Annals of Occupational Hygiene Advance Access originally published online on January 24, 2006
Annals of Occupational Hygiene 2006 50(4):385-393; doi:10.1093/annhyg/mei075
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
© 2006 British Occupational Hygiene Society Published by Oxford University Press
Original Article |
Determination of Airborne Isocyanates Generated During the Thermal Degradation of Car Paint in Body Repair Shops
1 McGill University, 3450 University Street, FDA Building, Room 31, Montreal, Quebec, Canada H3A 2A7; 2 Institut de recherche Robert-Sauvé en santé et en sécurité du travail, 505 de Maisonneuve Blvd. West, Montreal, Quebec, Canada H3A 3C2
* Author to whom correspondence should be addressed. e-mail: Andre.Dufresne{at}mcgill.ca
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTALRESULTS AND DISCUSSIONCONCLUSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Polyurethanes are widely used in car paint formulations. During thermal degradation, such polymeric systems can generate powerful asthmatic sensitizing agents named isocyanates. In body repair shops, the thermal degradation of car paint can occur during abrasive processes that generate enough heat to involve release of isocyanates in air. An environmental monitoring study was performed in two body repair training schools and in a body repair shop to evaluate the workers' exposure to isocyanates during cutting, grinding and orbital sanding operations. For sampling, cassettes containing two 1-(2-methoxyphenyl)piperazine (MOPIP)-coated glass fiber filters (MFs) (
5 mg of MOPIP per filter) and bubblers containing 15 ml of MOPIP solution in toluene (1.0 mg ml–1) backed at the outlet with cassettes containing two MFs were used. Tandem mass spectrometry was used to analyze the MOPIP derivatives of isocyanic acid (HNCO), all the linear aliphatic isocyanates ranging from methyl isocyanate (Me-i) to hexyl isocyanate, all the alkenyl isocyanates ranging from propylene isocyanate to hexylene isocyanate, 1,6-hexamethylene diisocyanate (HDI), trans- and cis-isophorone diisocyanate (IPDI), 2,4- and 2,6-toluene diisocyanate (TDI), 2,4'-; 2,2'- and 4,4'-methylenediphenyl diisocyanate (MDI), phenyl isocyanate (Ph-i) and p-toluene isocyanate (p-Tol-i). The instrumental detection limits (LOD) were in the 0.13–0.75 µg of NCO per m3 range for 15 l air samples converted into 3 ml liquid samples. The isocyanate concentrations detected in the workers' breathing zone were in the 1.07–9.80 µg of NCO per m3 range for cutting, 0.63–3.62 µg of NCO per m3 range for grinding and 0–1.29 µg of NCO per m3 range for sanding. However, a rapid decrease of the isocyanate concentration was observed while moving away from the emission source. Among the isocyanates detected the most abundant were the monomers (MDI, HDI, TDI and IPDI) and Me-i.
Keywords: body repair shop • car paint • cutting and orbital sanding • grinding • isocyanate • polyurethane • thermal degradation
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTALRESULTS AND DISCUSSIONCONCLUSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Polyurethane-based paints are readily used in car industry owing to their outstanding technical features such as durability, color stability and resistance to abrasion, chemicals and weather extremes. The typical polyurethane paint system has two components: polyol with pigments, solvent and additives; and a second component containing the isocyanate(s) (hardener) in an appropriate solvent. The primer applied between the metal and the paint and the clear-coat applied on the paint can also contain polyurethanes.
Upon thermal degradation, polyurethanes can regenerate isocyanates (Tinnerberg et al., 1997
; Karlsson et al., 2002; Pronk et al., 2005), which may cause bronchial asthma, bronchitis, rhinitis, conjunctivitis and dermatitis (Raulf-Heimsoth and Baur, 1998). Schedule 1 (Health & Safety Executive (UK) (HSE, 2005) of the Control of Substances Hazardous to Health (COSHH) Regulations specifies a Workplace Exposure Limit (WEL) of 20 µg m–3, 8 h time-weighted average (TWA) reference period, for isocyanates in air. A short-term exposure limit, 15 min reference period of 70 µg m–3 is also specified. These limits are expressed as ‘weight of equivalent NCO groups’. It has been observed that isocyanates do not have exactly the same toxicity. For example, the LC50 (death in 50% of animals) for male rats exposed 4 h to methyl isocyanate (Me-i) is 18 mg of NCO per m3, whereas it is 77 mg of NCO per m3 for propyl isocyanate (Prop-i) (Pauluhn, 1989). Moreover, some researchers provided evidence that workers sensitized to isocyanates could react to diisocyanate concentrations as low as 1.0 p.p.b. (3.4 µg of NCO per m3), which is significantly lower than the WEL (Mapp et al., 1999).
In car body repair shops, the emission of isocyanates has previously been observed during many thermal degradation processes such as cutting, grinding and welding (Karlsson et al., 2000; Henriks-Eckerman et al., 2002). The isocyanates generated during the thermal degradation of polyurethane-based car paints depend greatly on the thermal degradation conditions and the nature of the isocyanate(s) used in the formulation. For example, during welding operations in car repair shops, the main isocyanates detected by Karlsson et al. (2000) were Me-i, toluene diisocyanate (TDI) and 1,6-hexamethylene diisocyanate (HDI), whereas those detected by Henriks-Eckerman et al. (2002) were 4,4'-methylenediphenyl diisocyanate (MDI), TDI and HDI.
During paint spraying, workers are exposed to the isocyanates contained in the paint formulation. However, during thermal degradation, many new entities of isocyanates can be generated owing to secondary reactions such as chain breaking, isomerization and dehydrogenation. Boutin et al. (2004) studied the combustion of an HDI-based clear coating used in the car industry with a laboratory scale furnace. HDI, all the linear aliphatic isocyanates ranging from Me-i to hexyl isocyanate (Hex-i), all the alkenyl isocyanates ranging from propylene isocyanate (Propylene-i) to hexylene isocyanate (Hexylene-i) and many structural isomers of these compounds were detected. Because of the absence of a specific Occupational Exposure Standard for many of these isocyanates, the general HSE's WEL (HSE, 2005), including all organic compounds containing an NCO function was used as a reference value in the present study.
All isocyanates generated under thermal degradation in a gaseous effluent, even at low concentrations, can have a significant effect on workers' health if one hypothesizes their additive effect. Thus, the purpose of this research was to use the high sensitivity and the high selectivity of mass spectrometry to quantify the greatest possible number of isocyanates generated during the thermal degradation processes occurring in auto body shops and to evaluate the potential exposure of workers. For the present research, cutting, grinding and orbital sanding were retained as processes that potentially cause thermal degradation of car paint. Welding processes were not retained, because good work practices entail paint stripping in order to establish the arc between the electrode and the work piece.
|
EXPERIMENTAL |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTALRESULTS AND DISCUSSIONCONCLUSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Field measurements
Air measurements were performed at three different workplaces: two auto body repair training schools (A and B) and a body repair shop. For the sampling, each studied operation (cutting, grinding and orbital sanding) was performed continuously during 15 min in such a way so as to compare them under similar conditions. At each workplace, each operation was repeated twice except for School A where orbital sanding was performed only once. No painting operations were performed in the workplaces during the sampling to prevent air contamination.
The paints degraded during this study came from a green Honda Civic 2000 for School A, a gray Isuzu Rodeo 4WD 2001 for School B and a red Honda Civic 2000 for the body repair shop. All three cars had their original coating. The cutting operations were done with a Cut Off Tool model CP-861 from Chicago Pneumatic (Morton Grove, IL) equipped with a 3M Green Corps Cut Off Wheel, 3'' x 1/16'' x 3/8'', from 3M Automotive (St Paul, MN) (Fig. 1A). For the grinding operations, a High Speed Sander Heavy Duty model CP-778 from Chicago Pneumatic equipped with a 3M Green Corps Fibre Disc, grit 24, 5'' x 7/8'', from 3M Automotive was used (Fig. 1B). At School A and at the body repair shop, the orbital sanding was done using an Ingersoll Rand Air Angle Die Grinder from Northern Tool + Equipment (Burnsville, MN) equipped with a 3M Scotch Brite Roloc Surface Conditioning Disc, 3'', from 3M Automotive (Fig. 1C), whereas at School B it was a Hutchins 8650-Multi-Option Random-Orb Sander from S&J's Discount Tools (Lewis Center, OH) equipped with a 3M Imperial Hookit Dust-Free Disc, 6'', from 3M Automotive (Fig. 1D).
|
Air sampling
Air sampling was performed using two methods adapted and validated in the laboratory by Boutin et al. (2005)
for the sampling of the isocyanates generated during the thermal degradation of car paint. These methods use cassettes containing two 1-(2-methoxyphenyl)piperazine (MOPIP)-coated glass fiber filters (MFs) (5 mg of MOPIP per filter) and bubblers containing 15 ml of MOPIP solution in toluene (1.0 mg ml–1) backed with cassettes containing two MFs at the outlet. The connections between the bubblers and the cassettes were made using Fluran, an inert fluoroelastomer, tubing (Integra Companies, Inc., Devens, MA). Since particles <2 µm in diameter are not efficiently collected by bubblers (Streicher, 1994) and particles >10 µm, containing isocyanates, tend to polymerize on reagent-coated filters before the derivatization occurs (Brenner et al., 2003), the efficiency of these sampling devices may differ from the laboratory to the workplace. For this reason, the bubblers backed with cassettes at the outlet were used as reference to evaluate the sampling efficiency of the cassettes containing the two MFs. Indeed, contrary to the MFs, the sampling efficiency of the double stage sampling system (bubbler/MFs) is not affected by the aerosols size distribution.
During each experiment, two air samples were collected at 15 cm from the emission source and two others in the workers' breathing zone, using cassettes. Bubblers, containing a flammable solvent, were not used near the emission source owing to the explosion risks. Aircheck Samplers, model 224-52 from SKC Inc. (Eighty Four, PA), were used to produce flow rates of 1.0 l min–1. The sampling was also performed at 2 m from the emission source using a sampling tree containing three cassettes and three bubblers backed with cassettes. This tree, shown in Fig. 2, consisted of a Motor Mounted Rotary Vane model 1531 from Gast Manufacturing Inc. (Benton Harbor, MI) connected to a union cross itself connected to three Adjustable Low Flow Dual Tube Holders from SKC Inc. To compare their sampling efficiency, a cassette and a bubbler backed with a cassette were connected to each dual tube holder and the sampling was performed at 1.0 l min–1. The flow rate calibration for all the sampling devices used during this research was done at the work site using a DryCal DC2 Primary Flow Calibrator from Bios International Corporation (Butler, NJ).
|
Sample preparation
After the experiments, the sample solutions contained in the bubblers were evaporated to dryness in a TurboVap LV from Zymark (Hopkinton, MA). The dry residues were dissolved in 3.0 ml of dimethylformamide (DMF) (10% v/v) in acetonitrile (ACN). For the cassettes, the two MFs were put simultaneously in the same vessel containing 3.0 ml of the DMF/ACN solution immediately after sampling to support the contact between the isocyanates and the derivatizing reagent and, thus, to prevent the loss of isocyanates owing to side reactions. One additional hour of continuous lateral agitation (Eberbach Corporation, Ann Arbor, MI) was used in the laboratory to complete the desorption of the MFs.
Instrumental analysis
Tandem mass spectrometry (MS/MS) detection was chosen over ultraviolet, electrochemical or fluorescence detection for its selectivity to reduce the risks of interferences between the MOPIP derivatives of the isocyanates and other compounds generated during thermal degradation of car paint. A liquid chromatographic system (LC) series 1100 from Agilent Technologies, Inc. (Palo Alto, CA) equipped with an ion trap mass analyzer (MSD Trap) series VL, also from Agilent Technologies Inc., was used in the electrospray (ESI) mode monitoring positive ions. The compounds analyzed were the MOPIP derivatives of: HNCO, all the aliphatic isocyanates ranging from Me-i to Hex-i, all the alkenyl isocyanates ranging from Propylene-i to Hexylene-i, isophorone diisocyanate (IPDI), HDI, 2,4- and 2,6-TDI, MDI, phenyl isocyanate (Ph-i) and p-toluene isocyanate (p-Tol-i). For the MOPIP derivatives of the monoisocyanates, the parent ion was the molecular ion (M+H)+ whereas for the MOPIP derivatives of the diisocyanates it was the doubly-charged ion (M+2H)2+. In all cases, the daughter ion monitored was the MOPIP fragment at m/z = 193. According to the absence of pure standards for the MOPIP derivatives of the alkenyl isocyanates, it was assumed that response factors of these compounds were similar to those of the corresponding aliphatic isocyanate derivatives. Furthermore, similar response factors were also assumed for the MOPIP derivatives of the isomers of MDI (4,4'-; 2,4'-; and 2,2'-) and IPDI (cis- and trans-).
It was not possible to differentiate the daughter ions generated at m/z = 193 on the basis of their parent ions. Thus, it was necessary to separate all the MOPIP derivatives of the isocyanates using LC prior to mass spectrometry. Two LC/MSD Trap methods (I and II) were needed to separate and analyze the MOPIP derivatives of the 18 isocyanates monitored during this study. These methods are detailed in Table 1.
|
The calibration was performed using four solutions containing, respectively, 0.0050, 0.025, 0.10 and 1.0 µg of NCO per ml of the MOPIP derivative of each isocyanate analyzed, and the instrumental detection limits (LOD), defined as three times the standard deviation, were evaluated at the lowest point of the calibration curves (n = 8).
Chemicals
HPLC-grade toluene and water were from Merck (Darmstadt, Germany), HPLC-grade ACN was from Fisher Chemicals (Fair Lawn, NJ), and HPLC-grade dichloromethane and reagent-grade DMF were from J.T. Baker (Phillipsburg, NJ). Formic acid (96%), 2,4-TDI (98%), 2,6-TDI (97%), IPDI (mixture of cis- and trans-isomers, 98%), 4,4'-MDI (98%), Ph-i (98%) and p-Tol-i (99%) were from Aldrich (Milwaukee, WI).
The standard MOPIP derivatives of HNCO, HDI and Me-i to Hex-i were prepared in our laboratory as described earlier (Boutin et al., 2004). To synthesize the MOPIP derivatives of IPDI; 2,4- and 2,6-TDI; 4,4'-MDI; Ph-i and p-Tol-i, 4.5 mmol of MOPIP were dissolved in 10 ml of toluene, and 2.25 mmol of diisocyanate (IPDI; 2,4-TDI; 2,6-TDI or 4,4'-MDI) or 4.5 mmol of monoisocyanate (Ph-i or p-Tol-i) were then added. After 1 h of continuous stirring, the reaction mixture was put in the refrigerator overnight to optimize precipitation. The solid obtained was then filtered on a Bushner funnel and washed with cold toluene. Finally, recrystallization was performed in boiling toluene, except for the MOPIP derivative of 4,4'-MDI (4,4'-MDI-MOPIP) where dichloromethane was used. For the calibration curves, accurately weighed amounts of each urea derivative were dissolved in DMF and further diluted in ACN.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTALRESULTS AND DISCUSSIONCONCLUSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Figures 3 and 4 show examples of chromatogram obtained for samples collected at the emission source during cutting operations with the two LC/MSD Trap methods (I and II) (see Table 1 for more details). As expected, all the aliphatic and the alkenyl isocyanates previously observed by Boutin et al. (2004)
during the thermal degradation of car paint in the laboratory were also detected in the workplaces visited during this study. This is apparently the first report of alkenyl isocyanates becoming airborne in the workplace. However, these isocyanates were only observed close to the source and were not detected in any personal breathing zone samples.
|
|
The isocyanate concentrations measured during cutting and grinding operations in the three workplaces investigated and the corresponding limits of detection are shown in Tables 3 and 4. All the concentrations lower than half the LOD were considered as not detected (nd). The large relative standard deviations (RSDs) observed for the isocyanate concentrations could be explained by many factors that were out of our control. Such factors that could have affected the emission of isocyanates were, for instance, the pressure applied by the workers on the abrasive devices and the air-draught patterns that differed among workplaces.
HNCO was analyzed even though it is not an isocyanate because its emission during thermal degradation of car paint seems to be related to the emission of isocyanates and because there is very little information available about the toxicity of this compound. It is suggested that HNCO at high temperature will hydrolyze in air to form CO2 and NH3 (Aigner et al., 1995). Nevertheless, a better understanding of the behavior of HNCO following degradation of polyurethanes could be useful, given the observed concentrations of this compound (Tables 2 and 3).
|
|
The highest isocyanate concentrations were observed during the cutting process (Table 2). In spite of the high isocyanate concentrations observed at the emission source, the HSE's WEL (20 µg of NCO per m3) was never observed in the workers' breathing zone where the total isocyanate concentrations ranged between 5 and 49% of the WEL. Moreover, in real work situations, the studied processes are rarely performed continuously during periods longer than 15 min as observed in the present study. Owing to the low isocyanate concentrations measured 2 m from the emission source and the high RSD observed, the relative performance of the cassettes and the bubblers backed with cassettes could not be rigorously evaluated. Higher isocyanate concentrations would thus have been necessary to validate the sampling efficiency of the cassettes in the workplace. Even though they were observed at the emission source, Prop-i to Hex-i, the alkenyl isocyanates, Ph-i, p-Tol-i and 2,6-TDI were not detected in the workers' breathing zones. Surprisingly, high concentrations of IPDI were measured at the emission source during the cutting operations in the body repair shop, but not in the two painting schools. This could be the result of different primer, paint and/or clear-coat formulations. During the grinding process (Table 3), the total isocyanate concentrations measured in the workers' breathing zone ranged between 3 and 18% of the WEL and were thus lower then those detected for cutting. Finally, during the sanding (polishing) process, no isocyanate was detected in the workers' breathing zone. However, the total isocyanate concentrations, at the emission source, were, respectively 1.29, 0, 0.90 µg of NCO per m3 for School A, School B and the body repair shop.
|
CONCLUSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTALRESULTS AND DISCUSSIONCONCLUSIONACKNOWLEDGEMENTSREFERENCES |
|---|
The most abundant isocyanates generated during the thermal degradation of polyurethane-based car paints with the abrasive processes used in the present study were the monomers (MDI, HDI, TDI and IPDI) and Me-i. The workers' exposures to isocyanates were maximal during the cutting process and minimal during the orbital sanding process. The isocyanate concentrations rapidly decreased from the emission source to the sampling tree, which was located 2 m away. The use of a posture that maximizes the distance between the worker and the emission source could thus significantly decrease isocyanate exposures. The highest isocyanate concentration measured in the workers' breathing zone corresponded to approximately half the HSE's WEL. Based on current WEL, the risk to worker's health could be interpreted as being low for the exposures observed in our tests, except for workers already sensitized to isocyanates. However, a more accurate risk assessment would require dose–responses curves for each isocyanate, since the toxicity differs among isocyanates. Moreover, peak concentrations of isocyanates, which might contribute to the development of asthma, appear to exist. The use of a surrogate measure of isocyanate concentrations, such as ultrafine aerosol concentrations, could also be useful for estimating real-time air concentrations of isocyanates (HSE, 2002
). |
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTALRESULTS AND DISCUSSIONCONCLUSIONACKNOWLEDGEMENTSREFERENCES |
|---|
This study was supported by a grant from the Institut de recherche Robert-Sauvé en santé et en sécurité du travail (IRSST) (Montreal, Quebec, Canada). The authors also wish to thank the Centre de formation professionnelle de Verdun (Verdun, Quebec, Canada), the Centre de formation professionnelle Qualitech (Trois-Rivieres, Quebec, Canada) and Pièces d'auto M. Robert Inc. (Ste-Madeleine, Quebec, Canada) for having allowed environmental monitoring inside their establishments. Finally, we especially thank Mrs Martine Charette (Auto Prévention, Montreal, Quebec, Canada) and Mrs Lucie René (IRSST) for their contribution to the sampling in the workplaces.
Received September 19, 2005; in final form November 9, 2005
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTALRESULTS AND DISCUSSIONCONCLUSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Aigner M, Zeilinger M, Hofbaur H. (1995) Kinetic study of the hydrolysis of isocyanic acid in the gas phase. Chem Eng Proc; 34: 515–20.[CrossRef]
Boutin M, Lesage J, Ostiguy C et al. (2004) Identification of the isocyanates generated during the thermal degradation of a polyurethane car paint. J Anal Appl Pyrolysis; 71: 791–802.[CrossRef]
Boutin M, Lesage J, Ostiguy C et al. (2005) Validation of a solvent-free sampler for the determination of low molecular weight aliphatic isocyanates under thermal degradation conditions. J Occup Env Hyg; 2: 456–61.[CrossRef]
Brenner KS, Dharmarajan V, Maddison P. (2003) Sampling and analysis. In Allport DC, Gilbert DS, Outterside SM, editors. MDI & TDI safety, health and the environment. Chichester, England: John Wiley & Sons. pp. 358–430.
Henriks-Eckerman M-J, Välimaa J, Rosenberg C et al. (2002) Exposure to airborne isocyanates and other thermal degradation products at polyurethane-processing workplaces. J Environ Monit; 4: 717–21.[CrossRef][ISI][Medline]
HSE. (2002) The use of a direct reading instrument to obtain time-resolved exposure data for isocyanates. HSE: Toxic Substances Bulletin—Issue 48, May 2002.
HSE. (2005) Workplace exposure limits: containing the list of workplace exposure limits for use with the Control of Substances Hazardous to Health Regulations 2002 (as amended), EH40/2005. Health and Safety Commission, HSE Books 2005. ISBN 0717629775.
Karlsson D, Dahlin J, Skarping G et al. (2002) Determination of isocyanates, aminoisocyanates and amines in air formed during thermal degradation of polyurethane. J Environ Monit; 4: 216–22.[CrossRef][ISI][Medline]
Karlsson D, Spanne M, Dalene M et al. (2000) Airborne thermal degradation products of polyurethane coatings in car repair shops. J Environ Monit; 2: 462–69.[CrossRef][ISI][Medline]
Mapp CE, Butcher BT, Fabbri LM. (1999) Polyisocyanates and their prepolymers. In Bernstein IL, Chan-Yeung M, Malo J-L et al., editors. Asthma in the workplace. 2nd edn. New York: Marcel Dekker. pp. 457–78.
Pauluhn J. (1989) A mechanistic approach to assess the inhalation toxicity and hazard of methylisocyanate and related aliphatic monoisocyanates. In Mohr U, Bates DV, Dungworth DL et al., editors. Assessment of inhalation hazards. New York: Springer-Verlag. pp. 119–28.
Pronk A, Tielemans E, Skarping G et al. (2005) Inhalation exposure to isocyanates to car body repair shop workers and industrial spray painters. Ann Occup Hyg; (Advanced Access published August 26, 2005).
Raulf-Heimsoth M, Baur X. (1998) Pathomechanisms and pathophysiology of isocyanate-induced diseases-summary of present knowledge. Am J Ind Med; 34: 137–43.[CrossRef][ISI][Medline]
Streicher RP, Kennedy ER, Lorberau CD. (1994) Strategies for the simultaneous collection of vapours and aerosols with emphasis on isocyanate sampling. Analyst; 119: 89–97.[CrossRef][Medline]
Tinnerberg H, Spanne M, Dalene M et al. (1997) Determination of complex mixtures of airborne isocyanates and amines. Part 3. Methylenediphenyl diisocyanate, methylenediphenylamino isocyanate and methylenediphenyldiamine and structural analogues after thermal degradation of polyurethane. Analyst; 122: 275–78.[CrossRef][Medline]
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]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||