off diameter where 50% collection efficiency for a certain particle size is obtained. This value is denoted d50. The cut-off diameter (d50) for an impactor stage can be calculated using Stokes number and the formula:
 |
is the viscosity of the medium surrounding the particle, W is the nozzle width, 
p is the particle density, Cc is the Cunningham slip correction factor, U is the velocity through the nozzle and Stk50 is the Stokes number for 50% collection of the particles with the specified diameter d50. Other important parameters for impactor design are the Reynolds' number, which should be between 500 and 3000 to obtain a sharp cut-off, and the distance between the jet orifice and the impaction plate (jet to plate distance) (Marple et al., 2001). The purpose of this study was to investigate the possibility to use a denuder in combination with an impactor connected to an end filter for the separation of isocyanate aerosols in gas and different particle size fractions.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Apparatus
Liquid chromatography-mass spectrometry (LC-MS)/MS.
A triple quadrupole mass spectrometer, a Quattro Micro (Waters, Altrincham, Cheshire, UK) was used for analysis of air samples. Isocyanate–DBA derivatives were monitored using positive electrospray ionization (ES+) and multiple reaction monitoring (MRM). Quantitative measurements of isocyanates were made by monitoring the reaction [M+H]+
[DBA+H]+ for the isocyanates. Isocyanate–DBA derivatives with nona-deuterium-labeled DBA (d9-DBA) were used as internal standards. The reaction [M+H]+ [D9-DBA+H]+ was monitored for the internal standards. For the MS instrument, the capillary voltage was 4 kV. The temperature of the ion source was 130°C, and the desolvation gas temperature was 200°C. The desolvation gas flow was set to 500 l h–1. Argon was used as collision gas, and the pressure in the collision cell was 3 x 10–3 mbar. Entrance cone voltage and collision energies were optimized individually for the different derivatives. A more detailed description of the analysis of isocyanate–DBA derivatives using LC-MS/MS is available elsewhere (Karlsson et al., 2005). For separation of the DBA derivatives, a micro-LC system was used. Gradient elution was performed, using Shimadzu LC10ADVP micro-LC pumps (Shimadzu Corporation, Kyoto, Japan). The gradient was linear from 40/60 (v/v) acetonitrile/H2O to 95/5 (v/v) acetonitrile/H2O in 10 min. After analysis, the column was equilibrated for 4 min with 40/60 (v/v) acetonitrile/H2O before the next injection. Formic acid was also added to the mobile phase, with a total concentration of 0.05%. The total flow rate was 70 µl min–1.
Sample injection was made with an LC-Pal autosampler (CTC Analytics AG, Zwingen, Switzerland). Loop injection was performed; 2.5-µl sample surrounded by 17.5 µl of 5/95 acetonitrile/H2O was injected to achieve on-column concentration of the analytes. The analytical column used was a Waters Xterra, 50 x 1 mm with 2.5 µm C18 particles (Milford, Waters, MA, USA).
Particle measurements
An aerodynamic particle sizer (APS), TSI APS 3321 (TSI Incorporated, Shoreview, MN, USA), was used for particle measurements when the collection efficiency of the impactor stages was evaluated. The APS measures particles in the size range of 0.5–20 µm. For generation of a non-reactive test aerosol, an aerosol generator Palas BEG-1000 (Palas GmbH, Karlsruhe, Germany) was used.
Chemicals
DBA, ethyl isocyanate (EIC), 2,6-toluenediisocyanate (2,6-TDI) and isophorone diisocyanate (IPDI) were obtained from Merck-Schuchardt (Hohenbrunn, Germany). Urea for synthesis of isocyanic acid (ICA) was obtained from ICN Biomedicals Inc. (Aurora, OH, USA). 1,3-Dimethylurea for synthesis of methyl isocyanate (MIC), propyl isocyanate (PIC), phenyl isocyanate (PhI), 1,6-hexamethylene diisocyanate (HDI), 4,4'-methylene diphenyl diisocyanate (MDI) and technical grade TDI (20% 2,6-TDI and 80% 2,4-TDI) were obtained from Acros Organics (Morris Plains, NJ, USA). An HDI–isocyanurate-based paint system was obtained from Standox (Wuppertal, Germany). Reference solutions of ICA–DBA and MIC–DBA were prepared as described by Karlsson et al. (2001). 2,4-TDI was obtained from TCI (Tokyo Kasei Kogyo Co. Ltd., Tokyo, Japan). Toluene, acetonitrile, methanol and formic acid and acetic acid were obtained from Merck (Darmstadt, Germany). All solvents used were high pressure liquid chromatography (HPLC) grade or better. Deuterium-labeled DBA [NH(C4H9)(C4D9)] from Synthelec (Lund, Sweden) was used for synthesis of internal standards (Marand et al., 2005).
Standard solution for air samples.
A toluene solution containing isocyanate–DBA derivatives of ICA, MIC, EIC, PIC, PhI, 1,6-HDI, 2,6- and 2,4-TDI, IPDI and 4,4'-MDI with a concentration corresponding to 1 µg ml–1 of each isocyanate was used as standard solution for air samples.
Internal standards.
A toluene solution containing deuterium-labeled isocyanate–DBA derivatives of ICA, MIC, EIC, PIC, PhI, 1,6-HDI, 2,6- and 2,4-TDI, IPDI and 4,4'-MDI with a concentration corresponding to 4 µg ml–1 of each isocyanate was used as internal standard (IS).
Denuder–impactor sampling system
The denuder–impactor (DI) sampler consisted of three main parts, a denuder, a three-stage cascade impactor and an end filter. In the denuder gas-phase isocyanates are collected. Particle-borne isocyanates are collected in different size fractions in the cascade impactor and very small (<0.5 µm) particles are collected on the end filter. During sampling, the denuder releases DBA that continuously refreshes the impactor substrates (IPSs) and the end filter. This improves the derivatization of isocyanates in particles.
A schematic cross-cut view of the sampler is illustrated in Fig. 1. The same notation for the sampler parts has been used in the text.
- The multichannel-plate denuder consisted of eight parallel impregnated glass–fiber plates (filter type MGF, Munktell, Falun, Sweden) with a length of 72 mm and a width of 26 mm. The distance between the denuder plates was 2 mm. The total area facing the airstream was 300 cm2. The denuder plates were held together in a plastic holder constructed from polypropylene plastic. The denuder was manufactured in a local workshop.
- A single-orifice impaction stage with a diameter of 2.7 mm (d50 = 2.5 µm) and a jet to IPS distance of 3.6 mm.
- A second impaction stage with a diameter of 1.5 mm (d50 = 1.0 µm) and a jet to IPS distance of 2.4 mm.
- A third impaction stage with a diameter of 1.0 mm (d50 = 0.51 µm) and a jet to IPS distance of 1.9 mm. The IPSs consisted of impregnated thin glass–fiber plates (filter type MGC, Munktell) mounted in a recess on the impactor plates.
- An end glass fiber filter (type MG 160, Munktell) with pore size of 0.3 µm and a diameter of 25 mm. The end filter was placed in a polypropylene cassette (Swinnex 25 mm, Millipore, Bedford, MA, USA). The airflow through the DI sampler was 5 l min–1. A Laboport twin diaphragm vacuum pump (KNF Neuberger GmbH, Freiburg, Germany) was used for air sampling. The impactors were made from aluminum at a local workshop.
|
DBA impregnation of the DI sampler parts.
All impregnation solutions used contained equimolar amounts of DBA and acetic acid.
Each denuder plate was impregnated with 1.5 ml of a solution containing 1.4 M DBA and acetic acid in methanol. Impregnation was performed by adding 0.75 ml of the impregnation solution to one side of the plate and then the plate was dried. The other side of the plate was impregnated in the same way.
The impactor glass–fiber plates were placed in the IPS holder and impregnated with 65 µl of a saturated solution of DBA and acetic acid, obtained by mixing of 2 ml DBA, 673 µl of acetic acid and 50 µl of H2O. To keep this impregnation solution from becoming sticky during application, it was kept in a water bath with a temperature of about 50°C.
The end filter was impregnated with 300 µl of a 0.7 M solution of DBA and acetic acid in methanol. The filter was dried before it was mounted in the holder.
IF sampling system.
An IF sampling system was used as the reference method for comparison of air levels measured with the DI sampler. Midget impinger flasks (Werner Glas & Instrument, Stockholm, Sweden) were filled with 10 ml of 0.01 M DBA in toluene, and connected in series with a 13-mm glass–fiber filter (type AE, SKC Inc., Eighty Four, PA, USA) with a pore size of 0.3 µm. The filter was placed in a polypropylene holder (Swinnex 13 mm). The airflow through the IF system was 1 l min–1. The IF method using DBA as a reagent has been thoroughly described in the literature and adopted as an ISO standard for isocyanate measurements in 2006, by the International Organization for Standardization (2006).
Work-up procedures
Denuder.
After sampling, the glass–fiber filter plates were cut out with a sharp knife and folded using tweezers. Single-filter plates were transferred to test tubes, then 3 ml of 1 mM H2SO4, 3 ml methanol, 6 ml toluene and 50 µl IS were added. A sample was extracted for 5 min in a shaker, sonicated for 10 min in an ultrasonic bath and shaken once more for 20 min. The mixture was centrifuged and the toluene (5.5 ml) was transferred to a new test tube. The extraction was repeated by addition of an aliquot of 5.5 ml toluene. The toluene solution (
11 ml) was evaporated to dryness in a vacuum centrifuge, and the residue was dissolved in 0.5 ml acetonitrile.
Standards were prepared by adding isocyanate–DBA derivatives standard solution (0–0.7 ml) to 3 ml of 1 mM H2SO4, 1.5 ml methanol, 1.5 ml 1.4 M DBA–acetic acid in methanol, 6 ml toluene and 50 µl IS (4 µg ml–1) in test tubes. Standards were extracted in the same way as described for the samples.
IPS and end filter.
After sampling, the IPSs and the end filter were transferred to test tubes containing 5 ml of toluene. 50 µl IS and 2 ml of 1 mM H2SO4 was added. The samples were extracted by sonication for 10 min, shaking for 10 min and centrifugation for 10 min at 3000 r.p.m. The toluene phases were separated and transferred to new test tubes. The toluene was evaporated to dryness and the residues were dissolved in 0.5 ml acetonitrile before injection on the LC-MS.
Standards were prepared in 5 ml toluene, 50 µl IS, 2 ml of 1 mM H2SO4 and 150 µl of 1.4 M DBA–acetic acid in methanol by adding isocyanate–DBA standard solution (0–0.7 ml).
IF sampler.
After air sampling, the impinger solutions were transferred to test tubes before addition of 50 µl of IS. The end filters were transferred to test tubes containing 10 ml of 0.01 M DBA in toluene for separate analysis. The filter samples were shaken for 10 min and sonicated for 10 min to extract the isocyanate–DBA derivative from the filters. After centrifugation, the toluene was transferred to new test tubes. The toluene solutions were evaporated to dryness and dissolved in 0.5 ml acetonitrile before LC-MS analysis. Standards were prepared in 10 ml of 0.01 toluene–DBA by adding isocyanate–DBA standard solution (0–0.7 ml) and 50 µl IS.
Test chamber.
A test chamber with a total volume of 0.3 m3 made from stainless steel and glass was used for evaluation of the sampler. The chamber has been described earlier (Marand et al., 2005), but some modifications were made. Four small computer processor fans were placed at different positions in the chamber to reach an efficient mixing of the air. The chamber was equipped with a humidifier for conditioning of the inlet air. Briefly, the incoming air is humidified by mixing a dry airstream with an airstream that has been filtered through temperature-controlled water in a glass bottle. The humidifier makes it possible to generate a relative humidity (RH) between 10% and 95%, with a total flow rate of air between 30 and 100 l min–1.
During the studies, 1–4 DI samplers were placed on the floor of the chamber. The inlet of the sampler was positioned about 35 cm from the floor of the chamber. The inlet for the IF samplers was about 5 cm from the inlet of the DI sampler.
Generation of isocyanates in the test chamber
Gas-phase isocyanates.
An atmosphere of isocyanates was generated by injecting an isooctane solution containing isocyanates [HDI (6.3 mg ml–1), TDI (12.2 mg ml–1) and IPDI (10.6 mg ml –1)] with a 100-µl syringe, for about 10 s, in a wad of glass wool placed in a glass tube (10 cm long and 0.8 cm internal diameter). A flow of air (15 l min–1) was blown through the tube for 2 min to transfer the isocyanates to the chamber.
Thermal degradation products of PUR.
Thermal degradation products were generated by heating PUR (50 mg), surrounded by small glass wool wads and placed in a glass tube, for about 1.5–2 min in a hot airstream (400°C, 15 l min–1). The formed aerosol stream was mixed with the incoming humidified air and blown into the chamber. Soft TDI-based flexible foam, rigid MDI-based foam and HDI-based PUR coating from a car were studied. The generation of degradation products was started 1 min after the air sampling pumps was started. The air sampling time was 10 min. The generation technique produced an aerosol containing gas-phase isocyanates and small (<1 µm) condensation particles.
Spraying of PUR coating.
An isocyanate aerosol was generated by spraying, using a nebulizer. Components for a commercial coating system, containing isocyanate hardener, polyols and solvents, were mixed and sprayed to simulate the type of aerosol generated during spray painting in the industry. The coating system, of HDI–isocyanurate type, was mixed as described by the supplier. Nebulizing gas (nitrogen) was supplied with a pressure of 3.5 bars. The paint was sucked into the nebulizer by the nebulizing gas at a rate of 0.7 l min–1. The spray was directed towards a metal surface and paint particles that did not impact on the surface were transferred to the test chamber. This type of spraying generates an aerosol where larger particles (>2 µm) are dominant.
Particle collection efficiency of the impactor.
To determine the particle collection efficiency of the impactor, the different stages were connected to the inlet of the APS. The sampling flow for the APS was 5 l min–1, the same flow as the impactor was designed for. Single stages were evaluated by measuring the aerosol passing through the stage with the impaction plate and impregnated filter mounted to the stage and comparing this result with the aerosol passing through the stage without the plate. A non-reactive, poly-disperse glass aerosol was used to determine the collection efficiency of the two top stages. The glass aerosol consisted of very hard particles, representing a worst-case regarding particle bounce-off.
Extraction recovery.
The extraction efficiency for the denuder plates was checked by spiking impregnated denuder plates with known amounts of DBA derivatives. The amount of derivatives corresponded to 0.2 µg isocyanate/denuder plate. The denuder plates were extracted according to the work-up procedure. Results were compared to spiked mixtures of 3 ml of 1 mM sulfuric acid, 6 ml toluene, 1.5 ml of 1.4 M DBA–acetic acid in methanol and 1.5 ml methanol.
The extraction efficiency for the IPS and the end filter was tested by spiking with diisocyanate–DBA derivatives, corresponding to 0.2 µg isocyanate. Results were compared with spiked mixtures of 2 ml of 1 mM sulfuric acid, 5 ml toluene and 150 µl of 1.4 M DBA and acetic acid in methanol.
Comparison of DI sampler with IF sampler
Sampling of gas-phase isocyanates.
A comparison of total air levels for the DI sampler and the IF sampler for sampling of gas-phase isocyanates was performed. For each measurement, three IF samplers and three DI samplers were placed in the chamber. Sampling was performed at RH levels of 20%, 40%, 60% and 80%. The total sampling time was 10 min.
Sampling of thermal degradation products.
Total air levels obtained with the DI sampler and the IF sampler were compared for sampling of thermal degradation products of PUR. For each measurement, three IF samplers and three DI samplers were used. Sampling was performed at four different RH levels (20%, 40%, 60% and 80%) and the sampling time was 10 min.
Spraying of PUR-coating compounds.
The total air levels obtained for the DI sampler were compared in pairs (n = 8) with the IF sampler during spraying of PUR-coating compounds. The volume of paint used for spraying was between 0.5 and 6 ml. The RH in the sampling chamber was 50%.
Sampler variation.
Three DI samplers were used for simultaneous sampling in the test chamber to study the variation of the measured air levels between the samplers. For comparison, two measurements (n = 6) were performed at a RH of 80%.
Distribution in the DI sampler.
The distribution of isocyanates in the DI sampler was evaluated for thermal degradation products of PUR (n = 3), and for an aerosol generated by spraying of PUR coating (n = 3). For thermal degradation products, the total sampling time was 10 min and the RH was 60%. For the spraying, the sampling time was 15 min and the RH was 50%.
Distribution in the denuder.
The distribution of isocyanates in the denuder was studied by analyzing all the eight denuder plates separately. This was done to investigate if the isocyanates were spread symmetrically across the denuder.
Experiments where the denuder plates were combined (plates 1–4 and 5–8) and cut in four parts from top to bottom of the denuder were also performed to study the isocyanate distribution in the airflow direction of the denuder.
Storage of samplers
Storage of non-exposed samplers
The stability of samplers before sampling was investigated by storing impregnated samplers (n = 4) for a week before sampling. Prepared sampler parts were stored in the dark and at room temperature. Denuders and impactor plates were put in plastic bags. The end filters were mounted in their cassettes and the inlet and outlet were capped. Air sampling using the stored and freshly prepared samplers was performed simultaneously in the test chamber. For each sampling occasion, two stored and two fresh samplers were compared. Thermal degradation products of PUR were generated for the air sampling.
Storage of exposed samplers.
The stability of the samplers after sampling was investigated by storing samplers (n = 16) for different time after sampling. For each sampling occasion, four samplers were used simultaneously in the test chamber. The first sampler was extracted 12 h after sampling, the second 36 h after sampling, the third after 7 days and the fourth after 21 days. Sampling was performed four times for thermal degradation products of PUR, using the same amount of PUR for each aerosol generation. About 5 min after sampling, the denuders were put in separate plastic bags. The impaction plates were taken out of the impactor and stored in plastic bags. The end filters were sealed with plugs at both ends. The samples were stored in a refrigerator until analysis.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Particle collection efficiency of the impactor
Collection efficiency curves for the first and second impaction stages are presented in Fig. 2. The cut-off diameter d50 for the top stage was 2.5 µm. For the second stage, d50 was
0.9 µm. This value is smaller than the value the impactor was designed for (1.0 µm), but previous studies have shown that porous substrates may lower the cut-off compared to solid substrates (Kavouras and Koutrakis, 2001; Huang et al., 2005).
|
Aspiration efficiency of the sampler inlet
The inlet has a difficult geometry for calculation of the aspiration efficiency, but it was estimated using data presented by Hangal and Willeke (1990). The original model was developed for a thin-walled circular inlet sampler. Here, the aspiration efficiency was calculated for one rectangular slit of the denuder, with a height of 2 mm. For approximation, it was assumed that the slit could be regarded as thin walled, and the hydraulic diameter (4 x area/wetted perimeter) for the slit could be used, since the inlet is non-circular. For an ambient air velocity of 0.5 m s–1, an aspiration efficiency of 100% or above (over-sampling) was found for particles smaller than 20 µm, when the inlet angle increases from 0 to 70°. Above 70°, the aspiration efficiency decreases, and for an inlet angle up to 85°, a slight under-sampling was found. For an inlet angle of 90°, the aspiration efficiency falls quickly, and for particles with an aerodynamic diameter of 5 µm, it was found to be
80%.
Extraction recovery
A double-toluene extraction procedure was used for the denuder plates, in order to obtain a high extraction recovery of derivatives (Table 1). Within experimental errors (95% confidence interval), no losses were observed for 8 out of 10 isocyanate derivatives investigated. For IPDI–DBA, the recovery was slightly lower, but for ICA-DBA, it was surprisingly higher.
|
For the impactor plate filters and end filters, it was not necessary to use double extraction. One aliquot of toluene was sufficient to obtain a good recovery (Table 2). Within experimental errors (95% confidence interval), no losses were observed for the five isocyanate derivatives investigated.
|
Comparison of DI sampler with IF sampler
Sampling of gas-phase isocyanates.
When comparing the measurements of gaseous isocyanates using IF and DI, the two samplers show good agreement (Table 3). The experiments showed that the denuder part of the DI collected >96% of the isocyanates, so the breakthrough of the denuder for gaseous isocyanates is acceptable. For the IF samplers,
99% of the isocyanates were collected in the impinger part. For 16 out of 20 DI–IF pairs, there was no significant difference (P < 0.05). For 2,4-TDI (20% RH), IPDI (1) (20% RH), IPDI (2) (20% RH) and IPDI (2) (60% RH), there were some difference. Except for IPDI (1) (20% RH), the difference was <30%. The limited number of measurements indicates that the two methods overall show the same result.
|
Sampling of thermal degradation products.
When comparing the measurements of isocyanates in thermal degradation products of PUR using IF and DI, the two samplers show reasonable good agreement as the number of samples were low (Table 4). For 18 out of 36 DI–IF pairs, there was no significant difference (P < 0.05). For 32 out of 36 DI–IF pairs, the difference was <30%. For three EIC–DBA (RH 40%, 60% and 80%) pairs, there were differences of >30%, but the levels were low making the result inconclusive. For ICA–DBA, a considerable higher concentration was found with the DI sampler at RH = 80%. There was also a trend of higher concentration of ICA–DBA with increasing RH. Still, with the exception of ICA–DBA at high humidity, the limited number of measurements indicates that the two methods overall show the same result. In Table 4, it is clearly seen that the relative standard deviation for the DI sampler is in the range of 0.5–13% (average = 5%) and for the IF it is in the range of 1.3–19.5% (average = 7.5%). This demonstrates the experimental difficulties to obtain ideal sampling conditions. Calculated differences may therefore be misleading if the overall picture is not taken into account.
|
For five DI samplers, an additional end filter was connected. On the additional filter, an average level of 3.3% (range 1.1–7.9%) of the total collected amount of isocyanates was found indicating a low breakthrough.
Spraying of PUR coating.
Virtually linear relationships between results obtained with the DI and the IF samplers were observed. The concentration of HDI–biuret for the DI sampler (x; HDI–biuret, 1.2–11.7 µg m–3, n = 8) was plotted against the IF sampler (y). The linear expression was y = 1.17x – 0.69 (R2 = 0.983). The concentration of HDI–isocyanurate for the DI sampler (x; HDI–isocyanurate, 25.5–163 µg m–3, n = 8) was plotted against the IF sampler (y). The linear expression was y = 1.22x – 2.7 (R2 = 0.982). The concentration of HDI–diisocyanurate for the DI sampler (x; HDI–diisocyanurate, 7.4–44.7 µg m–3, n = 8) was plotted against the IF sampler (y). The linear expression was y = 0.97x – 3.2 (R2 = 0.965).
Sampler variation.
In Table 5, 18 triplets of isocyanates from DI thermal decomposition samples are presented. The average difference (in concentrations) for the 18 triplets is 10%. The difference ranged 3–19%. It was only possible to use three samplers simultaneously, due to limited space in the chamber. The limited number of samplers makes it difficult to apply proper statistical methods.
|
Distribution in the DI sampler.
Overall,
70% of the isocyanates formed during the thermal decomposition of PUR were found in the denuder part. It can be seen that different isocyanates have different distributions in the DI sampler (Fig. 3). ICA, MIC and HDI are predominantly present in the gas phase and collected to >95% in the denuder. About 40% of 2,4-TDI is associated to particles and is found in impactor stage 3 and the end filter. MDI is only found in the impactor stages and on the end filter.
|
The three HDI adducts from the coating system were found in all of the parts of the DI sampler (Fig. 4). The three HDI adducts show the same pattern and the composition reflects composition in the paint system. The isocyanates (90%) were mainly found on the two top stages of the impactor. The particles formed during spraying are much larger than particles formed during thermal degradation.
|
The variation in distribution over the impactor stages and end filter was low, as can be seen in Fig. 5.
|
Distribution in the denuder.
In a gas-phase study, the eight denuder plates were analyzed separately. It was found that the isocyanates were evenly distributed over the denuder plates.
The distribution of isocyanates in the airflow direction in individual denuder plates was studied. Soft TDI–PUR foam was thermally decomposed and samples were collected by four DI samplers. The denuder plates were cut into four parts. It can be seen in Fig. 6 that 70% of the isocyanates were collected in the top quarter part and 10–20% in the second part. About 2–6% of the isocyanates were collected in the third and the bottom parts.
|
Storage of samplers
Storage of non-exposed samplers.
Samplers (n = 2) that were prepared 1 week in advance of sampling showed only very small differences when they were compared with freshly prepared samplers (n = 2) (Table 6). The stability of the reagent on the different glass–fiber filters makes it possible to prepare the samplers and store them in the laboratory before sampling.
|
Storage of exposed samplers.
The DI samples (n = 16) were found to be stable during storage. The total average of all isocyanate derivatives between 12 h and 21 days, varied ±6%. As compared to samples of 12 h, the found amount, of all isocyanate derivatives, in the stored samples was 91%. For monoisocyanates, it was 96%, and for diisocyanates, it was 85% (Table 7). No trend of decreasing amounts was found in the samples from day 1 to day 21. It cannot be excluded that the amounts of isocyanate derivatives decreased during the first 24 h. The difference of amounts is, however, about the same as for the DI variation study and the number of samples was limited.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Particle collection efficiency of the impactor
The design of the impactor is an old principle for size-selective sampling of particles (Marple and Liu, 1974). The performance of the impactor was checked with a glass aerosol. For the stages one and two, the cut-off diameters were as expected. For the determination of the cut-off characteristics of the last stage, a scanning mobility particle sizer system will be used in future studies.
Aspiration efficiency
The calculation of the aspiration efficiency shows that there are differences in the collection efficiency of the sampler inlet for different polar angles. The calculation was extrapolated from data on sharp-edged samplers with cylindrical geometry, which is very different from that of the denuder. Dependence on azimuthal as well as the mentioned polar angle is likely significant. Aspiration may also be slot dependent. Therefore, further research into measuring or computing (e.g. via computational fluid dynamics) the aspiration efficiency into the denuder may be justified.
The angular dependence could be minimized with a preselector, designed according to aspiration efficiency models. A few attempts have been made to construct such a preselector, but so far no working solution has been found. A difficulty that was found was that the levels of gas-phase isocyanates were lower when a preselector was used, probably due to surface adsorption.
Extraction recovery of denuder samples
With the exception of ICA–DBA, the extraction recovery for the isocyanate DBA derivatives in denuder samples was very good. It was observed that the extraction of the DBA derivatives differed slightly as compared to the deuterium-labeled DBA derivatives in the IS. The double extraction enables a high extraction recovery for all of the derivatives.
In the impregnated denuder, a very high amount of DBA is present. Traces of impurities have a great effect. This is seen for ICA–DBA. The 122% recovery originates possibly from other compounds present. The formation of ICA–DBA is unaffected by an airflow through the denuder and hence it is possible to use a blank subtraction to correct the ICA values. The problem with formation of an ICA derivative has also been observed for other isocyanate reagents (Henneken et al., 2006).
Comparison of DI sampler with IF sampler
In general, a very good agreement between the DI samples and IF samples was obtained. For some of the compared DI and IF triplets, there were statistically significant differences. However, this is partially due to very low variation within the sampler type, that is, lower than the ‘normal’ variation of 10%. Experimental conditions made it possible only to use three IF samplers and DI samplers simultaneously. The total breakthrough of the DI sampler is low (<5%). At high humidity (>60%), the DI values greatly exceeds the IF values for ICA, and although this has been observed in a previous study (Marand et al., 2005), the reason for the high DI values is still not known.
Distribution in the DI sampler
From the samples collected in the sampling chamber, it is evident that the distributions of isocyanates vary for different types of aerosols. This is clearly seen in the sampling of thermal degradation products, where the volatile monoisocyanates are collected by the denuder, while aromatic diisocyanates are distributed over all sampler stages. The results from thermal degradation are interesting to compare with the results from gas-phase sampling of diisocyanates. In gas-phase sampling, the particle concentration is low, and the gaseous isocyanates generated are collected by the denuder. The same isocyanates are found in particle fractions when thermal degradation products are collected, indicating that the aromatic isocyanates condense onto particles formed during the thermal degradation process. It is also possible that aromatic isocyanates have greater affinity for particle surfaces, due to the electrons in the aromatic ring. No studies on the accurate partition of isocyanates between gas and particles have been found, but in studies of distribution of compounds in tobacco smoke, a similar situation has been found. For distributions of alkanes and polycyclic aromatic hydrocarbons (PAHs), it has been found that PAHs are more strongly associated to particles than aliphatic hydrocarbons, possibly by stronger interactions of electrons in the aromatic ring with organic particles (Liang and Pankow, 1996). If the air levels of HDI, TDI and MDI are compared to the equilibrium vapor pressure of the compounds, calculated as a concentration, HDI and TDI are both expected to be found in the gas phase, while MDI is expected to be distributed both between gas and particles (Streicher et al., 1994).
Spraying of coating compounds
The DI and IF samples were well in agreement regarding total isocyanates collected. The sampler is designed to collect particles larger than 2.5 µm on the first impaction stage. During spraying, 5% of the HDI adducts are collected in the denuder part. These are not in the gas phase. The result indicates that the flow before the denuder in not perfectly laminar.
For HDI–biuret and HDI–isocyanurate, the concentrations measured using IF sampler were slightly higher than for the DI sampler. Regarding the aspiration efficiency, there could be some under-sampling for the DI sampler when large particles in a spray aerosol are collected. The IF sampler may also under-sample larger particles, but the aspiration efficiency for this sampler do not drop off as fast as for the DI sampler.
Distribution in the denuder
About 90% of both 2,6-TDI and ICA is collected in the top-half of the denuder. This indicates that the differences in diffusion for the different isocyanates are small.
Storage of samplers
Stored-unexposed DI samplers are found to be stable, and stored-exposed DI samplers show no trend of degradation of the DBA derivatives. The variation between samplers is the same as for several freshly prepared exposed samplers. However, there may be losses during the first 12 h of storage, but in comparisons with the IF sampler, the result was about the same, regarding total concentrations of DBA derivatives.
General discussion
There are big differences in reaction rates for the monoisocyanates, aliphatic and aromatic isocyanates with the DBA reagent. The very reactive isocyanate group makes efficient derivatization essential. If not the isocyanates will react with moisture and other compounds present in air. In thermal decomposition products of PUR, polyols, aminoisocyanates and amines are often present; however, no side reactions with these species have been observed. The high efficiency of the derivatization in the DI sampler is achieved by the high amount of DBA present. In addition, the sampler is continuously flushed, with DBA from the denuder part, which refreshes the DBA on the IPSs and the end filter. The small breakthrough on the end filter is due to the high linear flow rate through the end filter.
The design of the presented DI sampler may look complicated, but it is based on three robust techniques: Denuder sampling with DBA–acetic acid (HAc) impregnation, the well known impactor technique that has been modified with DBA–HAc-impregnated substrates and the sampling on impregnated filters. The stable DBA reagent and the efficient technique to remove excess reagent in the work-up are essential as huge amounts of the reagent are used. Another advantage with DBA is the possibility to analyze other compounds such as aminoisocyanates, amines and anhydrides by the same sampling technique (Karlsson et al., 2002; Dahlin et al., 2004). Details on this will be presented in future studies.
Isocyanates are known to have a diverse toxicology. There are reasons to believe that the physical form may have a great effect. As the lungs are the one of the most interesting target organs for isocyanate exposure, it must be of importance to know to what concentrations the lungs are exposed to. It is not possible to take air samples in the lungs, but it is known that certain particle size fractions may penetrate into the alveolar part, whereas gas and large particles will only reach the upper respiratory system (nose, mouth and throat) (Witschi and Last, 1996). Exposure data from the DI sampler will be important for studies regarding the effect of isocyanates on different part of the human airways. According to models for particle deposition in the human airways, the particle penetration and deposition is dependent on particle size. The different parts of the DI sampler may reflect the exposure of different parts of the human airways. The denuder and to some extent the first impaction stage may represent fractions deposited in the nasopharyngeal region, the first and second impaction stage may reflect the tracheobroncial region, while the last impaction stage and the end filter may reflect the alveolar region.
When data from thermal degradation of PUR are related to deposition models, inhalation of thermal degradation products possesses a greater health risk than inhalation of isocyanates in gas phase, since a huge part of the reactive isocyanates reaches parts of the airways where disease may occur.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
A denuder cascade impactor sampler for isocyanates is presented. Compared with an impinger filter sampler, the total air levels are the same. The DI sampler provides information about the distribution of isocyanates between gas phase and particles size fractions. The DBA reagent used is stable, and samplers can be stored before and after sampling.
There are different distribution patterns for the isocyanates in the air depending on how the isocyanate aerosol is generated. During spraying of isocyanate coating compounds, large particle-borne isocyanate dominates. During thermal degradation of PUR, isocyanates are distributed between gaseous phase and small particles.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Monica Hassgård is acknowledged for skilful laboratory assistance. Ralf Gunnarsson is acknowledged for the manufacture of denuders and impactors.
Received August 27, 2007; in final form March 27, 2008
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Baur X, Marek W, Ammon J, et al. Respiratory and other hazards of isocyanates. Int Arch Occup Environ Health (1994) 66:141–52.[CrossRef][Web of Science][Medline]
Bernstein JA. Overview of diisocyanate occupational asthma. Toxicology (1996) 111:181–9.[CrossRef][Web of Science][Medline]
Dahlin J, Karlsson D, Skarping G, et al. Determination of airborne anhydrides using LC-MS monitoring negative ions of di-n-butylamine derivatives. J Environ Monit (2004) 6:624–9.[CrossRef][Web of Science][Medline]
De Santis F. Comment on wet effluent denuder coupled liquid ion chromatography systems—annular and parallel-plate denuders. Anal Chem (1994) 66:3503–4.
Hangal S, Willeke K. Aspiration efficiency—unified model for all forward sampling angles. Environ Sci Technol (1990) 24:688–91.
Hardy HL, Walker RF. Novel reagent for the determination of atmospheric isocyanate monomer concentrations. Analyst (1979) 104:890–1.[CrossRef]
Henneken H, Hayen H, Vogel M, et al. Validation of a diffusive sampling method for airborne low-molecular isocyanates using 4-nitro-7-piperazinobenzo-2-oxa-1,3-diazole-impregnated filters and liquid chromatography-tandem mass spectrometry. J Chromatogr A (2006) 1134:112–21.[CrossRef][Web of Science][Medline]
Heyder J, Svartengren MU. Basic principles of particle behavior in the human respiratory tract. In: Drug delivery to the lung—Bisgaard H, O'Callaghan C, Smaldone GC, eds. (2001) New York: Marcel Dekker, Inc. 37–8.
Huang CH, Chang CS, Chang SH, et al. Use of porous foam as the substrate of an impactor for respirable aerosol sampling. J Aerosol Sci (2005) 36:1373–86.[CrossRef][Web of Science]
IOS. Determination of organonitrogen compounds in air using liquid chromatography and mass spectrometry—part 1: isocyanates using dibutylamine derivatives (2006) ISO 17734–1 Geneva: International Organization for Standardization.
Karlsson D, Spanne M, Dalene M, et al. Airborne thermal degradation products of polyurethene coatings in car repair shops. J Environ Monit (2000) 2:462–9.[CrossRef][Web of Science][Medline]
Karlsson D, Dalene M, Skarping G, et al. Determination of isocyanic acid in air. J Environ Monit (2001) 3:432–6.[CrossRef][Web of Science][Medline]
Karlsson D, Dahlin J, Skarping G, et al. Determination of isocyanates, aminoisocyanates and amines in air formed during the thermal degradation of polyurethane. J Environ Monit (2002) 4:216–22.[CrossRef][Web of Science][Medline]
Karlsson D, Dahlin J, Marand A, et al. Determination of airborne isocyanates as di-n-butylamine derivatives using liquid chromatography and tandem mass spectrometry. Anal Chim Acta (2005) 534:263–9.[CrossRef][Web of Science]
Karol MH. Respiratory effects of inhaled isocyanates. Crit Rev Toxicol (1986) 16:349–79.[Medline]
Kavouras IG, Koutrakis P. Use of polyurethane foam as the impaction substrate/collection medium in conventional inertial impactors. Aerosol Sci Technol (2001) 34:46–56.[CrossRef][Web of Science]
Kloskowski A, Pilarczyk M, Namiesnik J. Denudation—a convenient method of isolation and enrichment of analytes. Crit Rev Anal Chem (2002) 32:301–35.[CrossRef]
LaPuma PT, Schilke RA, Kauth DA, et al. Chromate dissociation from three types of paint particles. Regul Toxicol Pharmacol (2002) 36:325–30.[CrossRef][Web of Science][Medline]
Lesage J, Goyer N, Desjardins F, et al. Workers exposure to isocyanates. Am Ind Hyg Assoc J (1992) 53:146–53.[Web of Science][Medline]
Liang CK, Pankow JF. Gas/particle partitioning of organic compounds to environmental tobacco smoke: partition coefficient measurements by desorption and comparison to urban particulate material. Environ Sci Technol (1996) 30:2800–5.
Marand A, Dahlin J, Karlsson D, et al. Determination of technical grade isocyanates used in the production of polyurethane plastics. J Environ Monit (2004) 6:606–14.[CrossRef][Web of Science][Medline]
Marand A, Karlsson D, Dalene M, et al. Solvent-free sampling with di-n-butylamine for monitoring of isocyanates in air. J Environ Monit (2005) 7:335–43.[CrossRef][Web of Science][Medline]
Marple VA, Liu BYH. Characteristics of laminar jet impactors. Environ Sci Technol (1974) 8:648–54.
Marple VA, Olson BA, Rubow KL. Inertial, gravitational, centrifugal, and thermal collection techniques. In: Aerosol measurement: principles, techniques and applications—Baron PA, Willeke K, eds. (2001) 2nd edn. Hoboken, NJ: John Wiley & Sons, Inc. 229–60.
Nordqvist Y, Melin J, Nilsson U, et al. Comparison of denuder and impinger sampling for determination of gaseous toluene diisocyanate (TDI). Fresenius J Anal Chem (2001) 371:39–43.[CrossRef][Medline]
Ott MG, Diller WF, Jolly AT. Respiratory effects of toluene diisocyanate in the workplace: a discussion of exposure-response relationships. Crit Rev Toxicol (2003) 33:1–59.[Web of Science][Medline]
Rando RJ, Poovey HG. Dichotomous sampling of vapor and aerosol of methylene-bis-(phenylisocyanate) [MDI] with an annular diffusional denuder. Am Ind Hyg Assoc J (1994) 55:716–21.[Web of Science][Medline]
Rando RJ, Poovey HG. Development and application of a dichotomous vapor/aerosol sampler for HDI-derived total reactive isocyanate group. Am Ind Hyg Assoc J (1999) 60:737–46.[Web of Science][Medline]
Redlich CA, Karol MH. Diisocyanate asthma: clinical aspects and immunopathogenesis. Int Immunopharmacol (2002) 2:213–24.[CrossRef][Web of Science][Medline]
Sangö C, Zimerson E. A new reagent for determination of isocyanates in working atmospheres by HPLC using UV or fluorescence detection. J Liq Chromatogr (1980) 3:971–90.[CrossRef][Web of Science]
Spanne M, Tinnerberg H, Dalene M, et al. Determination of complex mixtures of airborne isocyanates and amines.1. Liquid chromatography with ultraviolet detection of monomeric and polymeric isocyanates as their dibutylamine derivatives. Analyst (1996) 121:1095–9.[CrossRef]
Spanne M, Grzybowski P, Bohgard M. Collection efficiency for submicron particles of a commonly used impinger. Am Ind Hyg Assoc J (1999) 60:540–4.[Web of Science]
Streicher RP, Kennedy ER, Lorberau CD. Strategies for the simultaneous collection of vapors and aerosols with emphasis on isocyanate sampling. Analyst (1994) 119:89–97.[CrossRef][Medline]
Streicher RP, Arnold JE, Ernst MK, et al. Development of a novel derivatization reagent for the sampling and analysis of total isocyanate group in air and comparison of its performance with that of several established reagents. Am Ind Hyg Assoc J (1996) 57:905–13.[Web of Science][Medline]
Streicher RP, Reh CM, Key-Schwartz RJ, et al. Determination of airborne isocyanate exposure: considerations in method selection. AIHAJ (2000) 61:544–56.[CrossRef][Web of Science][Medline]
Tsai CJ, Cheng KC, Aggarwal SG, et al. Simultaneous sampling of gas- and aerosol-phase TDI with a triple filter system. J Air Waste Manag Assoc (2003) 53:1265–72.[Web of Science][Medline]
Vincent JH. The inhalation of aerosols. Aerosol science for industrial hygienists (1995) 1 edn. Oxford: Pergamon Press. 136–58.
Warwick CJ, Bagon DA, Purnell CJ. Application of electrochemical detection to the measurement of free monomeric aromatic and aliphatic isocyanates in air by high-performance liquid-chromatography. Analyst (1981) 106:676–85.[CrossRef]
Witschi HR, Last JA. Toxic responses of the respiratory system. In: Casarett and Doull's toxicology—Klaassen CD, ed. (1996) 5 edn. New York: McGraw Hill. 446–9.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. Dahlin, M. Spanne, M. Dalene, D. Karlsson, and G. Skarping Size-Separated Sampling and Analysis of Isocyanates in Workplace Aerosols--Part II: Aging of Aerosols from Thermal Degradation of Polyurethane Ann. Hyg., July 1, 2008; 52(5): 375 - 383. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||