Atmospheric Environment 34 (2000) 4639}4647

Determination of VOC-components in the exhaust of gasoline

and diesel passenger cars
Thomas Schmitz *, Dieter Hassel
, Franz-Josef Weber

Institut fu( r Chemie der Belasteten Atmospha( re (ICG-2), Forschungszentrum Ju( lich, 52425 Ju( lich, Germany

TU$ V Rheinland Sicherheit und Umweltschutz GmbH, Testzentrum Energietechnik, Am Grauen Stein, 51105 Ko( ln, Germany
Received 3 September 1999; accepted 14 June 2000

Abstract
The composition of VOC emissions from in-use passenger cars with di!erent engine types, i.e. cars with diesel engines,
cars with petrol engines equipped with three-way-catalysts, and cars with petrol engines without catalysts was
determined. Five cars of each engine type have been measured on a chassis dynamometer under conditions of the US
FTP 75 test procedure and the `Autobahna test developed by TUG V Rheinland. Measurements of C }C hydrocarbons
 
were made with a GC-FID system. In addition, samples on DNPH cartridges were taken and analysed by means of
a HPLC-system for the determination of aldehydes and ketones.The in#uence of cold/warm-conditions on the VOC
composition was determined. In the case of cars with diesel engines as well as for the petrol-driven cars without exhaust
treatment, the e!ect caused by the cold start only led to minor changes in the VOC composition. A similar behaviour was
observed for these car types at higher speeds. In contrast to the cars without catalysts, the cars with three-way-catalysts
showed a great variability of the VOC composition. During the cold start phase the aromatic compounds and the alkenes
yielded the main fraction of the VOC. During the warm phase the less reactive alkanes were predominant. With
increasing mean velocities the VOC composition changed in favour of the more reactive compounds.  2000 Published
by Elsevier Science Ltd.
Keywords: Passenger car emissions; C }C hydrocarbons; Aldehydes; Ketones; Ozone forming potential
 

1. Introduction
The volatile organic compounds (VOC) in the atmosphere comprise many di!erent species with various
properties. Some of these compounds directly in#uence
human health due to their toxicity. VOC are also important precursors of photochemically formed secondary
pollutants, such as ozone, which poses a serious air
pollution problem under speci"c summertime conditions. Since the individual VOC react with di!erent rates
and di!erent mechanisms, they also di!er in their contribution to photochemical ozone formation. For the development of e$cient abatement strategies to lower the
VOC emissions in order to improve air quality in urban

* Corresponding author.
E-mail address: th.schmitz@fz-juelich.de (T. Schmitz).

areas with regard to direct toxic e!ects or photochemical

air pollution, reliable data of VOC are required.
The emission factors of the total VOC from road tra$c
are known (e.g. De Vlieger, 1996; Hassel et al., 1994;
Lenaers, 1996), but for typical European in-use passenger
cars there are only few data concerning the VOC composition (e.g. Bailey et al., 1990; Hoekman, 1992; Jemma
et al., 1995; Rijkeboer and Hendriksen, 1993; Siegl et al.,
1999). The data cited above are based on measurements
of individual cars.
Another approach to obtain emission data from road
tra$c is to perform measurements at roadsides or in
street tunnels (e.g. Zweidinger et al., 1988; Haszpra and
Szilagyi, 1994; Pierson et al., 1996; Staehelin et al., 1998).
The main issue of these studies was the evaluation of
existing emission factor models for speci"c tra$c situations rather then the extension of those models by
emission factors of individual VOC.

1352-2310/00/$ - see front matter  2000 Published by Elsevier Science Ltd.

PII: S 1 3 5 2 - 2 3 1 0 ( 0 0 ) 0 0 3 0 3 - 4

4640

T. Schmitz et al. / Atmospheric Environment 34 (2000) 4639}4647

In recent years new methods based on chassis dynamometer studies for determining emission data of motor
vehicles have been developed (Hassel et al., 1994). From
these data the dependence of the emissions of VOC, CO
and NO on the driving behaviour is known. In the
V
present work the dependence of the VOC-composition
on di!erent driving situations was investigated. Also the
in#uence of cold-/warm-conditions was quanti"ed.
2. Experimental
2.1. Vehicles, fuels and driving conditions
The measuring programme included 15 di!erent passenger cars. Three engine types were investigated: petrol
driven cars equipped with a fuel injection system and
a closed-looped controlled three-way-catalyst (TWC),
petrol-driven cars without any catalyst and cars with
diesel engines. Five cars of each engine type were measured. Three of the cars with diesel engines were equipped
with oxidative catalytic converters, one with a direct fuel
injection.
Because of the correlation between exhaust emissions
of individual hydrocarbons and fuel composition it was
necessary to use test fuels which are representative for the
average composition of German market fuels. The test
fuels, two unleaded fuels of di!erent quality for the catalyst cars and the cars without catalyst and diesel fuel,
were obtained from the ARAL research centre in
Bochum (Germany). The characteristics of the fuels used
in this programme are given in Table 1.
The emission measurements were performed on the
chassis dynamometer of TUG V Rheinland. A detailed description of the test facility is given in Hassel et al. (1994).
The measurements were carried out under the conditions
of the United States federal test procedure (US FTP 75)
and the `Autobahna test presented in Fig. 1.
The FTP test is divided into three phases. The test is
starting with a cold start. For this cold start procedure
the cars have been conditioned at a temperature of 203C

over a time period of 12 h. In the "rst test phase an

average speed of 41 km h\ is driven. In the second test
phase the average speed is only 26 km h\. After the
second test phase the engine is shut o! for 10 min and
then the test restarts with a warm start. The driving
pattern of this third phase is identical with the driving
pattern in the "rst phase. By comparing the emissions
obtained in the "rst and in the third test phase, the e!ects
caused by the cold start procedure can be derived from
the data.
The `Autobahna cycle was developed by TUG V Rheinland. This cycle represents driving behaviour on German
motorways. Again this cycle consists of three phases.
From phase one to phase three the average velocity is
increasing. In the third phase of the test cycle the cars
have to accelerate from 120 km h\ to the maximum
speed of 162 km h\ in about 60 s. Under these conditions most of the cars tested in this programme were
running at their limits. The emissions which were
observed under these conditions "gure out the situation
for high engine loads which are not covered in the US
FTP 75.
2.2. Sampling and analysis
The exhaust gas is diluted by a factor of 10}15 with
"ltered ambient air using a constant volume sampler
(CVS) device. The concentrations measured in the exhaust gas are corrected by the values measured in the
dilution air which is sampled simultaneously. For the
measurement of CO, NO , CO and the total hydrocarV

bons (THC), diluted exhaust gas is sampled in Tedlar
bags over the time periods of the individual test phases.
CO and CO are measured by means of non-dispersive

infrared spectroscopy. In case of NO a chemiluminesV
cence detector and for THC a #ame ionisation detector
(FID) are used. Calibration of these components was
achieved using di!erent gas standards (supplied from
Linde AG, Germany). The FID was calibrated with
a propane-standard.

Table 1
Characteristics of the test fuels, analysis done by ARAL (Germany)
Fuel quality

Density (153C)
(g ml\)

ROZ/cetan
number

MOZ

Lead
(mg l\)

Sulphur
(mass %)

Super-Plus
Super
Diesel

0.7614
0.7574
0.8304

99.3/
96.4/
/51.8

88.1
85.1

1
1

0.0064
0.0124
0.0325

Super-Plus
Super

Benzene
(vol %)
0.8
1.9

Toluene
(vol %)
14.0
10.3

Methanol
(vol %)
0.0
0.0

Aromatics
(vol %)
46.0
45.8

MTBE
(vol %)
8.79
2.74

Para$ns
(vol %)
41.8
46.9

Ole"ns
(vol %)
3.2
4.3

T. Schmitz et al. / Atmospheric Environment 34 (2000) 4639}4647

4641

Fig. 1. Driving cycles of the measuring programme.

For the measurement of the C }C hydrocarbons

 
a gas chromatographic system with a #ame ionisation
detector (HP 5890A) and a specially designed sampling
device was used (Schmitz et al., 1997). The sampled air
passes a cooling trap at !253C for removal of water
before being sucked through a sample loop, which is
kept at about !1903C. When sampling is completed,
the sample loop is heated up to 803C and the hydrocarbons are injected on the capillary column (DB-1;
90 m;0.32 mm ID, "lm thickness: 3 lm). After injection
the column is kept isothermal at !503C for 2 min
and then heated up to 2003C with a rate of 53C/min.
Subsequently, the column temperature is kept at
2003C for 15 min. A chromatogram taken from a car
with TWC in the "rst phase of the US FTP 75 is shown in
Fig. 2.
The diluted exhaust gas which is taken for the analysis
of the hydrocarbons is sampled in glass vessels, because

tests with Tedlar bags showed substantial blank values

for some hydrocarbons (Schmitz et al., 1999).
The GC-system was calibrated using a gas standard
containing n-butane (810 ppb), benzene (439 ppb) toluene
(416 ppb) and o-xylene (270 ppb) (British Oxygen Company (BOC)). This standard mixture is diluted dynamically with nitrogen (99.999% purity) to mixing ratios of
4}13 ppb, equivalent to the mixing ratios measured in the
diluted exhaust gas. The response factors of the individual hydrocarbons were calculated on a per molecular
weight basis.
Aldehydes and ketones were trapped on dinitrophenylhydrazine (DNPH) cartridges. The hydrazones
formed on the cartridge were analysed by means of highperformance liquid chromatography and detected with
a diode array detector.
The determination of phenol and the sum of o-, m- and
p-cresols was performed by means of gas chromatography.

4642

T. Schmitz et al. / Atmospheric Environment 34 (2000) 4639}4647

Fig. 2. Chromatogram taken from vehicle with TWC: 1"ethene; 2"ethyne; 3"ethane; 4"propene; 5"propane; 6"i-butane;
7/8"i-/1-butene; 9"1,3-butadiene; 10"n-butane; 11"trans-2-butene; 12"cis-2-butene; 13"i-pentane (coelution with acetone);
14"1-pentene; 15"n-pentane; 16"isoprene; 17"trans-2-pentene; 18"cis-2-pentene; 19"2,2-dimethylbutane; 20/21"cyclopentane/2,3-dimethybutane; 22"2-methylpentane; 23"3-methylpentane; 24"1-hexene; 25"n-hexane; 26"2,4-dimethyl-pentane;
27"benzene; 28"cyclohexane; 29"2-methylhexane; 30"2,3}dimethylpentane; 31"3-methylhexan; 32"i-octane; 33"n-heptane; 34"methylcyclohexane; 35"toluene; 36"2}methylheptane; 37"3-methylheptane; 38"n-octane; 39"ethylbenzene;
40/41"m/p-xylene; 42"styrene; 43"o-xylene; 44"n-nonane; 45"n-propylbenzene; 46"1,3,5-trimethyl-benzene; 47"1,2,4trimethylbenzene; 48"n-decane.

Samples were taken on Silicagel-adsorption tubes which

then were thermally desorbed and analysed by GC-MS
(column used: HP-MS-5 25 m, ID 0.25 mm; "lm thickness
0.25 lm).
For both methods calibration was achieved using
liquid standards which were injected into the analytical
system.
In Table 2 the analytical methods are listed. For more
analytical details see Schmitz et al. (1999).

3. Results
3.1. Regulated emissions
The mean emissions of CO, THC, and NO obtained
V
for the single test phases are presented in Fig. 3. As it is
expected, the highest emissions were measured for the
cars without catalysts (Hassel et al., 1994; Lies et al.,
1988). For CO and THC the di!erences between diesel

engines and catalyst cars in comparison to the cars without catalyst span a factor of 10}100. NO emissions from
V
cars with diesel engines and cars without catalysts di!er
by a factor of "ve in all test phases, in case of the catalyst
cars the di!erences are between a factor of 15}20.
Looking into detail two points have to be noted: the
e!ect of cold start conditions, which can be derived by
comparison of the results obtained in phase 1 and phase
3 of the US test, and the e!ect of high engine loads, which
is best represented in the third phase of the `Autobahna
cycle.
In general, the relative increase of emissions caused by
the cold start process is less pronounced for cars with
diesel engines and cars without catalysts than for catalyst
cars. For TWC cars the highest increase is observed for
the hydrocarbon emissions. For cold start conditions the
THC emissions are higher by about a factor of 10 compared to warm start conditions.
In the `Autobahna cycle for all engine types the increase of NO emissions is about 25% between the
V

T. Schmitz et al. / Atmospheric Environment 34 (2000) 4639}4647

4643

Table 2
Measured components and detection methods
Component
CO
CO

THC
NO
V
CH
L K
CH O
L K

Measuring method
Carbon monoxide
Carbon dioxide
Total hydrocarbons
Nitrogen oxides (NO,NO )

C }C hydrocarbons
 
Carbonyls; phenol and cresols

second and third phase. In case of CO and THC emissions the behaviour is comparable to that for cars without catalyst and cars with diesel engines. In contrast, for
the TWC cars the emissions of CO increase by a factor of
6 and for THC the emissions are higher by a factor of 4.
The observed emission behaviour of TWC vehicles
indicates that in case of CO and THC the conversion
e$ciency of the catalyst strongly depends on the operating conditions of the engine. During the cold phase the
catalytic converter does not reach its optimum work and
the emission reduction is diminished.
Under driving conditions with high mean speeds and
high engine loads the conversion e$ciency is low too,
although the catalyst is warmed up completely. For the
acceleration phases of the third phase of the `Autobahna
cycle rich air/fuel mixtures are injected. Under these
engine operating conditions the oxidation capacity of the
catalytic converter for CO and the hydrocarbons is diminished. For NO a decreasing e$ciency of the catalyst at
V
conditions with high engine loads does not occur due to
the fact that the reducing conditions in the catalyst promote the conversion of NO .
V
3.2. VOC-composition
The relative composition of the VOC measured for the
di!erent engine types is shown in Fig. 4. In order to get
a better overview the single compounds were grouped
into "ve substance classes. The alkynes ethyne and
1-butyne are listed together with the alkenes.
The relative VOC pattern which is obtained for the
TWC cars shows strong variations in dependence of
the di!erent test phases. Under cold start conditions the
VOC pattern is dominated by the fraction of the aromatic compounds followed by the alkanes and the
alkenes. The amount of carbonyls and phenols is less
than 5%. In the corresponding warm phase (US FTP 75
phase 3) the VOC pattern is changing completely. The
more reactive alkenes and aromatic compounds are reduced more e$ciently. Therefore, the main fraction of the
VOC is given by the alkanes, a fact which also is reported
in an Australian study (Du!y et al., 1999). Looking

Non-dispersive infrared
Non-dispersive infrared
Flame ionization detector
Chemiluminescence detector
Gas chromatography/#ame ionization
High-pressure liquid chromatography/diode
array detector; gas chromatography/mass
spectroscopy

towards higher average velocity, the fraction of alkenes

and aromatics is increasing continuously. In the last
phase of the `Autobahna cycle the aromatic fraction
nearly reaches 60% of the VOC pattern. This again is due
to the e!ect of rich air/fuel mixtures which are injected
under high load conditions existing in the `Autobahna
cycle. Therefore, conversion e$ciency of the catalyst for
the hydrocarbons is diminished as described above. This
e!ect leads to an increase in THC emission (see Fig. 3)
and to a change in the VOC pattern especially in the
third phase of the test cycle.
For the cars without exhaust treatment the changes in
the VOC-pattern are small. The pattern measured in the
cold start phase is nearly identical to that obtained in the
warm phase. Over the whole speed range phenols, carbonyls and the aromatics are nearly constant. Only at
higher average speed a slight increase for the alkenes and
hence a slight decrease of the alkane compounds is observed.
In the case of the cars with diesel engines alkenes and
carbonyl compounds yield the main fraction of the VOC.
A similar VOC fraction in the exhaust gas of in-use
light-duty diesel vehicles was found in the emission data
represented by Rijkeboer and Hendriksen, 1993, and
Siegl et al., 1999. The di!erence between the cold and
warm start conditions is very small. The fraction of the
alkanes and the aromatics is nearly constant over the
single phases. Only for the alkenes and the carbonyls
there are some minor changes.
The measured mass emissions in mg km\ of the VOC
grouped in di!erent substance classes and the mean con"dence levels in % calculated for the small number
of vehicles (5 cars of each engine type) are given in
Table 3(a) and (b). The detailed results for the individual
compounds are given in Schmitz et al. (1999).
The detection limits for the measurements of carbonyl
compounds and for the measurements of phenol and the
cresols is an order of magnitude higher than that for the
measurements of the individual hydrocarbons. Therefore,
the second and the third test phase of the `Autobahna
cycle had to be measured together in order to reach the
detection limits. Nevertheless, for some of the tested

4644

T. Schmitz et al. / Atmospheric Environment 34 (2000) 4639}4647

Fig. 3. Emissions of CO, THC and NO for the di!erent vehicle

V
categories.

vehicles, especially for the cars with TWC the measured

values were below detection limit. Then the value of the
detection limit was used for the calculation of mean
emissions as an upper limit. In these cases no con"dence
levels were calculated.
The mean values of the individually measured VOC
show large con"dence levels of about 80%. In the US
FTP 75 the con"dence levels are smaller than that calculated for the `Autobahna cycle. The con"dence intervals which were found for the mean values of the TWC
&&&&&&&&&&&&&&&&&&&&&&&&&&&

Fig. 4. VOC pattern obtained for the three vehicle categories in
the di!erent test cycles.

T. Schmitz et al. / Atmospheric Environment 34 (2000) 4639}4647

4645

Table 3
Emission rates of the VOC in mg km\ measured in the di!erent phases
Substance

US 75
phase 1
mean values

Con"dence
level (%)

US 75
phase 2
mean values

Con"dence
level (%)

US 75
phase 3
mean values

Con"dence
level (%)

(a) US FTP 75
Cars with TWC
Alkanes
Alkenes
Alkynes
Aromatics
Carbonyls
Phenol and cresols

130.20
62.63
21.00
250.81
12.26
1.66

74.25
90.11
105.45
81.76
*
*

5.47
0.38
0.04
2.92
3.59
1.54

92.99
110.02
93.38
82.40
*
*

18.79
1.89
0.08
10.23
2.96
1.66

52.09
95.47
108.33
87.68
*
*

Cars without TWC

Alkanes
Alkenes
Alkynes
Aromatics
Carbonyls
Phenol and cresols

661.87
344.27
139.12
1266.70
90.42
6.05

36.86
42.71
38.90
35.47
*
46.58

650.88
341.32
100.90
1138.44
113.58
15.80

65.32
48.21
43.75
45.07
*
65.02

567.60
291.71
77.32
975.62
95.08
11.52

63.24
50.97
30.41
43.71
*
43.12

Diesel cars
Alkanes
Alkenes
Alkynes
Aromatics
Carbonyls
Phenol and cresols

8.23
44.35
8.83
9.14
24.36
1.61

83.91
60.79
79.21
63.07
*
*

7.87
35.41
5.94
7.49
25.67
1.49

95.54
67.47
90.33
59.52
*
*

5.16
20.23
4.15
4.56
17.07
1.59

80.85
80.82
94.54
61.72
*
*

Substance

Autobahn
phase 1
mean values

Con"dence
level (%)

Autobahn
phase 2
mean values

Con"dence
level (%)

Authobahn
phase 3
mean values

Con"dence
level (%)

(b) Autobahn cycle

Cars with TWC
Alkanes
Alkenes
Alkynes
Aromatics
Carbonyls
Phenol and cresols

18.02
4.16
0.12
16.68
3.06
1.22

102.43
107.70
89.86
95.61
*
*

11.81
7.49
0.21
18.76

80.50
106.75
154.47
105.40

28.81
60.71
0.94
129.48
3.73
0.77

90.13
124.74
133.41
92.32

Cars without TWC

Alkanes
Alkenes
Alkynes
Aromatics
Carbonyls
Phenol and Cresols

214.96
292.54
86.34
580.16
67.75
9.58

51.13
65.08
40.77
54.53
*
29.86

259.14
366.57
118.60
723.96

48.90
86.52
46.02
66.16

224.48
502.86
175.92
846.70
40.61
13.17

67.12
90.20
43.98
90.51

Diesel cars
Alkanes
Alkenes
Alkynes
Aromatics
Carbonyls
Phenol and cresols

1.57
6.48
1.66
1.86
8.23
1.11

99.67
132.72
141.41
94.16
*
*

1.46
8.11
2.32
2.16

1.22
9.35
4.34
3.02
7.55
0.71

133.94
135.08
141.79
150.48

106.19
120.51
136.09
102.35

25.96

The second and the third test phase of the `Autobahna cycle were measured as one sample in order to reach the detection limits of the
methods.

4646

T. Schmitz et al. / Atmospheric Environment 34 (2000) 4639}4647

cars and cars with diesel engines are signi"cantly higher

than that for cars without exhaust gas treatment. The
large scatter is due to di!erent reasons: a small number of
cars of each engine type, di!erences in the mass emissions
due to vehicle age, standard of maintenance and type of
manufacture. It was found that the sum of the identi"ed
VOC and the corresponding THC measurements agree
better than 20% (Schmitz et al., 1999). Therefore, a normalisation of the VOC composition to the measured
THC was performed resulting in smaller con"dence intervals than for the mass emissions. In principle, the
normalised VOC patterns can be adapted to existing
emission factors of THC which are available for a large
number of driving situations. However, it has to be carefully studied to which extend the driving patterns of the
investigated test phases are representative for real-word
driving.

Fig. 5. Ozone formation potential as function of mean speed in

di!erent test phases. Note the logarithmic scale. The "tted lines
only show the trend between OFP and mean velocity.

3.3. Ozone formation potential (OFP)

In order to assess the measured VOC emissions concerning the photochemical ozone production, the ozone
formation potential (OFP) using the method by Carter
(1994) was calculated for the di!erent engine types. This
method is based on a model scenario in which ozone
formation is calculated under optimum conditions: i.e.
high solar #uxes and a base scenario with VOC/NO V
ratios which yield a maximum ozone formation. For
single hydrocarbons a maximum incremental reactivity
(MIR) can be determined in terms of grams ozone formed
per gram hydrocarbon added to the scenario. The ozone
forming potential of a certain VOC mixture is calculated
by summing up the products of measured VOC concentrations and corresponding MIR factors

emissions of reactive hydrocarbons. The e$ciency of the

catalyst decreases with higher average velocities. At an
average speed of 140 km h\ there is an increase of the
ozone formation potential of a factor of 200 compared to
the OFP values calculated for low average speed.
In contrast, the OFP calculated for the emissions
of the diesel cars decreases with increasing speed. The
highest OFP values are observed at moderate average
speed, whereas lowest OFP values are found in the Autobahn cycle. This dependency is caused by the fact that the
mass emissions of diesel cars are continuously decreasing
with increasing mean speed (cf. Fig. 3) without major
changes in the reactivity of the VOC composition.

OFP" VOC MIR .

(1)
G
G
G
One important point has to be noted: these conditions
represent typical California conditions and are signi"cantly di!erent from typical summertime conditions
in Central Europe. Nevertheless, for the ranking of
VOC mixtures with regard to ozone production these
MIR-factors are very useful.
When calculating the ozone forming potential for the
5 phases with warmed-up engines a speed dependency
can be derived (see Fig. 5). As expected the highest ozone
formation potential is observed for cars without catalyst
due to the highest mass emission of THC in connection
with a high share of reactive substances like alkenes and
aromatics.
The ozone formation potential for the TWC cars is
found to be up to two orders of magnitude lower than for
cars without catalyst and is increasing with increasing
speed. The low values in the ozone formation potential at
moderate speeds are caused by the high e$ciency of the
catalysts resulting in low mass emissions of THC and low

4. Summary and conclusions

The VOC compositions of three di!erent engine types
were measured under conditions of the US FTP 75 and
the `Autobahna-test cycle. The variations of the VOC
compositions in the di!erent phases observed for the
gasoline driven cars without catalysts and the diesel cars
are moderate. For the cars with three-way-catalyst there
is a strong dependency of the VOC composition on the
operating conditions of the engine and the temperature
of the catalyst. There is a large change in the VOCpattern from cold start phase to the phases with
warmed-up engine. A signi"cant change can be observed
at high engine loads and high speeds compared to moderate driving conditions with warmed-up engines.
The resulting ozone formation potential of the VOC
emissions shows a dependency with respect to the
average speed. Especially, for the cars with three-waycatalysts the speed dependency of the ozone formation
potential has to be taken into account. In this context, it
should be stressed that there is a lack of information with

T. Schmitz et al. / Atmospheric Environment 34 (2000) 4639}4647

regard to tra$c situations like stop and go or other

emission relevant parameters like road inclination.
Using the emission results of this study it should be
mentioned that the measurements are based on mean
market fuels in Germany of the reference year 1997. It is
known that VOC composition and therefore OFP will
change (Decker et al., 1996) when fuel composition is
signi"cantly modi"ed (for example, according to the new
EU Fuel Directive).

Acknowledgements
This work is supported by BMBF as part of the `Forschungsschwerpunkt TroposphaK renforschunga (TFS),
principal subject 2: `Anthropogenic and biogenic emissions as well as deposition of trace gasesa in the course of
the joint project A.2: `Analysis, development and evaluation of emission models and data sets for road tra$c in
Germany and Europea. The authors thank the ARAL
research centre for providing and for the detailed analysis
of the test fuels. They acknowledge Dr. U. Metzner
for the analysis of the carbonyl compounds and
Dr. G. Kuropka for the measurements of phenol and
cresols. They also thank Dr. D. Klemp for the critical
review of the manuscript.

References
Bailey, J.C., Schmidl, B., Williams, M.L., 1990. Speciated hydrocarbon emissions from vehicles operated over the normal
speed range on the road. Atmospheric Environment 24A,
43}52.
Carter, W.P.L., 1994. Development of ozone reactivity scales for
volatile organic compounds. Journal of Air Waste Management Association 44, 881}899.
De Vlieger, I., 1996. On-board emission and fuel consumption
measurement campaign on petrol-driven passenger cars. Atmospheric Environment 31, 3753}3761.
Decker, G., Beyersdorf, J., Schulze, J., Wegener, R., Weidmann,
K., 1996. Das Ozonbildungspotential unterschiedlicher
Fahrzeug- und Kraftsto!konzepte. Automobiltechnische
Zeitschrift 98, 280}289.
Du!y, B.L., Nelson, P.F., Ye, Y., Weeks, I.A., 1999. Speciated
hydrocarbon pro"les and calculated reactivities of exhaust
and evaporative emissions from in-use light-duty Australian
vehicles. Atmospheric Environment 33, 291}307.
Hassel, D., Jost, P., Weber, F.-J., Dursbeck, K., Sonnborn, S.,
Plettau, D., 1994. Abgasemissionsfaktoren von Pkw in der
Bundesrepublik Deutschland. In: Umwelt Bundesamt (Ed.),
Abgasemissionen von Fahrzeugen der Baujahre 1986 bis
1990. Erich Schmidt Verlag, Berlin.

4647

Haszpra, L., Szilagyi, I., 1994. Non-methane hydrocarbon composition of car exhaust in Hungary. Atmospheric Environment 28, 2609}2614.
Hoekman, S.K., 1992. Speciated measurements and calculated
reactivities of vehicle exhaust emissions from conventional
and reformulated gasolines. Environmental Science and
Technology 26, 1206}1216.
Jemma, C.A., Shore, P.R., Widdicombe, K.A., 1995. Analysis of
C }C hydrocarbons using dual-column capillary GC: ap 
plication to exhaust emissions from passenger car and motor
cycle engines. Journal of Chromatographic Science 33,
34}38.
Lenaers, G., 1996. On-board real life emission measurements on
a 3 way catalyst gasoline car in motor way-, rural- and city
tra$c and on two Euro1- diesel city buses. Science of Total
Environment 189/190, 139}147.
Lies, K.H., Schulze, J., Winnecke, H., Kuhler, M., Kraft, J.,
Hartung, A., Postulka, A., Gring, H., SchroK ter, D., 1988.
Nicht limitierte Automobil-Abgaskomponenten. Studie der
Volkswagen AG.
Pierson, W.R., Gertler, A.W., Robinson, N.F., Sagebiel, J.C.,
Zielinska, B., Bishop, G.A., Stedman, D.H., Zweidinger, R.B.,
Ray, W.D., 1996. Real-world automotive emissions }
Summary of studies in the Fort McHenry and Tuscarora
mountain tunnels. Atmospheric Environment 30,
2233}2256.
Rijkeboer, R.C., Hendriksen, P., 1993. Regulated and unregulated exhaust gas components from LD vehicles on
petrol, diesel, LPG and CNG. TNO-report 93.OR.VM:029/
1/PHE/PR TNO. Road Vehicles Research Institute, Delft,
the Netherlands.
Schmitz, Th., Hassel D., Weber F.-J., 1999. Zusammensetzung
der Kohlenwassersto!e im Abgas unterschiedlicher Fahrzeugkonzepte. Berichte des Forschungszentrum JuK lich,
JUG L-3646, ISSN 0944}2952, JuK lich.
Schmitz, Th., Klemp, D., Kley, D., 1997. Messung der Immissionskonzentrationen verschiedener OzonvorlaK ufersubstanzen in Ballungsgebieten und an Autobahnen } Charakterisierung der EmissionsverhaK ltnisse des Stra{enverkehrs
unter verschiedenen Verkehrssituationen durch Messungen
in QuellnaK he. Berichte des Forschungszentrums JuK lich
JUG L-3457, ISSN 0944}2952, JuK lich.
Siegl, W.O., Hammerle, R.H., Herrmann, H.M., Wenclawiak,
B.W., Luers-Jongen, B., 1999. Organic emissions pro"le for
a light-duty diesel vehicle. Atmospheric Environment 33,
797}805.
Staehelin, J., Keller, C., Stahel, W., SchlaK pfer, K., Wunderli, S.,
1998. Emission factors from road tra$c from a tunnel study
(Gubrist tunnel, Switzerland). Part iii: results of organic
compounds, SO and speciation of organic exhaust emis
sions. Atmospheric Environment 32, 999}1009.
Zweidinger, R.B., Sigsby, J.E., Tejada, S.B., Stump, F.D.,
Dropkin, D.L., Ray, W.D., Duncan, J.W., 1988. Detailed
hydrocarbon and aldehyde mobile source emissions from
roadway studies. Environmental Science and Technology 22,
956}962.