Assessment of occupational exposure to diesel fumes—parameter optimization of the thermal coulometric measurement method for carbon† Vincent Perret,* Cong K. Huynh, Pierre-Olivier Droz, Trinh Vu Duc and Michel Guillemin IST , Institute of Occupational Health Sciences, Rue du Bugnon 19, CH-1005 Lausanne, Switzerland Received 31st March 1999, Accepted 17th June 1999 ‘Elemental’ carbon (EC) is used as a surrogate to assess occupational exposure to diesel soot.EC thermal analysis needs complete desorption of organic compounds from the soot particles prior to analysis in order to minimize positive interferences and artefacts. The desorption of the organic compounds can be considered as the major step which influences the reliability of the EC determination. A systematic study was carried out to investigate the diVerent parameters of influence such as desorption temperature, desorption duration, heating rate and type of the sample on the desorption eYciency.It was found that temperature and duration are the major parameters of influence on the desorption eYciency. The influence of the sample load can be seen as a measure of the pyrolysis susceptibility of the sample.An optimized temperature program is proposed. EC, respectively). The organic fraction of the diesel particulate Introduction is first removed in helium, at ramps and steps going up to Diesel exhaust was classified in 1989 as probably carcinogenic about 850 °C. The evolved carbon is catalytically oxidized to to humans by the International Agency for Research on CO2, which is reduced (catalytically) to methane and quantified Cancer1 and is classified in category 2b because there is using a flame ionization detector (FID).The temperature is suYcient evidence for its carcinogenicity in animals and some lowered and an oxygen–helium mixture is introduced into the trends in epidemiological studies. Since that time, more results sample oven.The temperature is then raised to about 940 °C from additional epidemiological studies have confirmed that to remove the remaining carbon, which is again converted to Diesel exhaust represents a significant occupational and public methane for quantification. If the diesel particulate sample health problem.2,3 contains materials that pyrolyze (form char), a portion of the Although the gaseous phase of the exhaust of diesel engines carbon removed during the second part of the analysis is contains irritant gases such as oxides of nitrogen, sulfur dioxide assigned to the organic fraction.This correction for pyrolysis and aldehydes, leading to acute eVects on workers, the chronic is accomplished through continuous monitoring of the filter hazard related to lung cancer comes from the particulate phase transmittance.The correction assumes that a decrease in filter of the exhaust and is considered as a priority in the field of transmittance during the first part of the analysis (in helium) occupational health.1,4 Present toxicological knowledge sup- is caused by charring. To correct for the thermally deposited ports the use of the carbon core of the diesel particles as the char, the split between the organic and elemental carbon is most relevant index of exposure to assess the health risk of not made until enough is oxidatively removed to bring the exposed people.5–7 filter transmittance back to its initial value. The second method Taking into account that diesel particles are mainly com- was developed in Germany and is called the coulometric posed of a core of ‘elemental’ carbon (EC) coated by organic method.12 This is a simple thermal method were the ‘elemental’ compounds such as oil residues, polycyclic aromatic hydrocarbon is defined as the carbon not removed from the sample carbons and other derivatives, the measurement methods after a given thermal treatment under an inert gas.The oYcial developed up to now have been focused on the ‘elemental’ German method has defined a thermal treatment of 8 min at carbon, which is not pure carbon but something more complex 500 °C under nitrogen.After that, oxygen is introduced and with a high proportion of carbon.8,9 The analytical challenge the temperature is increased to 800 °C; the CO2 formed is here is diVerent from the usual one aimed at the analysis of a determined by coulometry in a solution of barium perchlorate.specific molecule. In our case, the aim is to be as close to the A round robin test has been carried out to compare these diesel carbon core as possible, knowing that this core is not a two methods.13 The tested samples were either filters spiked specific and well determined chemical and mineralogical with organic compounds such as sucrose and EDTA, field species.10,11 samples taken in diVerent environments or field samples spiked Two types of methods, based on the same basic principle, with cigarette smoke.The results showed that the simple have been developed. In both methods, diesel soot samples thermal method gave results biased towards higher levels of are collected on quartz or glass fiber filters according to usual ‘elemental’ carbon for the samples containing a high pro- occupational hygiene techniques. The first method (NIOSH portion of organic compounds compared with the thermal– analytical method No. 5040) is a thermal–optical technique optical method. The positive bias in the analysis of the OC for determination of organic and elemental carbon (OC and standards is attributed largely to inadequate removal of OC during the first part of the analysis.Lack of correction for pyrolytically formed carbon (char) is also a factor. †Presented at AIRMON ’99, Geilo, Norway, February 10–14, 1999. J. Environ. Monit., 1999, 1, 367–372 367By the end of 1998, there was still a limited availability of to the equation instruments on the market to perform NIOSH method 5040.Ba2++CO2+2 OH-�BaCO3+H2O On the other hand, the coulometer is an instrument which has been on the market for a long time and has been used, for The OH- ions consumed are electrolytically reformed accordinstance, for the determination of total organic compounds in ing to water. Its availability is still assured despite the bankruptcy of 2 H2O+2 e-�H2+2 OH- the manufacturer, and production is now continued by another representative.It was therefore decided to investigate the in the back-titration in parallel with this reaction. The carbon performance of the simple thermal method further by using content of the sample is determined from the number of the coulometer, looking at the main parameters of influence: coulombs required to restore the initial alkalinity (pH#9.7) desorption temperature and duration, sample loading, heating of the absorption solution.The integrated electrolysis current rate and type of samples (pure organic compounds, diesel soot is indicated by the coulometer in the form of counts, each and cigarette smoke). The objective of the study was to count corresponds to 0.2 mg of equivalent carbon (see the improve our knowledge of the influence of these factors on Coulomat response and linearity section).the determination of ‘elemental’ carbon by a simple thermal Organic carbon is determined by using an inert gas (N2) to method. desorb adsorbed compounds at high temperature (500– 800 °C). Formation of CO2 is assured by the catalytic oxidation of the desorbed compounds. The determination of elemental Experimental carbon is performed after the desorption by combustion of Products and materials the sample under a stream of pure O2.The drift of the Coulomat due to the atmospheric CO2 Coronene (purum), barium perchlorate (puriss p.a.) and infiltration was checked before each measurement. A drift barium carbonate (puriss p.a.) were obtained from Fluka below 2 counts min-1 was considered acceptable and sub- (Buchs, Switzerland). Analytical-reagent grade anthracene, tracted as background during the measurements. The response saccharose, oxalic acid and calcium carbonate and extra-pure of the Coulomat was tested every 10 measurements with two activated charcoal were obtained fMerck (Dietikon, oxalic acid controls (50 ml of aqueous oxalic acid, 3.8 and Switzerland).Activated charcoal was conditioned at 800 °C 0.4 mg L-1 carbon concentration). The reproducibility was for 4 h under an N2 atmosphere. Nitrogen 60 (99.9999%), tested using triplicates of increasing amounts of oxalic acid. oxygen 55 (99.9995%) and helium 60 (99.9999%) were obtained from Carbagas (Liebefeld, Switzerland).Sampling Coronene and anthracene were chosen as representative of PAHs adsorbed on diesel soot14 and saccharose was chosen All quartz fiber filters (Whatman, Maidstone, Kent, UK; as a representative of highly pyrolyzable organic compounds. QM-A) were conditioned under O2 at 1000 °C for 30 min. Saccharose samples were prepared by deposition of an aqueous Instrumentation solution (10 g L-1) on a 43 cm2 quartz fiber filter. Coronene and anthracene samples were prepared by deposition and A Stro�hlein 720 DR/C Coulomat was used, with the modifi- evaporation (15 min, 120 °C) of a toluene solution on a 43 cm2 cation that the pump was replaced by carrier gas overpressure quartz fiber filter.Passive cigarette smoke samples were in order to prevent any entry of air into the system through sampled using a cyclone (CasellaTM respirable dust) in a leaks (Fig. 1). laboratory-made sampling chamber. Diesel exhaust samples were obtained using a laboratory-made mixing chamber con- Measurement method nected to the exhaust of a Fiat Ducato 2.5 diesel powered The determination of carbon is based on converting carbon- light duty truck, using a cyclone (CasellaTM respirable dust).containing compounds to CO2 and H2O by catalytic oxidation or using an oxidizing atmosphere at high temperature. Calibrations To ensure complete conversion to CO2, the actual desorp- Temperature measurement. In the Coulomat (Fig. 1), tion/combustion unit (oven I ) is followed by a CuO/Pt catalyst measurements of oven temperature are made by using a at 900 °C (oven II ) for post-oxidation.Any sulfur oxides and thermocouple placed outside the desorption quartz tube. To hydrogen chloride formed are removed from the gas stream check the relationship between the measured temperature and in oven III, which is kept at 400 °C and packed with silver the temperature inside the tube (where the sample is intro- wool. As a safety measure, a PerhydritA absorber is further duced), a calibration has been performed with a reference connected upstream of the following absorption vessel to scrub thermocouple (chromel–alumel ) placed at the same location out any sulfur dioxide residues not removed by oven III, which as the sample.The relationship was linear from 200 to 900 °C could influence the pH-dependent CO2 determination. The (n=8, slope=1.04±0.01, intercept=-8.29±4.58, r= CO2 passes together with the gas stream into the absorption 0.99984).All temperatures cited in this paper are temperatures vessel and is absorbed in alkaline Ba(ClO4)2 solution according inside the oven tube and were calculated using the regression slope determined above. Coulomat response and linearity.To check the Coulomat response, a calibration was performed using increasing amounts of oxalic acid. The linearity was checked up to 500 mg of carbon (n=30, r=0.99989). A count is equal to 0.2000±0.0006 mg of carbon, which corresponds to the manufacturer’s specifications. Data analysis Statistical analysis of the results was carried out using Systat Fig. 1 Schematic diagram of the modified Stro� hlein 702 DR/C Coulomat (pump unit removed). 5.2 for multifactorial analysis and variance analysis. 368 J. Environ. Monit., 1999, 1, 367–372Results and discussion Factors of influence The desorption of the organic compounds can be considered as the major step which influences the reliability of the EC determination. Many factors are susceptible to influence the desorption eYciency.In this paper, factors to be studied were chosen based on two criteria: factors from the analysis process and technical feasibility. The factors selected were desorption temperature, desorption duration, sample load, heating rate and sample type. Influence of temperature The influence of temperature on the desorption eYciency of two types of organic compounds, saccharose and coronene, is Fig. 3 Desorption thermogram at (solid line) 550 °C under N2 and illustrated in Fig. 2. (dotted line) 750 °C under N2 of 50 ml of aqueous saccharose solution Saccharose is a compound with a high proportion of oxygen (2 mg of carbon per ml of solution). The curves are the desorption (C12H22O11) and coronene is a PAH (C30H14). Both are strictly kinetic in counts min-1 of the CO2 detector during desorption (1 count=0.2 mg of carbon).organic compounds (the carbon content is 100% organic carbon) and should lead to an EC response of zero within experimental error (EC/TC=0%). Fig. 2 shows that the posidesorption eYciency is increased compared with a lower tive interference decrease with increase in temperature for both temperature. types of compounds.As shown in Fig. 2, the coronene desorption eYciency seems The pyrolysis of organic compounds (also known as charto reach a maximum after 800 °C (EC/TC=2–3%). The ring) during the desorption process is a major cause of artefacts saccharose desorption eYciency reaches its maximum at the and this phenomenon increases with increase in temperature. highest temperature, without any step.This could be explained Another possible cause of artefacts is the non-desorption of by a diVerent sensitivity to pyrolysis. Coronene, a PAH, seems the organic compounds because of an insuYcient temperature not to be very sensitive to thermal degradation. of desorption. Although higher temperatures increase the desorption These two phenomena are opposed: a high temperature eYciency of organic compounds, an excessive desorption tem- increases the interference by pyrolysis but also decreases it by perature could desorb the ‘elemental’ carbon present in the increasing the desorption eYciency.The increase in desorption diesel samples ( leading to a positive interference during the eYciency with increase in temperature is illustrated in Fig. 3. OC determination) and inorganic carbonaceous compounds Fig. 3 clearly shows a diVerence in the kinetic behavior such as carbonates. Fig. 4 shows the desorption of elemental during the first 3 min. At 750 °C, more organics are volatilized carbon and calcium carbonate versus temperature. during the very beginning of the desorption process as com- Fig. 4 suggests that the loss of elemental carbon will not pared with 550 °C, and more thermal decomposition products occur below 800–850 °C under N2.However, some internal are removed. This is why the total count (total area under the oxidants (e.g., metal oxides) present in the sample could curves) related to the OC removed is higher at 750 °C. oxidize part of the EC fraction during desorption under a Depending on the kind of sample, thermal degradation neutral atmosphere.15 could take place at relatively low temperatures.A higher Calcium carbonate, which occurs naturally in environmental temperature increases the thermal degradation but, on the samples, could cause a positive interference during the desorp- other hand, also enhances the desorption eYciency of nontion and/or oxidation processes. Fig. 4 suggests that desorption volatile species.Both eVects act in opposite directions. The of CaCO3 will occur at temperatures higher than 800 °C. Other results in Fig. 3 tend to show that at a higher temperature the carbonates, such as MgCO3 and Na2CO3, could decompose at lower temperatures but are less common in occupational environments with the exception of specific situations where they are used in the working process.The interferences due to the presence of carbonates could be eliminated by acid treatment of the sample.16 Influence of the duration of desorption The influence of the duration of desorption on the desorption eYciency of an organic compound is illustrated in Fig. 5 for saccharose. The behavior of other pure organic compounds is expected to be similare behavior of mixtures such as diesel fuel is more complex and very sample dependent.Fig. 5 shows how long desorption should be to be as complete as possible. An essential point of the coulometric method is the necessity to ensure complete desorption of the organic compounds prior to the determination of the EC fraction. Excessive desorption causes many problems, such as Fig. 2 Ratio of elemental carbon (EC) to total carbon (TC) versus desorption of EC during OC determination, drift of the temperature of desorption of (&) saccharose (84 mg OC) and ($) apparatus and excessive cost of the analysis. A compromise coronene (100 mg OC). EC was determined by the combustion of the has to be found.We consider that the best strategy is to desorb sample for 8 min at 800 °C under O2 (100%) after 40 min of desorption until the signal does not increase over a period of 2 min, under N2 for saccharose and 15 min for coronene.TC is the sum of OC+EC. instead of waiting for a certain fixed time. This usually J. Environ. Monit., 1999, 1, 367–372 369contributes to a higher relative desorption eYciency. This can be easily explained.At the beginning of the desorption, a sample with a higher sample load will desorb relatively more compounds. This means that at the end of the desorption, a relatively smaller amount could be aVected by thermal degradation. This eVect does not appear for anthracene, which is less sensitive than saccharose to thermal degradation. Influence of the heating rate The influence of the heating rate on the desorption eYciency of saccharose is illustrated in Fig. 7. It shows a weak improvement of the desorption eYciency with increasing rate of heating for saccharose but not for anthracene. Saccharose is more sensitive to thermal degradation than anthracene (a PAH). At low heating rates, samples are heated longer at lower temperatures, which can be suYcient to pyrolyze the compound but Fig. 4 Desorption thermogram under N2 of activated charcoal, used not high enough to ensure its desorption. as an elemental carbon source, and of calcium carbonate under N2 and O2. Carbonates decompose on heating with CO2 emission. Each temperature was stepped during 5 min after reaching the target Hierarchy of influences temperature. The fitted curves were constructed by mathematical B-spline interpolation. To check the relative importance of the various parameters of influence, a combinatory based analysis was conducted.For four diVerent kinds of samples, four parameters at two levels (low and high) were tested. The types of samples, factors of influence and levels used in the combinatory analysis were as follows Samples Saccharose, coronene, cigarette smoke, diesel soot Desorption temperature 500, 800 °C Desorption duration 15, 30 min Heating rate 50, 500 °Cmin-1 Sample load Saccharose 50, 260 mg total control (TC) Coronene 40, 160 mg TC Cigarette smoke 40, 240 mg TC Diesel 24, 48 mg TC Fig. 5 Desorption at 500 °C under N2 of saccharose (630 mg OC) This combination leads to 64 analyses to be complete. versus time. 100% desorption corresponds to the total recovery of OC Table 2 presents the crude results for cigarette smoke and (OC=TC).Desorption kinetic refers to Dcount/Dtime= [count(time2)-count(time1)]/(time2-time1). diesel soot samples. Table 3 summarizes the crude analysis of variance of the results and Table 4 summarizes the crude analysis of variance of the interaction between factors. corresponds to about 20 min for the occupational hygiene The F-ratio can be used as a measure of the intensity of an samples treated in our laboratory.influence and therefore as a means to establish a hierarchy between the factors. Influence of the sample load It is interesting to note from Tables 1 and 3 that the The influence of the sample load on the desorption eYciency temperature and the duration of desorption are the main of organic compounds, saccharose and anthracene, is illustrated in Fig. 6, and suggests for saccharose that a higher sample load Fig. 7 Desorption at 800 °C under N2 40 min (total time meaning including heating time) of saccharose versus rate of heating. EC was Fig. 6 Desorption of saccharose and anthracene at 800 °C under N2 determined at 800 °C for 8 min under O2.The heating rate was adjusted manually. The graphs show the heating rate of the oven and for 20 min as a function of the equivalent carbon content of the sample (TC). EC was determined at 800 °C for 8 min under O2. not of the sample itself. Errors bars represent standard deviations on triplicates. Errors bars represent standard deviations on triplicates. 370 J.Environ. Monit., 1999, 1, 367–372Table 1 Summary of factors of influence on the EC recovery versus type of sample Factor Saccharose Coronene Cigarette smoke Diesel Temperature XXXX XXXX XXXX XXX Desorption duration XXXX XXX XX XXXX Rate of heating n.s.a n.s. n.s. X Sample load XX n.s. XXX XX an.s., non-significant factors (P>0.05). Table 2 Crude results for cigarette smoke and diesel soot samples.The ratio EC/TC is presented as a function of the diVerent parameters levels (1=high level, 0=low level; e.g., last line is temperature 800 °C, duration 30 min, heating rate 500 °Cmin-1, sample load 240 and 48 mg TC for cigarette smoke and diesel soot, respectively) Sample type Temperature Duration Heating rate Sample load Ratio EC/TC (%) Cigarette smoke 0 0 1 0 48 0 0 0 0 46 0 1 0 0 37 0 0 0 1 34 0 1 1 0 31 1 0 0 0 28 0 0 1 1 22 1 0 0 1 21 0 1 1 1 16 0 1 0 1 16 1 1 0 0 8 1 0 1 0 8 1 1 1 0 6 1 0 1 1 4 1 1 0 1 1 1 1 1 1 1 Diesel soot 0 0 0 0 73 0 0 1 0 70 1 0 0 0 65 0 0 0 1 56 1 0 1 0 54 0 1 0 0 54 0 0 1 1 53 1 0 0 1 49 0 1 1 0 49 0 1 0 1 44 0 1 1 1 43 1 0 1 1 37 1 1 0 1 31 1 1 0 0 31 1 1 1 0 22 1 1 1 1 21 Table 3 Analysis of variance of the temperature, desorption duration, heating rate and sample load influence on the EC/TC ratio.Bold figures indicate that the F-ratio does not reach a significance level of 5% (P>0.05) Cigarette smoke Saccharose Coronene Diesel Factor F-ratio P F-ratio P F-ratio P F-ratio P Temperature 45.677 0.000 72.246 0.000 81.308 0.000 33.707 0.000 Desorption duration 14.004 0.003 32.049 0.000 27.875 0.000 52.236 0.000 Rate of heating 4.163 0.066 1.587 0.234 3.216 0.100 5.373 0.041 Sample load 16.261 0.002 10.757 0.007 0.433 0.524 13.332 0.004 Table 4 Analysis of variance of the interaction between factors.Bold figures indicate that the F-ratio has reached a significance level of 5% (P<0.05) Cigarette smoke Saccharose Coronene Diesel Interaction F-ratio P F-ratio P F-ratio P F-ratio P Temperature–time 3.413 0.124 12.569 0.016 8.478 0.033 53.512 0.001 Temperature–rate 1.520 0.272 4.228 0.095 0.237 0.647 20.400 0.006 Temperature–load 1.321 0.302 1.719 0.247 5.543 0.065 0.312 0.601 Time–rate 3.741 0.111 4.162 0.097 0.339 0.585 1.761 0.242 Time–load 3.262 0.131 2.623 0.166 4.783 0.080 14.090 0.013 Rate–load 0.082 0.786 0.176 0.692 0.000 0.986 0.330 0.590 J.Environ. Monit., 1999, 1, 367–372 371Table 5 Temperature program proposed to determine EC/OC content of diesel soot using the thermal method. A first step (pre-analysis) of 3 min under an N2 atmosphere is used to purge the system of the ambient air. The second step (OC determination) is a fast increase in the temperature of desorption to 800 °C.This temperature is stepped until no count is observed during 2 min. The third step (EC determination) is a switch to an O2 atmosphere and a temperature of 800 °C stepped until no count is observed during 2 min Analysis Gas Temperature/°C Duration/min Pre-analysis High-purity N2 Ambient 3 Organic carbon (OC) High-purity N2 800 Stepped Elemental carbon (EC) High-purity O2 800 Stepped Table 6 EC recovery for pure organic samples (saccharose and eYciency of the Diesel organic fraction.To improve the EDTA). Saccharose solution 50 ml, carbon content 2 g l-1. EDTA desorption, we propose to adopt the highest possible temperasolution 50 ml, carbon content 2.2 g l-1. For the thermal method, ture whilst still avoiding EC desorption. Using the results desorption temperature and duration are indicated.A stepped temillustrated in Fig. 4, we propose to use 800 °C as a good perature is maintained until no count on the CO2 detector is observed compromise and to continue the desorption until there is no during 2 min count detected during 2 min (Table 5). EC recovery (%) Simple modifications of the method can lead to a substantial improvement of the organic desorption (Table 6).Method Saccharose EDTA NIOSH 5040a <0.1 <0.1 Acknowledgements Thermal, 550 °C, 8 min 66 87 Thermal, 550 °C, stepped 59 82 We thank Mr. J.-N. Lepdor for technical assistance. This Thermal, 800 °C, 10 mina 22 11 project was funded by the Swiss National Science Foundation, Thermal, 800 °C, steppedb 8 15 project number 32–42315–94. aThe ‘NIOSH 5040’ and the ‘Thermal, 800 °C, 10 min’ values are from an interlaboratory comparison.13 bThe step duration was around 35 min.References 1 IARC Monographs on the Evaluation of Carcinogenic Risks to factors of influence for all types of sample. The other factors, Humans, Vol. 46, Diesel and Gasoline Exhausts and Some especially the sample load, may act in a diVerent way according Nitroarenes.International Agency for Research on Cancer, Lyon, to the sample type. Cigarette smoke, for instance, is more 1989. sensitive to the load than to the duration of desorption. This 2 Health EVects Institute, Diesel Exhaust: a Critical Analysis of confirms that this very common pollutant in workplaces may Emissions, Exposure, and Health EVects, Special Report of the Institute’s Diesel Working Group, Health EVects Institute (HEI), cause significant interferences in the EC determination, as Cambridge, MA, 1995.already shown earlier,13 and should be looked for in any field 3 B. Bhatia, P. Lopipero and A. H. Smith, Epidemiology, 1998, 9, 84. sampling. Charring probably explains this interference, which 4 International Programme on Chemical Safety, Diesel Fuel and suggests that the influence of the sample load on the desorption Exhaust Emissions, Environmental Health Criteria No. 171, eYciency could be used as a measure of the charring sensitivity World Health Organization, Geneva, 1996. of the sample, but more data are needed to support this 5 U.Heinrich, R. Fuhrst and S. Rittinghausen, Inhal. Toxicol., 1995, 7, 533. conclusion. Saccharose is an organic compound representative 6 K.Steenland, D. Silvermann and D. Zaebst, Am. J. Ind. Med., of highly pyrolyzable species but probably not representative 1992, 21, 887. of common workplace pollutants. Saccharose readily decom- 7 E. Brich and R. A. Cary, Analyst, 1996, 121, 1183. poses. The charring process continues as the temperature is 8 J. Perez and R. L. Williams, SAE (Soc.Automot. Eng.) Tech. increased, but almost all of the decomposition products Pap., 1989, No. 894291. (including char) can be removed at high enough temperatures. 9 D. M. Smith, M. S. Akhter, J. A. Jassim, C. A. Sergides, W. F. Welch and A. R. Chugghtai, Aerosol Sci. Technol., 1989, 10, 311. Therefore, the char formed during analyses for sucrose and 10 B. K. Cantrell and W. F. Watts, Jr., Appl. Occup. Environ. Hyg., similar compounds is probably not a major source of EC 1997, 12, 1019. interference when a higher temperature is employed. 11 S.-J. Chen, S.-H. Lioa, W.-J. Jian and C.-C. Lin, Environ. Int., PAHs or the other compounds tested in this study are 1997, 4, 475. representatives of the fraction of the organic compounds 12 Von den Berufsgenossenschaften anerkannte Analysenverfahren zur suspected to be mutagenic which are coated on the carbon Festellung der Konzentrationen krebserzeugender ArbeitstoVe in der Luft in Arbeitsbereichen. Method No 44: Diesel Engine Emission, core of diesel soot. They are almost insensitive to the rate of ZH 1/120, Carl Heymanns Verlag, Cologne, 1995. heating and to the sample load for the ranges used in our 13 E. Birch, Analyst, 1998, 123, 851. experiments. This confirms that charring is not a problem for 14 N. R. Khalili, P. A. ScheV and T. M. Holsen, Atmos. Environ., these PAHs. 1995, 29, 553. 15 E. Birch, ECDSEMeeting, Berlin, personal communication, 1998. 16 H. Cachier, M.-P. Bremond and P. Buat-Me�nard, Tellus, 1989, Conclusion 41B, 379. The temperature and the duration of desorption seem to be the parameters of most influence acting on the desorption Paper 9/02622C 372 J. Environ. Monit., 1999, 1, 367–3