Arch Toxicol (2011) 85 (Suppl 1):S11–S20
DOI 10.1007/s00204-011-0678-1
ANALYTICAL TOXICOLOGY
Air sampling and determination of vapours and aerosols
of bitumen and polycyclic aromatic hydrocarbons
in the Human Bitumen Study
Dietmar Breuer • Jens-Uwe Hahn • Dieter Höber • Christoph Emmel • Uwe Musanke •
Reinhold Rühl • Anne Spickenheuer • Monika Raulf-Heimsoth • Rainer Bramer •
Albrecht Seidel • Bernd Schilling • Evelyn Heinze • Benjamin Kendzia • Boleslaw Marczynski
Peter Welge • Jürgen Angerer • Thomas Brüning • Beate Pesch
Received: 12 January 2011 / Accepted: 10 February 2011 / Published online: 1 March 2011
Ó Springer-Verlag 2011
Abstract The chemical complexity of emissions from
bitumen applications is a challenge in the assessment of
exposure. Personal sampling of vapours and aerosols of
bitumen was organized in 320 bitumen-exposed workers
and 69 non-exposed construction workers during
2001–2008. Area sampling was conducted at 44 construc-
tion sites. Area and personal sampling of vapours and
aerosols of bitumen showed similar concentrations between
5 and 10 mg/m3, while area sampling yielded higher con-
centrations above the former occupational exposure limit
(OEL) of 10 mg/m3. The median concentration of personal
bitumen exposure was 3.46 mg/m3 (inter-quartile range
D. Breuer
J.-U. Hahn
Institute for Occupational Safety and Health of the German
Social Accident Insurance (IFA), Sankt Augustin, Germany
D. Höber
C. Emmel
U. Musanke
R. Rühl
BG BAU, München and Frankfurt, Germany
A. Spickenheuer
M. Raulf-Heimsoth
R. Bramer

E. Heinze
B. Kendzia
B. Marczynski
P. Welge

J. Angerer
T. Brüning
B. Pesch (&)
Institute for Prevention and Occupational Medicine
of the German Social Accident Insurance, Institute
of the Ruhr-Universität Bochum (IPA),
Bürkle-de-la-Camp-Platz 1, 44789 Bochum, Germany
e-mail: pesch@ipa-dguv.de
A. Seidel
Biochemical Institute for Environmental Carcinogens,
Prof. Dr. Gernot Grimmer-Stiftung, Grosshansdorf, Germany
B. Schilling
Eurofins GfA GmbH, Hamburg, Germany
J. Angerer
Institute and Outpatient Clinic of Occupational,
Social- and Environmental Medicine,
University of Erlangen-Nuremberg, Erlangen, Germany
•
1.80–5.90 mg/m3). Only few workers were exposed above
the former OEL. The specificity of the method measuring
C–H stretch vibration is limited. This accounts for a
median background level of 0.20 mg/m3 in non-exposed
workers which is likely due to ubiquitous aliphatic
hydrocarbons. Further, area measurements of polycyclic
aromatic hydrocarbons (PAHs) were taken at 25 construc
tion sites. U.S. EPA PAHs were determined with GC/MS,
with the result of a median concentration of 2.47 lg/m3
at 15 mastic asphalt worksites associated with vapours
and aerosols of bitumen, with a Spearman correlation
coefficient of 0.45 (95% CI -0.13 to 0.78). PAH exposure
at mastic-asphalt works was higher than at reference
worksites (median 0.21 lg/m3), but about one order of
magnitude lower compared to coke-oven works. For a
comparison of concentrations of vapours and aerosols of
bitumen and PAHs in asphalt works, differences in sam-
pling and analytical methods must to be taken into account.
Keywords Air sampling
Asphalt
Bitumen

Exposure assessment
Polycyclic aromatic hydrocarbons
Introduction
Bitumen (CAS# 8052-42-4) is produced as a residue during
the distillation of crude oil. It is used as industrial bulk
material in the construction industry, especially for paving
and roofing. In European nomenclature, the term ‘asphalt’
refers to the mixture of bitumen as organic binder with
sand or gravel as filler. The majority of asphalt is applied
mechanically as rolled asphalt in road construction at
temperatures between 150 and 180°C and contains
approximately 4–6% bitumen. Mastic asphalt is a special
mixture with 6–8% bitumen for certain surface applications
123
S12
at higher temperatures, limited to 230°C in Germany in
2008. The worksites comprise also indoor settings where
mastic asphalt is transported and applied manually.
Hot bitumen applications lead to emission of a variety of
volatile compounds, mainly aliphatic and fewer aromatic
hydrocarbons. These compounds are released as vapours,
which condensate to aerosols in the air. Potential health
hazards associated with exposure to vapours and aerosols
of bitumen have been attributed to small quantities of
polycyclic aromatic hydrocarbons (PAHs), and some of
which have been classified as human carcinogens (Straif
et al. 2005). Due to the low concentration of PAH com-
pounds, the U.S. National Institute for Occupational Safety
and Health (NIOSH) recommended sensitive analytical
methods for monitoring exposure in asphalt works (NIOSH
2000).
The measurement of exposure to bitumen emissions is
an analytical challenge due to the chemical complexity of
this mixture and the presence of bitumen-derived matters
either in the gas phase or as aerosols of small size. Various
methods have been developed for air sampling and
extraction of bitumen-derived from material. NIOSH has
developed a method No. 5042 for the determination of total
particulate matter (TPM) and benzene-soluble material
(BSM) from bitumen aerosols (Eller 1998). The German
Institute for Occupational Safety and Health of the
Germany Social Accident Insurance (IFA, formerly BGIA)
has developed a method No. 6305 for monitoring exposure
to vapours and aerosols of bitumen (Breuer 2008) for
monitoring workplaces against the background of a former
German occupational exposure limit (OEL). This OEL was
reduced from 15 to 10 mg/m3 in 2000, except for mastic
asphalt works as reported by Raulf-Heimsoth et al. (2011a, b)
in this issue. In 2005, all German OELs based on technical
feasibility were discontinued, and efforts are being made to
derive health-based OELs. The present nationwide cross-
shift study was initiated to conduct a comprehensive survey
among workers exposed to vapours and aerosols of bitu-
men. Here, we present the results of air monitoring of
vapours and aerosols of bitumen and PAHs with a special
focus on sampling and analytical methods. The detailed
analysis of factors influencing the personal exposure
to vapours and aerosols of bitumen is presented by
Spickenheuer et al. (2011) in this issue.
Materials and methods
Study design
Details of the study design and conduct were described by
Raulf-Heimsoth et al. (2011a, b) in this issue. In brief, a
nationwide cross-sectional study was conducted at 57
123
Arch Toxicol (2011) 85 (Suppl 1):S11–S20
construction sites with and without bitumen application
from 2001 to 2008. Personal sampling was used for the
determination of airborne exposure to vapours and aerosols
of bitumen in 320 exposed workers and in a subgroup of 69
construction workers without bitumen application during a
working shift. Exposure was not only measured during
working with hot bitumen but also during the exposure-free
periods. Therefore, dependent on the numbers of unex-
posed periods, the shift-related exposure may be lower than
the task-related exposure. In addition, area sampling was
performed with a corresponding sampling device to mea-
sure vapours and aerosols of bitumen and PAHs at several
construction sites with and without bitumen application.
The construction site was specified independently by the
occupational hygienist in charge of the project and by the
technicians who took the measurements for this project.
Air sampling
Air sampling of vapours and aerosols of hot bitumen
application was performed according to BGIA method No.
6305 (Breuer 2008). Both personal and area sampling was
conducted with the German GGP sampler (Fig. 1). This
sampling system has been adopted for sampling of vapour
and aerosol mixtures. The conical sampling head is suited
to collect the inhalable particle fraction (Kenny et al. 1997)
at a flow rate of 3.5 l/min. A 37-mm glass-fibre filter is
mounted in the filter holder for filtration of particles.
Vapours are adsorbed on 3 g purified AmberliteTM XAD-2
resin inserted into the GGP cartridge. The pumps comply
with the standards EN 1232 and EN 2919 for the volu-
metric flow to be constant with a maximum deviation
of ±5%.
All 320 bitumen-exposed workers and a subgroup of 69
non-exposed workers were equipped with the GGP sampler
in the breathing zone as shown in Fig. 2. Area measure-
ments of PAH were taken with four GGP samplers attached
to one stationary device to increase the mass for analytical
determinations and another GGP sampler for bitumen
Fig. 1 German GGP sampler for vapours and aerosols of bitumen
Arch Toxicol (2011) 85 (Suppl 1):S11–S20
Fig. 2 Personal sampling of vapours and aerosols of bitumen
Fig. 3 Area sampling of vapours and aerosols of bitumen
measurements (Fig. 3). These four samples were pooled for
the analysis of the PAH compounds. Area measurements
were taken by moving the sampler following the hot
asphalt application to the respective point of the highest
anticipated exposure.
Sampling was conducted to assess shift exposure with a
median duration of measurements in mastic asphalt work-
ers of 315 min (inter-quartile range 240–410 min). There
was no correlation between the duration of sampling and
the concentration of vapours and aerosols of bitumen
(rs = -0.04, P = 0.47). The glass-fibre filter and the
XAD-2 cartridge were replaced after 2–4 h. The average
shift concentration was calculated as time-weighted aver-
age of the concentrations derived from each sample.
Determination of bitumen-derived vapours and aerosols
The samples of both personal and area measurements were
shipped to the central laboratory at IFA for the extraction
S13
of bitumen-derived matter with hydrocarbon-free tetra-
chloroethylene as solvent and then analysed by Fourier
Transform Infrared spectrometry (Thermo FTIR Avatar
370, C–H stretch region 2,800–3,000 cm-1, 10 mm quartz
glass cuvette, 169 spectrum addition, resolution 4 cm-1).
The signals were calibrated against a mineral-oil standard
to derive the concentrations of vapours and aerosols of
bitumen (Breuer 2008). A conversion to values related to a
bitumen-condensate standard can be performed by multi-
plying these concentrations with a factor of 1.47 (DFG
2011). Concentrations are presented for the mineral-oil
standard and for both calibrations where appropriate. Limit
of quantitation (LOQ) was 0.2 mg/m3 depending on the
sampled mass over the sampling period. The uncertainty
over the measurement range varied between 20 and 25%
satisfying requirements for the performance of methods for
chemical agents (EN 482 2006).
Determination of PAH compounds
The determination of PAHs in samples from 25 construc-
tion sites was carried out at Eurofins ERGO Ltd., Hamburg,
according to a previously published method of the German
Research Society (DFG 2003). In brief, subsequent to
sampling, a solution of 15 deuterated internal standards (16
EPA-PAH without acenaphthylene and fluoranthene,
additionally benzo[b]naphtho[2,1-d]thiophene) was added
to the XAD-2 cartridge. PAHs and their labelled internal
standards were extracted by soxhlet extraction with 25 ml
hexane/toluene (7:3/v:v) from the XAD-2 resin and the
filter of the sampling cartridge. The extract was reduced to
1 ml at 30°C in a stream of nitrogen and used for GC-MS
analysis. Further concentration to a volume of 100 ll
allowed achieving lower LOQs (1–5 ng per sample).
Deuterated benzo[e]pyrene was used as recovery standard.
Two ll of the extracts was injected to the analytical sys-
tem. Gas chromatographic separation was achieved on a
HP Ultra 1 column (50 m 9 0.2 mm 9 0.25 lm), and a
mass spectrometer was used in EI-SIM mode for detection.
Quantitative determination was carried out using 16
deuterated PAHs as internal standards.
According to this method, the samples were analysed for
the following 16 U.S. EPA PAHs: acenaphthene, ace-
naphthylene, anthracene, fluoranthene, fluorene, naphtha-
lene, phenanthrene, pyrene, benzo[ghi]perylene, as well as
EPA PAH compounds that were evaluated as carcinogenic
(cPAHs) comprising benzo[a]anthracene, benzo[a]pyrene
(BaP), dibenzo[a,h]anthracene and indeno[1,2,3-cd]pyrene
(Straif et al. 2005). With regard to two other carcinogenic
EPA PAH compounds, chrysene was co-eluted with
triphenylene, and benzo[b]fluoranthene, benzo[j]fluoran-
thene and benzo[k]fluoranthene could not be separated
into benzo[b]fluoranthene and benzo[k]fluoranthene. We
123
S14 Arch Toxicol (2011) 85 (Suppl 1):S11–S20

Table 1 Personal and area exposure to vapours and aerosols of bitumen in workers enrolled for the Human Bitumen Study
mg/m3 Construction workers without bitumen applicationa Bitumen-exposed workers
b b
N N\LOQ Min P25 Median P75 Max N N\LOQ Min P25 Median P75 Max

Personal sampling
Vapours and aerosols
Mineral-oil standard 69 23 LOQc LOQ 0.20 0.30 2.50 320 9 LOQ 1.80 3.46 5.90 41.68
Bitumen-condensate standard LOQ LOQ 0.29 0.44 3.67 LOQ 2.64 5.08 8.67 61.22
Aerosols
Mineral-oil standard 69 55 LOQ LOQ LOQ LOQ 0.41 320 22 LOQ 0.97 2.08 4.00 33.02
Bitumen-condensate standard LOQ LOQ LOQ LOQ 0.60 LOQ 1.42 3.05 5.87 48.51
Area sampling
Vapours and aerosols
Mineral-oil standard 13 9 LOQ LOQ LOQ 0.08 0.50 44 0 0.40 2.00 3.85 6.30 34.20
Bitumen-condensate standard LOQ LOQ LOQ 0.12 0.73 0 0.59 2.94 5.66 9.25 50.24
Aerosols
Mineral-oil standard 13 11 LOQ LOQ LOQ LOQ 0.30 44 0 0.20 1.12 2.65 4.30 29.00
Bitumen-condensate standard LOQ LOQ LOQ LOQ 0.44 0 0.29 1.64 3.89 6.32 42.60
a
Background exposure due to the lack of a specific method for the chemical complexity of bitumen emissions
b
N\LOQ, number of measurements below limit of quantitation (LOQ)
c
All measurements below LOQ

summarized the concentrations as EPA PAHs and cPAHs, Results

respectively, with and without co-eluted compounds. Fur-
ther, benzo[e]pyrene, 1-benzothiophene, dibenzothiophene
and benzo[b]naphtho[2,1-d]thiophene were determined.
Statistical methods
The distributions of exposure variables (vapours and
aerosols of bitumen, PAH compounds) were presented with
median, inter-quartile range (IQR), minimum and maxi-
mum and stratified by exposure group (with and without
exposure to mastic asphalt). For the calculations of the
summary concentrations of PAHs and regression analyses,
values below LOQ were set to ‘ LOQ. Spearman corre-
lation coefficients (rs) were calculated to describe the rank
associations between variables. For more detailed analyses
of associations, the exposure variables were log-trans-
formed due to their skewed distributions. A linear trend and
a LOESS fit of the association between area and personal
measurements of bitumen fume were estimated to describe
the changes and shape of the associations with 95% con-
fidence intervals (CI) (Cleveland et al. 1988; Cleveland and
Grosse 1991). Smoothing parameters were obtained using
the method developed by Hurvich et al. in a given range of
0.2–0.8 (Hurvich et al. 1998). A tricube weight function
was used for modelling. The local polynomials were of first
degree, and the interpolating local polynomials were cubic.
All calculations were performed with the statistical soft-
ware SAS/STAT, version 9.2 (SAS Institute Inc., Cary,
NC, USA) (PROC LOESS).
Personal exposure to vapours and aerosols of bitumen
The distributions of the concentrations of vapours and
aerosols of bitumen in 320 exposed and 69 non-exposed
workers are described in Table 1. Vapours and aerosols of
bitumen measured with personal devices resulted in a
median concentration of 3.46 (IQR 1.80–5.90) mg/m3 in
exposed workers (calibration according to the mineral-oil
standard). Median level of the background exposure was
0.20 (IQR \ LOQ-0.30) mg/m3 in the reference group. The
corresponding concentrations of aerosols were 2.08
(0.97–4.00) mg/m3 in exposed workers and mostly below
LOQ in the reference group. Aerosols of bitumen and
vapours and aerosols of bitumen were closely associated (rs
0.96; 95% CI 0.95–0.97), and the regression was linear at
log-transformed scales with an adjusted R2 of 0.92. After
re-transformation, the regression equation was presented as
yaerosol = 0.52
x1.11
vapours and aerosols. In the following, we
focus on vapours and aerosols of bitumen. Due to their
strong association, the results hold for aerosols of bitumen
as well.
Area versus personal measurements of vapours
and aerosols of bitumen
Table 1 presents also the distribution of area measurements
of vapours and aerosols of bitumen at 44 with and at 13
worksites without bitumen exposure. At most of the

123
Arch Toxicol (2011) 85 (Suppl 1):S11–S20
Fig. 4 Association between concentrations of vapours and aerosols
of bitumen from area sampling and personal sampling
construction sites of the reference group, the determined
bitumen-derived matter was below LOQ, and a maximum
background concentration of 0.50 mg/m3 was determined.
At 44 construction sites with hot bitumen application, the
median shift concentration from area sampling was 3.85
(IQR 2.00–6.30) mg/m3. Figure 4 compares the measure-
ments of vapours and aerosols of bitumen with stationary
and personal devices at construction sites with mastic
asphalt application. A LOESS model was applied in order
to evaluate the shape of this association. In the range
between 5 and 10 mg/m3, a linear relation was observed
with overall similar concentrations derived from stationary
and personal samplers. In the range below 3 mg/m3, sam-
pling with personal devices yielded higher concentrations
than area sampling. This relation is reversed above 10 mg/
m3 where area sampling resulted in higher concentrations
than personal sampling. Twenty-one workers were exposed
to concentrations above 10 mg/m3 when measuring expo-
sure with personal devices and 27 workers when assessing
exposure with area measurements.
Area measurements of PAHs
Table 2 shows the distributions of area measurements of
PAH compounds at 15 worksites with bitumen application
and at 10 construction sites without asphalt works. The
median shift concentration of U.S. EPA PAHs was 2,473
(IQR 1,847-4,289) ng/m3 at construction sites with bitumen
application and 210 (IQR 172–258) ng/m3 at construction
sites without exposure to bitumen. This summary con-
centration included few imputed data for measurements
\LOQ and the non-EPA PAHs triphenylene and
S15
benzo[j]fluoranthene that co-eluted together with the EPA
PAHs chrysene and benzo[k]fluoranthene, respectively.
The average fraction of compounds with co-elution
(chrysene and benzofluoranthenes) constituted less than
10% of the summary concentration of all EPA PAHs. The
most prominent EPA PAHs were naphthalene, followed by
phenanthrene and fluorene.
The total concentration of cPAHs was heavily influ-
enced by co-elution. At mastic asphalt worksites, the
median concentration of cPAHs was 376.5 ng/m3 with
co-eluted compounds and 107.0 ng/m3 when excluding
benzo[b]fluoranthene, benzo[j]fluoranthene and benzo[k]
fluoranthene and chrysene co-eluted with triphenylene.
Exposure to cPAH differed by bitumen exposure where the
majority of compounds could be measured in mastic
asphalt applications while the measurements were usually
below LOQ at construction sites without bitumen appli-
cation. For example, all BaP measurements were above
LOQ and presented with a median concentration of 45
(range 6–460) ng/m3 at construction sites with bitumen
exposure, while nine out of ten measurements at reference
worksites were below LOQ.
Figure 5 shows the association between concentrations
of vapours and aerosols of bitumen with the sum of U.S.
EPA PAHs from area measurements at construction sites
with mastic asphalt application. There was no significant
correlation between both groups of compounds (rs = 0.45,
95% CI -0.13 to 0.78), but the statistical power was lim-
ited by the few area measurements of PAH.
Discussion
The German Bitumen Forum has initiated this nationwide
study in workers exposed to vapours and aerosols of bitu-
men and in non-exposed construction workers to monitor
exposure to vapours and aerosols of bitumen with regard to
the former technically based German OEL and to assess
potential health effects (Rühl et al. 2006). The complex
chemical composition of bitumen makes it difficult to
develop a specific method for the analysis of emissions of
bitumen. Hence, OELs differ between countries according
to the monitoring methods (Table 3). In Germany, the
BGIA method 6305 has been specifically developed to
measure bitumen-derived matter in both the vapour and
aerosol component (Breuer 2008). When beginning this
study, the German technically based OEL of 10 mg/m3 was
not valid for mastic asphalt works, because of the higher
processing temperature compared to the application of
rolled asphalt. Our study revealed that only few mastic
asphalt workers were exposed during the shift above this
former OEL. Overall, 50% of the measurements ranged
between 2.0 and 6.3 mg/m3.
123
S16 Arch Toxicol (2011) 85 (Suppl 1):S11–S20

Table 2 Concentrations of polycyclic aromatic hydrocarbons from air sampling at construction sites with and without bitumen application in the
Human Bitumen Study
ng/m3 Construction sites without bitumen application Bitumen application
b a
N N\LOQ Min Median Max N N\LOQ Min Median Max

EPA PAHs (with co-elution)b 10 – 122 210 445 15 – 1,256 2,473 18,448
EPA PAHs (without co-elution)c 10 – 121 209 443 15 – 789 2,318 16,759
cPAHs (with co-elution)d 10 – 1.3 3.6 5.3 15 – 42.5 376.5 2,906.0
cPAHs (without co-leution)e 10 – 0.6 2.2 3.3 15 – 16.5 107.0 1,217.0
EPA PAHs
Naphthalene 10 0 53.0 146.0 403.0 15 0 172.0 1,010.0 9,480.0
Phenanthrene 10 0 11.0 28.5 38.1 15 0 95.4 267.0 1,441.0
Fluorene 10 0 6.9 17.0 20.6 15 0 67.3 183.0 1,019.0
Acenaphthene 10 0 2.6 13.0 23.0 15 0 31.0 75.4 839.0
Pyrene 10 1 \LOQ 3.70 7.43 15 0 26.6 95.4 1,017.0
Anthracene 10 4 \LOQ 1.31 17.0 15 0 16.0 46.0 502.0
Acenaphthylene 10 0 1.00 3.65 13.1 15 0 17.5 49.0 642.0
Fluoranthene 10 1 \LOQ 3.47 10.0 15 0 23.0 58.2 505.0
Benzo[ghi]perylene 10 9 \LOQ \LOQ 1.0 15 0 7.2 45.4 301.0
Carcinogenic EPA PAHs (cPAHs)
Chrysene co-eluted with triphenylene 10 9 \LOQ \LOQ 1.50 15 0 17.0 221.0 1,222.0
Benzo[b,j,k]fluoranthene 10 8 \LOQ \LOQ 1.50 15 0 9.0 69.0 467.0
Benzo[a]anthracene 10 6 \LOQ \LOQ 1.79 15 0 4.0 48.0 519.0
Benzo[a]pyrene 10 9 \LOQ \LOQ 1.00 15 0 6.0 45.0 460.0
Dibenzo[a,h]anthracene 10 10 \LOQ \LOQ \LOQ 15 1 \LOQ 17.0 118.0
Indeno[1,2,3-cd]pyrene 10 9 \LOQ \LOQ 1.00 15 0 1.5 11.5 120.0
Other compounds
Benzo[e]pyrene 1 0 – 0.40 – 5 0 23.0 99.0 223.0
1-Benzothiophene 10 3 \LOQ 5.25 37.7 15 0 8.7 81.0 2,610.0
Dibenzothiophene 10 1 \LOQ 3.17 14.0 15 0 24.5 120.0 1,290.0
Benzo[b]naphtho[2,1-d]thiophene 10 8 \LOQ \LOQ 1.77 15 0 11.0 171.0 888.0
a
N\LOQ, number of measurements below limit of quantitation (LOQ)
b
Sum of U.S. EPA PAHs including compounds with co-elution (naphthalene, acenaphthene, acenaphthylene, phenanthrene, fluorene, anthra-
cene, fluoranthene, pyrene, benzo[a]anthracene, chrysene co-eluted with triphenylene, benzo[b,j,k]fluoranthene, benzo[a]pyrene, di-
benz[a,h]anthracene, indeno[1,2,3-cd]pyrene, benzo[ghi]perylene)
c
Sum of U.S. EPA PAHs without benzo[b,j,k]fluoranthene and chrysene co-eluted with triphenylene
d
Sum of carcinogenic PAHs including compounds with co-elution (benzo[a]anthracene, chrysene co-eluted with triphenylene,
benzo[b,j,k]fluoranthene, benzo[a]pyrene, dibenz[a,h]anthracene, indeno[1,2,3-cd]pyrene)
e
Sum of carcinogenic PAHs without chrysene co-eluted with triphenylene and benzo[b,j,k]fluoranthene

So far, no health-based OEL has been derived in
countries where health assessments have been conducted,
including Germany, the United States and The Nether-
lands. In 2000, the American Conference of Governmental
Industrial Hygienists (ACGIH) recommended 0.5 mg/m3
as threshold limit value (TLV) for the shift average but
measured with a different method as benzene-soluble
matter (BSM) in the inhalable fraction of particulates. A
side-by-side comparison of the IFA method 6305 and the
NIOSH method 5042 in the field revealed a good correla-
tion (R2 = 0.88) between inhalable aerosols and benzene-
soluble matter (BSM) (Kriech et al. 2010). A rough
conversion of 3 mg/m3 as average concentration of aero-
sols calibrated against the bitumen-condensate standard
would result in about 2 mg/m3 BSM according to the
equation estimated from this field comparison.
Various differences between the BGIA method and the
NIOSH method 5042 make the comparison of the German
exposure levels against this recommend exposure limit of
ACGIH and with measurements in other studies difficult.
Differences in methods are therefore a challenge in har-
monizing exposure data in large international databases. In
an approach to assess exposure during laying of mastic
asphalt for the historical IARC cohort study in European

123
Arch Toxicol (2011) 85 (Suppl 1):S11–S20
Fig. 5 Association between concentrations of vapours and aerosols
of bitumen and polycyclic aromatic hydrocarbons from area sampling
asphalt workers (Burstyn et al. 2003), methodological
differences in sampling and analysis are apparent with the
inclusion of data from various laboratories into the AWE
database (Burstyn et al. 2003). NIOSH has presented a
comprehensive review on the methods to assess exposure
to vapours and aerosols of bitumen (NIOSH 2000). Vola-
tile compounds are emitted during hot applications of
bitumen, which constitute a complex mixture of mostly
aliphatic hydrocarbons and few other compounds,
depending on the source of the crude petroleum, the pro-
cessing temperature of bitumen and other factors. These
vapours condense during cooling and thus generating
aerosols containing bitumen-derived matters. Accordingly,
an exposure assessment has to take into consideration that
asphalt workers are exposed to both, vapours in the gas
phase and aerosols as particulate matter of small size. The
aerosols from bitumen applications are small in size, and
various devices such as GGP collect the inhalable fraction.
Larger dust particles can be trapped and may confound
such measurements (Kriech et al. 2004). The commonly
used 37-mm cassette collects less aerosols than a sampler
for the inhalable fraction (Deygout and Le 2010). The
chemical complexity of emissions of bitumen is another
methodological problem that adds to uncertainties in the
Table 3 Occupational exposure limits for average shift exposure to bitumen in selected countries and methods for monitoring bitumen exposure
Country OEL (TWA) Method
Germany At present no OEL; 10 (14.7) mg/m3 BGIA 6305
discontinued in 2005.
USA No federal OEL; BSM 0.5 mg/m3 NIOSH 5042
recommended by ACGIH.
Finland No OEL; bitumen fume have been Modified
referred to 5 mg/m3 as OEL for NIOSH 5042
oil mist
Netherlands 5 mg/m3 as ‘asphalt smoke’ Not specified
until 2006. At present no OEL.
S17
comparison of exposure data from different studies, having
used different samplers and analytical methods (Burstyn
et al. 2002a, b). A comparison of various methods applied
in personal sampling of emissions of bitumen was made
using a laboratory fume generator (Ekstrom et al. 2001).
The German method yielded lower concentrations for the
aerosol component than the US method, but higher values
for the combined concentrations of vapours and aerosols.
While gas chromatography with flame ionization is cur-
rently used in various laboratories, Fourier Transform
Infrared Spectroscopy is being employed in Germany. The
use of different solvents to extract bitumen-derived mate-
rial may add to the differences between methods (tetra-
chloroethylene in the German BGIA method, benzene in
the NIOSH method). Kriech et al. compared IFA method
6305, NIOSH method 5042 and the Heritage method for
total organic matter (TOM) in the field with side-by-side
measurements (Kriech et al. 2010). To improve the com-
parability of the methods, the German data were calibrated
according to the US standard derived from four common
bitumen types (bitumen-condensate standard). A factor of
1.47 can be applied for conversion of values calculated on
the basis of the German mineral-oil standard into corre-
sponding values on the basis of the US standard, because
bitumen condensates have a lower infrared sensitivity in
the range of 2,800–3,000 cm-1 than mineral oils (Kriech
et al. 2010). Overall, the German aerosol measurements
correlated with BSM concentrations, but were higher.
Vapours and aerosols of bitumen correlated with TOM and
were also higher. Similar concentrations were observed for
aerosols and total particulate matter.
Aerosols and vapours of bitumen were measured during
the shift with personal and stationary devices at construc-
tion sites, with and without asphalt works. Concerns have
been raised with regard to additives like rubber or waste
products (Vaananen et al. 2006). The mastic asphalt mix-
tures used in this study did not contain such additives.
Exposure concentrations followed skewed distributions
that we described with percentiles. Median concentrations
Air sampling Bitumen-derived matter
Vapours and aerosols; Extraction with tetrachloroethylene;
GGP sampler FTIR-analysis; C–H stretch region
(2,800–3,000 cm-1)
Particulate matter, 37-mm Extraction with benzene (BSM),
sampler and PTFE filter. gravimetric analysis
Total particulates; 37-mm Extraction with tetrachloroethylene,
sampler and PTFE filter. FTIR-analysis; C–H stretch region
(2,800–3,000 cm-1).
123
S18
of vapours and aerosols of bitumen were 3.46 mg/m3
measured with personal samplers in exposed workers and
3.85 mg/m3 with stationary devices. The corresponding
concentrations of aerosols were 2.08 and 2.65 mg/m3,
respectively. Influences of processing temperature, work-
site characteristics and job tasks on the exposure levels are
presented by Spickenheuer et al. (2011) in this issue. A
comparison of area and personal sampling showed similar
concentrations between 3 and 10 mg/m3. In the low-dose
range, personal sampling resulted in higher concentrations.
In contrast, area sampling yielded higher concentrations
above 10 mg/m3 because the sampler was sometimes
moved along with the laying of asphalt to the site of the
highest anticipated exposure. At construction sites where
area sampling yielded concentrations C10 mg/m3, nine out
of 320 mastic asphalt workers were above this former OEL
also with personal measurements. At construction sites
where area measurements showed concentrations \10 mg/
m3, 12 workers had exposures above OEL with personal
sampling. For monitoring OELs, personal sampling in
breathing zone should be preferred to better assess the
airborne exposure of workers. For reasons of prevention, it
is also a common practice to assess task-related exposure
scenarios (Rühl et al. 2006).
Area measurements were also taken at certain con-
struction sites without bitumen application, mainly in road
construction and fewer in the construction of special
buildings such as car parks. The specificity of method 6305
measuring C–H stretch vibration is limited. This accounts
for a low background level in non-exposed workers, which
might be due to ubiquitous aliphatic hydrocarbons. Motor
exhaust might be a potential source of organic matter
(Kriech et al. 2004). When concentrations of vapours and
aerosols of bitumen measured in this study are compared
with levels reported from other studies, differences in
sampling and analytical methods have to be taken into
account.
Chemical analyses demonstrate that bitumen consists
only of small quantities of PAH compounds (for review
NIOSH 2000). The content of carcinogenic PAHs is below
the substance-specific concentration limit of 100 mg/kg of
the European classification of mixtures as carcinogenic
(European Commission 2008). The International Agency
for Research on Cancer (IARC) considered the low content
of PAH compounds in emissions from hot applications
when classifying the carcinogenicity of bitumen fume as
Group 2B (IARC 1985). We found an average concentra-
tion of 2.5 lg/m3 EPA PAHs that represents about 0.05%
of the mass of bitumen emissions from hot mastic asphalt
applications. The correlation of EPA PAHs with the con-
centrations of vapours and aerosols from working area
measurements at the same sampling points being although
positive, lacked statistical power and significance. The
123
Arch Toxicol (2011) 85 (Suppl 1):S11–S20
poor association between vapours and aerosols of bitumen
and PAH exposure was demonstrated by observations at a
construction site where rolled and mastic asphalt were
subsequently applied as presented by Raulf-Heimsoth et al.
(2011a, b) in this issue. Here, pronounced differences in
airborne vapours and aerosols of bitumen concentrations
were not reflected in the levels of PAH metabolites.
However, a poor association was observed for the urinary
concentrations of PAH metabolites in all workers with
personal measurements of vapours and aerosols of bitumen
as shown by Pesch et al. (2011) in this issue.
This weak or even lacking association of vapours and
aerosols of bitumen with airborne PAH exposure may also
reflect methodological difficulties in determining PAHs as
reviewed by NIOSH (2000). In brief, HPLC/fluorescence
techniques according to NIOSH method 5506 can result in
uncertain data for PAHs, while GC with flame ionization
detector according to method 5515 and GC/MS yields
more meaningful data. Emissions of bitumen are mainly
composed of aliphatic hydrocarbons and their isomers, and
of aromatic compounds with two or more fused aromatic
rings in small amounts that impair a separation into single
PAH compounds. In our analysis, GC/MS was applied
according to a method where separation of chrysene from
triphenylene and separation of specific benzofluoranthenes
could not be achieved. Co-elution contributed to a small
fraction among the total sum of EPA PAHs but constituted
a large methodological shortcoming with regard to car-
cinogenic PAHs.
BaP is a commonly recognized carcinogenic compound
applied in monitoring airborne PAH exposures. At con-
struction sites with mastic asphalt applications, the BaP
concentrations ranged between 6 and 460 ng/m3 with a
median of 45 ng/m3, whereas most measurements at con-
struction sites were below LOQ and thus in the range
of environmental concentrations that are usually between
0.3 and 0.5 ng/m3 (URL: www.lanuv.nrw.de). However,
ambient air sampling in the general environment is per-
formed with higher volumes over a longer period, for
example 50 m3/24 h. The average concentration of EPA
PAHs at bitumen workplaces was 2.5 and 1.0 lg/m3 for
naphthalene as most prominent EPA PAH compound.
These concentrations were slightly higher than 1.42 and
0.75 lg/m3 reported as geometric means for EPA PAHs
and naphthalene, respectively, from personal samplers in
Finnish pavers laying polymer-modified asphalt at tem-
peratures between 145 and 165°C (Vaananen et al. 2006).
In contrast, we determined a much higher median con-
centration of 30.2 lg/m3 for EPA PAHs in coke-oven
works and other PAH settings (Marczynski et al. 2009).
This is also reflected in the urinary concentrations of PAH
metabolites in bitumen-exposed workers compared with
coke-oven workers as presented by Pesch et al. (2011) in
Arch Toxicol (2011) 85 (Suppl 1):S11–S20
this issue. A large difference in PAH metabolite concen-
trations has been also demonstrated between bitumen and
accidental tar exposure during this German study (Raulf-
Heimsoth et al. 2008). These results support the view that
current PAH exposure during mastic-asphalt applications is
not equal to levels observed during historical tar use. These
results might be of interest for the assessment of exposure
to mastic asphalt and tar in epidemiological studies
(Burstyn et al. 2000, 2002a, b). The methodological chal-
lenge in measuring PAHs in the low-dose range has criti-
cally to be taken into account.
In summary, our results support that measurements of
the bitumen-derived matter in aerosols and vapours and the
appropriate determination of carcinogenic PAH com-
pounds from hot bitumen applications remain a challenge
with regard to the complex chemical composition of bitu-
men and the low concentrations of PAHs. PAH exposure at
mastic asphalt worksite was higher than at construction
sites without bitumen application, but more than one order
of magnitude lower than in settings with well-known high
PAH exposure like coke-oven works. When concentrations
of vapours and aerosols of bitumen measured with the
method used in this study are compared with limit values or
levels reported from other studies using the NIOSH method
or other methods, differences in sampling and analytical
methods have to be taken into account. Bitumen-derived
matter should be measured in both the gas and particulate
component of the air at workplaces with hot bitumen
applications.
Acknowledgments The Human Bitumen Study was initiated by the
German Committee for Hazardous Substances (AGS) and the German
Bitumen Forum and financially supported by the German Social
Accident Insurance (DGUV), Eurobitume, Deutscher Asphaltverband
e.V., Concawe, Zentralverband des Deutschen Dachdeckerhandwerks,
Industrieverband Bitumen-, Dach- und Dichtungsbahnen e.V., Ber-
ufsgenossenschaft Rohstoffe und Chemische Industrie (BG RCI) and
Aksys GmbH, Beratungsstelle für Gussasphaltanwendungen (bga),
BG BAU and Arbeitsgemeinschaft der Bitumenindustrie (Arbit). We
thank all workers having participated in the Human Bitumen Study.
We gratefully acknowledge the support of the field team, especially
Anne Flagge, Anja Molkenthin, Bianca Wachter, Gerd Zoubek, and in
part of the study Klaus Schott ( ) and Hans-Jürgen Schicker.
References
Breuer D (2008) Bitumen (Daempfe und Aeorosole, Mineralöl-
Standard). BGIA Method 6305-1 Berlin, Erich Schmidt. BGIA
Working Folder
Burstyn I, Kromhout H, Kauppinen T, Heikkila P, Boffetta P (2000)
Statistical modelling of the determinants of historical exposure
to bitumen and polycyclic aromatic hydrocarbons among paving
workers. Ann Occup Hyg 44:43–56
Burstyn I, Boffetta P, Burr GA, Cenni A, Knecht U, Sciarra G,
Kromhout H (2002a) Validity of empirical models of exposure in
asphalt paving. Occup Environ Med 59:620–624
S19
Burstyn I, Randem B, Lien JE, Langard S, Kromhout H (2002b)
Bitumen, polycyclic aromatic hydrocarbons and vehicle exhaust:
exposure levels and controls among Norwegian asphalt workers.
Ann Occup Hyg 46:79–87
Burstyn I, Boffetta P, Kauppinen T, Heikkila P, Svane O, Partanen T,
Stucker I, Frentzel-Beyme R, Ahrens W, Merzenich H, Heederik
D, Hooiveld M, Langard S, Randem BG, Jarvholm B, Bergdahl
I, Shaham J, Ribak J, Kromhout H (2003) Estimating exposures
in the asphalt industry for an international epidemiological
cohort study of cancer risk. Am J Ind Med 43:3–17
Cleveland WS, Grosse E (1991) Computational methods for local
regression. Stat Comput 1:47–62
Cleveland WS, Devlin SS, Grosse E (1988) Regression by local
fitting. J Econ 37:87–114
Deygout F, Le CP (2010) Field sampling investigations within the
road paving industry. J Occup Environ Hyg 7:103–108
DFG (2003) Analytische Methoden zur Prüfung gesundheitsschädli-
cher Arbeitsstoffe. Luftanalysen. Polycyclicsche aromatische
Kohlenwasserstoffe (PAH). Methoden-Nr. 3 PAH. Wiley-VCH,
Weinheim
DFG (2011) The MAK Collection for Occupational Health and
Safety, Part III: Air Monitoring Methods, Bitumen (vapour and
aerosol), Vol 12. Wiley-VHC, Weinheim, pp 37–47
Ekstrom LG, Kriech A, Bowen C, Johnson S, Breuer D (2001)
International studies to compare methods for personal sampling
of bitumen fumes. J Environ Monit 3:439–445
Eller PM (1998) Analytical method 5042. Benzene solubles and total
particulate (asphalt fume). Cinicinatti, NIOSH. NIOSH manual
of analytical methods
EN 482 (2006) Workplace atmospheres—general requirements for
the performance of procedures for the measurement of chemical
agents, German version. Beuth, Berlin
European Commission (2008) Classification, labelling and packaging
of substances and mixtures. Regulation EC No. 1272/2008.
Brussels, Belgium
Hurvich CM, Simonoff JS, Tsai CL (1998) Smoothing parameter
selection in non-parametric regression using an improved Akaike
information criterion. J R Stat Soc B 60:271–293
IARC (1985) Polynuclear aromatic compounds. Bitumens. Monogr
Eval Carcinog Risks Hum 35:39–81
Kenny LC, Aitken R, Chalmers C, Fabries JF, Gonzalez-Fernandez E,
Kromhout H, Liden G, Mark D, Riediger G, Prodi V (1997) A
collaborative European study of personal inhalable aerosol
sampler performance. Ann Occup Hyg 41:135–153
Kriech AJ, Osborn LV, Wissel HL, Kurek JT, Sweeney BJ, Peregrine
CJ (2004) Total versus inhalable sampler comparison study for
the determination of asphalt fume exposures within the road
paving industry. J Environ Monit 6:827–833
Kriech AJ, Emmel C, Osborn LV, Breuer D, Redman AP, Hoeber D,
Bochmann F, Ruehl R (2010) Side-by-side comparison of field
monitoring methods for hot bitumen emission exposures: the
German IFA Method 6305, U.S. NIOSH Method 5042, and the
Total Organic Matter Method. J Occup Environ Hyg 7:712–725
Marczynski B, Pesch B, Wilhelm M, Rossbach B, Preuss R, Hahn JU,
Rabstein S, Raulf-Heimsoth M, Seidel A, Rihs HP, Adams A,
Scherenberg M, Erkes A, Engelhardt B, Straif K, Kafferlein HU,
Angerer J, Bruning T (2009) Occupational exposure to polycy-
clic aromatic hydrocarbons and DNA damage by industry: a
nationwide study in Germany. Arch Toxicol 83:947–957
NIOSH (2000) Health effects of occupational exposure to asphalt.
NIOSH Cincinnati (Hazard Review)
Pesch B, Spickenheuer A, Kendzia B, Schindler BK, Welge P,
Marczynski B, Rihs H-P, Raulf-Heimsoth M, Angerer J, Brüning
T (2011) Urinary metabolites of polycyclic aromatic hydrocar-
bons in workers exposed to vapours and aerosols of bitumen.
Arch Toxicol (for submission)
123
S20
Raulf-Heimsoth M, Angerer J, Pesch B, Marczynski B, Hahn JU,
Spickenheuer A, Preuss R, Rühl R, Rode P, Bruning T (2008)
Biological monitoring as a useful tool for the detection of a coal-
tar contamination in bitumen-exposed workers. J Toxicol Envi-
ron Health A 71:746–750
Raulf-Heimsoth M, Marczynski B, Spickenheuer A, Pesch B, Welge
P, Rühl R, Bramer R, Kendzia B, Heinze E, Angerer J, Brüning
T (2011) Bitumen workers handling mastic versus rolled asphalt
in a tunnel—assessment of exposure and biomarkers of irritation
and genotoxicity. Arch Toxicol (for submission)
Raulf-Heimsoth M, Pesch B, Rühl R, Brüning T (2011) The human
bitumen study: executive summary. Arch Toxicol (for
submission)
Rühl R, Musanke U, Kolmsee K, Priess R, Zoubek G, Breuer D
(2006) Vapours and aerosols of bitumen: exposure data obtained
by the German Bitumen Forum. Ann Occup Hyg 50:459–468
123
Arch Toxicol (2011) 85 (Suppl 1):S11–S20
Spickenheuer A, Rühl R, Höber D, Raulf-Heimsoth M, Marczynski B,
Welge P, Breuer D, Gabriel S, Musanke U, Rode P, Heinze E,
Kendzia B, Bramer R, Knecht U, Hahn J-U, Brüning T, Pesch B
(2011) Levels and determinants of exposure to vapours and
aerosols of bitumen. Arch Toxicol (for submission)
Straif K, Baan R, Grosse Y, Secretan B, El Ghissassi F, Cogliano V
(2005) Carcinogenicity of polycyclic aromatic hydrocarbons.
Lancet Oncol 6:931–932
Vaananen V, Elovaara E, Nykyri E, Santonen T, Heikkila P (2006)
Road pavers’ occupational exposure to asphalt containing waste
plastic and tall oil pitch. J Environ Monit 8:89–99