Occupational exposure to fluorinated hydrocarbons during

refrigeration repair work

Merete Gjølstad,*a Dag G. Ellingsen,a Oscar Espeland,b Karl-Christian Nordby,a

Harald Evenseth,b Syvert Thorud,a Nils Petter Skaugseta and Yngvar Thomassena

aNational Institute of Occupational Health, P.O. Box 8149 Dep., N-0033 Oslo, Norway.E-mail: Merete.Gjolstad@stami.no; Fax: 47 23195206; Tel: 47 23195335

bTelemark Hospital, Department of Occupational and Environmental Medicine, N-3710Skien, Norway

Received 3rd January 2003, Accepted 20th February 2003

First published as an Advance Article on the web 28th February 2003

This study describes refrigeration repair workers’ occupational exposures to halogenated refrigerants, focusing

on difluorochloromethane (HCFC 22), tetrafluoroethane (HFC 134a) and a mixture of tri-, tetra- and

pentafluoroethane (R404A) in 30 work operations. Unlike earlier reported studies, the present study includes

working procedures involving welding in order to measure possible occupational exposure to decomposition

products. The measurements included hydrogen fluoride (HF), hydrogen chloride (HCl), phosgene (COCl2) and

volatile organic compounds (VOC). The exposures were assessed during work operations on small-scale cooling

installations like refrigerators and freezers. The repair workers’ occupational exposures to refrigerants were

moderate, and the major part of the exposures were associated with specific working procedures lasting for

relatively short periods of time (v20 min). During these exposure events the concentrations were occasionally

high (up to 42434 mg m23). Although welding operations lasted only for short periods of time, HF was

detected in 9 out of 15 samples when HCFC 22, HFC 134a or R404A had been used. Hydrogen chloride was

detected in 3 out of 5 samples in air polluted with HCFC 22. Phosgene was not detected. A large number of

VOCs in various concentrations were found during welding. Except for the applied refrigerants, halogenated

compounds were only found in one sample.

Introduction

Chlorinated and fluorinated hydrocarbons have been used asrefrigerants since the 1930’s.1 For the last fifty years or so theyhave been extensively used as refrigerants in smaller coolerinstallations like refrigerators, freezers and air conditioners.The first generation of the short-chained refrigerants wascompletely halogenated. Due to their harmful effect to theenvironment, import and use of these compounds are nowrestricted in many countries.2

Today, the partially halogenated methane and ethane deriva-tives are the commonly used refrigerants.3,4 These chlori-nated compounds may, however, gradually be phased out andreplaced by products considered to be less harmful to theenvironment.

The need for protective actions against occupational expo-sure to refrigerants depends on the toxicity of individualcompounds and the levels of exposure. Several studies haveinvestigated their toxic effects.5–7 At high concentrations, refri-gerants can induce suffocation due to displacement of air. Bothchlorinated fluorocarbons and the more recent fluorinatedhydrocarbons are assumed to induce heart arrhytmias and

heart arrest at concentration levels above 2500–15000 ppm.8,9

There are no reports to support the assumption of carcinogeniceffects of halogenated hydrocarbons used in refrigeration.4,10

The toxic effects of the refrigerants under study are sum-marized in Table 1.

The halogenated hydrocarbons are colourless, almostodourless and non-flammable. They have very low boilingpoints and occur either as gases or fluids, and their physicalproperties make them highly suitable as refrigerants.2–4 Coolersare normally operating as closed systems, but in connectionwith maintenance and repair work the installations are opened,and the repair workers may thus be exposed to refrigerants.

There are few studies on occupational exposure amongrefrigeration repair workers. The documented knowledge isinsufficient, both with respect to levels of exposure, the patternof exposure and exposure to potential degradation productsdue to welding. Permonius11 measured dichlorodifluoromethane(CFC 12), difluorochloromethane (HCFC 22) and R502 (amixture of 48.8% HCFC 22 and 51.2% chloropentafluoro-ethane (CFC 115)), and observed high short time exposure.Andersson12 performed measurements of CFC 12 on 7occasions and concluded that although peak concentrations

Table 1 Toxic effects of the refrigerants HCFC 22, HFC 134a, HFC 143a and HFC 125

Toxic effect HCFC 22 HFC 134a HFC 143a HFC 125

Cardiac senzitationa (dogs) Yes Yesa Yes YesCNS depression (rats) — Yesb — —Carcinogenicity No No No NoHepatitis-like allergic response No Yesc No Yesc

aAt concentrations above 75000 ppm (313000 mg m23), CFC compounds may lead to arrhytmias, especially when stress is applied simulta-neously. Heart arrest deaths in humans after inhalation of CFC has been encountered. bAt concentrations above 40000 ppm (167000 mg m23).cDue to the formation of trifluoroacetic acid protein adducts in hepatocytes. — no study found.

236 J. Environ. Monit., 2003, 5, 236–240 DOI: 10.1039/b212612e

This journal is # The Royal Society of Chemistry 2003

Publ

ishe

d on

28

Febr

uary

200

3. D

ownl

oade

d by

Geo

rge

Was

hing

ton

Uni

vers

ity o

n 26

/10/

2014

17:

29:2

2.

View Article Online / Journal Homepage / Table of Contents for this issue

up to 1200 ppm were measured, the over day exposure wasgenerally low. In another study, Johansson13 refered to peakconcentrations above 10000 ppm during 15 min workingperiods, while the 8 h exposure in general was below therecommended national occupational exposure limit values(OEL’s). Ljungkvist et al.14 presented similar results. Only oneof the above studies used a direct reading infrared spectro-photometer (MIRAN 1A).11 There are some indications thathigh short time exposure levels of refrigerants duringrefrigeration repair work are associated with a few specifictasks.11 Although welding is a routine part of refrigerationrepair work, none of these investigations reported measure-ments of possible thermal decomposition products from therefrigerants.

This study focuses on exposure to some of the mostfrequently used halogenated refrigerants in Norway, includingthe second generation HCFC 22, the non-chlorinated thirdgeneration compounds tetrafluoroethane (HFC 134a) and amixture of 52% tri-, 4% tetra- and 44% pentafluoroethane(R404A). The study was also designed to assess the exposure topossible thermal decomposition products of the refrigerants, asit is likely that such compounds may degrade by increasingtemperature. Dekant15 has shown that trifluoroacetic acid(TFA) is the main urinary metabolite in rodents when exposedto fluorinated refrigerants, and measurement of TFA in urinewas also included in the study.

Materials and methods

Occupational exposure was measured during 30 maintenance/repair occasions. Changing the oil on compressors was a part ofthe job on 22 occasions. Seventeen work procedures includedoperations such as replacements of filters, vibration absorbersand compressors. On four occasions the refrigeration unit itselfwas partly or completely removed. Fifteen of the 30 job eventsinvolved welding. Twenty-two male refrigeration repair work-ers with mean age 37.7 years (range 24–64; median 34.5),participated in the study. They were all service engineersemployed by cooler-unit manufacturers to maintain/serviceunits in various commercial locations. On average, they hadbeen working as refrigeration repair men for 13 years (range0.5–49; median 8). An informed written consent to participatein the study was obtained from all subjects.

Sampling and measurements of refrigerants in workroom air

Direct reading measurements. The measurements of refrig-erants in the workroom air was performed using a directreading photoacoustic infrared gas analyser, (Model 1302,Bruel & Kjær, Nærum, Denmark).

The instrument was calibrated beforehand using industrialrefrigerants from gas cylinders in order to measure HCFC 22,HFC 134a, pentafluoroethane (HFC 125) and trifluoroethane(HFC 143a). Correction was applied during measuring ofR404A in order to compensate for any interference from eachgas on the other filters mounted. The calibration of theinstrument was accepted when the variation of 4–6 readingswas less than ¡0.5% RSD. The filter configuration is shown inTable 2.

During the measurements the sampling tube of the instrument

was kept as close as possible to the breathing zone of theworkers. The length of the sampling tube was 10 or 20 m. Eachmeasurement lasted 30 to 110 s depending on the complexity ofthe calculation matrix and the length of the sampling tube. Thetotal measuring time was equal to the working periods of therepairmen.

Sorbent tube measurements. The concentrations of refriger-ants in the workroom air were also measured using samplingtubes with Carboxen 1000 as sorbent (Supelco, Bellefonte,Pennsylvania, USA). The personal pumps were Casella SP2(Casella London Ltd., Bedford, UK), SKC series 224 or SKCPocket Pump (SKC Ltd., Dorset, UK) operated at a flow rateof 50 ml min21 ¡ 5%. The flow was measured by means of amass flowmeter, Top-Trak, Model 822-2, with range 0 to200 ml min21 (Sierra instruments Inc., Monterey, California,USA). The sampling tubes were placed at the collar of theworkers close to the breathing zone. The tubes were exposedparallel to the direct reading instrument. The sorbent tubeswere stored at 220 uC prior to desorption.

The tubes were desorbed with 3.0 ml carbon disulfide (CS2)for 2 h followed by capillary gas chromatographic determina-tion. A HP 5890 Series 2 gas chromatograph (HP 5890 Series 2,Hewlett Packard Company, California, USA) equipped with aflame ionisation detector was used. The column was a CPPoraplot Q-HT (Chrompack, Middelburg, The Netherlands).Calibration curves were obtained by spiking the refrigerants onCarboxen 1000 sorbent tubes. The spiked tubes were desorbedand analysed similarly to the exposed samples 1 to 2 h afterpreparation. Assuming an air sampling volume equal to 1 l, thedetection limit of the method was 5 mg m23 for HFC 125 and4 mg m23 for the other refrigerants, respectively.

Sampling of volatile organic compounds (VOC) from thermaldecomposition. Sampling of VOC from thermal decompositionof refrigerants or compressor oils was performed duringwelding only, and on 15 welding operations using pumpedsampling with a flow rate of 50 ml min21 ¡ 5% measured bymeans of a mass flowmeter, Top-Trak, Model 822-2. Stainlesssteel or glass PerkinElmer thermal desorption tubes (ATD-tubes) (PerkinElmer Ltd, Buckinghamshire, UK) were used.The sorbents used were Tenax TA (60/80 mesh) (Supelco) infront and Carbosieve (35/60 mesh) (Supelco) or Spherocarb(60/80 mesh) (Phase Separation Ltd., Queensferry, UK) asbackup. Tenax TA is known to be a good adsorbent forcompounds with boiling points 100–400 uC, while Carbosieveand Spherocarb are stronger adsorbents for very volatilecompounds.16 During measurements, the sampling tubes werekept as close as possible to the breathing zone of the workers.

The ATD-tubes were analysed with a PerkinElmer ATD 400thermal desorption unit connected to a Fisons MD800 gaschromatograph with mass spectrometric (MS) detector withelectron impact ionisation (Fisons Instruments, Milan, Italy).The Tenax TA tubes were desorbed at 250 uC and theCarbosieve and Spherocarb tubes at 350 uC. The GC columnwas a BPX-5, 30 m, 0.25 mm id, 0.25 mm film (SGE Inc.,Austin, Texas, USA). The MS was run in full scan mode. Usinga toluene standard, the concentrations of VOC were calculatedas toluene equivalents.

Sampling and measurement of hydrogen fluoride (HF) andhydrogen chloride (HCl). Hydrogen fluoride was collectedduring 15 welding operations whereas HCl was collectedduring 5 welding operations with HCFC 22 as refrigerant. Thesampling was performed using 37 mm filter cassettes (MilliporeCorp., Bedford, MA, USA) equipped with 0.8 mm aerosolcellulose membrane filters and cellulose pads impregnated with10% potassium hydroxide. The pumps used were in-house unitsconstructed at The National Institute of Occupational Healthoperated at a constant flow rate of 2.0 l min21. The flow was

Table 2 Filter specifications for Bruel & Kjær Model 1302 photo-acoustic IR gas analyser

RefrigerantFilterposition Filter no.

Wavelength/mm

Difluorochloromethane A UA0973 9.1Tetrafluoroethane A UA0973 9.1Pentafluoroethane B UA0972 8.8Trifluoroethane C UA0969 8.0

J. Environ. Monit., 2003, 5, 236–240 237

Publ

ishe

d on

28

Febr

uary

200

3. D

ownl

oade

d by

Geo

rge

Was

hing

ton

Uni

vers

ity o

n 26

/10/

2014

17:

29:2

2.

View Article Online

measured by means of an in-house calibrated rotameter, ModelV 100–80.12 (Vogtlin, Basel, Switzerland). The filters weremounted on the workers collar close to the breathing zone.Fluoride was measured with an ion selective F2 electrode, andchloride with ion chromatography (Dionex DX-500, DionexCorp., Sunnyvale, California, USA) equipped with an AS9 HCcolumn (Dionex). The detection limit of fluoride was 2 mg F2

per filter, and of chloride 130 mg Cl2 per filter.

Measurement of phosgene (COCl2). As phosgene is apossible thermal decomposition product from chlorinatedaliphatic hydrocarbons, phosgene was measured during weld-ing in the presence of HCFC 22 (n ~ 5). The measurementswere performed with Drager pump (20 strokes) (DragerwerkAG, Lubeck, Germany) and indicator tubes (Drager 8101521Phosgen 0.02/a). The detection limit for the method is 0.02 ppm.

Sampling and measurement of TFA in urine. Three urinesamples were collected from each person. The first sample wasvoided in the evening of the day of exposure, the second one thenext morning, and the third one on the morning after that. Allsamples were stored at 220 uC until analysis. TFA wasmeasured in urine from 9 workers with the highest cumulativeexposure to HFC 134a or R404A by ion chromatography. ADionex DX-500 instrument equipped with an electrochemicaldetector (Dionex ED 40) operating in conductive mode and aDionex AS-11 column and Dionex AG11 pre-column wasused. The detection limit of the method was calculated as threetimes the standard deviation when integrating the baselinefrom analysis of 19 samples without detectable amounts ofTFA. The established detection limit was 6.4 mg TFA ml21

urine. Calibration curve was obtained by spiking SeronormTM

Trace Elements, Urine Blank, lot FE 1113, (Sero, Asker,Norway) with TFA.

Reagents and chemicals

Carbon disulfide (with low benzene content) was purchasedfrom Rathburn Chemicals Ltd., (Walkerburn, Scotland) andtoluene (y99.5%) and potassium hydroxide were purchasedfrom Merck, (Darmstadt, Germany). The refrigerants weresupplied by Du Pont de Nemours International S. A. (Geneva,Switzerland).

Statistics

Several of the variables had skewed distributions. Thesevariables were log-transformed if the skewness exceeded 2.0.The geometric mean concentrations were calculated for thesevariables. Generally the measured concentrations of workroomcontaminants had skewed distributions. The statistical packageSPSS1, version 8.0, was used on a personal computer.

Results

Table 3 shows the concentrations of refrigerants in theworkroom air collected with sampling tubes. The duration ofthe exposure periods varied considerably (6–390 min.), and thecorresponding air concentrations measured within theseperiods varied from 1 mg m23 HFC 143a (one of thecomponents in R404A) to 2171 mg m23 HCFC 22. Altogether,the air concentrations were low, with the average concentrationof HCFC 22 being the highest (geometric mean 235 mg m23).The air concentrations of HFC 134a or R404A were evenlower; the geometric mean concentration of HFC 134a was85.6 mg m23. For R404A the mean concentrations arepresented for each component in the mixture with geometricmeans of 42.7 mg m23 (HFC 134a), 53.0 mg m23 (HFC 125)and 45.0 mg m23 (HFC 143a), respectively.

Although the cumulative air concentrations were generallylow, the direct reading photoacoustic IR analyser showed thatshort periods of higher concentrations were quite prevalent.The concentrations were arbitrarily defined to be ‘‘high’’ if theconcentration measured during a specific period was at least 3times higher than the concentration level previously measured.This implies that such periods of peak exposure may representquite low concentrations if the exposure in general is low.The average duration of these peaks was 5 to 8 min (range1–21 min), depending on the compound (Table 4). On average,there were 2.7 ‘‘high’’ concentration periods when HCFC 22was the refrigerant, approximately 2.8 periods with HFC 134aand approximately 3.6 periods with R404A (not tabulated).

The maximum concentrations measured during high peakexposure periods are shown in Table 4. The geometric meanpeak concentration was 1627 mg m23 HCFC 22, and 25% ofthe peaks were higher than 10308 mg m23, while the highestmeasured concentration was 42434 mg m23. For the remainingrefrigerants the peak concentrations were generally lower(Table 4). Periods with high exposure were associated with few

Table 3 Concentrations of refrigerants in work room air. Personal sampling by sorbent tubes

Refrigerant

Sampling time/min Concentration/mg m23

n Arithmetic mean Range Geometric mean Range

Difluorochloromethane 12 98 6–390 235 10.6–2171Tetrafluoroethane 8 54 20–120 85.6 12.5–442R404A 117 20–210 — —Pentafluoroethane 10 — — 53.0 4.9–182Tetrafluoroethane 10 — – 42.7 1.3–513Trifluoroethane 10 — — 45.0 1.0–210

Table 4 Peak concentrations measured during high exposure periods to HCFC 22, HFC 134a or R404A at 30 exposure incidents

Refrigerant

Sampling time/min Peak concentration/mg m23

n AMa Range GMb Minimum Maximum 25 percentile 50 percentile 75 percentile

Difluorochloromethane 32 4.9 1–21 1627 39 42434 262 1450 10308Tetrafluoroethane 22 5.0 1–11 1285 133 5842 522 1669 3539R404APentafluoroethane 36 8.3 3–18 771 64 5891 403 888 1468Tetrafluoroethane 31 5.7 1–16 46 6.3 1448 21 46 75Trifluoroethane 36 8.3 3–18 646 38 4950 368 746 1299aAM ~ arithmetic mean. bGM ~ geometric mean.

238 J. Environ. Monit., 2003, 5, 236–240

Publ

ishe

d on

28

Febr

uary

200

3. D

ownl

oade

d by

Geo

rge

Was

hing

ton

Uni

vers

ity o

n 26

/10/

2014

17:

29:2

2.

View Article Online

specific tasks such as evacuating the coolers and draining andrefilling of the compressor oil. The air concentrations duringthe periods between such work operations were considerablylower, and some times even hardly detectable. Typical exposurepatterns for the three refrigerants under study are illustrated inFigs. 1–3.

Table 5 shows the concentrations of HF and HCl measuredduring welding. Welding was not a continuous operation, andthe tabulated welding sampling times are from the start of the

first welding to the end of the last episode. Each weldingepisode lasted, on average, for 11.5 min (range 4–24 min)(not tabulated). Even though the welding periods were short,HF was detected in 9 out of 15 collected samples when HCFC22, HFC 134a or R404A were the refrigerants in theinstallation where welding was carried out. Hydrogen chloridewas detected in 3 out of 5 collected samples in air polluted withHCFC 22. The concentration range for HF and for HCl was0.27–0.65 mg m23 and 0.7–11.5 mg m23 respectively, in thesamples where these gases were detected. Phosgene was notdetected in connection with welding in the presence of HCFC22 in the atmosphere (not tabulated).

Air samples collected during welding in order to identifydecomposition products showed a large number of vola-tile organic compounds in concentrations between v5 and1400 mg m23 (not tabulated). Roughly, the componentsdetected can be divided into three main groups: halogenatedcompounds (refrigerants and some chlorinated solvents);aliphatic hydrocarbons (alkanes/alkenes); and aromatic hydro-carbons (toluene, xylenes). In one sample small amounts of3-(chloromethyl)heptane was detected (10 mg m23), and aceticacid (99 and 143 mg m23), phthalic anhydride (2 and 6 mg m23)and benzothiazole (2 and 6 mg m23) were each detected in twosamples.

TFA in urine was only found in one out of 27 samples andin a concentration (7.1 mg ml21) close to the detection limit(6.4 mg ml21).

Discussion

This study shows that the occupational exposure to refrigerantsamong the refrigeration repair workers is moderate. A mainwork environment characteristic is that the major part of therepair workers’ cumulative occupational exposure to refriger-ants is associated with specific tasks lasting for relatively shortperiods of time, up to approximately 20 min. During theseperiods the exposure can be considerable.

Other studies11–13 have suggested that peak exposures areassociated with specific tasks such as draining used oil from thecompressors. The used compressor oil contains residues ofrefrigerants, making handling of the oil a major contributingfactor to the refrigerant exposure. In the present study, the usedoil was often drained out into open cans which to some extent,resulted in spilling. This work operation can probably beperformed in a ‘‘closed loop’’, thereby reducing the emission ofrefrigerants to the workroom air. This becomes even moreimportant since the compressor rooms usually are small andinadequately ventilated.

The study could not verify the occurrence of organicfluorocontaining thermal decomposition products in theatmosphere, originating from the applied refrigerants. Severalof the hydrocarbons identified are normally present in indoorair and are likely to be indicators for a general backgroundlevel more than decomposition products of refrigerants. Someof the organic compounds detected may, however, originatefrom thermal degradation of other products present in theworkroom, such as pipeline insulation materials. Whenconsidering the fact that no organic decomposition productsfrom refrigerants were found with certainty in our study, itshould be pointed out that sampling was only carried out

Fig. 1 Concentration of HCFC 22 in workroom air measured with aphotoaccustic IR gas analyser during changing of the compressor oil.

Fig. 2 Concentration of HFC134a in workroom air measured with aphotoaccustic IR gas analyser during changing of the compressor oil.

Fig. 3 Concentration of R404A in workroom air measured with aphotoaccustic IR gas analyser during changing of the compressor oil onthree compressors.

Table 5 Concentrations of HF and HCl measured during welding. Only measurements above the detection limit are considered for the calculations

Refrigerant

Sampling time/min HF/mg m23 HCl/mg m23

n Arithmetic mean Range Arithmetic mean Range Arithmetic mean Range

Difluorochloromethane 5 38.0 9–114 0.27 0.03–0.66 5.0 0.7–11.5Tetrafluoroethane 5 61.2 15–118 0.27 0.11–0.39 Not measured Not measuredR 404A 5 57.2 17–157 0.65 0.08–1.59 Not measured Not measured

J. Environ. Monit., 2003, 5, 236–240 239

Publ

ishe

d on

28

Febr

uary

200

3. D

ownl

oade

d by

Geo

rge

Was

hing

ton

Uni

vers

ity o

n 26

/10/

2014

17:

29:2

2.

View Article Online

during the short welding periods, and the concentrations ofrefrigerants were low. This implicates that the amount ofdecomposition products probably would be small and conse-quently difficult to detect. The heavier aliphatic hydrocarbonsthat were found, may originate from mineral oils that are usedas compressor oils, and the small amounts of 3-(chloromethyl)-heptane detected in one sample could be a reaction productformed from decomposition products from the refrigerants andthe compressor oil.

Furthermore, the detection of HF and HCl during welding,clearly indicates that degradation of the refrigerants duringwelding in fact does occur. This suggests that even organicdegradation products may be present. Neither the samplingprocedure nor the analytical methodology used to determineF2 and Cl2 in air are specific to HF and HCl, but we are stillconfident that the main chemical form of the collected F2 andCl2 was HF and HCl as decomposition products.17 Fluorideswere detected in 9 out of 15 samples, chlorides in 3 out of 5samples. We are not familiar with other studies reportingsimilar measurements of degradation products in the workenvironment of refrigeration repair workers. The fact thatphosgene was not detected in our study does not exclude thatphosgene may occur in the work atmosphere, but rather thatthe welding periods were too short or the concentrations of therefrigerants were too low for the creation of phosgeneconcentrations above the detection limit of the applied method.

TFA in urine was only detected in one sample out of 27collected from 9 workers with the highest cumulative exposureto HFC 134a or R404A, at a concentration slightly above thedetection limit. This indicates that the exposure to refrigerantsat these exposure levels results in a marginal urine excretion ofthe TFA. This suggests that TFA in urine is not a relevantindicator for the occupational exposure to fluorinated refrig-erants at moderate to low exposure levels.

Conclusions

This study has shown that refrigeration repair workers’occupational exposure to refrigerants is moderate, but withtypically short time high peak exposure work procedures.Welding is frequently performed, but only during short workperiods. Even so, thermal decomposition products such as HFand HCl were detected in major numbers of the samples.

Acknowledgements

The authors acknowledge Receiving Institute for Gas (SRGNorway) for financially supporting this study and therefrigeration repair workers who voluntarily participated inthe study.

References

1 WHO: International Programme on Chemical Safety (ICPS), FullyHalogenated ChloroFluorocarbons, Environmental Health Cri-teria, EHC No 113, WHO, Geneva, 1990.

2 United Nations Environment Programme. The Montreal Protocolon Substances that Deplete the Ozone Layer, Montreal, 1987.

3 WHO: International Programme on Chemical Safety (ICPS),Partially Halogenated ChloroFluorocarbons (Methane Deriva-tives), Environmental Health Criteria; EHC No 126, WHO,Geneva, 1991.

4 WHO: International Programme on Chemical Safety (ICPS),Partially Halogenated ChloroFluorocarbons (Ethane Derivatives),Environmental Health Criteria; EHC No 139, WHO, Geneva,1992.

5 W. J. Brock, H. J. Trochimowicz, C. H. Farr, R.-J. Millischer andG. M. Risch, Fundam. Appl. Toxicol., 1996, 31, 200.

6 D. G. Clark and D. J. Tinston, Hum. Toxicol., 1982, 1, 239.7 F. Leuschner, B. M. Neumann and F. Huebschner, Arzneim.-

Forsch., 1983, 33, 1475.8 C. Edling and P. Søderkvist, Nordic Criteria Document for

Occupational Exposure Limits: Fluorocarbons, Arbete och halsa,1982, p. 26, (in Swedish).

9 W. J. Brady, Jr, E. Stremski, L. Eljaiek and T. P. Aufderheide, Am.J. Emerg. Med. (USA), 1994, 12(5), 533.

10 Re-evaluation of some Organic Chemicals, Hydrazine and HydrogenPeroxide, IARC Monographs on the Evaluation of CarcinogenicRisks to Humans, 71, part 3, WHO, Lyon, 1999.

11 B. Permonius, Examination paper, Bygghalsan Safety and HealthCentre Corporation, Stockholm, 1980, (in Swedish).

12 K. Andersson, Examination paper, IVL Swedish EnvironmentalResearch Institute, Gothenburg, 1983, (in Swedish).

13 A. Johansson, Examination paper, The Swedish Telecommunica-tion Corporation, Stockholm, 1987, (in Swedish).

14 G. Ljungkvist, A. Oliv and M. Torgen, 37th Nordic Conference onOccupational Health, Gothenburg, August 29–31, 1988, NationalInstitute for Working Life, Solna, 1988, pp. 162–163.

15 W. Dekant, Environ Health Perspect., 1996, 104, 75.16 ISO/DIS, Indoor, Ambient and Indoor Air—Sampling and Analysis

of Volatile Organic Compounds by Sorbent Tube/Thermal Desorp-tion/Capillary Gas Chromatography—Part 1: Pumped Sampling,Draft International Standard ISO/DIS 16017-1, ISO, Berlin, 1999.

17 A. Kasper and H. Pauxbaum, Anal. Chim. Acta, 1994, 291, 297.

240 J. Environ. Monit., 2003, 5, 236–240

Publ

ishe

d on

28

Febr

uary

200

3. D

ownl

oade

d by

Geo

rge

Was

hing

ton

Uni

vers

ity o

n 26

/10/

2014

17:

29:2

2.

View Article Online