Annals of Occupational Hygiene Advance Access originally published online on July 20, 2006
Annals of Occupational Hygiene 2006 50(8):805-812; doi:10.1093/annhyg/mel048
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Collection, Validation and Generation of Bitumen Fumes for Inhalation Studies in Rats Part 2: Collection of Bitumen Fumes from Storage Tanks
Fraunhofer Institute Toxikology und Experimental Medicine (Fh-ITEM) Nikolai-Fuchs-Strasse 1, 30625 Hannover, Germany
*Author to whom correspondence should be addressed. Tel: +(49) 511 5350 116; fax: +(49) 511 5350 155; e-mail: pohlmann{at}item.fhg.de
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISSCUSIONSUMMARY AND CONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Objectives: The objective of the study described in this and an accompanying series of papers was to develop a laboratory generated exposure atmosphere to be used for chronic inhalation toxicity studies in rats that resembles, as closely as possible, personal exposures seen by workers during road paving operations.
Methods: To achieve this objective, atmospheric workplace samples were collected at road paving worksites and compared analytically with bitumen fume samples collected from the headspace of hot bitumen storage tanks. In Preiss et al. (2006) the collection and analysis of workplaces samples is described. This contribution describes the strategy for the in-line extraction of a suitable fraction of bitumen fume collected from the headspace of a bitumen storage tank and the comparison of the collected condensate to workplace samples.
Results: Results show that is possible to develop a collecting procedure that allows sampling from hot bitumen storage tanks in an operational asphalt mixing plant. The sampling procedure has been optimized to collect material that matches the workplace samples as closely as possible. The comparison to workplace samples has been performed using parameters that can be analyzed in both the workplace samples and the bitumen fume condensate collected from the tanks. Boiling point distribution (BPD), UV fluorescence (UV-Fl) and content of individual polycyclic aromatic hydrocarbons (PAH) were selected as parameters. The BPD of the final collected bitumen fume condensate did not differ by more than 17°C from any point on the average BPD curve of the workplace samples, in the range from 5 to 95%. UV-Fl of the bitumen fume condensate nearly exactly matched the average UV-Fl of the workplace samples. However, the sum of the 17 PAHs analyzed in the test samples, compared to the mass of the condensate, is lower by a factor of 
3 than the sum of the 17 PAHs in some personal samples compared to the mass of Total Organic Matter (TOM). It has to be recognised that during the collection of the workplace samples, despite all efforts a number of the workers who carried a personal sampler could not be prevented from smoking.
Keywords: asphalt • bitumen • fume • workplace • chemical analysis • sampling methods • polycyclic aromatic hydrocarbons
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISSCUSIONSUMMARY AND CONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Bitumen fumes emitted during road paving operations comprise a variable, complex mixture of chemical species. It is difficult therefore to validate artificially generated bitumen fumes for inhalation studies against workplace atmospheres. In the first paper of this series (Preiss et al., 2006) we described the development of validation criteria for a laboratory-generated exposure atmosphere to be used for chronic inhalation toxicity studies in rats. To achieve this, three different laboratories [Shell Global Solutions, Int. (Shell), Fraunhofer Institue of Toxicology and Experimental Medicine (Fh-ITEM, Germany) and Berufsgenossenschaftliches Institut für Arbeitssicherheit' (BIA, Germany)] analyzed personal and static samples taken at road paving worksites by different methods. Parameters determined were: total particulate matter (TPM), benzene soluble matter (BSM), semivolatiles (SV) and total organic matter (TOM), boiling point distribution (BPD), polycyclic aromatic hydrocarbons (PAH), UV fluorescence (UV-Fl), aerosol and total fume (TF). Different sampling methods were used to allow also comparison of the standard German sampling method with the most commonly used industrial methods.
The results of the analysis of the workplace samples enabled us to setup validation criteria involving parameters that can be analyzed in samples from the workplace and in bitumen fume condensate collected from bitumen storage tanks: BPD, UV-Fl and content of individual PAHs were selected as parameters.
Generation of a stable inhalation atmosphere at three different concentrations, but equal in terms of relative composition, that can be maintained for two years is a prerequisite for inhalation carcinogenicity studies following OECD guideline 451. Binet et al. (2002) describes a generation system for bitumen fumes suitable for flow rates of 23 l min–1. In the current inhalation study, a total of 0.8 m3 bitumen fume min–1 for the three different dose groups running in parallel is needed and the exposure atmosphere has to be generated for 6 h a day, 5 days a week for 2 years. To accomplish this with a system comparable to that described by Binet et al. (2002), three enlarged fume generators running in parallel consuming huge amounts of bitumen would have been needed. Using such an evaporation-condensation apparatus was also not considered adequate to produce a representative bitumen fume at the concentrations and amounts needed for the inhalation study. Based on the results of the Heritage Research Group (Kriech, 1999), it was proposed to collect bitumen fume from the head space of hot bitumen storage tanks and to generate the test atmosphere under controlled conditions using a laboratory set-up developed at the Fh-ITEM. In this apparatus, liquid tank condensate is evaporated and re-condensed on a large number of condensation nuclei, also generated by the apparatus. This leads to a highly dispersed stable aerosol phase in equilibrium with the corresponding vapour phase.
From the parameters measured at the workplace, acceptance criteria were established (Preiss et al., 2006), for comparison with the bitumen fume condensate sampled from the headspace of the bitumen storage tanks. It was calculated that 16 kg of bitumen fume condensate would be needed for the carcinogenicity study. This amount could only be collected from a large amount of hot bitumen and consequently sampling had to be carried out at a busy asphalt mixing plant with sufficient bitumen throughput.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISSCUSIONSUMMARY AND CONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Development of the sampling procedure
The methodology to obtain bitumen condensate similar to workplace fumes as described by the Heritage Research Group (Kriech, 1999) was tested in an initial trial. This method consist of sampling the condensate from the headspace of a heated storage bitumen tank approx. half-filled (Fig. 1). The fume is sampled with a dropping funnel (250 ml, inner diameter 4.5 cm) filled with
70 g XAD-2 (Fig. 2). Pre-cleaned glass wool is used to keep the resin in place. A peristaltic pump collects the fumes at a rate of 5 l min–1. The relative vapor concentration before and behind the XAD cartridge is monitored using a flame ionization detector (FID).
|
|
Using this initial setup, after a short time the XAD cartridge became blocked with water condensing in the funnel. Therefore, the setup was modified to use a stainless steel spiral condenser working at ambient temperature, together with a peltier condenser (5°C) in front of the XAD cartridge. The condensers served to remove the water vapor before the fume entered the funnel (Fig. 3). Before entering the stainless steel spiral, the fume is passed through a heated tube to prevent condensation in the upper section. This setup was equipped with two XAD cartridges to monitor vapour breakthrough. Using this equipment setup, it was possible to sample bitumen fume condensate for up to 8 h.
|
Comparison of the PAH content and the BPD of the condensates sampled with the setup described above, with workplace samples showed considerable differences. For example the PAH content was much lower than the concentrations found at workplaces (Ekström, 1990). The BPDs also showed a higher concentration of light boiling components in the tank samples compared to workplace samples collected at different locations. It was, therefore, agreed to develop an in-situ method to enrich the PAH content of the bitumen samples and achieve a BPD that more closely matched exposures found at workplaces. In addition bitumen fume had to be collected at higher sample flow rates.
The rational behind the approach adopted was that light boiling point compounds elute in a shorter time through a XAD–cassette than heavier boiling point compounds. Therefore, compounds with higher boiling points will be enriched near the entrance of the cassette, whereas lighter boiling point compounds will pass through the cassette. To test this hypothesis, we used tubes with the same inner diameters as the dropping funnels but with 10-fold greater length. These tubes were filled with XAD, subdivided by stainless steel meshes into four sections of equal length. Since we used a 10-fold increase in tube length, we used a flow rate that was also 10 times higher. Figure 4 provides a schematic representation of the idea. The tubes were used in a set-up similar to that described in Fig. 3.
|
Chemical analysis confirmed the enrichment of semi-volatile compounds in the foremost section of the tube. However, the enrichment was not sufficient to fulfill the needs. Since the flow rate was 10 times the flow rate of the small XAD cassette, we found not only water in the receiver bottle of the peltier condenser, but also bitumen condensate in an amount sufficient to perform chemical analysis. It turned out that this condensate was much closer to the workplace samples with respect to the BPD and the PAH content than the XAD samples. Therefore we decided not to use XAD cassettes but sample the bitumen fume condensate directly using the peltier cooler.
Sampling system for collection of the final bitumen fume condensate
The final sampling setup used for collecting the bitumen fume condensate is shown in Figs 5 and 6. Sampling of the condensate for the 2-years study took place at an operating asphalt mixing plant, supplied directly with bitumen from a nearby refinery. Eight identical sampling units sampling from two different heated bitumen storage tanks were installed to shorten sampling time.
|
|
The fume was sampled
10 cm above the liquid bitumen surface. Since the bitumen level in the tanks changes continuously during workdays, floats were used to keep the sampling pipe entrance at a constant level above the bitumen surface. The fume was sucked in through a stainless steel pipe (3 m) and directed through a heated Teflon tube (5 m in length) into a tool house located near the bitumen tanks (refer to Fig. 7), where the sampling equipment was located.
|
For each tank, a separate setup as shown in in Fig. 6 was used. To prevent the sampling system from being blocked by bitumen that may be accidentally sucked into the sampling tube, a pre-separator was installed at the entrance to the sampling system. From the insulated pre-separator the fume was conducted into four separate sampling units. These units consist of a cooling spiral, a Peltier condenser, collection bottles and a peristaltic pump.
After completion of all workplace samples of the Strassburg road trial (Preiss et al., 2006) and the corresponding chemical analysis, the bitumen fume condensate sampling system was optimized in regard to the parameters already mentioned before.
The fume was conducted through the cooling spiral into the Peltier condenser running at a temperature of 5°C. The condensed bitumen fume and water were collected in two 10-l vacuum tied polyethylene collection bottles at a pressure level in the bottle of 800 mbar. Bitumen condensate was only found in small amounts in the first bottle. This condensate was not used for the inhalation study. A total amount of 16.2 kg bitumen fume condensate was collected from the second bottle through the whole sampling time.
Characterization of the bitumen fume condensate samples
Sample preparation. The condensate sampled from the tanks consisted of two phases: an organic phase (fume condensate) and an aqueous phase. The aqueous phase was separated and extracted twice with dichloromethane. The organic phases were combined and dried using sodium sulphate. The dichloromethane was evaporated from the sample using a rotary evaporator at a temperature of 40°C and an initial vacuum pressure of 800 mbar. When the dichloromethane had been evaporated almost completely, the vacuum pressure was reduced stepwise to 400 mbar to remove the remaining dichloromethane. After 30 min, when no more dichloromethane remains, the procedure is finished.
Analysis. The analytical methods used for the determination of BPD, UV-Fl and PAH are described in Preiss et al. (2006).
|
RESULTS AND DISSCUSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISSCUSIONSUMMARY AND CONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
Test samples
To ensure that the composition of the bitumen fume condensate for the inhalation study corresponds to that of the bitumen fume at workplaces, six test samples were collected from 04 to 17 July 2002 from one storage tank and analyzed according to the methods described in Preiss et al. (2006).
Figure 8 shows the average BPD of the test samples, of the Shell personal workplace samples of the Strassburg campaign (Preiss et al., 2006) and of samples from a US field study (Kriech et al., 2002). The sample curve starts at 204°C and is 22°C above the corresponding curve of the workplace samples. At 90%, the tank and workplace curves cross and >90% the workplace samples exhibit a higher boiling point.
|
The concentrations of the 17 PAHs analyzed in the condensate are listed in Table 1. The concentrations of individual PAHs as well as the sum of all PAHs analyzed are relatively constant in the samples. An average (sum of) 506 ± 78 µg PAHs g–1 of bitumen fume condensate was found.
|
Finally, the UV-Fl intensities of the test tank samples and of the Shell personal samples are compared in Table 2. The UV-Fl values of the test samples refer to the mass of the bitumen fume condensate. As an average, 2.09 ± 1.02 EU (emission units) was measured. The mean UV-Fl of the Shell personal samples, which refers to the mass of TOM, was 2.05 ± 0.95 EU. Both, the values of the test samples and of the Shell personal samples, show a wide spread.
|
For the parameters UV-Fl and BPD, reasonable agreement between the test samples and the Shell personal samples was observed. However, the sum of the concentrations of the 17 PAHs analyzed in the test samples (506 ± 78 µg g–1 condensate) is lower by a factor of
3 than the sum of the 17 PAHs in the Shell personal samples when expressed as concentration relative to TOM (1475 µg g–1 TOM) (note: all workers who carried a personal sampler were smokers). On the other hand, the concentrations of individual 4–6 ring PAHs (which are assumed to have a carcinogenic potential) are so low that they are very close to the limit of detection. Whilst there were differences in the sum of PAHs between the personal monitoring samples and the tank condensates, the PAH profiles were similar and it might also be possible that the differences in the absolute amount may be caused by the two different collection techniques used (fume condensation procedure and sampling on a combined filter/adsorbent).
Nevertheless, the test samples were agreed to be suitable to be used in the 2-year carcinogenicity study. The actual concentration of bitumen fume in the inhalation atmosphere in the high dose group of the animal study (nominal 100 mg m–3) is expected to be 166 mg m–3. Nominal concentrations are given in so-called BIA equivalents and have to be multiplied by a factor of 1.66 to give the true concentrations; this factor was calculated by division of the gravimetrically determined concentration by the concentration determined using the BIA method 6305. Therefore, the expected total PAH content in the inhalation atmosphere for the 2-year study will be 50.6 µg x 1.66 = 84 µg m–3. (The factor 1.66 results from the conversion of the so called BIA-Standard (BIA Method 6305) to absolute mass.) This is 77 times greater than the average PAH concentration of 1.1 µg m–3 in the workplace atmosphere measured by the Shell personal samples (Preiss et al., 2006).
Final tank fume condensate sample
The amount of bitumen fume condensate collected for the 2-year carcinogenicity study amounts to 16 kg in total. For each of the seven 1-week sampling periods, the PAH content, the UV Fl, and the BPD were measured for both tanks separately to be able to reject individual samples not fulfilling the acceptance criteria. Finally, all acceptable samples were mixed, filled into brown glass bottles (1 l) and stored in the dark under nitrogen in PTFE bags at –18°C. The pooled bitumen fume condensate was then analyzed again. The BPD from this survey is given in Table 3. The concentrations of 18 individual PAHs were determined by the Fh-ITEM, Hannover, and for comparison also by the Biochemisches Institut für Umweltkarzinogene' (BIU), Großhansdorf. The results are listed in Table 4. The two data sets do not deviate for most of the PAH's by more than ±15% from the mean value. Higher deviations are only observed for acenaphthylene (±30.4%), fluorene (±22.4%) and anthracene (±18.7%). The danger of overlapping of these PAH's with disturbing components is particularly large. The deviations can therefore be explained presumably by the fact that both labs used different clean-up procedures. The UV-Fl value of the pooled sample is 1.9 EU µg–1.
|
|
Comparison of bitumen fume from workplaces with bitumen fume condensate
Figure 9 shows a comparison of the average BPD curves of bitumen fume condensate from workplaces (from Preiss et al., 2006) and the BPD curve of the final bitumen fume condensate (pooled) used for the 2-year study. There is a reasonable level of overall agreement between the two distributions. Nevertheless, the workplace BPD starts at a somewhat lower temperature and is to some extent steeper, leading to a slightly higher end temperature. This behaviour is also reflected in the comparison of the relative PAH content of the samples shown in Fig. 10. PAHs with lower boiling points (and hence higher relative content) seem to be slightly more pronounced in the pooled bitumen fume condensate than in the averaged personal workplace samples.
|
|
The UV-Fl value of the pooled sample is 1.9 EU µg–1. The average UV-Fl determined from the workplace samples (Part 1) is 2.05 ± 0.95 and, therefore, considerably comparable to the bitumen fume condensate.
|
SUMMARY AND CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISSCUSIONSUMMARY AND CONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
The aim of the study was to develop a laboratory-generated exposure atmosphere to be used for chronic inhalation toxicity studies in rats, which resembles as closely as possible personal exposures seen in workers during road paving operations. To achieve this, bitumen fume condensate samples were collected from hot bitumen storage tanks and compared analytically with atmospheric workplace samples collected by different sampling devices and strategies.
Three different laboratories (Shell, Fh-ITEM, BIA) analyzed personal and static samples taken at road paving worksites by different methods. Parameters determined were: TPM, BSM, SV and TOM, BPD, PAH, UV Fl, Aerosol and TF. Different sampling methods were used to allow comparison of the standard German sampling method with the most commonly used industrial methods.
A collecting procedure was developed that allowed sampling from hot bitumen storage tanks in an operational asphalt mixing plant. The sampling procedure has been optimized to collect material that matches the workplace samples as closely as possible. The validation procedure has been performed using parameters that could be analyzed in both the workplace samples and the bitumen fume condensate collected from the tanks. BPD, UV-Fl and content of individual PAHs were selected as parameters.
The BPD of the collected sample did not differ by >17°C from any point on the the average BPD curve of the workplace samples, in the range from 5 to 95%. UV-Fl of the bitumen fume condensate nearly exactly matched the average UV-Fl of the workplace samples. However, the sum of the 17 PAHs analyzed in the test samples, compared to the mass of the condensate, is lower by a factor of 3 than the sum of the 17 PAHs in the Shell personal samples compared to the mass of TOM (but note: all workers who carried a personal sampler were smokers). On the other hand, no correlation was found between cancer incidence and the sum of the EPA PAHs in workplace samples (Kriech, 2002), and the concentrations of individual 4–6 ring PAHs (which are assumed to have a carcinogenic potential) are found to be so low that they are very close to the limit of detection. Whilst there were differences in the sum of PAHs between the personal monitoring samples and the condensate, the PAH profiles were similar (refer to Fig. 10), and it is possible that the differences in absolute amounts may be caused by the two different collection techniques used (fume condensation procedure and sampling on a combined filter/absorbent). Nevertheless, the test samples showed that the tank fume condensate is suitable to be used in the 2-year carcinogenicity study, since the PAH concentration in the 100 mg m–3 group can be calculated to be 84 µg m–3 and, therefore, 77 times higher than the average PAH concentration of 1.1 µg m–3 found in the workplace atmospheres measured in the Shell personal samples (refer to Part 1). As a result, 16 l of bitumen fume condensate have been collected from the storage tanks in the asphalt mixing plant.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISSCUSIONSUMMARY AND CONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
We would like to thank ARBIT and EUROBITUME for the financial funding of the study and all the members of the project steering committee for their engagement.
Received February 7, 2006; in final form May 16, 2006
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISSCUSIONSUMMARY AND CONCLUSIONSACKNOWLEDGEMENTSREFERENCES |
|---|
BIA Method 6305: Messverfahren für Gefahrstoffe (Analysenverfahren). Kennzahl 6305: Bitumen–Dämpfe und Aerosole. (Berufsgenossenschaftliches Institut für Arbeitssicherheit, Germany).
Binet S, Bonnet P, Brandt H, et al. (2002) Development and validation of a new bitumen fume generation system which generates polycyclic aromatic hydrocarbon concentrations proportional to fume concentrations. Ann. occup. Hyg 46:617–28.
Ekström LG. (1990) Bitumen fumes, exposure to solvents in connection with the laying of asphalt. (NYNAS, Nynashamn, Sweden) NYNAS BITUMEN AB, Report no. BiTech 625.
Kriech AJ. (1999) Collection and analysis of bitumen fumes for use by the Fraunhofer Institute. Draft-Report to ARBIT(Heritage Research Group, Germany).
Kriech AJ, Kurek JT, Wissel HL, et al. (2002) Evaluation of worker exposure to asphalt paving fumes using traditional and non-traditional techniques. AIHA J 63:628–635.
Preiss A, Pohlmann G, Koch W, et al. (2006) Collection, validation and generation of bitumen fumes for inhalation studies in rats, part 1: workplace samples and validation criteria. Ann. occup. Hyg in press.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||