Joint EUCAR/JRC/CONCAWE Study on:

Effects of Gasoline Vapour Pressure and

Ethanol Content on Evaporative
Emissions from Modern Cars

Edited by G. Martini

Institute for Environment and Sustainability

2007

EUR 22713 EN
The mission of the Institute for Environment and Sustainability is to provide scientific-technical
support to the European Union’s Policies for the protection the environment and sustainable
development of the European and global environment.

European Commission
Directorate-General Joint Research Centre
Institute for Environment and Sustainability

Contact information
Address:Giorgio Martini – T.P. 441- Via Fermi, 1 – 21020 Ispra (Italy)
E-mail: [email protected]
Tel.: +39 0332 789293
Fax: +39 0332 785869

http://ies.jrc.cec.eu.int
http://www.jrc.cec.eu.int

Legal Notice
Neither the European Commission nor any person acting on behalf of the Commission is
responsible for the use which might be made of this publication.

A great deal of additional information on the European Union is available on the Internet.
It can be accessed through the Europa server
http://europa.eu.int

EUR 22713 EN
ISBN 978-92-79-05249-1
ISSN 1018-5593
Luxembourg: Office for Official Publications of the European Communities

© European Communities, 2007

Reproduction is authorised provided the source is acknowledged

Printed in Italy
 Joint EUCAR/JRC/CONCAWE Programme on:

Effects of gasoline vapour pressure and

ethanol content on evaporative emissions
from modern cars

Authors:
G. Martini, U. Manfredi, G. Mellios, V. Mahieu, B. Larsen,
A. Farfaletti, A. Krasenbrink, G. De Santi
European Commission – DG Joint Research Centre, Institute for
Environment and Sustainability

S. McArragher, N. Thompson, J. Baro, P. J. Zemroch, J. Rogerson,

J. Farenback-Brateman, J. Canovas, I. Dijs, K. Rose, R. Bazzani
CONCAWE

F. Boggio, A. Celasco, C. Cucchi, G.F. (Barry) Cahill

EUCAR

Acknowledgments
The authors would like to acknowledge the help and input from others who advised on
the work, specifically T. Gameson (eBIO) and W. Mirabella (EFOA).
In particular the authors would like to acknowledge the essential contribution of the staff
of the JRC VELA laboratory: P. Le Lijour. G. Lanappe, M. Sculati, R. Colombo.
Special thank to V. Forcina and M. Douane for the analysis of unregulated pollutants.
INDEX

1. EXECUTIVE SUMMARY
2. BACKGROUND
3. OBJECTIVES
4. SCOPE OF TEST PROGRAMME
4.1. Test vehicles
4.2. Fuels
4.3. Programme structure
4.4. Test protocol
4.5. Issues not covered
5. TEST RESULTS
5.1. Test facility correlation (Phase 1)
5.2. Main test programme (Phase 2-3)
5.2.1. Total Evaporative Emissions
5.2.2. Evaporative VOCs Speciation and Ethanol Content
5.2.3. Exhaust Emissions and Fuel Consumption
5.2.4. Aldehyde Emissions
5.3. Extra tests
5.3.1. Tests performed with the canister vented out of the VT SHED
5.3.2. Consecutive multiple tests on the same fuel
6. DISCUSSION OF THE RESULTS
6.1. Evaporative emissions
6.2. Results of the extra tests
6.3. Effect of ethanol on canister efficiency
6.4. VOC Speciation data
7. CONCLUSIONS FROM EUCAR/CONCAWE/JRC EVAPORATIVE EMISSIONS TEST
PROGRAMME
8. REFERENCES

APPENDIX 1 - EVAPORATIVE EMISSIONS: OVERVIEW

APPENDIX 2 – FUEL ANALYTICAL DATA
APPENDIX 3 – DETAILS OF TEST PROCEDURE
APPENDIX 4 – EVAPORATIVE EMISSIONS AND TEST PROTOCOL RELATED EFFECTS
APPENDIX 5 – EVAPORATIVE EMISSIONS VOC SPECIATION AND ETHANOL
CONTENT: ANALYTICAL METHOD
APPENDIX 6 – EXHAUST EMISSIONS RESULTS
APPENDIX 7 – ALDEHYDE EMISSIONS: ANALYTICAL METHOD AND STATISTICAL
ANALYSIS

2
List of Special Terms and Abbreviations

ACEA European Automobile Manufacturers Association

ASTM American Society for Testing and Materials
BWC Butane Working Capacity of activated carbon
CARB California Air Resource Board
CEN European Committee for Standardization
CO Carbon Monoxide
CO2 Carbon Dioxide
CONCAWE The Oil Companies’ European Association for Environment, Health and
Safety in Refining and Distribution
DG Directorate General
DI Direct Injection
DISI Direct Injection Spark Ignition
DVPE Dry Vapour Pressure Equivalent at a temperature of 37.8 °C (kPa)
EC European Commission
EMS Engine Management System
EN European standard issued by CEN
EPA Environmental Protection Agency (USA)
EPEFE European Programme on Emissions, Fuels and Engine Technologies
ETBE Ethyl Tertiary Butyl Ether
EU European Union
EUCAR European Council for Automotive R & D
EUDC Extra Urban Driving Cycle = Part 2 of the NEDC driving cycle
EURO # European emission standard
FID Flame ionization detector
GC Gas chromatography
GC-MS Gas chromatography mass spectrometry
kPa 1 KiloPascal = 1000 N/m2
HC Hydrocarbon
HPLC High performance liquid chromatography
IES Institute of Environment and Sustainability
IPCS The international programme on chemical safety
IR Infrared red spectroscopy
JRC Joint Research Centre
LD Light Duty
LHV Lower Heating Value
MPI Multi Point Injection
MS Mass spectrometry
MTBE Methyl Tertiary Butyl Ether
m/z Mass to charge ratio for detected ions in mass spectrometry
NEDC New European Driving Cycle = Type I test
NOx Oxides of Nitrogen
PC Passenger Car
RVP Reid Vapour Pressure
Scan mode When operated in the scan mode the MSD collects data by scanning and
detecting through a range of ions
TAEE Tertiary Amyl Ethyl Ether

3
Type I Type of emission test as laid down in the Directive 70/220/CEE and
subsequent amendments
UDC Urban Driving Cycle= Part 1 of the NEDC driving Cycle
US The United States of America
VELA Vehicles Emission Laboratory
VOC Volatile Organic Compound
VT SHED Variable Temperature Sealed House for Evaporative Determination
WCOT Wall Coated Open Tubular column, a type of capillary column

4
1. EXECUTIVE SUMMARY

CONCAWE, EUCAR and the Joint Research Centre of the European Commission jointly
carried out a major test programme specifically designed to investigate the influence of
gasoline vapour pressure and ethanol content on evaporative emissions from modern
passenger cars as determined using the current European regulatory test procedure.
Breathing losses through the tank vent and fuel permeation are in general the most important
sources of evaporative emissions in a vehicle. Breathing losses are due to evaporation of
gasoline in the tank during driving, hot soak and normal diurnal temperature variation. In
current vehicles vapour emissions are controlled by means of an activated carbon canister
connected to the fuel tank vent to the atmosphere. Fuel permeation can also occur through
plastic and rubber components of the fuel system.

Evaporative emissions depend mainly on ambient temperature, fuel volatility and fuel system
design [8]. As far as the fuel is concerned, Vapour Pressure (Dry Vapour Pressure Equivalent
- DVPE) is a key parameter: higher values of DVPE, which is measured at 37.8 °C, mean that
the fuel is more volatile or, in other words, that more fuel will evaporate at that temperature.
When blended with gasoline, ethanol is known to cause an increase of the ethanol/gasoline
blend DVPE compared to the base gasoline.

As a consequence of the European policy aiming to promote the use of ethanol and other
biofuels [21] [22], the question has arisen whether the vapour pressure limit for
ethanol/gasoline blend should be relaxed with respect to the current fuel specifications laid
down in the Directive 98/70/EC.
The programme described in this report was conceived to assess specifically the effect of
ethanol/gasoline blends on evaporative emissions with the objective of providing a technical
basis for discussion on this issue.

A range of seven gasoline passenger cars representative of current EURO 3-4 technology
were tested for evaporative emissions with ten different test fuels. The test fuel matrix
comprised 60 and 70 kPa hydrocarbon base fuels with 5 and 10% ethanol splash blends and 5
and 10% ethanol matched volatility blends. The evaporative emission tests were carried out
according to a test protocol agreed by the partners of the programme and based on the
European homologation test procedure.

This test procedure turned out to have a considerable influence on the results; in particular
repeating the test procedure, without any additional conditioning between tests above that
prescribed by the legislation, resulted in an increase of the carbon canister weight with
successive tests. In other words, the test protocol was not able to return the vehicle to a
consistent condition at the start of each test. The increase of the canister weight, which is
indicative of the accumulation of gasoline/ethanol vapours probably due to insufficient
canister purging, may not represent real-world operating conditions as the canister load
depends on the typical driving patterns. Running at these increased canister weights is a very
severe test of the evaporative control system. The canister weight problem made it more
difficult both to obtain representative emission measurements from the various vehicles and
to clearly determine fuel effects. However the programme has provided valuable information
and several clear conclusions can still be drawn from the results.

5
The vehicles tested differed in their level of evaporative emissions and in the extent of their
response to fuel changes. All cars met the 2 g/test emission limit on the first test on fuel A, the
reference fuel with DVPE of 60 kPa. Some vehicles slightly exceeded the limit on subsequent
tests on fuel A, probably related to increased canister loading in later tests.

The test results confirmed that vapour pressure (DVPE) is a key fuel variable for evaporative
emissions; in general, increasing fuel vapour pressure above that of the 60 kPa DVPE
reference fuel used for system development increased evaporative emissions. However the
effect of vapour pressure is strongly non-linear, as expected for a process in which a vapour
breakthrough effect may occur. The ethanol blends with final DVPE around 75 kPa gave
considerably higher evaporative emissions than the other lower volatility fuels in most of the
vehicles. Differences between fuels with DVPE in the range 60-70 kPa were small.
Furthermore, due to the combination of DVPE variations, the presence or absence of ethanol,
and to significant changes of canister weight it is difficult to draw any reliable conclusions on
the influence of individual parameters. The results obtained in a few tests where extra purging
of the canister was carried out suggest that differences in evaporative emission measurements
on fuels in this volatility range could be reduced if a more extensive canister conditioning
procedure was adopted. The engineering margin built into the system may also explain the
reduced fuel effect. The evaporative emission control system is designed for the DVPE of the
reference fuel (60 kPa) used in the homologation test but, as for other emission control
devices, the manufacturer introduces a certain margin to take into account the production
variability.

Ethanol might influence evaporative emissions also via different mechanisms than the
increased vapour pressure of ethanol/gasoline blends [19]. For example ethanol is known to
be more difficult to purge from carbon canisters (as are heavy hydrocarbons), so could reduce
their working capacity. To explore this possibility ethanol/gasoline blends matching the
vapour pressure of the pure hydrocarbon base fuel were included in the fuel matrix. The
increase in canister loading noticed during the programme could be due to increased
adsorption of hydrocarbons or ethanol in the canister, or both. Unfortunately the poor
repeatability of the main data set does not allow us to quantify the relative size of these
effects. However multiple additional tests on one vehicle showed that ethanol containing fuels
with matched volatility gave higher emissions than the hydrocarbon fuels. Ethanol was also
found in the VT SHED vapour of tests on pure hydrocarbon fuels following use of ethanol-
containing fuels. Further research would be required to clarify these effects.

Extra diurnal emission tests were carried out on two vehicles with the canister vented outside
of the VT SHED. The results of these tests suggest that fuel permeation through plastics and
rubbers could be a significant contributor to evaporative emissions. As demonstrated by other
studies, ethanol does increase the fuel permeation rate [3] [19] [20].

Measurements of regulated exhaust emissions showed few statistically significant differences

between fuels. However, the test programme was not designed to look at exhaust emissions
and this could explain the noticeable variability of the exhaust emission measurements.
Specific conditioning between tests and multiple testing is essential for such studies.
Nevertheless the data show clearly that volumetric fuel consumption (litres/100 km) increased
with increasing ethanol content. This increase was roughly proportional to the oxygen content
of the fuel. However there was no effect of ethanol on energy consumption.

6
The test programme was designed to explore only the effects of ethanol and fuel vapour
pressure on evaporative emissions from a range of latest generation canister-equipped
gasoline cars using the EU Evaporative Emissions test procedure. Other parameters like test
temperature profile, presence of ethers in the fuel, fuel permeation and the long term effect of
ethanol and water on carbon canister working capacity have not been addressed by this
programme.

7
2. BACKGROUND

One of the measures envisaged by the EU policies in the field of energy and transport is the
promotion of biofuels usage in order to increase their market penetration in response to the
need for energy supply security and implementation of the Kyoto protocol [21] [22]. In
particular, the Directive 2003/30/EC has set a target market share for biofuels, to be achieved
by 2010, of 5.75% in terms of energy content. Ethanol and biodiesel are at the moment the
only biofuels available in sufficient quantity that can allow the achievement of this target.

The EU Directive 2003/17/EC, in addition to provisions on gasoline and diesel fuel maximum
sulphur content in 2005 and beyond, requires the European Commission to review a number
of other fuel specifications for possible amendments. One specific requirement is to assess the
current gasoline summer vapour pressure limits with respect to ethanol directly blended into
gasoline. Ethanol is known to increase fuel DVPE when blended with gasoline [2]. The
increase of DVPE is roughly constant at ~7 kPa for ethanol contents between 2% and 10%
(see Fig. 1, [23]). The fuel Directive 98/70/EC defines gasoline volatility classes and their
vapour pressure limits. Each European country applies one or more volatility classes
depending on its climate and on the season, and all gasoline, including gasoline/ethanol
blends, must comply with the relevant DVPE limits. A vapour pressure waiver for
gasoline/ethanol blends has been proposed in order to facilitate the spread of ethanol usage
and consequently to increase its market penetration. Ethanol is normally distributed separately
to gasoline, and only blended at the terminal into road tankers for final distribution.
However, there is concern about the possible consequences of the increased vapour pressure
of the ethanol/gasoline blends on evaporative emissions from gasoline cars.

Figure 1: Vapour pressure (DVPE) increase of ethanol/gasoline blends as a function of

ethanol content [23]
Ethanol/gasoline blends: vapour pressure increase vs ethanol content

25

20

15
Delta P (kPa)

10

-5

-10
0 2 4 6 8 10 12 14 16 18 20
Ethanol (% vol.)

This is not a new issue, as a similar debate has already taken place in the USA especially in
California. Some experimental and theoretical studies [2] [3] [4] [5] [6] [19] [20] to address

8
this issue have been carried out in the USA but those results cannot be easily extended to
Europe for a number of reasons (different vehicles, differences in fuel system materials,
evaporative emissions control systems, etc.).

In Europe, most of the data available on evaporative emissions had been obtained in studies
carried out in the late 1980’s on “uncontrolled” vehicles and early model carbon canister-
equipped cars [24] mainly using conventional fuels. For this reason CONCAWE, EUCAR
and the DG-JRC decided to investigate the influence of vapour pressure and ethanol content
on evaporative emissions with a range of the current generation vehicle technologies. This
study provides technical input to broader analyses of the potential air quality impact of
ethanol fuels that will provide guidance to the Commission.

3. OBJECTIVES

The objectives of the joint CONCAWE/EUCAR/JRC programme on evaporative emissions

were the following:

o To assess the effects of ethanol content and vapour pressure on evaporative

emissions as determined using the current European test procedure from a range of
latest generation canister-equipped gasoline cars.

o To provide a technical basis for discussion on gasoline vapour pressure limits in

relation to ethanol blending for the Fuels Directive Review.

9
4. SCOPE OF TEST PROGRAMME

The test programme was designed to investigate the influence of gasoline vapour pressure and
ethanol content on evaporative emissions from gasoline passenger car models marketed in
Europe. Evaporative emissions were measured according to the European legislative test
procedure; therefore in this report the term “evaporative emissions” refers to the sum of all
the Volatile Organic Compounds (VOCs), not deriving from fuel combustion, emitted by the
vehicle during the Hot Soak and the Diurnal test as defined in the related European legislation
(Directive 98/69/EC, Annex IV). An overview of sources and control of evaporative
emissions from vehicles and of the related European legislation is given in Appendix 1.

A representative range of vehicles covering 8 cars, some of which were provided by the
European Auto manufacturers and others hired, was tested over a fuel matrix consisting of
fuels differing in ethanol content and vapour pressure (DVPE) provided by CONCAWE.

The programme was divided into three different phases.

In Phase 1 a vehicle was tested with the specific objective of correlating evaporative
emissions measured in the JRC VT SHED with an existing VT SHED owned by a European
car manufacturer. In this case the tests were performed using only a certified reference fuel
and just following the legislative procedure.

After this first phase, seven cars were tested for evaporative emissions on the different test
fuels according to an agreed procedure (Phase 2 and Phase 3, as described in section 4.3). The
details of the test procedure are given in Appendix 3.

4.1. Test vehicles

The test fleet included conventional Multi-Point Injection (MPI) vehicles and one Direct
Injection Spark Ignition (DISI), with both fuel return to tank and returnless systems. Six
vehicles had plastic fuel tanks and one was equipped with a metal tank.
Test vehicle data are shown in Table 1: all are modern European vehicles meeting either
Euro 3 or Euro 4 emission limits. Engine size varied from 1.2 to 3.0 litres and only two of the
vehicles had a fuel return line. Most of the vehicles were supplied directly by ACEA, but two
cars were rented locally and vehicle 6 with metal tank was also rented.
For six vehicles the carbon canister used for the tests was the one present on the vehicle at the
time of delivery. In one case (vehicle 4), the carbon canister originally installed on the vehicle
did not work properly and was replaced with a new one (original OEM spare part). The
vehicle was then driven on the road for more than 3000 km before starting the tests.

10
Table 1: Test Vehicle data
Vehicle No. 1 2 3 4 5 6 7
Emission Std Euro3 Euro4 Euro4 Euro4 Euro4 Euro4 Euro4
(homologation)
Engine Size (litres) 1.8 1.4 1.8 1.2 2.0 1.8 3.0
Fuel System MPI MPI DISI MPI MPI MPI MPI
Fuel Return Yes No No No No No Yes
Fuel tank material. Plastic Plastic Plastic Plastic Plastic Metal Plastic
Year of Registration 2003 2003 2004 2002 2004 2005 2004
or Production
Mileage (km) 2640 10127 6973 61453 10566 4500 20002

4.2. Fuels

The test fuel matrix was provided by CONCAWE and comprised 60 and 70 kPa hydrocarbon
base fuels with 5 and 10% ethanol splash blends and matched volatility blends.
The two hydrocarbon base fuels with vapour pressures of 60 (fuel A) and 70 (fuel B) kPa
represent (A) the current standard European summer grade gasoline and (B) summer grade
gasoline specified in regions with “arctic conditions”. The composition of these two fuels was
kept similar, the main difference being use of more butane in fuel B to increase the DVPE, as
would be normal refinery practice. No oxygenated compounds were used to blend these fuels.

Synthetic ethanol at 5 and 10% v/v was splash blended into fuels A and B to make fuels A5S,
A10S, B5S and B10S. The ethanol used was denatured with cyclohexane.

Ethanol blends with DVPE matched to the 60 kPa (A5E and A10E) and 70 kPa (B5E and
B10E) hydrocarbon base fuels were also included in the fuel matrix to investigate if there was
an effect of ethanol at constant vapour pressure. Distillation properties E70, E100 etc. were
allowed to move upwards to reflect what a real-world refinery would do, but in a manner not
to exceed the EN228 values. Fuel composition and properties other than distillation were
matched as closely as possible, but in all cases within the EN 228 specifications. Essentially
the “E-fuels” were based on the hydrocarbon fuels with butane and C5 molecules removed to
control DVPE.

A summary of key properties is given in Table 2. Most data are averages of test results from
three laboratories. (ethanol and oxygen are averages from 2 laboratories while LHV is based
on the HC speciation at 1 laboratory). Distillation curves and more detailed test data (Table
A2.1) are given in Appendix 2.

In addition the detailed chemical composition of the fuels was determined by Gas
Chromatography. This data is also given in Appendix 2, Table A2.2 and was used to calculate
the carbon, hydrogen and oxygen content of the fuels. This information is required to
calculate fuel consumption from CO2 emission data, see section 5.2.3. In addition the Lower
Heating value (LHV) was calculated so that energy consumption could be calculated. Note
that the GC calculated oxygen contents used to determine fuel consumption and LHV (shown
in lower half of Table 2 and Table A2.2) differ slightly from those calculated from measured
ethanol contents (shown upper half of Table 2 and Table A2.1).

11
Table 2: Fuels Inspection data

Fuel Unit A A5E A10E A5S A10S B B5E B10E B5S B10S
DVPE kPa 60.1 59.7 59.9 67.1 66.8 69.0 69.9 66.5 75.4 75.6
E70 %v/v 38.3 40.2 44.6 42.7 51.8 38.9 42.0 46.3 44.0 53.1
E100 %v/v 54.7 61.3 54.8 56.6 59.4 54.8 61.8 58.0 56.8 60.0
Ethanol %v/v 0.0 5.0 10.4 4.7 9.7 0.0 5.4 9.9 5.0 10.2
Oxygen %m/m 0.00 1.84 3.79 1.71 3.52 0.00 1.97 3.64 1.83 3.70
3
Density kg/m 755.5 747.1 756.0 757.2 758.7 753.3 747.1 750.0 754.3 756.0
Carbon %m/m 87.48 85.09 83.50 85.75 83.99 87.39 84.90 83.38 85.58 83.74
Hydrogen %m/m 12.50 13.08 12.75 12.52 12.51 12.55 13.10 12.98 12.56 12.61
Oxygen %m/m 0.01 1.82 3.74 1.72 3.49 0.06 2.00 3.63 1.86 3.64
LHV MJ/kg 42.80 42.29 41.25 42.05 41.24 42.81 42.23 41.41 42.00 41.22

4.3. Programme structure

The evaporative emissions tests were performed at the EC DG-JRC VELA laboratories
located in Ispra (Italy).

Due to the amount of testing and the fact that only two tests per week were possible, the
programme was run in three phases.

Phase 1: Shakedown and correlation of the newly commissioned JRC VT SHED with an
existing one owned by a car manufacturer and currently used for vehicle homologation. One
vehicle (same model as test car no. 2 in Table 1 above) supplied by ACEA was used only for
that purpose. This car was not used for the testing programme. The evaporative emission
tests, performed initially at the car manufacturer’s laboratory and then at the JRC, were done
using the same certification fuel.

Phase 2: The tests were carried out according to the scheme shown in Figure 2.
Four vehicles as two pairs, were initially tested on fuels A (Test A1) and B (Test B1) then
A5S and B5S, followed by a repeat of A (A2). The subsequent tests were decided jointly by
the consortium on the basis of an agreed test protocol, based on the results of the first five
tests.

Phase 3: This was originally intended as a repetition of the Phase 2 on the remaining vehicles,
unless the partners of the project decided to amend either the test sequence or the test
procedure after review of Phase 2 data. Eventually Phase 3 was performed on the three
remaining vehicles following the same general scheme as in Phase 2.

12
Figure 2: Flow-chart of the test sequence

TEST A1&B1

TEST A5S & B5S

TEST A2

YES
Is A2>1.25A1 or <0.8A1 ? Arrange phone conf to discuss
and decide next steps

NO
Is A5S >1.25*A1 or
B5S > 1.25*B1? NO

YES TEST A10S & B10S

TEST A5E & B5E

TEST A3
YES
Is A3>1.25A1 or <0.8A1 ? Arrange phone conf to discuss
and decide next steps
TEST A3

Is A10S > 1.25*A2 or

B10S > 1,25*B1 ?

NO
YES TEST A10E & B10E

TEST A4 End

4.4. Test protocol

Each vehicle/fuel combination was tested using the current European regulatory evaporative
HC emissions VT SHED test procedure (see Directive 98-69-EC Annex VI, p. 27). A flow-
chart and details of the procedure is given in Appendix 3. The current procedure comprises
the following steps:

• Load canister to breakthrough with butane.

• Drain and refill fuel tank with test fuel
• Preconditioning Drive (NEDC + one further EUDC)
• Soak for 12 – 36 hours
• Evaporative Emission System conditioning drive (NEDC + one further UDC)
• Hot soak test for 1 hour at 20 -30 ºC
• Soak for 6 – 36 hours

13
 • 24 hour Diurnal test with 20 - 35 ºC diurnal variation.

For this programme the hot soak temperature range was controlled more tightly, at a constant
value of 27 ºC over the whole test. The procedure allows two alternative ways to load the
canister, with butane or with gasoline vapour. For this programme the butane loading
procedure was used. The full details of the test procedure used in this programme are reported
in Appendix 3.

Each vehicle was first tested on fuel A, which is equivalent to the EU evaporative emissions
test reference fuel RF-02-03 (apart from sulphur content), to ensure compliance with the EU
standard.

The following measurements were made for each test:

o VT SHED VOC mass Hot Soak and Diurnal emissions

o Speciation of VT SHED emissions including ethanol content (see Appendix 5 for
details on the analytical methods used )
o Record of canister weight changes where possible (see Appendix 4 for details)
o Record of gasoline temperature during hot soak and diurnal test (except for vehicle
7)
o Exhaust emissions measured during Evap system conditioning drive (see Appendix
6)
o Exhaust HC speciation and Aldehyde emissions.

No additional conditioning cycle was run between tests, in order to follow the EU evaporative
emission test procedure exactly. This was originally thought unnecessary as the focus was on
Evaporative Emissions, and the canister “load to breakthrough” part of the test was expected
to ensure constant canister starting weight for each test.

4.5. Issues not covered

There are a number of issues not covered by the programme as it was designed:

o Fuel Tank Permeation: the legislative procedure does not allow splitting the
evaporative emissions into breathing losses and fuel permeation contributions.
o Influence of canister working capacity and canister purging strategy.
o The programme was not designed to investigate ether effects, so Bio-Ethers (i.e. bio-
ETBE and bio-TAEE) were not included in the programme and were not present in
the test fuels.
o Testing at different temperatures: this test programme was run according to the
European legislative procedure that prescribes a temperature profile having a
minimum and maximum temperature respectively of 20 °C and 35 °C. This profile
can be considered representative, to a certain extent, only of summer temperature in
southern Europe.
o Longer-term effects of ethanol and water on the working capacity of carbon
canisters: it is understood that ethanol and water may be preferentially adsorbed, then
remain in the canister to reduce its working capacity. This programme was not
designed to investigate this aspect.

14
5. TEST RESULTS

5.1. Test facility correlation (Phase 1)

As already described, in Phase 1 a vehicle was tested in the JRC laboratory and in another
ACEA member’s lab in order to correlate the newly commissioned JRC VT SHED with an
existing one. The scope of this exercise was limited to checking the correct functioning of the
JRC’s SHED. The results are shown in Table 3 and confirm good correlation of JRC’s test
results with those of a recognised vehicle certification laboratory.

Table 3: Correlation between the JRC VT SHED and a second laboratory VT SHED

HOT SOAK DIURNAL TEST TOTAL

Test Number HC g/test HC g/test HC g/test
1 0.031 0.360 0.391
2 0.015 0.371 0.386
JRC VELA 3 0.027 0.296 0.324
Laboratory 4 0.035 0.381 0.416
Average 0.027 0.352 0.379
StdDev 0.0088 0.038 0.039
CV % 33% 11% 10%

1 0.016 0.349 0.365

Car 2 0.043 0.280 0.323
Manufacturer 3 0.025 0.380 0.405
laboratory Average 0.028 0.336 0.364
StdDev 0.014 0.051 0.041
CV % 49% 15% 11%

Difference % -4% 5% 4%

5.2. Main Test Programme (Phase 2-3)

Limited experimental resources (each evaporative emission test takes 3 days), lack of prior
knowledge about the likely responses of the various cars to fuel changes, and concerns about
carryover effects meant that an adaptive exploratory test sequence was adopted rather than a
statistically-designed experiment with the same fuels repeat tested in random order in each
car. As described in Section 4.3 fuel A was used as a reference and tested on several
occasions during each vehicle test sequence. During the course of the programme it emerged
that for most vehicles results on fuel A increased with time, and there were large changes in
canister weight from test to test. These variations were of such a magnitude, and the test order
so non random, that averaging repeat results on the various fuels in any particular vehicle
would be misleading. Therefore it has been decided to report the raw emission measurements
and to plot their evolution in time.
Statistical techniques, such as multiple regression analysis, have been tried in conjunction
with various visualisations to separate fuel and canister weight effects. However multiple
regression models did not explain the variations in emissions in the various cars in a clear
and consistent way and so will not be reported here.

15
Statistical outlier tests were also used to identify possible aberrant test results. During the test
programme, some technical problems occurred (e.g. interruption of the test due to VT SHED
equipment malfunctioning) resulting in some invalid tests that were repeated. In a few cases
the technical problem was judged not to significantly affect the test result, so it was retained.
In one case (vehicle 5, Fuel B) the FID analyzer measuring the VOC concentration in the VT
SHED became saturated because of a very high value. The result has been included in Figures
3, 4 and 5 although the evaporative emissions measured in that tests are clearly
underestimated. All these tests are clearly marked in the charts by comments.

5.2.1. Total Evaporative Emissions

A bar chart showing the time series of total evaporative emissions results has been generated
for each vehicle, as shown in Figure 3. In addition, the weight of the carbon canister recorded
at the start of each test procedure, after butane loading, has been plotted as a line
superimposed on the evaporative emissions bar chart. Total evaporative emissions are also
shown as a function of fuel DVPE in Figure 4.

The first test carried out with Fuel A almost always gave the lowest evaporative emissions
value and generally started at the lowest canister weight, even after loading to breakthrough
with butane. This is believed to be due to vehicles having a well-purged canister after normal
vehicle operation on the road prior to the first test (no specific vehicle pre-conditioning was
performed prior to the first test beyond the one prescribed by the evaporative emissions test
procedure).

The line referred to canister weight after butane loading clearly shows an upward trend for
most of the vehicles tested. This behaviour of the canister, most probably due to the
accumulation of hydrocarbon/ethanol vapours in the activated carbon, was not expected and
raised the question whether to change the test procedure and introduce additional driving
between tests. This was originally thought unnecessary, as the canister “load to breakthrough”
part of the test should ensure constant conditions. However as testing progressed it became
clear that canister weights, including weight after butane loading to breakthrough, were not
constant but increased with test number. Generally canister weight increased rapidly during
the first few tests then reached a plateau or increased more gradually. The extent of the
increase varied between vehicles, possibly due to differences in the vehicle canister purging
strategy. A few tests were run with additional purging of the canister between tests (using dry
air) to a constant initial weight, marked as “Extpg” in Figures 3 and 4. However after some
discussion it was decided to complete the test programme with the originally agreed
procedure, so that all results would be comparable. Starting the test at such high canister
weights is in fact a very severe test of the vehicles’ evaporative control system. The sharp
weight decrease that can be observed in the charts of vehicles 3 and 4 after the second test on
Fuel A is due to this extra-purging procedure. The situation is however different for vehicle 5:
after the completion of the planned tests it was decided to run more tests after fully purging
the canister with hot air to dryness, as described in Section 5.3.2.

16
Figure 3: Time series of evaporative emissions and canister weights
Bars: Total Evaporative Emissions (g/test; left-hand axis)
Line: Canister weight at start of test after loading to breakthrough with butane (g; right-hand
axis)
Note: For clarity, different scales are used for each vehicle.

Vehicle 1 Vehicle 2

4.5 940 1.6 920

4.0 930 1.4
910
3.5 920 1.2
3.0 900
910 1.0
g/test

2.5

g/test
900

g
0.8 890

g
2.0
890 0.6
1.5
880
1.0 880
0.4
0.5 870 870
0.2
0.0 860
0.0 860
A

A5S

B5S

A10S

B10S

A10E

B10E

B10E

A5E

B5E

B
A B B A5S B5S A A A10S B10S A A10S B10S A

Fuel Fuel

Vehicle 3 Vehicle 4

6.0 670 9.0 740

660 8.0 720
5.0
7.0 700
650 6.0
4.0 680
g/test

5.0
g/test

640
660

g
3.0
g

4.0
630 640
2.0 3.0
620 2.0 620
1.0 610 1.0 600
0.0 580
0.0 600
A

A5S

B5S

A5S

B5S

B5S

A5E

B5E

B10E

A10E

A10S

B10S

B5S
A B A5S B5S A A B5S B B5S A5E B5E A A5S B5S
Ext pg Ext pg Ext pg Ext pg Ext pg
Fuel Fuel

Vehicle 5 Vehicle 6

9.0 720 1.6 720

8.0 710 1.4 710
7.0 700 700
690 1.2
6.0 690
680 1.0
g/test

5.0
g/test

670 680
g

0.8

g
4.0 670
660
3.0 650 0.6
660
2.0 640 0.4 650
1.0 630 0.2 640
0.0 620
0.0 630
A

A5S

B5S

A5E

B5E

B10S

B10E

A5S

B5S

A B A5S A5S B5S A A5S A10S A10S B10S A5E B5E A

Intruptd FID Aborted Aborted
Fuel Fuel

Vehicle 7

1.6 2720
1.4 2715 Key:

1.2 2710 ExtPg: Extra purging carried out prior to test

1.0 2705
g/test

Intruptd: Diurnal test interupted 30 mins before the

0.8 2700
g

scheduled end
0.6 2695
0.4 2690 FID: FID analyzer saturated, result invalid
0.2 2685 Aborted: Test aborted due to technical problems
0.0 2680
A A B B A5S B5S A5S B10S A
Fuel

17
Figure 4: Total evaporative emissions vs fuel DVPE
Note: For clarity, different scales are used for each vehicle

Vehicle 1 Vehicle 2

1.6
4.5 1.4
4.0 A
1.2 A
3.5 A5E A5S
3.0 1.0
A10E

g/test
A10S
g/test

2.5 A5S 0.8

B
2.0 A10S 0.6 B5S
1.5 1st B 1st
0.4 B10S
1.0 B5E
0.5 0.2
B10E
0.0 0.0
B5S
55 60 65 70 75 80 B10S 55 60 65 70 75 80
DVPE, kPa DVPE, kPa

Vehicle 3 Vehicle 4

6.0 9.0
8.0
5.0 A
A 7.0 ExtPg A5E
4.0 A5E 6.0 A10E

g/test
g/test

A5S 5.0 A5S

3.0 ExtPg
B 4.0 A10S
2.0 ExtPg ExtPg B5E 3.0 B
1st ExtPg
B5S 2.0 1st B5E
1.0
1.0 B10E
0.0 0.0 B5S
55 60 65 70 75 80 55 60 65 70 75 80 B10S
DVPE, kPa DVPE, kPa

Vehicle 5 Vehicle 6

9.0 1.6
8.0 FID A 1.4 A
7.0 A5E 1.2 A5E
6.0 A5S A5S
1.0
g/test
g/test

5.0 B A10S
0.8
4.0 B5E B
B10E 0.6 B5E
3.0 Intrupt'd
B5S 0.4 1st B5S
2.0
1.0 1st B10S 0.2 B10S
0.0 0.0
55 60 65 70 75 80 55 60 65 70 75 80
DVPE, kPa DVPE, kPa

Vehicle 7

1.6
Key:
1.4
1.2 1st: First test on fuel A
A
1.0 A5S
g/test

ExtPg: Extra purging carried out prior to test

0.8 1st B
0.6 B5S Intruptd: Diurnal test interupted 30 mins before the
0.4 B10S scheduled end

0.2 FID: FID analyzer saturated, result invalid

0.0
55 60 65 70 75 80
DVPE, kPa

18
 5.2.2. Evaporative VOCs Speciation and Ethanol Content

Besides the measurement of the mass of evaporative emissions performed following the
legislative procedure, the composition (speciation) of evaporative emissions was analysed via
gas-chromatography. A sample of the internal atmosphere of the VT SHED was taken at the
beginning and at the end of the diurnal test and then analysed by means of a suitable gas-
chromatograph coupled with FID detectors (the detailed analytical method is described in
Appendix 5).

Evaporative emission speciation is not available for all the tests; in some cases it was not
possible to perform the analysis due to failures or unavailability of the instrument.

The results of the VOC speciation analyses from all test vehicles are shown in Fig. 5.
In this plot the identified VOCs are aggregated in the following way:
- C4- Alkanes : all the saturated VOCs having 4 or less atoms of Carbon
- C4- Unsaturates: includes olefins, dienes and alkynes having 4 or less atoms of Carbon
- C5+ Alkanes: all the saturated VOCs having more than 4 atoms of Carbon
- C5+ Unsaturates: includes olefins, dienes and alkynes having more than 4 atoms of Carbon
- Benzene
- Aromatics: all the higher aromatic VOCs, Benzene excluded.
- Ethanol

Speciated hydrocarbon emissions from the VT SHED Diurnal tests generally show relatively
high levels of light hydrocarbons (C3 – C5) and low levels of ethanol. Moreover, once
ethanol has been used, it appears in the vapour of all subsequent tests.

Fig. 5 – VOC speciation of evaporative emissions

Fig.5a – Vehicle 1
VEHICLE 1

4.5
Ethanol
Benzene
4.0
Aromatics
C5+ Unsaturates
3.5
C5+ Alkanes
C4- Unsaturates
3.0 C4- Alkanes
VOC (grams)

2.5

2.0

1.5

1.0

0.5

0.0
A B A5S A A10 S B10S A A10 E B10E B5E B

19
Fig.5b – Vehicle 2
Ethanol
VEHICLE 2
Benzene
Aromatics
1.4
C5+ Unsaturates
C5+ Alkanes
C4- Unsaturates
1.2
C4- Alkanes

1.0
VOC (grams)

0.8

0.6

0.4

0.2

0.0
A B A5S B5S A A A10S B10S A A10S A

Fig.5c – Vehicle 3
Ethanol VEHICLE 3
Benzene
5.0 Aromatics
C5+ Unsaturates
4.5
C5+ Alkanes
C4- Unsaturates
C4- Alkanes
4.0

3.5

3.0
VOC (grams)

2.5

Extp Extp
2.0
Extp

1.5

1.0

0.5

0.0
A B A5S A A B5S B B5S A5E B5E A A5S B5S

20
Fig.5d – Vehicle 4
VEHICLE 4

7.0 Extpg
Ethanol
Benzene
6.0 Aromatics
C5+ Unsaturates
C5+ Alkanes
5.0 C4- Unsaturates
C4- Alkanes
Extpg
VOC (grams)

4.0

3.0

2.0

1.0

0.0
A B A5S A B5S B5S A5E B5E B10E A10E A B

Fig.5e – Vehicle 5
VEHICLE 5
Ethanol
12.0 Benzene
Aromatics
FID
C5+ Unsaturates
C5+ Alkanes
10.0
C4- Unsaturates
C4- Alkanes

8.0
VOC (grams)

6.0

4.0

2.0

0.0
B10S A A5S B5S B A A A
Extra test Extra test

21
Fig.5f – Vehicle 6
VEHICLE 6 Ethanol
Benzene
1.4
Aromatics
C5+ Unsaturates
C5+ Alkanes
1.2 C4- Unsaturates
C4- Alkanes

1.0
VOC (grams)

0.8

0.6

0.4

0.2

0.0
A10S B10S A5E B5E

Fig.5g – Vehicle 7

Ethanol
VEHICLE 7
Benzene
Aromatics
1.4 C5+ Unsaturates
C5+ Alkanes
C4- Unsaturates
1.2
C4- Alkanes

1.0
VOC (grams)

0.8

0.6

0.4

0.2

0.0
A B B5S B10S

22
There is special concern over evaporative emissions of benzene, and whether ethanol could
increase them by formation of azeotropes. Benzene emissions are plotted against ethanol
content in Figure 6. Ethanol was found to have no noticeable nor statistically significant effect
on benzene emissions in any vehicle considered in isolation, nor when the data were pooled
and the vehicles considered together. The analysis was rerun omitting the abnormal results
indicated in Figure 6 and again no significant effect was seen. Ethanol also has no effect on
benzene emissions expressed as a percent of total evaporative emissions rather than as g/test.

Figure 6: Benzene Evaporative Emissions from all cars

23
Note: For clarity, different scales are used for each vehicle

Benzene emissions vs Ethanol in fuel

Vehicle 1 Vehicle 2

0.025 0.014
A 0.012
0.020
A5E A
0.010
A10E A5S
0.015 1st
g/test

g/test
A5S 0.008 A10S
A10S 0.006 B
0.010
B B5S
0.004
0.005 1st B5E B10S
B10E 0.002
0.000 B5S 0.000
0 2 4 6 8 10 12 B10S 0 2 4 6 8 10 12
Ethanol %v/v Ethanol %v/v

Vehicle 3 Vehicle 4

0.035 0.030
ExtPg
0.030 0.025 1st A
ExtPg
A A5E
0.025
A5E 0.020 A10E
1st
g/test
g/test

0.020 A5S A5S

0.015
0.015 B A10S
ExtPg
ExtPg B5E 0.010
ExtPg B
0.010
B5S B5E
0.005 0.005
B10E
0.000 0.000 B5S
0 2 4 6 8 10 12 0 2 4 6 8 10 12 B10S
Ethanol %v/v Ethanol %v/v

Vehicle 5 Vehicle 6

0.040 0.008
0.035 A 0.007 A
0.030 A5E 0.006 A5E
0.025 High evap A5S 0.005 A5S
g/test
g/test

B A10S
0.020 0.004
B5E B
0.015 B10E 0.003 B5E
0.010 Ext pg B5S 0.002 B5S
0.005 B10S 0.001 B10S
0.000 0.000
0 2 4 6 8 10 12 0 2 4 6 8 10 12
Ethanol %v/v Ethanol %v/v

Vehicle 7

0.007 Key:
0.006
ExtPg: Extra purging carried out prior to test
0.005 A
A5S 1st: First test on fuel A - low canister weight
g/test

0.004
B
0.003 B5S High evap: Abnormally high evaporative emissions in
B10S diurnal test
0.002
0.001
0.000
0 2 4 6 8 10 12
Ethanol %v/v

24
 5.2.3. Exhaust Emissions and Fuel Consumption

Exhaust regulated and unregulated emissions were measured in addition to evaporative

emissions, although the programme had not been designed to investigate this aspect. Exhaust
emission measurement was performed over the legislative European driving cycle (NEDC)
during the conditioning phase of the vehicle; this is carried out just prior to the start of the Hot
Soak test (see Appendix 3). Vehicle conditioning consists of a complete NEDC cycle
followed by an additional urban driving cycle.

Investigating fuel quality effects on exhaust emissions requires a strict test protocol to ensure
that any emission variation observed is really due to the different properties of test fuels.
Normally an absolute minimum of 3 tests on each fuel in each vehicle is necessary to see
significant fuel effects [25] [26] [27]. Furthermore the fuel change and vehicle conditioning
procedure is critical. The tank and the fuel system must be thoroughly flushed with the new
fuel in order to reduce the possibility of carry-over effects. Even the carbon canister could
influence exhaust emissions as when engine is running part of the combustion air is drawn
through the canister itself; hydrocarbons desorbed from activated carbon are then burnt in the
cylinders and therefore, to avoid any carry-over effect, they should reflect the composition of
the test fuel.

Most importantly, the test vehicle needs to be conditioned to let the Engine Management
System (EMS) adapt to the new fuel. This is particularly important in case of test fuels
differing for oxygen content as it directly influences the air/fuel ratio. Finally, the test
sequence must be designed to avoid as much as possible systematic errors and to minimize
the risk of confusing emission variation due to the specific test order with fuel effects. The
test protocol used in the EPEFE programme (Auto/Oil I) [27] is one of the most important
examples as it was developed with the specific aim of investigating the effect of some fuel
properties on exhaust emissions. Further improvements in the sequence of testing were made
in later CONCAWE studies; see [26].

However, as the programme described in this report aimed to investigate fuel effects on
evaporative emissions, the test protocol was bases on the relevant legislative test procedure.
For this reason no additional vehicle conditioning procedure to minimize carry-over effects
on exhaust emission was included in the protocol. Moreover, the long test duration imposed
by the evaporative emissions test limited the opportunity for repeat tests. As a consequence,
the exhaust emissions data is very variable and allows few conclusions to be drawn. For
information, the regulated emissions results are illustrated in Appendix 6. Some statistical
analysis was done but showed little, so is not reported.

The effects of vehicle conditioning however are less critical for CO2 emissions, and fuel
consumption that is calculated from it using the carbon balance method. Thus the results were
analysed to see if any fuel effects could be seen on CO2 emissions, energy consumption and
fuel consumption. The extra tests carried out on vehicle 5 (see section 5.3) are included in this
analysis. While CO2 emissions are measured, energy and fuel consumption must be calculated
from this data. This requires knowledge of the fuels carbon, hydrogen and oxygen content and
heating value. This data was calculated for the test fuels from the detailed GC analysis carried
out (see Appendix 2). Fuel consumption is often calculated using default values for C/H/O
from a typical reference fuel, but this is not appropriate when different fuels are tested,
especially oxygenated fuels.

25
The fundamental equation for the carbon balance calculation of fuel consumption is:

FCm = (CWFexh x HC + 0.429 x CO + 0.273 x CO2)/CWFfuel g/km

Where FCm is the calculated fuel consumption in g/km

CWFexh is the Carbon mass (weight) fraction of the hydrocarbon emissions
HC is hydrocarbon emission in g/km
0.429 is the carbon mass fraction of CO
CO is Carbon Monoxide emission in g/km
0.273 is the carbon mass fraction of CO2
CO2 is Carbon Dioxide emission in g/km
CWFfuel is the Carbon mass (weight) fraction of the fuel

CWFexh is realatively unimportant (and very hard to calculate) as hydrocarbon emissions are
small, but the correct CWFfuel is critical, so CWFexh is assumed to equal CWFfuel. Fuel
consumption in l/100km than then be calculated from:
FCl/100km = (FCm x 100)/(SGfuel x 1000)
where SGfuel is fuel specific gravity in kg/litre

Energy Consumption in MJ/100km is calculated from:

ECMJ/100km = FCl/100km x SGfuel x LHVfuel
where LHVfuel is the Lower Heating Value of the fuel in MJ/kg

Very few significant fuel effects on CO2 emissions were seen in the urban, extra-urban or
combined test cycles in any of vehicles. However, the data do show a consistent increases in
volumetric fuel consumption with increasing ethanol which are statistically significant at
P<5% or better in all vehicles (see Figure 8). This is consistent with what would be expected
due to the lower heating value of ethanol.

In order to quantify this effect, the full data set from all seven cars was analysed as a whole
using multiple regression analysis. 1. 2 The analysis shows that fuel oxygen content (%m/m)
has a highly significant effect (P<0.1%) on fuel consumption (l/100km). On average, fuel
consumption (l/100km) increases by a factor

1 + 0.0109 × fuel oxygen content (%m/m)

relative to a test on oxygen-free fuel in the same car.

Fuel ethanol content can be used equally well as a predictor. On average, fuel consumption
(l/100km) increases by a factor

1 + 0.00397 × fuel ethanol content (% v/v)

1
Strictly speaking, multiple regression analysis should not be performed across cars as Bartlett’s test shows that
the levels of variability (on a log scale) vary significantly from car to car (P<5%). However the variations
are relatively small, so the analysis can be used to try and understand trends in the data
2
There is a significant downward trend over time (P<0.1%) in CO2 and hence FCl/100km, and EC measurements
in Vehicle 6. This is adjusted for in both the individual vehicle and fleet analysis.

26
relative to a test on ethanol-free fuel in the same car. This is very close to the theoretical loss
in economy expected from the oxygen content of the fuel, which is shown in Figure 7 below.
For these vehicles a 10% ethanol blend increases fuel consumption by 3.97%, slightly more
than the theoretical energy loss of 3.4%.

Another aspect that has been analysed is the energy consumption of the vehicles on different
fuels. It has been claimed that oxygen-containing fuels result in a better efficiency of the
engine and therefore less energy should be consumed to complete the test cycle. The energy
consumed for each vehicle on each fuel is plotted in Figure 9. These plots do not show any
difference between the test fuels. Statistical analysis also revealed hardly any statistically
significant fuel effects on energy consumption.

Figure 7: Calculated theoretical energy penalty due to ethanol in gasoline

0.0%
Energy Penalty relative to Gasoline (%)

-0.5%

-1.0%

-1.5%

-2.0%

-2.5%

-3.0%

-3.5%

-4.0%
0.0% 2.0% 4.0% 6.0% 8.0% 10.0% 12.0%
Ethanol Volume Fraction

27
Figure 8: Calculated volumetric fuel consumption (l/100km) vs. fuel ethanol content
The fuel consumption has been calculated from exhaust emissions using the actual carbon
content and density of the fuels.
Vehicle 1 Vehicle 2

8.2 6.9
A
8.1 6.8
A5E
8.0 A10E 6.7 A
A5S A5S
l/100km

l/100km
7.9 6.6 A10S
1st A10S
7.8 B 6.5 1st B
B5E B5S
7.7 6.4
B10E B10S
7.6 B5S 6.3
7.5 B10S
6.2
0 2 4 6 8 10 12 0 2 4 6 8 10 12
Ethanol %v/v Ethanol %v/v

Vehicle 3 Vehicle 4

7.6 7.0
A
7.4 ExtPg 6.9 A5E
7.2 A A10E
ExtPg A5E 6.8 A5S

l/100km
l/100km

7.0 A5S A10S

6.7
6.8 ExtPg B ExtPg B
B5E 6.6 B5E
6.6 B5S B10E
ExtPg
6.4 1st 6.5 B5S
1st
B10S
6.2 6.4
0 1 2 3 4 5 6 0 2 4 6 8 10 12
Ethanol %v/v Ethanol %v/v

Vehicle 5 Vehicle 6

9.0 8.2
8.8 8.1
A A
1st A5E 8.1 A5E
8.6
A5S 8.0 A5S
l/100km
l/100km

8.4 B A10S
8.0
Ext pg B5E 1st B
8.2
Ext pg B10E 7.9 B5E
8.0 Ext pg B5S 7.9 B5S
7.8 B10S 7.8 B10S

7.6 7.8
0 2 4 6 8 10 12 0 2 4 6 8 10 12
Ethanol %v/v Ethanol %v/v

Vehicle 7

11.3
11.2
11.1
A
11.0
l/100km

A5S
10.9 B
1st B5S
10.8
B10S
10.7
10.6
10.5
0 2 4 6 8 10 12
Ethanol %v/v

Notes: For clarity, different scales are used for each vehicle
1st = first test which usually has the lowest evaporative emissions
Extpg = Extra purging carried out
Vehicle 5 plot: extra tests are included (see section 5.3)

28
Figure 9: Energy consumed (MJ/100km) vs. fuel ethanol content
The energy consumption has been calculated from exhaust emissions using the actual carbon
content and density of the fuels

Vehicle 1 Vehicle 2

258 A 220
256 A5E 218
A10E 216 A
254 1st
A5S 214
MJ/100km

MJ/100km
A5S
252 A10S 212 A10S
250 B 210 1st B
B5E 208 B5S
248
B10E 206 B10S
246 B5S
204
244 B10S
202
0 2 4 6 8 10 12 0 2 4 6 8 10 12
Ethanol %v/v Ethanol %v/v

Vehicle 3 Vehicle 4

240 216
215
235 ExtPg A
214
A A5E
230 213
ExtPg

MJ/100km
MJ/100km

A5E A10E
225 212
A5S A5S
211 ExtPg
220 ExtPg B A10S
210
B5E 209 1st B
215
B5S 208 B5E
210 B10E
1st 207 ExtPg
205 206 B5S
0 1 2 3 4 5 6 0 2 4 6 8 10 12 B10S
Ethanol %v/v Ethanol %v/v

Vehicle 5 Vehicle 6

290 260
285 259
A A
258
280 1st A5E A5E
257 1st
275
MJ/100km
MJ/100km

A5S 256 A5S

270 B 255 A10S
265 B5E 254 B
Ext pg 253
260 Ext pg B10E B5E
B5S 252 B5S
255 Ext pg 251
250 B10S B10S
250
245 249
0 2 4 6 8 10 12 0 2 4 6 8 10 12
Ethanol %v/v Ethanol %v/v

Vehicle 7

358
356
354
352 A
MJ/100km

1st A5S
350
B
348
B5S
346
B10S
344
342
340
0 2 4 6 8 10 12
Ethanol %v/v

Notes: Notes: For clarity, different scales are used for each vehicle
1st = first test which usually has the lowest evaporative emissions
Extpg = Extra purging carried out
Vehicle 5 plot: extra tests are included (see section 5.3)

29
 5.2.4. Aldehyde Emissions

As already described in the previous chapter, regulated and aldehyde exhaust emissions were
measured over the NEDC cycle during the conditioning phase of the vehicle just prior to the
start of the Hot Soak test.
The results have been analysed using statistical techniques; in general they show a quite high
degree of variability probably due to the test protocol not designed to investigate exhaust
emissions.
Fuel effects on formaldehyde and acetaldehyde emissions have been examined on a vehicle-
by-vehicle basis for the five cars for which data was available (vehicles 3, 4, 5, 6, 7). Details
on the analytical method and on the statistical analysis can be found in Appendix 7.

Formaldehyde emissions

There is little evidence of fuel effects on formaldehyde emissions. Ethanol has no significant
effect in any of the five cars tested. A time trend, significant at P < 5%, was included in the
model for the vehicle 4, but the ethanol effect remained insignificant.

Noting that the regression coefficients b in the model

Formaldehyde = a + b x EtOH

were positive for all five cars, the data was from the five cars was pooled and examined as a
whole. However the effect of ethanol remained insignificant.

(Note: it is difficult to compute fleet averages as not every fuel was tested in every vehicle
and the tested-fuel subsets varied from car to car)

Acetaldehyde emissions

Ethanol effects on acetaldehyde emissions varied from car to car and were statistically
significant at P < 5% for vehicles 5, 7 and also, after adjusting for the time trend, in vehicle 4.
The regression models for the first two cars were

Acetaldehyde = 160 + 12.0 x EtOH (vehicle 5)

Acetaldehyde = 184 + 30.1 x EtOH (vehicle 7)

where acetaldehyde is measured in μg/km and EtOH as %v/v. The EtOH term in the more
complex trend-adjusted model for the vehicle 4 had a regression coefficient of 10.9. The
coefficients were also positive, but not significant, for vehicles 3 and 6.

A fleet analysis was inappropriate because of the differing response patterns and the different
fuel subsets tested in the different vehicles.

30
 5.3. Extra Tests

Besides the tests carried out according to the test protocol based on the legislative procedure
and described in chapter 4.4, additional tests were performed, when possible, to address
specific issues. In particular, two series of extra-tests were carried out:

- Diurnal tests with the vehicle canister vented out of the VT SHED with the objective of
estimating the fuel permeation contribution to total evaporative emissions.

- Consecutive multiple tests on the same test fuel. The aim of these tests, performed
following the legislative procedure, was to try and understand the reason for the increase in
emissions with time seen in the main programme. The plan was to explore whether pure
hydrocarbon fuels exhibit a different behaviour than ethanol containing fuels.

5.3.1. Tests performed with the canister vented out of the VT SHED.

In standard evaporative emission tests the whole vehicle is placed into the airtight measuring
chamber (VT SHED) and all the hydrocarbons emitted by the vehicle itself are released into
the VT SHED. Regardless of the source (tank, fuel hoses, construction materials,..), all the
hydrocarbons emitted by the vehicle contribute to the total evaporative emissions and thus to
the final result of the test.

To try to estimate the contribution of fuel permeation through fuel hoses and tank walls a few
diurnal tests were carried out on two vehicles (vehicle 5 and vehicle 6) with the canister
vented out of the VT SHED via an opening in the VT SHED wall. Any breathing losses
through the carbon canister are in this way released in the atmosphere outside the VT SHED
and do not contribute to the VOC detected by FID in the VT SHED. Only hydrocarbons
coming from fuel permeation or leaks and from other non-fuel related sources should
therefore contribute to the test result.

Each of these special tests was carried out after a full standard evaporative emission test,
without changing fuel in the tank. The detailed procedure is described below:

• Completion of a full evaporative emission test (on one of the test fuels) comprising
vehicle and canister conditioning, hot soak test and diurnal test.
• At the end of the standard diurnal test, the VT SHED door was opened and the
measuring chamber purged with dry air.
• The vehicle canister was disconnected, weighed and then reconnected to the fuel
system. The vent of the canister was connected to an opening on the VT SHED wall
by means of a metallic tube.
• A new diurnal test was started (the vehicle was left in the VT SHED between the two
consecutive diurnal tests, no conditioning of the vehicle was performed)
• At the end of this second diurnal test, after recording its weight, the canister was
purged with dry air down to the weight recorded after the standard diurnal test.

The two consecutive diurnal tests differed not only because in the second test the canister was
vented out of the VT SHED but also for the initial conditions of the canister and the fuel in
the tank:

31
 - The canister weight was much higher at the beginning of the second diurnal test as it
had already been loaded with gasoline vapours generated during the previous diurnal
test. This means it had a reduced working capacity, but this did not affect the result of
the test as any breathing loss through the canister was not released into the VT SHED
but outside of it.
- Furthermore, during the standard evaporative emission test the gasoline in the tank
might have lost part of its lighter fractions, especially having been heated up to 35 °C
in the diurnal test. This might have resulted in a less severe test due to the lower
vapour pressure. However, it is not clear whether this might have had any influence
on the result of the test as in any case only sources different from breathing losses
contributed to it.

The results of these tests are presented in the following Tables 4 and 5 where the standard
diurnal test and subsequent test with the canister vented out of the VT SHED are compared.

Table 4 –Vehicle 5 - Comparison of standard Diurnal test results with Modified Diurnal
tests carried out with the canister vented out of the VT SHED

Vehicle 5
FUEL A B10S

Standard Diurnal (legislative

procedure) 1.573 3.201

Modified Diurnal (canister vented

out of the VT SHED) 1.351 2.525
Ratio (Mod. Diurnal/Std.Diurnal) 86% 79%

Table 5 – Vehicle 6 - Comparison of standard Diurnal test results with Diurnal tests
carried out with the canister vented out of the VT SHED

Vehicle 6
FUEL A B A5S B5S A B10S B5E

Standard Diurnal (legislative

procedure) 0.391 0.394 0.457 1.098 0.4590 1.289 0.911

Modified Diurnal (canister vented

out of the VT SHED) 0.486 0.539 0.508 1.059 0.595 1.024 0.850
Ratio (Mod. Diurnal/Std.Diurnal) 124% 137% 111% 96% 130% 79% 93%

In both vehicles the results of the diurnal tests carried out with the canister vented out of the
VT SHED are very close (somewhat higher or lower) to the values obtained in the standard
diurnal tests.
As there is no contribution of breathing losses to the measured evaporative emissions when
the canister is vented out of the VT SHED, it would suggest that other VOC sources are the
main contributors.
However it is very difficult to identify the real source; due to the low emission levels it is not
possible to use a point-source detector like a FID to look for the different sources.

32
The results of vehicle 6 are even more difficult to understand, unless there was a very small
leak somewhere. As vehicle 6 had a metallic tank that is not permeable to hydrocarbons, fuel
permeation should be very limited as it can occur only through fuel hoses.

5.3.2. Consecutive multiple tests on the same fuel.

Additional tests were carried out on vehicle 5 once the test series originally planned according
to the test protocol had been completed. There is no technical or scientific reason why these
additional tests were carried out only on vehicle 5; this choice was simply driven by the
longer availability for testing of this vehicle. At the end of the main test series, three different
blocks of tests were carried out.

The main objective of these additional tests, as already described, was to try to understand the
reason for the increase in emissions with time seen in the main programme. The aim was to
explore whether pure hydrocarbon fuels exhibit a different behaviour than ethanol containing
fuels and to investigate whether the observed increase in emissions is due to increased
canister loading, to an ethanol “memory” build up in the canister or to other possible effects.
From a theoretical point of view, ethanol is a polar molecule and is known to be harder to
purge from the active carbon in the canister. This ethanol characteristic might result in a
progressive reduction of carbon canister working capacity leading to an increase of emissions
[19] [15]. However, the presence of ethanol in the fuel could also progressively increase the
fuel permeation rate and, as a consequence, the total evaporative emissions.

Each of these blocks was preceded by a thorough purging operation of the canister in order to
eliminate any “memory” of fuels previously tested. The initial idea was to run multiple NEDC
cycles with fuel A (at least 5 but preferably 10) with the aim to purge completely the ethanol
from the canister. The amount of air flowing through the canister during the NEDC cycles
was measured, as well as the weight of the canister after each cycle. However, after ten cycles
it was decided to disconnect the canister from the vehicle and purge it to dryness with hot dry
air (40 °C). This decision was taken as the weight reduction of the canister after each driving
cycle was progressively becoming smaller and it was clear that ethanol and any other
hydrocarbons would not have been completely removed from the canister without completing
a huge number of driving cycles.
The canister was therefore purged for 2 hours with dry air at 40 °C and then for 4 hours with
air at 80 °C. Additionally the concentration of ethanol in the air exiting from the canister was
measured to monitor the residual amount of ethanol still adsorbed on activated carbon. When
the canister weight had stabilised, purging was stopped although a very small level of ethanol
was still detectable in the airflow coming out of the canister. The canister was then
reconnected and the first block of tests was started.

The first block of tests comprised four consecutive tests on fuel A, followed by one test on
fuel B and a final test on A, as shown in Table 6. The final test had to be repeated as the
temperature profile of the diurnal test was not correct due to a malfunctioning of the
thermocouple control module.

At the end of the first block, the canister was purged again with dry air down to the same
initial weight of the first block. Then three tests on fuel A5E and one test on B5E were
performed to look at the effect of ethanol at constant vapour pressure. The result of the

33
diurnal test on fuel B5E is suspect as the temperature in the VT SHED dropped too quickly
during the cooling down phase of the test due to a power supply failure. As a consequence the
test resulted to be less severe than a standard one. However, the result of the test has been
retained, as usually VOC concentration remains almost stable during the decreasing
temperature phase of the diurnal test.

After a third purging operation of the canister, a third block of tests, consisting of three tests
on fuel A5S followed by two on fuel A, was carried out. The aim of the last two tests on fuel
A was to check emission levels with the base fuel after the vehicle had been using ethanol
containing fuels for a substantial period of time.

The results of these tests are illustrated in Figure 10 and in table 6.

Figure 10: Extra tests performed on Vehicle 5 (consecutive multiple tests on same fuel)

Vehicle 5 - Extra Tests After Purging the Canister to Dryness

Bars: total evaporative emissions - Line: canister weight before N2/butane saturation

3.00 700

680
2.50
660
Total Evap. Emiss. (g/test)

640
2.00

Canister Weight (g)

620

1.50 600

580
1.00
560

540
0.50
Not
520
valid
test
0.00 500
5E

5E

5E

5S

5S

5S
A

A
B

5E
ng

ng
A

A
i

i
rg

rg
Pu

Pu

Fuel

Table 6 – Canister weights and evaporative emissions during consecutive multiple tests
on Vehicle 5

34
 BLOCK 1 Fuel A Fuel A Fuel A Fuel A Fuel B Fuel A Fuel A
Canister weight before
butane saturation (g) 606.80 635.38 649.11 663.74 662.50 667.30 665.37
Canister weight after
butane saturation (g) 632.81 657.96 667.28 680.95 678.66 681.51 681.51
Diurnal test initial
canister weight (g) 613.55 628.80 640.90 645.25 647.66 652.03 648.38
Diurnal test final canister
weight (g) 634.30 648.50 659.56 662.69 667.09 667.27 665.78
Evaporative emissions
Hot Soak test (g) 0.150 0.162 0.156 0.187 0.157 0.170 0.149
Diurnal test (g) 1.23 1.32 1.36 1.23 1.27 Not Valid 1.27
Total Evaporative
Emissions (g) 1.38 1.48 1.52 1.42 1.43 1.42

BLOCK 2 Fuel A5E Fuel A5E Fuel A5E Fuel B5E

Canister weight before
butane saturation (g) 605.81 644.30 651.74 662.08
Canister weight after
butane saturation (g) 646.50 664.11 667.21 680.77
Diurnal test initial
canister weight (g) 620.02 633.56 637.84 649.52
Diurnal test final canister
weight (g) 642.58 652.42 661.27 670.50
Evaporative emissions
Hot Soak test (g) 0.241 0.284 0.272 0.325
Diurnal test (g) 1.76 2.02 1.97 2.17*
Total Evaporative
Emissions (g) 2.00 2.30 2.24 2.49
*During the diurnal test, after reaching the maximum value the temperature dropped more quickly than prescribed by the
legislative profile. This might have led to a slightly lower emission value for this test.

BLOCK 3 Fuel A5S Fuel A5S Fuel A5S Fuel A Fuel A

Canister weight before
butane saturation (g) 605.85 644.60 654.74 666.10 669.00
Canister weight after
butane saturation (g) 648.63 664.82 673.22 680.34 685.56
Diurnal test initial
canister weight (g) 622.59 633.48 641.63 646.00 649.21
Diurnal test final canister
weight (g) 644.50 654.85 665.48 668.97 673.27
Evaporative emissions
Hot Soak test (g) 0.397 0.421 0.357 0.378 0.267
Diurnal test (g) 2.26 2.18 2.41 1.74 1.91
Total Evaporative
Emissions (g) 2.65 2.60 2.77 2.12 2.17

These results show clearly that fuels A, A5E and A5S behave in a very different way.

With fuel A total evaporative emissions are essentially constant at 1.4 – 1.5 g/test. There is no
significant variation of the emission levels between the first and the last test even though the
weight of the canister before butane saturation increased considerably over the first tests.
After the fourth test the canister weight levelled off at around 665 g and no further increase

35
was observed. In this case it seems that the canister initial weight did not affect significantly
the measured evaporative emissions. Fuel B, having a DVPE of 70 kPa gave emission levels
very close to those obtained with fuel A.

With fuels A5E and A5S total evaporative emissions were higher than on fuel A, and showed
a clear increase with time. Fuel A5S gave the highest emissions. However the weight of the
canister showed a very similar trend to that observed with fuel A: after a noticeable increase
in the first tests, the weight stabilized around a value ~665 g. In absolute terms, only small
differences for the canister weight were observed among the three test fuels. This seems to
suggest that the different emission levels measured with fuel A, A5E and A5S are not due
simply to a different canister loading, but to other factors including a change in working
capacity.

The final two tests on fuel A were substantially higher than the first block of tests, but lower
than the tests on ethanol fuels.

These results suggest that the influence of ethanol on evaporative emissions is not only linked
to the increase of gasoline vapour pressure; in fact, with fuel A5E evaporative emissions
increased even though its vapour pressure is identical to that of fuel A. However, vapour
pressure seems to play a role as evaporative emissions increased further when fuel A5S,
differing only for the vapour pressure from A5E, was tested.

From these data it is difficult to identify the reason why fuel A5E behaves differently from
fuel A; a possible explanation might be a greater role of fuel permeation in determining
evaporative emissions than initially thought. This is discussed further in Section 6.1.2.

Statistical analysis of fuel effects in Vehicle 5 (consecutive multiple tests)

Statistical techniques can be applied to analyze the results of the repeated extra tests carried
out on Vehicle 5 in order to investigate the relative performance of fuels A, A5E and A5S in
terms of evaporative emissions.
A significant upward trend in evaporative emissions (P<0.1%) was seen in these additional
tests conducted after the end of the main programme using weighted regression analysis (this
analysis is based on the extra test only and excludes the first test in each block where the
initial canister weight was low). The trend corrected mean emission values for the three fuels
were:

Fuel Mean g/test SE

A 1.745 0.042
A5E 2.166 0.081
A5S 2.449 0.104

Significant differences are seen between the 3 fuels in an F-test (P<0.1%). Investigating these
differences further using the Tukey-Kramer multiple comparison procedure, fuel A has lower
emissions than A5E (P<1%) and A5S (P<0.1%) with no significant difference between A5E
and A5S.

36
6. DISCUSSION OF RESULTS

6.1. Evaporative emissions

The results clearly showed that evaporative emissions measured according to the legislative
procedure are dominated by the contribution of the Diurnal test which accounts for 80 to 99%
of the total value, depending on the vehicle and on the fuel (see Appendix 4 for more details
on Hot Soak results). The Hot Soak contribution was not only small but also quite variable
with no clear pattern and no correlation with fuel quality.
Due to the low levels of Hot Soak emissions and despite their variability, total evaporative
emissions reflect the trend observed for the diurnal test data; therefore there is no substantial
difference in conclusions between considering the total evaporative emissions or only the
diurnal test data.

The key objective of the programme was to establish to what extent evaporative emissions
were influenced by gasoline vapour pressure and ethanol content. However it emerged during
the programme that the data presented a problem of comparability. As already discussed, an
upward trend in emission levels was noticed for each vehicle, as clearly demonstrated by the
multiple tests carried out on Fuel A. The first test with this fuel gave consistently lower
emissions than subsequent tests on the same fuel. The only exceptions are the tests carried out
after extra purging of the canister (vehicles 3 and 4).

This is likely to be due to the increased loading of the canister, as the canister weight
increased with test number, especially during the first few tests on each vehicle. Most
probably the conditioning phase prescribed by the legislative procedure was not sufficient (for
most of the vehicles) to return the canister to a constant condition before each test. It is not
clear whether this behaviour of the canister during the test sequence is due to, or has been
exacerbated by the presence of ethanol in the test fuels as, being a polar molecule, it is known
to be harder to purge from activated carbon than light hydrocarbons. The conditions in which
the tests have been carried out may not correspond to typical on-road vehicle usage and
therefore cannot be considered fully representative of average real-world conditions.
Nevertheless, the high canister loading due to the limited conditioning made the test
conditions very severe and therefore more likely to accentuate any fuel effects.

As a consequence, the tests carried out at the beginning of each vehicle test sequence might
not be directly comparable with those carried out later. This is particularly true for fuel A and
raised the question whether the first test on Fuel A should be excluded and not considered in
evaluating the influence of the fuel properties on evaporative emissions. It is also clear that
simply averaging repeat results on the various fuels in any particular vehicle would be
misleading.
In this study, time constraints together with the length of each evaporative emission test
limited the number of repeat tests. Moreover, an adaptive test sequence was used that could
not be randomised properly because of concerns about carry over effects (see section 5.2). As
a consequence, once unexpected variations in canister weight appeared, it became difficult to
disentangle fuel (ethanol, vapour pressure,…), canister weight and other time-related effects
even using sophisticated multiple regression techniques. No consistent causal model(s)
emerged which properly quantified the sizes of the various effects. In particular, it was not
possible to determine from the main programme data whether ethanol had any effect over and

37
above its effect on fuel vapour pressure. Therefore it was decided not to report these analyses
and to focus only on those conclusions that can be straightforwardly drawn from the data.

Evaporative emissions clearly appear to be influenced by fuel vapour pressure as shown by

Figures 3 and 4. However the effect of vapour pressure is not linear and only the fuels having
a DVPE close to 75 kPa (B5S, B10S) gave clearly higher evaporative emissions than the base
fuel A. This non-linearity of the vapour pressure influence on evaporative emissions was
expected and can be easily explained. Carbon canisters are very efficient at trapping gasoline
vapours until they become saturated (known as breakthrough). Once the breakthrough
condition is reached, the canister can no longer adsorb all the vapour generated in the tank
and some is emitted to atmosphere. In particular, if the breakthrough condition is reached
during the heating up phase of the diurnal test, from the instant in which it occurs the vehicle
acts like a vehicle without a canister and evaporative emissions can reach very high levels.

The homologation test for evaporative emissions is carried out, as prescribed by the relevant
legislation, with a reference fuel having a DVPE of 60 kPa at typical summer temperatures in
central/south Europe. Evaporative emission control systems are therefore developed to cope
with fuels having a vapour pressure close to 60 kPa, together with some engineering margin.
However, such systems may be not able to efficiently control evaporative emissions in
conditions different from those envisaged by the homologation procedure. Even if an
engineering margin is built into the system, fuels with a significantly higher vapour pressure
than 60 kPa may easily give higher evaporative emissions if canister breakthrough occurs.
The results of this programme suggest that, for most of the vehicles tested, fuels having a
DVPE around 75 kPa are beyond the capability of the carbon canister to effectively control
evaporative emissions and in some cases lead to breakthrough at temperatures that may be
encountered in summer.

Regarding the test fuels having a DVPE in the range of 60-70 kPa, the picture is less clear.
For all the vehicles tested, the differences in evaporative emissions between the various fuels
were quite small, and not always in the same direction. This does not necessarily mean that
there is no fuel effect on evaporative emissions for the fuels having a DVPE in the range of
60-70 kPa, only that the fuel effects are too small to be discerned in this particular series of
experiments.
As already mentioned, the increasing loading of the canister with successive test could have
altered the test results either hiding or magnifying possible fuel effects. Furthermore each
vehicle exhibited a different sensitivity to fuel quality and in particular to fuel vapour pressure
changes making even more difficult to draw any reliable conclusions.

n two vehicles (3 and 4), during the main programme, a few tests were carried out with a
modified procedure, purging the carbon canister back to its initial weight on the first test on
fuel A before each test. In these tests the evaporative emissions decreased substantially,
making the differences among the fuels even smaller. This suggests that the canister
conditioning and the initial canister loading is a key parameter that has a large influence on
the test results. It seems likely, therefore, that breathing losses only contribute significantly to
total evaporative emissions once canister breakthrough occurs, and are therefore more or less
independent of vapour pressure up to this point. However it is clear that a higher vapour
pressure speeds up breakthrough.

Ethanol might influence evaporative emissions also via different mechanisms than the
increased vapour pressure of ethanol/gasoline blends [19]. To explore this possibility

38
ethanol/gasoline blends matching the vapour pressure of the pure hydrocarbon base fuel were
included in the fuel matrix. However the main data set does not allow establishing whether
ethanol has any effect above and beyond the effect of increased fuel vapour pressure. Some
extra tests were therefore carried out as discussed below.

6.2. Results of the extra tests

The extra tests carried out after the main programme and not originally planned (see chapter
5.3.) suggest a particular interpretation of the evaporative emission data.
First of all, the tests with the canister vented out of the VT SHED confirm that, unless the
canister reaches breakthrough, the main contribution to evaporative emissions measured in a
diurnal test comes from sources other than canister breathing losses. A potential candidate as
main source is fuel permeation, although one of these vehicles had a metal fuel tank and it
would be expected to be less sensitive to permeation.

Secondly, the repeated tests on the same fuel showed that, at least for vehicle 5, there is a
statistically significant difference in terms of emissions among the fuels A, A5S and A5E. As
discussed in section 5.3.2, emissions on fuel A were relatively constant at 1.4 – 1.5
grammes/test. However all results on A5E were higher at 2 – 2.5 g/test, emissions on A5S
were slightly higher. The fact that fuels A and A5E had the same DVPE implies that there
must be other fuel parameters, different from vapour pressure, contributing to the higher
evaporative emissions of fuel A5E. This is most likely due to ethanol either reducing canister
working capacity or increasing fuel permeation [19].
If the latter is the case, the higher emission noticed with fuel A5E may be explained by the
fact that the vehicle had finished the main programme on fuel A, so had a hydrocarbon fuel in
its tank for several weeks before and during the first block of tests on fuel A. It was then
operated on ethanol containing fuels for over a month during which time emissions increased
continuously. Such effect of ethanol on fuel permeation has been already demonstrated by
other studies (CRC/CARB study [3]), which showed that it takes several weeks for a fuel
system to reach equilibrium permeation rates when ethanol is present in fuels.

The results of the tests carried out during the main programme according to the original
protocol apparently do not show any similar effect. The matched volatility ethanol fuels do
not behave differently from the pure hydrocarbon fuels. However the effect of ethanol on fuel
permeation is not instantaneous and therefore it is likely that the test programme, due to its
design, was not able to put in evidence this effect.

6.3. Effect of ethanol on canister efficiency

One of the potential issues associated with the use of ethanol/gasoline blends is the effect of
ethanol on canister efficiency [15] [19].
The working capacity of a canister is typically around 50% of its total equilibrium adsorption
capacity and it is heavily dependant on several parameters like canister design and purge
conditions. During normal operation a “heel” of material that cannot be easily desorbed builds
up within the carbon bed reducing the working capacity of the canister [14]. The magnitude of
the heel depends also on the carbon properties. Larger hydrocarbon molecules are less easily
desorbed than smaller ones, so over time the average molecular weight of the heel increases.
Ethanol is a polar molecule and it is known to be less easily desorbed from activated carbon;
therefore the use of a fuel containing ethanol could significantly increase the heel and reduce
canister working capacity. This would result in an increase of evaporative emissions.

39
As already mentioned, this programme was not designed to specifically address the influence
of ethanol on carbon canister performance. However, VOC speciation data showed that when
evaporative emission tests with pure hydrocarbon fuels followed tests carried out with ethanol
containing fuels, ethanol was found in the samples of the internal atmosphere of the VT
SHED. The canister is the most likely source of ethanol in these cases and this means that the
canister purging prescribed by the test procedure was not sufficient to remove all the ethanol
from the activated carbon when going back to pure hydrocarbon fuels. Experience of purging
the canister of vehicle 5 to dryness for the extra tests also showed that it was very difficult to
remove all the ethanol.

However from this programme it is not possible to establish if the residual ethanol does
irreversibly reduce the working capacity of the canister. Literature data confirm that the
presence of ethanol in the fuel leads to an increased heel in the canister but it is not clear
whether this impairs the canister performance in the long term. According to data provided by
activated carbon manufacturers the heel does increase as a result of ethanol use but the
gasoline working capacity of the activated carbon is not significantly affected by that.

6.4. VOC Speciation data

In order to have a better understanding of the influence of gasoline properties on evaporative

emissions, VOC speciation analysis was performed on samples of the internal atmosphere of
the VT SHED collected at the beginning and at the end of the diurnal test. Although
evaporative emission composition varied noticeably depending on the vehicle, some
consistent general trends can be identified from the data (see chapter 5.2.2).

In all the cases ethanol was found to be present in low concentrations, always lower than its
level in the test fuel. However ethanol was also found when pure hydrocarbon fuels were
tested after tests carried out with the ethanol containing fuels. As already discussed, in these
tests the most likely source of ethanol was the canister heel; most probably, ethanol was not
completely desorbed during the purging operations carried out prior to the start of the test.
The low levels of ethanol seem also to confirm that it is efficiently trapped by the canister and
that the contribution of ethanol breathing losses to the total evaporative emission is quite low.

Figure 5 shows that the major part of the evaporative emissions consisted of C4- and C5+
alkane hydrocarbons. Light hydrocarbons are expected to be the main contributors to
evaporative emissions because of their low boiling point and their higher diffusion rate in the
carbon bed. The main sources of these hydrocarbons are probably canister bleed emissions
and breathing losses.

Heavier hydrocarbons like aromatics were also found in significant concentrations; for
example, aromatics contributed an average of 32% to total evaporative emissions of vehicle 2.
These hydrocarbons are less likely to be emitted through the canister vent as their
concentration in the gasoline vapours above the liquid surface in the tank is expected to be
low due to the high boiling point. In this case the main source of emissions is likely to be fuel
permeation. The ratio between aromatics and the other hydrocarbons is quite different from
vehicle to vehicle; Table. 7 shows that even if the mass of aromatics emitted is similar (e.g.
vehicle 2 and vehicle 3), their contribution to the total average emissions is different. This is
due to the very different total emissions and probably to the different relative contribution of

40
the various evaporative emission sources (fuel permeation, bleed emissions, breathing
losses,…).

Benzene emissions were consistently around 1% for vehicles 2 – 6, and about half that for
vehicles 1 and 7, which also had low aromatic emissions. In all cases this was similar to, or
lower than the benzene content of the fuels.

Tab. 7 – Average emissions of benzene and aromatic hydrocarbons compared to total

evaporative emissions
Average Vehicle 1 Vehicle 2 Vehicle 3 Vehicle 4 Vehicle 5 Vehicle 6 Vehicle 7
Aromatics
(g/test) 0.23 0.29 0.34 0.56 0.51 0.21 0.12
Aromatic/Tota
l (%) 13% 32% 17% 21% 20% 21% 14%
Benzene
g/test 0.011 0.009 0.02 0.022 0.020 0.006 0.005
Benzene %
Total 0.6 1.1 1.0 0.9 0.8 0.9 0.5

41
7. CONCLUSIONS FROM EUCAR/JRC/CONCAWE EVAPORATIVE EMISSIONS
TEST PROGRAMME

While disentangling fuel, canister weight and other time related effects was difficult, several
clear conclusions can still be drawn from the results.

1. The vehicles tested differed in their level of evaporative emissions and in the extent of
their response to fuel changes. All cars met the 2 g/test emission limit on the first test on
fuel A, the reference fuel with DVPE of 60 kPa. Some vehicles slightly exceeded the limit
on subsequent tests on fuel A, probably related to increased canister loading in later tests.

2. A key fuel variable that affects evaporative emissions is vapour pressure (DVPE). In
general, increasing fuel DVPE above that of the reference fuel used for system
development increased evaporative emissions. The effect appeared to be non-linear (as
expected for a canister breakthrough effect). The ethanol blends with final DVPE around
75 kPa gave considerably higher evaporative emissions than the other fuels in several tests
over most of the vehicles.

3. Differences between the other fuels with DVPE in the range 60-70 kPa were small.
Furthermore, due to the combination of DVPE variations, the presence or absence of
ethanol, and to significant changes of canister weight, it is difficult to draw any reliable
conclusions on the influence of each single parameter. The engineering margin built into
the system may also explain the reduced fuel effect in this volatility range.

4. A limited number of tests carried out using modified test procedures suggest that the
increase of fuel vapour pressure alone may not satisfactorily explain the influence of
ethanol on total evaporative emissions. An increase of fuel permeation rate or a reduction
in canister working capacity due to ethanol could explain the different emissions
measured with fuel A and A5E in repeated tests carried out on one of the test vehicles
after the main programme.

5. This programme has shown that the test protocol used for this fuel evaluation was not able
to return the vehicle to a consistent condition at the start of each test. For most of the
vehicles, the canister weight increased with successive tests. This meant that emissions on
base fuel A increased with test number after testing on other fuels. The main weight
increase was noticed over the first two to three tests; after this the canister weight either
flattened out or increased less steeply. This is most likely due to increased vapour loading
on the canister, which is not sufficiently purged during the vehicle conditioning
prescribed by the regulatory test procedure. Extra purging carried out on some vehicles
between tests reduced this effect. Starting the test at such high canister weights may be
not representative of real-world operating conditions, but is a very severe test of the
evaporative control system. For any further work, a more extensive canister conditioning
procedure is needed to ensure that the canister system is properly conditioned to the new
fuel at the start of each test.

6. The increase in canister loading could be due to two factors, increased hydrocarbon
loading, a build up of ethanol in the canister, or both. It is not clear from this work what is
the relative size of these effects. Heavy hydrocarbons are known to be harder to purge
from the canister. Ethanol is a polar molecule and is also known to be harder to purge

42
 from the active carbon in the canister. Ethanol was found in the VT SHED vapour of tests
on pure hydrocarbon fuels following use of ethanol-containing fuels.

7. Measurements of regulated exhaust emissions showed no clear effects of fuel properties or

other parameters such as different canister loadings. However, the test programme was not
designed to look at exhaust emissions. Specific conditioning between tests and multiple
testing is essential for such studies.

8. Volumetric fuel consumption (litres/100 km) increased with increasing ethanol content.
This increase was roughly proportional to the oxygen content of the fuel. However neither
CO2 emissions nor energy consumption (MJ/100km) showed any effect of ethanol
content.

9. Speciated hydrocarbon emissions from the VT SHED Diurnal tests generally show
relatively high levels of light hydrocarbons (C4-C6) and low levels of ethanol. The main
sources of the light hydrocarbons are probably canister bleed emissions and breathing
losses. Heavier hydrocarbons like aromatics were also found in significant concentrations.
In this case the main source of emissions is likely to be fuel permeation.

43
8. REFERENCES

1. “A Technical Study on Fuels Technology related to the Auto/Oil II Programme –

Volume II – Alternative Fuels”, December 2000 – Prepared for the European
Commission Directorate General for Energy
2. Report No. SR00-01-01 “Potential Evaporative Emission Impacts Associated with the
Introduction of Ethanol-Gasoline Blends in California”, Sierra Research, Inc., January
11, 2000
3. “Fuel Permeation from Automotive Systems” - Final Report, CRC Project No. E-65,
September 2004
4. “Proposed Determination Pursuant to Health and Safety Code Section 43830(g) of the
Ozone Forming Potential of Elevated RVP Gasoline Containing 10 Percent Ethanol”,
November 1998, CEPA, CARB
5. “Issues Associated with the Use of Higher Ethanol Blends (E17-E24)”, National
Renewable Energy Laboratory (NREL), October 2002
6. “Comparison of the Effects of a Fully-Complying Gasoline Blend and a High RVP
Ethanol Gasoline Blend on Exhaust and Evaporative Emissions”, CARB, November
1998
7. Directive 98/69/EC of the European Parliament and of the Council of 13 October 1998
relating to measures to be taken against air pollution by emissions from motor vehicles
and amending Council Directive 70/220/EEC
8. “Relative Importance of 22 Parameters to Evaporative Emissions – A Sensitivity
Analysis of EVAP 2.0”, SAE Technical Paper No. 881594
9. “Studies on Carbon Canisters to Satisfy LEVII EVAP Regulations”, SAE Technical
Paper No. 2000-01-0895
10. “Speciation of Evaporative Emissions from Plastic Fuel Tanks”, SAE Technical Paper
No. 981376
11. “Impact and Control of Canister Bleed Emissions”, SAE Technical Paper No. 2001-
01-0733
12. “Identifying Sources of Evaporative Emissions – Using Hydrocarbon Profiles to
Identify Emission Sources”, SAE Technical Paper No. 2000-01-1139
13. “Activated Carbon Canister Performance During Diurnal Cycles: An Experimental
and Modelling Evaluation”, SAE Technical Paper No. 971651
14. “Development of a Gasoline-Fueled Vehicle with Zero Evaporative Emissions”, SAE
Technical Paper No. 2000-01-2926
15. “Gasoline Evaporative Emissions: Ethanol Effects on Vapor Control Canister Sorbent
Performance”, SAE Technical Paper No. 952748
16. “Carbon Canister Modeling for Evaporative Emissions: Adsorption and Thermal
Effects”, SAE Technical Paper No. 961210
17. “A Fuel Vapor Model (FVSMOD) for Evaporative Emissions System Design and
Analysis”, SAE Technical Paper No. 982644
18. “VOC running losses from canister-equipped vehicles”, CONCAWE Report No.
92/53
19. “Air Quality Impacts of the Use of Ethanol in California Reformulated Gasoline” –
Final Report to the California Environmental Policy Council - California
Environmental Protection Agency –Air Resources Board – 1999
20. “Effects of Gasoline Ethanol Blends on Permeation Emissions Contribution to VOC
Inventory From On-Road and Off-Road Sources” - The American Petroleum Institute
-2005

44
21. Directive 2003/30/EC of the European Parliament and of the Council of 8 May 2003
on the promotion of the use of biofuels or other renewable fuels for transport
22. Council Directive 2003/96/EC of 27 October 2003 restructuring the Community
framework for the taxation of energy products and electricity
23. Alcohols and Ethers, a Technical Assessment of Their Application as Fuels and Fuel
Component; American Petroleum Institute (API); publication 4261, rev. 1, June 2001;
p 23.
24. “The effects of temperature and fuel volatility on vehicle evaporative emissions”,
CONCAWE Report No. 90/51
25. “Fuel effects on emissions from modern gasoline vehicles: Part 1 - sulphur effects”,
CONCAWE Report No. 5/03
26. “Fuel effects on emissions from modern gasoline vehicles: Part 2 – aromatics, olefins
and volatility effects”, CONCAWE Report No. 02/04
27. “European Programme on Emissions, Fuels, Engine Technologies - EPEFE Report”,
ACEA, EUROPIA
28. “Fuel Permeation from Automotive Systems : E0, E6, E10 and E85”, Interim Report
CRC Project No. E-65-3

45
APPENDIX 1: EVAPORATIVE EMISSIONS: OVERVIEW

Evaporative emissions: definition and sources

Evaporative emissions from a vehicle can be defined, in a very generic way, as all the
Volatile Organic Compounds (VOCs) emitted by the vehicle itself and not deriving from fuel
combustion. For gasoline vehicles most of evaporative emissions are due to a loss of
hydrocarbons from the fuel system; more specifically the major contributions to evaporative
emissions come from fuel evaporation from the tank and fuel permeation through fuel hoses,
fuel tank, connectors, etc.

VOC compounds may also come from materials used for vehicle construction like plastics,
interior trim or from other system fluids (e.g. windshield detergent); these emissions are
usually very low in modern cars and in any case do not depend on fuel quality. Finally, an
important source of VOC emissions is the refuelling operation when the tank is open and
gasoline is pumped in. Refuelling emissions are regulated in the USA, and vehicles must
control them using a larger carbon canister. In Europe refuelling emissions are controlled in
the service station using so-called “Stage 2” systems to return vapour to the underground
storage tank.

Current European legislation sets a limit for evaporative emissions of 2.0 grammes/test and
defines the procedure to measure them [7]. There is similar legislation in USA and Japan. In
order to comply with the relevant emission standard, modern vehicles rely on an evaporative
emission control system consisting of an activated carbon canister, which adsorbs fuel
vapours and prevents their release to the air.

Figure 11: Typical layout of an evaporative emission control system

Fuel to Injectors

Vapors to Engine
Vapors to Canister

C Fuel Pump

Fuel Tank
Canister vent

When the vehicle is parked and the engine is switched off, the carbon canister traps gasoline
vapours produced in the tank by gasoline evaporation. When the vehicle is running, in certain
operating conditions and under the control of the EMS (Engine Management System), part of
the combustion air is drawn through the canister and into the engine; so that the activated
carbon is purged and the fuel vapours burned in the engine. The amount of air drawn through
the canister is managed by the Engine Management System (EMS) and controlled by means

46
of a valve (purge valve) located on the line connecting the canister with the air intake
manifold. A typical layout of an evaporative emission control system is shown in Figure 11.

In general, a carbon canister will trap hydrocarbons very efficiently, until it becomes
saturated and canister breakthrough occurs. The “gasoline working capacity” (GWC) of a
canister is defined as the amount of vapour it can absorb in normal use before breakthrough
occurs. A standard “Butane Working Capacity” (BWC) is specified by ASTM D 2652 and
used in the industry. In automotive application, breakthrough is usually defined as the point
when 2.0 grams of hydrocarbons have been emitted by the canister or when the total
hydrocarbon concentration reaches 5000 ppm.

When breakthrough is reached, the canister can no longer trap all hydrocarbons, so some will
pass through the canister via a diffusion based mechanism and be released into the air.
Eventually the carbon will become saturated and can adsorb no more hydrocarbons. Usually
there is a hydrocarbon concentration gradient within the carbon bed: higher concentration in
the part closer to the tank, lower towards the vent of the canister. The adsorbed hydrocarbons,
especially the lightest ones, tend to diffuse from the high concentration areas to the low
concentration ones. In this way light hydrocarbons can reach the canister exit port and escape
from it. These emissions are called “bleed emissions” [11].

In the case of a modern carbon canister equipped vehicle in good general condition, three
main sources of fuel related evaporative emissions can be identified [3]:

“Breathing losses”
Breathing losses are due to fuel vapours escaping from the tank through its vent. The fuel
tank, by design, is vented to the atmosphere through a pressure relief valve, so that tank
pressure is maintained slightly above atmospheric. In modern vehicles the tank is vented to
atmosphere through an activated carbon canister which adsorbs and stores the hydrocarbons
preventing emissions to atmosphere, as shown in Figure 11.

Breathing losses are influenced by a number of factors [8] [9] [14], mainly fuel volatility,
temperature and fuel system design in addition, of course, to the size and design of the
canister. The most important fuel property is the vapour pressure, a property measured
according to the EN 13016-1 method. This method measures the Dry Vapour Pressure
Equivalent (DVPE) of the older Reid Vapour Pressure (RVP) method. The difference
between RVP and DVPE is the measuring method. DVPE is measured “dry” after removing
all moisture from the test chamber prior to injection of the sample. This overcomes the
unpredictability of results experienced when testing samples containing oxygenates by the
conventional Reid method. The DVPE is measured at a temperature of 37.8° C. The CEN
standard EN 228:2004 sets out the gasoline volatility classes applied in Europe. Class A
(max. 60 kPa DVPE) is required by the EU fuels Directive during summer in Europe (1 May
to 30 September). Class B (max. 70 kPa DVPE) is required for countries “with arctic
conditions” from 1 June to 31 August. It is exactly the influence of fuel volatility on
evaporative emissions that is the main subject of the CONCAWE/EUCAR/JRC joint study
described in this report.

Residual HC concentration in the canister after purging has a certain influence as well.
Canister breakthrough will occur more easily when the residual HC concentration increases,
because this reduces the working capacity of the canister. Polar molecules like ethanol (or
water) or heavier hydrocarbons are usually harder to purge from the carbon. It has been

47
shown that activated carbon affinity for ethanol vapours is greater than for olefins and
aliphatic hydrocarbons [15]. Therefore it is possible that ethanol’s propensity to be tightly
held by activated carbon [19], in conjunction with its hygroscopic nature may decrease the
working capacity of the canisters used to control evaporative emissions and result in
increased diurnal emissions. For these reasons, in the USA a revised vehicle certification
procedure including also the use of an ethanol-containing gasoline has been adopted.

Breathing losses depend also on the fuel temperature inside the tank. The following plot [3]
shows the HC concentration in the vented vapour space above liquid gasoline as a function of
the temperature. For example, if temperature is increased from 70 °F to 95 °F (approximately
from 20 to 35 °C), the HC concentration will increase from 20 to 50% due to evaporation.
This leads to an increase of pressure and therefore some vapour must be expelled.

Temperature in the tank obviously depends on the ambient temperature but can vary
significantly during vehicle operation. If the engine has a fuel return, part of the fuel supplied
to the engine by the fuel pump is returned to the tank and can be significantly hotter than the
fuel in the tank, so tank temperature can increase substantially. However most modern
vehicles no longer use fuel return lines. Also the position of the tank may have some
influence as its temperature can be affected by the proximity of the exhaust tailpipe. When the
vehicle is parked the temperature of fuel in the tank can be either lower or higher than
ambient temperature. The temperature profile over time of fuel in the tank is different from
the profile of ambient temperature and this is due to the fuel thermal inertia. In particular, the
fuel temperature profile is time shifted compared to the ambient temperature profile and both
the maximum and the minimum temperature will be reached later compared to ambient.
Furthermore, the range from minimum to maximum temperature in the tank is narrower
(about 87% of the range of ambient temperature variation) [8]; in particular, maximum
temperature in the tank is somewhat lower than ambient maximum temperature.

Some other factors are related to the correct functioning of the evaporative emission control
system: the capacity and condition of the carbon canister, the quality and condition of
activated carbon, the condition of lines, connectors, the correct functioning of the purge valve,
the presence of a leak, etc.

48
- “Fuel permeation”
Hydrocarbons also escapes the vehicle’s fuel system by permeation through the plastic and
rubber components; e.g., hoses, seals, and in vehicles with a non-metallic tank, the fuel tank
itself. Permeation does not occur through an identifiable opening; instead individual fuel
molecules penetrate (i.e. they effectively mix with) the walls of the various components and
eventually find their way to the outside. Fuel permeation is significant only for plastic or
elastomeric materials.
Fuel permeation rate depends on the material used for the fuel system and on the chemical
species contained in the gasoline; in particular, alcohols like methanol and ethanol can
increase significantly the permeation rate.
A study co-sponsored by the California Air Resources Board (CARB) and Georgia (US)-
based Coordinating Research Council was recently carried out with the objective of
investigating the effect of gasoline/ethanol blends on permeation [3]. This study has
demonstrated that in many cases the permeation rate of ethanol/gasoline blends is higher
compared both to non oxygenated gasoline and to MTBE containing gasoline.
As follow-up of this work, the CRC began a new testing programme to investigate the
influence on permeation rate of several factors and in particular of the ethanol content in the
gasoline [28]. The interim report recently published has confirmed the higher permeation
rates of the low-level ethanol blends (E6, E10) compared to the base gasoline (E0); however
it seems that the permeation rate dose not increase between E6 and E10.

- “Running losses”
Running losses are those evaporative emissions that occur during normal driving of the
vehicle and include both fuel permeation and any breathing loss from the carbon canister.
These emissions are not currently regulated in Europe.

- Other sources
An unintended source of HC emissions may occur from leaks in the system. Leaks may occur
in the vapour and/or the liquid system as a result of deterioration and/or faulty service
techniques. Examples of deterioration are corrosion of metallic components (e.g., fuel lines,
tanks), cracking of rubber hoses, hardening of seals, and mechanical failures.

- Refuelling emissions
Another important source of evaporative emissions is the refuelling operation. These
emissions depend mainly on the fuel vapour pressure and on the design of the nozzle of the
fuel dispenser. They can be controlled in two ways. One way is with a so-called “Stage 2”
vapour recovery system that recycles emitted fuel vapours to the underground storage tank.
This system is used in Europe and was used in some US states. The alternative is an
“Onboard Vapour Recovery System” (ORVR), which is essentially a larger carbon canister
that can adsorb refuelling vapour. These systems are required in USA.

49
APPENDIX 2: FUEL ANALYTICAL DATA

Figure 12: Fuel distillation profiles

EUCAR EVAP PROGRAMME "A" FUELS

200.0

180.0

160.0

140.0
Temperature (°C)

120.0

100.0
C Fuel A
80.0 A5E
A10E
60.0
A10S
40.0 A5S

20.0

0.0
0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0
Evaporated (% v/v)

EUCAR EVAP PROGRAMME "B" FUELS

200.0

180.0

160.0

140.0
Temperature (°C)

120.0

100.0
B
80.0 B5E
B10E
60.0
B5S
40.0 B10S

20.0

0.0
0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0
Evaporated (% v/v)

50
Table A2.1 Detailed Fuel Properties
All distillation properties are mean of three different labs. Aromatics, olefins and ethanol mean of two labs, octane and sulfur one lab only.
Fuel name A A5E A10E A5S A10S B B5E B10E B5S B10S

DVPE kPa 60.1 59.7 59.9 67.1 66.8 69.0 69.9 66.5 75.4 75.6
E70 % v/v 38.3 40.2 44.6 42.7 51.8 38.9 42.0 46.3 44.0 53.1
E100 %v/v 54.7 61.3 54.8 56.6 59.4 54.8 61.8 58.0 56.8 60.0
E150 %v/v 87.3 87.2 84.1 87.7 88.5 84.6 87.9 85.5 84.8 85.8
IBP (0.5 % v/v) (°C) 35.1 39.2 37.8 35.1 36.4 31.9 35.4 38.1 33.0 33.5
5 % v/v (°C) 45.5 47.6 46.8 43.9 44.9 43.1 44.6 47.8 42.2 42.5
10 % v/v (°C) 50.1 50.4 52.9 47.0 48.3 46.9 48.3 51.9 45.8 46.7
20 % v/v (°C) 56.6 53.8 58.7 51.2 52.5 54.7 53.2 57.3 50.4 52.0
30 % v/v (°C) 63.1 58.7 63.5 56.1 56.8 62.1 57.9 62.0 54.6 56.7
40 % v/v (°C) 72.2 69.7 67.5 66.3 61.2 71.6 67.5 66.0 64.5 61.5
50 % v/v (°C) 88.5 83.0 88.2 83.6 67.1 88.3 82.1 80.8 83.5 66.4
60 % v/v (°C) 119.3 97.9 111.4 112.6 104.9 120.3 97.5 105.1 113.0 103.1
70 % v/v (°C) 139.1 114.5 125.1 138.2 136.6 140.3 115.1 120.8 138.9 136.7
80 % v/v (°C) 144.7 133.5 143.6 144.6 144.0 146.8 133.2 139.5 146.4 145.8
90 % v/v (°C) 154.0 156.8 164.3 153.6 153.5 159.0 155.6 162.0 158.8 157.1
95 % v/v (°C) 168.7 170.2 182.7 167.4 168.3 175.3 169.4 181.2 175.7 174.8
FBP (99.5 % v/v) °C 189.0 195.6 200.9 188.4 187.9 193.7 192.7 197.3 192.3 192.6
Density g/l @15 °C 755.5 747.1 756.0 757.2 758.7 753.3 747.1 750.0 754.3 756.0
RON 98.4 97.6 98.4 99.0 100.0 97.8 99.2 98.4 99.0 99.9
MON 86.1 85.8 86.0 86.5 86.7 86.1 86.9 86.4 86.3 86.7
Sulfur mg/kg 13.0 10.0 12.0 10.0 10.0 11.0 11.0 12.0 10.0 10.0
Aromatics %v/v 36.3 29.5 33.8 34.5 32.3 35.2 30.7 31.2 34.0 32.8
Olefins %v/v 8.7 9.3 10.8 8.5 8.3 9.1 7.8 10.4 8.7 7.9
Ethanol %v/v 0.0 5.0 10.4 4.7 9.7 0.0 5.4 9.9 5.0 10.2
Oxygen %m/m 0.00 1.84 3.79 1.71 3.52 0.00 1.97 3.64 1.83 3.70

51
Table A2.2: Detailed Hydrocarbon Speciation Data - %m/m

Fuel Ref A A5S A10S A5E A10E B B5S B10S B5E B10E
Propene 0.00 0.00 0.00 0.00 0.00 0.01 0.00 0.00 0.00 0.00
Propane 0.03 0.02 0.02 0.01 0.06 0.17 0.16 0.13 0.06 0.07
Isobutane 0.34 0.32 0.28 0.19 0.28 0.60 0.57 0.52 0.42 0.44
Methanol 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
But-1-ene 0.12 0.12 0.10 0.11 0.13 0.13 0.12 0.11 0.11 0.13
n-Butane 1.17 1.11 0.97 0.82 1.02 1.83 1.71 1.60 1.43 1.46
tBut-2-ene 0.18 0.17 0.15 0.17 0.20 0.19 0.17 0.16 0.16 0.19
2,2-Dimethylpropane 0.02 0.02 0.01 0.01 0.01 0.02 0.02 0.02 0.01 0.01
cBut-2-ene 0.18 0.17 0.15 0.16 0.19 0.18 0.17 0.16 0.16 0.19
Ethanol 0.00 4.90 10.00 5.20 10.70 0.12 5.30 10.40 5.70 10.40
3-Methylbut-1-ene 0.08 0.07 0.07 0.07 0.09 0.08 0.07 0.07 0.07 0.09
Isopentane 7.29 6.92 6.38 5.35 5.97 7.42 6.97 6.71 5.64 5.62
Pent-1-ene 0.31 0.30 0.27 0.84 1.05 0.31 0.29 0.28 0.33 0.41
2-Methylbut-1-ene 0.54 0.52 0.48 0.50 0.59 0.53 0.50 0.48 0.48 0.59
n-Pentane 3.88 3.68 3.38 1.20 1.63 3.87 3.63 3.48 2.10 1.67
2-Methyl-1,3-butadiene 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01
tPent-2-ene 0.85 0.81 0.75 0.78 0.93 0.83 0.78 0.75 0.75 0.93
3,3-Dimethylbut-1-ene 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
cPent-2-ene 0.47 0.44 0.41 0.43 0.51 0.45 0.43 0.41 0.41 0.51
2-Methylbut-2-ene 1.12 1.06 0.99 1.02 1.22 1.09 1.02 0.98 0.98 1.22
t1,3-Pentadiene 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01
c1,3-Pentadiene 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.00 0.00 0.01
2,2-Dimethylbutane 1.31 1.24 1.17 0.40 0.38 1.27 1.20 1.16 0.74 0.54
Cyclopentene 0.18 0.17 0.16 0.16 0.19 0.17 0.16 0.15 0.15 0.19
4-Methylpent-1-ene 0.05 0.05 0.05 0.05 0.06 0.05 0.05 0.05 0.04 0.06
3-Methylpent-1-ene 0.08 0.07 0.07 0.07 0.08 0.07 0.07 0.07 0.06 0.08
Cyclopentane 11.64 11.04 10.28 9.21 0.59 10.33 9.69 9.26 10.43 1.52
2,3-Dimethylbutane 1.29 1.23 1.16 2.24 1.24 1.27 1.20 1.16 1.87 1.83
4-Methyl-cpent-2-ene 0.05 0.04 0.04 0.04 0.05 0.05 0.04 0.04 0.04 0.05
MTBE 0.08 0.07 0.07 0.07 0.08 0.08 0.07 0.14 0.07 0.09
2-Methylpentane 3.92 3.73 3.52 8.12 4.08 3.85 3.62 3.49 6.31 6.25
4-Methyl-tpent-2-ene 0.14 0.13 0.13 0.13 0.16 0.14 0.13 0.12 0.12 0.15
3-Methylpentane 2.48 2.36 2.24 2.50 1.85 2.44 2.29 2.21 2.43 2.43
2-Methylpent-1-ene 0.23 0.22 0.21 0.21 0.25 0.23 0.21 0.20 0.20 0.26
Hex-1-ene 0.13 0.13 0.12 0.12 0.14 0.13 0.12 0.13 0.11 0.14
n-Hexane 2.10 2.00 1.88 1.01 1.09 2.07 1.95 1.87 1.42 1.20
c+tHex-3-ene 0.62 0.59 0.56 0.56 0.67 0.59 0.56 0.54 0.53 0.68
tHex-2-ene 0.45 0.43 0.41 0.41 0.50 0.44 0.41 0.39 0.39 0.50
3-Methyl-cpent-2-ene 0.27 0.25 0.24 0.24 0.29 0.26 0.24 0.23 0.23 0.29
cHex-2-ene 0.20 0.19 0.18 0.18 0.22 0.19 0.18 0.18 0.17 0.22
3-Methyl-tpent-2-ene 0.33 0.32 0.30 0.30 0.36 0.32 0.30 0.29 0.28 0.36
2,2-Dimethylpentane 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Methylcyclopentane 2.71 2.58 2.44 1.08 1.48 2.67 2.51 2.42 1.66 1.51
2,4-Dimethylpentane 0.18 0.17 0.16 0.32 0.35 0.17 0.17 0.16 0.31 0.35
2,2,3-Trimethylbutane 0.00 0.01 0.01 0.00 0.02 0.01 0.00 0.01 0.02 0.02
Benzene + 1-
Methylcyclopentene 0.49 0.47 0.44 0.70 0.81 0.48 0.46 0.44 0.69 0.78
3-Methylhex-1-ene 0.01 0.01 0.01 0.02 0.02 0.02 0.02 0.01 0.01 0.02
3,3-Dimethylpentane 0.05 0.04 0.04 0.12 0.13 0.05 0.05 0.04 0.12 0.13
Cyclohexane 0.70 0.76 0.83 0.54 0.67 0.74 0.81 0.88 0.54 0.55
5-Methylhex-1-ene 0.02 0.01 0.01 0.02 0.02 0.01 0.01 0.01 0.01 0.02
4-Methyl-c/thex-2-ene 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.00 0.00 0.01

52
 2,3-Dimethylpentane 0.18 0.18 0.17 0.48 0.51 0.20 0.19 0.16 0.48 0.51
2-Methylhexane 0.55 0.56 0.50 1.10 1.27 0.57 0.54 0.52 1.14 1.20
1,1-Dimethylcyclopentane 0.02 0.01 0.01 0.02 0.02 0.02 0.02 0.02 0.02 0.02
Cyclohexene 0.03 0.03 0.03 0.03 0.04 0.03 0.03 0.00 0.00 0.03
3-Methylhexane 0.54 0.52 0.50 1.23 1.39 0.57 0.54 0.52 1.28 1.33
c1,3-Dimethylcyclopentane 0.11 0.10 0.10 0.10 0.13 0.11 0.10 0.10 0.11 0.13
t1,3-Dimethylcyclopentane 0.09 0.09 0.08 0.09 0.11 0.09 0.09 0.08 0.09 0.11
t1,2-Dimethylcyclopentane 0.16 0.13 0.12 0.20 0.24 0.14 0.13 0.12 0.22 0.23
Isooctane 0.02 0.02 0.02 7.01 6.54 0.00 0.02 0.01 6.47 6.69
Hept-1-ene 0.01 0.03 0.03 0.00 0.00 0.18 0.17 0.16 0.00 0.00
tHept-3-ene 0.07 0.06 0.06 0.07 0.08 0.07 0.06 0.06 0.07 0.08
n-Heptane 0.29 0.28 0.27 0.81 0.90 0.32 0.30 0.29 0.84 0.85
cHept-3-ene 0.03 0.02 0.02 0.03 0.04 0.02 0.02 0.02 0.03 0.04
tHept-2-ene 0.01 0.01 0.01 0.02 0.02 0.01 0.01 0.01 0.02 0.02
cHept-2-ene 0.08 0.08 0.08 0.06 0.07 0.05 0.05 0.05 0.06 0.07
2,3-Dimethylpent-2-ene 0.03 0.02 0.02 0.04 0.04 0.03 0.03 0.02 0.04 0.04
Methylcyclohexane 0.13 0.12 0.12 0.00 0.00 0.00 0.00 0.13 0.12 0.12
1,1,3-Trimethylcyclopentane 0.02 0.02 0.02 0.29 0.28 0.03 0.03 0.02 0.28 0.29
2,2-Dimethylhexane 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Ethylcyclopentane 0.02 0.02 0.02 0.02 0.03 0.02 0.02 0.02 0.02 0.02
2,5-Dimethylhexane 0.02 0.02 0.02 0.41 0.39 0.03 0.03 0.03 0.38 0.39
2,4-Dimethylhexane 0.03 0.03 0.02 0.40 0.39 0.04 0.03 0.03 0.38 0.39
1,t2,c4-Trimethylcyclopentane 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.00 0.01 0.01
3,3-Dimethylhexane 0.01 0.01 0.01 0.04 0.04 0.01 0.01 0.01 0.04 0.04
1,t2,c3-Trimethylcyclopentane 0.01 0.01 0.00 0.01 0.02 0.01 0.01 0.00 0.01 0.01
2,3,4-Trimethylpentane 0.01 0.01 0.01 1.14 1.07 0.05 0.04 0.04 1.05 1.09
Toluene + 2,3,3-TrimeC5 2.80 2.70 2.56 9.96 10.90 3.28 3.12 2.98 9.63 9.81
1,1,2- Trimethylcyclopentane 0.02 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
2,3-Dimethylhexane 0.00 0.02 0.02 0.35 0.33 0.04 0.03 0.03 0.30 0.34
2,5-Dimethylhex-2-ene 0.00 0.00 0.00 0.01 0.02 0.00 0.00 0.00 0.00 0.01
2-Methylheptane 0.05 0.04 0.04 0.19 0.21 0.05 0.05 0.05 0.19 0.21
4-Methylheptane 0.03 0.02 0.02 0.10 0.11 0.02 0.03 0.02 0.07 0.11
3,4-Dimethylhexane 0.01 0.00 0.01 0.07 0.07 0.00 0.00 0.00 0.00 0.00
c1,3-Dimethylcyclohexane 0.00 0.00 0.00 0.00 0.01 0.00 0.00 0.00 0.00 0.01
3-Methylheptane 0.06 0.06 0.05 0.24 0.26 0.07 0.06 0.06 0.25 0.25
3-Ethylhexane 0.04 0.03 0.03 0.06 0.07 0.03 0.03 0.00 0.07 0.07
t1,4-Dimethylcyclohexane 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.00 0.00 0.01
1,1-Dimethylcyclohexane 0.01 0.00 0.00 0.00 0.00 0.01 0.01 0.00 0.00 0.00
2,2,5-Trimethylhexane 0.00 0.00 0.00 0.11 0.10 0.01 0.01 0.00 0.11 0.11
t1,3-Ethylmethylcyclopentane 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
t1,2-Ethylmethylcyclopentane 0.01 0.00 0.00 0.01 0.01 0.01 0.00 0.00 0.00 0.01
2,2,4-Trimethylhexane 0.00 0.00 0.00 0.01 0.01 0.00 0.00 0.00 0.00 0.01
t1,2-Dimethylcyclohexane 0.01 0.01 0.01 0.00 0.01 0.01 0.01 0.00 0.00 0.01
1,c2,c3-
Trimethylcyclopentane 0.00 0.00 0.00 0.01 0.01 0.00 0.00 0.00 0.00 0.01
n-Octane 0.06 0.06 0.05 0.20 0.22 0.06 0.06 0.05 0.21 0.10
tOct-2-ene 0.37 0.27 0.25 0.55 0.63 0.48 0.42 0.22 0.32 0.46
2,4,4-Trimethylhexane 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01 0.01
2,3,5-Trimethylhexane 0.00 0.00 0.00 0.02 0.02 0.00 0.00 0.00 0.02 0.00
2,2-Dimethylheptane 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.01
2,4-Dimethylheptane 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.00 0.01 0.01
2,3,4-Trimethylhexane 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
n-Propylcyclopentane 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
2,6-Dimethylheptane 0.03 0.02 0.02 0.01 0.01 0.01 0.01 0.00 0.00 0.02
1,1,3-Trimethylcyclohexane 0.01 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00 0.00

53
 2,5-Dimethylheptane 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
3,3-Dimethylheptane 0.01 0.01 0.01 0.02 0.03 0.01 0.01 0.00 0.02 0.03
Ethylbenzene 5.71 5.47 5.23 2.06 2.24 5.23 4.99 4.68 2.06 2.10
1,t2,t4-Trimethylcyclohexane 0.02 0.02 0.02 0.01 0.01 0.01 0.01 0.00 0.00 0.01
m-Xylene 15.71 14.93 14.12 5.10 5.47 14.21 13.60 12.64 5.03 5.14
p-Xylene 5.48 5.26 5.15 1.93 2.15 5.26 4.97 4.76 1.94 2.00
2,3-Dimethylheptane 0.00 0.00 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.01
3,4-Dimethylheptane 0.00 0.00 0.00 0.01 0.01 0.00 0.00 0.00 0.01 0.01
4-Methyloctane 0.01 0.01 0.02 0.03 0.03 0.01 0.01 0.01 0.03 0.03
2-Methyloctane 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
1,t2,c4-Trimethylcyclohexane 0.00 0.00 0.00 0.01 0.01 0.00 0.00 0.00 0.01 0.01
3-Methyloctane 0.03 0.00 0.02 0.04 0.04 0.01 0.01 0.01 0.04 0.04
o-Xylene 5.59 5.34 5.12 2.98 3.27 5.18 4.94 4.64 3.01 3.08
Isobutylcyclopentane 0.01 0.00 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.01
n-Nonane 0.09 0.09 0.09 0.04 0.05 0.01 0.01 0.01 0.04 0.05
Isopropylbenzene 0.08 0.08 0.07 0.25 0.27 0.09 0.08 0.08 0.25 0.26
3,3-Dimethyloctane 0.01 0.01 0.01 0.01 0.01 0.00 0.00 0.00 0.00 0.00
n-Propylbenzene 0.25 0.25 0.24 0.78 0.86 0.26 0.24 0.23 0.80 0.82
m-Ethyltoluene 0.61 0.58 0.56 1.86 2.07 0.68 0.66 0.62 1.92 1.97
p-Ethyltoluene 0.28 0.27 0.26 0.83 0.93 0.31 0.30 0.28 0.86 0.88
1,3,5-Trimethylbenzene 0.29 0.28 0.27 0.75 0.84 0.32 0.30 0.29 0.77 0.79
o-Ethyltoluene 0.27 0.25 0.24 0.79 0.88 0.31 0.30 0.28 0.82 0.84
1,2,4-Trimethylbenzene 1.29 1.23 1.20 3.01 3.35 1.47 1.40 1.33 3.09 3.18
n-Decane 0.28 0.27 0.26 0.10 0.24 0.08 0.08 0.08 0.10 0.18
1,2,3-Trimethylbenzene 0.46 0.44 0.43 0.89 0.99 0.55 0.52 0.49 0.91 0.94
Indane 0.15 0.14 0.14 0.22 0.25 0.17 0.17 0.15 0.23 0.24
o-Isopropyltoluene 0.03 0.03 0.02 0.09 0.09 0.03 0.03 0.03 0.08 0.09
1,3-Diethylbenzene 0.11 0.10 0.10 0.16 0.17 0.13 0.12 0.11 0.16 0.16
m-n-Propyltoluene 7.48 7.00 6.74 2.18 2.34 8.20 7.86 7.22 1.94 2.17
n-Butylbenzene 0.35 0.33 0.32 0.50 0.56 0.40 0.38 0.36 0.51 0.53
1,3-Dimethyl-5-ethylbenzene 0.32 0.30 0.29 0.37 0.42 0.38 0.36 0.34 0.38 0.39
o-n-Propyltoluene 0.26 0.25 0.24 0.31 0.34 0.32 0.30 0.28 0.31 0.32
1,3-Dimethyl-4-ethylbenzene 0.25 0.23 0.23 0.27 0.30 0.30 0.29 0.27 0.28 0.29
1,2-Dimethyl-4-ethylbenzene 0.48 0.45 0.44 0.49 0.55 0.58 0.54 0.52 0.50 0.51
n-Undecane 0.07 0.07 0.07 0.38 0.93 0.36 0.33 0.31 0.04 0.68
1,2,4,5-Tetramethylbenzene 0.31 0.29 0.29 0.25 0.29 0.38 0.36 0.34 0.26 0.27
1,2,3,5-Tetramethylbenzene 0.45 0.43 0.41 0.37 0.41 0.55 0.52 0.49 0.37 0.39
1,3-Di-isopropylbenzene 0.01 0.01 0.02 0.01 0.01 0.01 0.01 0.02 0.01 0.01
1,2-Di-isopropylbenzene 0.60 0.57 0.54 0.85 1.21 0.79 0.76 0.70 0.65 1.03
Naphthalene 0.29 0.28 0.27 0.25 0.28 0.36 0.35 0.33 0.26 0.27
n-Dodecane 0.04 0.04 0.04 0.38 0.88 0.33 0.32 0.30 0.03 0.66
2-Methylnaphthalene 0.05 0.05 0.04 0.11 0.12 0.06 0.05 0.05 0.11 0.11
1-Methylnaphthalene 0.02 0.02 0.02 0.06 0.07 0.03 0.03 0.02 0.06 0.06
TOTAL 99.99 99.99 99.99 99.99 99.99 99.99 99.99 99.99 99.99 99.99
Density 755.5 757.2 758.7 747.1 756.0 753.3 754.3 756.0 747.1 750.0
% m/m Carbon 87.48 85.75 83.99 85.09 83.50 87.39 85.58 83.74 84.90 83.38
% m/m Hydrogen 12.50 12.52 12.51 13.08 12.75 12.55 12.56 12.61 13.10 12.98
%m/m Oxygen 0.01 1.72 3.49 1.82 3.74 0.06 1.86 3.64 2.00 3.63
LHV MJ/kg (calc from HC) 42.80 42.05 41.24 42.29 41.25 42.81 42.00 41.22 42.23 41.41
LHV MJ/litre 32.34 31.83 31.29 31.60 31.18 32.24 31.68 31.16 31.55 31.06
Fuel Ref A A5S A10S A5E A10E B B5S B10S B5E B10E

54
Figure 13: Fuels hydrocarbon speciation by key hydrocarbons

70 ethanol
1,2,4-Trimethylbenzene
1,3,5-Trimethylbenzene
o-xylene
60 m-Xylene
Ethylbenzene
Toluene + 2,3,3-TrimeC5
Benzene + 1-Methylcyclopentene
50 n-Heptane
n-Hexane
2-Methyl-1,3-butadiene
3-Methylpentane
40 2-Methylpentane
cPent-2-ene
tPent-2-ene
n-Pentane
30 1,3 butadiene
Isopentane
propyne
cBut-2-ene
20 isobutene
But-1-ene
tBut-2-ene
n-Butane
10 Isobutane
acetylene
Propene
Propane
0 ethene
A A5S A10S A5E A10E B B5S B10S B5E ethane

55
APPENDIX 3: DETAILS OF TEST PROCEDURE.

Both the limit and the detailed measurement procedure for evaporative emissions are laid
down in the Directive of the European Parliament and the Council no. 98/69/EC.

The evaporative emission test (Type IV) is designed to determine hydrocarbon evaporative
emissions as a consequence of diurnal temperatures fluctuation and hot soaks during parking
after urban driving. The test consists of these phases (see Fig. 14):

— test preparation (canister and vehicle conditioning)

— hot soak loss determination (Hot Soak test).
— diurnal loss determination (Diurnal test).

Evaporative emissions are measured using a gas-tight chamber (VT SHED) able to contain
the vehicle under test. The VOC concentration inside the chamber is monitored by means of a
FID analyzer. The mass emissions of hydrocarbons from the hot soak and the diurnal loss
phases are added up to provide an overall result for the test.

The Hot Soak test simulates the condition of a vehicle parked after having been driven for a
certain distance. At the end of the conditioning drive (one NEDC driving cycle + one urban
driving cycle) the vehicle is then pushed into the measuring chamber. The engine must be
turned off before any part of the vehicle enters the measuring chamber. The test lasts 60
minutes and the temperature must not be less than 296 °K (23 °C) and more than 304 °K (31
°C) during the 60-minute hot soak period.

The Diurnal test lasts instead 24 hours and simulates the situation of a vehicle parked for one
day and one night in the summer period; the temperature in the VT SHED is varied according
to a profile defined by the Directive to ideally reflect the fluctuations occurring during day
and night time. The starting temperature is 20 °C while the maximum value is 35 °C, reached
after 12 hours. Then, during the subsequent 12 hours the temperature decreases slowly and
goes back to 20 °C again.

The final result of the test is given by the sum of the emissions measured during the Hot Soak
and the Diurnal tests.
The European limit for evaporative emissions (Hot Soak + Diurnal) is currently 2
grammes/test (Directive 70/220/EC and subsequent amendments).

56
Figure 14 : Standard EU Evaporative Emissions Test Procedure (EU Directive 98-69-
EC)

57
Detailed Test Procedure for this Programme

The whole programme was carried out in the VELA laboratory of the Joint Research Centre
located in Ispra (Italy). The JRC’s VT SHED used for this programme complies with the
98/69/EC and EPA requirements. The measuring chamber is a fixed-volume type one.
Volume compensation is achieved by continuous withdrawal of internal atmosphere and re-
filling with ambient air. Outgoing and incoming flows are controlled by means of mass flow
controllers.

Vehicle preparation

The vehicle was mechanically prepared before the test as follows:

— The exhaust system of the vehicle was checked for any leaks
— The engine was checked for any leaks of the gasoline/lubricant circuit.
— The ageing (minimum 3 000 km) of the canister was verified. When the mileage
accumulated by the canister was less than 3000 km, the vehicle was driven on the road
to reach this target value.
— When necessary, the vehicle was cleaned before the test.
— When necessary, additional fittings, adapters or devices were fitted to the fuel system
in order to allow a complete draining of the fuel tank. In general, the draining of the
tank was accomplished by means of the vehicle fuel pump.
— When possible, the engine was equipped with suitable thermocouples to monitor the
lubricant and coolant temperature.
— When possible, a thermocouple was fitted to the tank in order to monitor gasoline
(liquid) temperature. In some case, an additional thermocouple was used to monitor
vapour temperature in the volume of the tank above the liquid surface.

Canister load to breakthrough

According to the legislation, canister load to breakthrough can be accomplished using normal
fuel by means of repeated heat builds or using butane. The second option was chosen.

The canister was loaded with a mixture composed of 50 % butane and 50 % nitrogen by
volume at a rate of 40 grams butane per hour. Two flow-meters, calibrated with a special
equipment (rotameter), were used to accurately set the required flow.

Breakthrough was determined using an auxiliary carbon canister connected downstream of

the vehicle’s canister. In the legislative procedure the breakthrough is defined as the point
corresponding to an increase of 2.0 grams of the auxiliary canister. The auxiliary canister was
purged with dry air prior to start the butane loading of the main canister.

As soon as the canister reached breakthrough, the nitrogen and butane lines were immediately
shut off. The evaporative emission canister was then reconnected to the fuel system of the
vehicle.

58
In order to accomplish the butane loading of the canister and to simplify the disconnection
and connection of the canister itself, this was removed from the vehicle and equipped with
special quick release connectors incorporating automatic shut-off valves.
Special care was taken during this step to avoid damage to the components and the integrity
of the fuel system.

Canister weight

The weight of the vehicle carbon canister was measured and recorded before and after the
main steps prescribed by the legislative test procedure.
In particular, the following weights were recorded:

1 Initial weight of vehicle canister before saturation

with N2/Butane
2 Final weight of vehicle canister after saturation
with N2/Butane (at breakthrough)
3 Final weight of vehicle canister after the
preconditioning cycle UDC+2EUDC
4 Weight of vehicle canister before conditioning
drive (after pause > 12 hours)
5 Final weight of vehicle canister after conditioning
drive (emissions cycle UDC+EUDC+UDC)
6 Weight of vehicle canister before the UDC-
Diurnal 24h test
7 Weight of vehicle canister after the Diurnal 24h
test

The automatic shut-off valves fitted to the lines connecting the carbon canister to the fuel
system allowed to perform weighing operation without significant loss of vapour from the
fuel system itself.
The complete set of canister weights mentioned above is not available for all the tests
performed on the first two vehicles as the automatic shut-off valves were adopted only after a
few tests had been already carried out.

Vehicle conditioning (Roller bench)

Fuel drain and refill

The tank of the vehicle was emptied using the fuel pump located in the tank. This was
done taking care not to abnormally purge or abnormally load the evaporative control
canister fitted to the tank. Removal of the fuel cap is normally sufficient to achieve this.

After the butane loading of the canister and after it had been reconnected to the fuel
system, the tank was refilled with test fuel at a temperature of about 287 °K (14 °C) to 40 ±
2 % of the tank’s normal volumetric capacity. The fuel cap of the vehicle was fitted at this
point.

Preconditioning drive

59
 Within one hour from the completing of canister loading the vehicle was placed on the
chassis dynamometer and driven through one Part One and two Part Two driving cycles of
Type I test (NEDC cycle). Exhaust emissions were not sampled during this operation.

Soak (before Hot Soak test)

Within five minutes of completing the preconditioning drive, the vehicle was driven off the
chassis dynamometer and parked in the soak area. The vehicle was left parked for a
minimum of 12 hours and a maximum of 36 hours. In most of the cases the soak period
lasted about 14-16 hours.

Emission test
After conclusion of the soak period the vehicle was driven through a complete Type I test
drive (cold start urban and extra urban test). Then the engine was shut off. Exhaust
emissions were measured during this operation. In addition to the regulated emissions, a
sample of the diluted exhaust gas was taken to determine the speciation of unburned HC
and aldehydes.

Conditioning drive
Within two minutes of completing the Type I test, the vehicle was driven a further
conditioning drive consisting of one urban test cycle (hot start) of a Type I test. Then the
engine was shut off again.
Exhaust emissions were not sampled during this operation.

Hot soak evaporative emissions test

— VT SHED conditioning
The ambient temperature of the VT SHED was set at a temperature of 300 ± 1 °K
Before the completion of the conditioning drive the measuring chamber was purged for
several minutes with ambient air.
The hydrocarbon analyzer was zeroed and spanned immediately prior to the test.

— Hot soak test

At the end of the conditioning drive the vehicle was pushed out from the chassis
dynamometer cell with the engine idling. The vehicle was then pushed into the measuring
chamber with the engine switched off. The enclosure door had to be closed and gas-tight
sealed within two minutes of the engine being switched off and within seven minutes of
the end of the conditioning drive. The vehicle’s windows and luggage compartments were
opened just before this stage, if not already opened.
The start of a 60 ± 0.5 minute hot soak period began when the chamber was sealed. The
hydrocarbon concentration, temperature and barometric pressure were measured to give
the initial readings for the hot soak test. These figures are used in the evaporative emission
calculation.
The ambient SHED temperature was kept constant at a temperature of about 300 °K during
the 60-minute hot soak period. Hydrocarbon concentration was measured and recorded
every minute.
The hydrocarbon analyzer was zeroed and spanned immediately before the end of the 60 ±
0.5 minute test period and then the final hydrocarbon concentration in the chamber was
measured.

60
 Soak (after Hot Soak test)
At the end of the Hot Soak test, the VT SHED was opened and purged in order to reach
equilibrium with the ambient air. The temperature of the VT SHED was then set to 293 °K
± 2 °K (20 °C ± 2 °C) and the vehicle left into the VT SHED itself for not less than 6 hours
and not more than 36. In this programme, in all the cases except when explicitly said, the
soak period lasted about 6-7 hours.

Diurnal test
The test vehicle was exposed to one cycle of ambient temperature according to the profile
specified in Appendix 2 of the Directive 98/69/EC with a maximum deviation of ± 2 °K at
any time.

— VT SHED conditioning
The measuring chamber was purged for several minutes immediately prior to the star of
the Diurnal test.

— Diurnal test
The mixing fans of the VT SHED were adjusted in such a way as to maintain a minimum
air circulation of 8 km/h under the fuel tank of the test vehicle.
The hydrocarbon analyser was zeroed and spanned immediately before the test. The
enclosure doors were then closed and gas-tight sealed.
Within 10 minutes of closing and sealing the doors, the hydrocarbon concentration,
temperature and barometric pressure were measured to give the initial readings for the
diurnal test. Hydrocarbon concentration was measured and recorded every minute during
the whole test.
The hydrocarbon analyser was zeroed and spanned immediately before the end of the test
and the final hydrocarbon concentration measured after 24 hours ± 6 minutes after the
beginning of the initial reading.

HC Speciation
A Tedlar bag was filled with a sample of the internal atmosphere of the chamber at the
start and at the end of each Diurnal test.
The sample was taken by means of a sample probe located very close to the one used to
measure, with the FID, the hydrocarbon concentration in the chamber.
The sample was then analyzed using a GC.

61
APPENDIX 4: EVAPORATIVE EMISSIONS AND TEST PROTOCOL RELATED
EFFECTS

Canister Conditioning and Canister Weights

As laid down in the legislated test procedure (see Appendix 3), the carbon canister of each
test vehicle was conditioned before each evaporative emission test following the butane
loading procedure, one of the two alternative procedures permitted. Canisters were loaded to
breakthrough with a mixture of butane (50%) and nitrogen (50%). The weight of the canister
was recorded before and after the butane loading operation to monitor the behaviour of the
canister itself.

An unexpected result of this test programme was the fact that, for a given canister, the
breakthrough condition did not correspond to a constant canister weight. The initial weight of
the canister at the beginning of the butane loading operation is very close to the weight
recorded at the end of the previous test, as nothing was done to the vehicle or canister
between these two steps. Thus the initial weight progressively increased, reflecting the same
pattern as the canister weight recorded at the end of each evaporative test.

As expected, the amount of butane loaded into the canister generally decreased with
increasing initial weight, as illustrated in Figure 15, due to a reduced working capacity of the
canister caused by the higher mass of residual HC vapour in the carbon.
However canister weight recorded after reaching breakthrough also increased noticeably,
especially after the first test, and stabilized only after a certain number of tests as shown in
Fig. 3. In other words, low initial canister weights resulted, once breakthrough had been
reached, in low final canister weights.

The reason that breakthrough weight was not constant could be due to a combination of
factors. The progressively increasing canister hydrocarbon loading having a high molecular
weight could be one reason. This will have a higher mass than an equivalent amount of butane
in the canister pores. Furthermore, adsorption is an exothermic process as demonstrated by
the fact that during butane loading the canister surface becomes warm. The temperature
increase of the activated carbon bed depends on the amount of butane adsorbed and therefore,
when more butane is adsorbed the temperature will be higher. The adsorption capacity of
active carbon is known to be lower at higher temperatures.

Hot Soak Test

The hot Soak test is meant to measure evaporative emissions from a vehicle in a condition
simulating the situation in which the vehicle is driven for a certain distance and then parked
on a warm day. For this reason, before starting the Hot Soak test the vehicle is conditioned by
driving it over a complete New European Driving Cycle (NEDC) followed by an additional
urban driving cycle. As a result of this conditioning drive, the engine is still hot when the test
vehicle is pushed into the VT SHED as shown in Figure 16 where, as an example, a lubricant
temperature profile recorded during Hot Soak test is plotted for each vehicle.

62
Figure 15: Gain in canister weight (mass of butane loaded) during N2/Butane saturation

Gain in canister weight Gain in canister weight

VEHICLE 1 VEHICLE 2
Initial canister weight Initial canister weight
Final canister weight Final canister weigth
30 940 30 920

910

25 920 25
900

890
20 900 20

Canister weight, g

Canister weight, g
880
Weight gain, g

Weight gain, g
15 880 15 870

860

10 860 10
850

840
5 840 5

830

0 820 0 820
Fuel A Fuel B Fuel Fuel Fuel A Fuel Fuel Fuel A Fuel Fuel Fuel Fuel A Fuel Fuel Fuel B Fuel B Fuel A Fuel B Fuel B Fuel Fuel Fuel A Fuel A Fuel Fuel Fuel A Fuel Fuel Fuel A
A5S B5S A10S B10S A10E B10E B10E A5E B5E A5S B5S A10S B10S A10S B10S

Gain in canister weight

VEHICLE 3 VEHICLE 4
Initial canister weight
Final canister weight
70 680 60 750
Gain in canister weight
Initial canister weight
ExtPg ExtPg ExtPg ExtPg ExtPg
660 Final canister weight
60
50 700

640
50

40 650
Canister weight, g

620

Canister weight, g
Weight gain, g

Weight gain, g
40

600
30 600
30
580
20 550
20
560

10 10 500
540

0 520
0 450
Fuel A Fuel B Fuel Fuel Fuel A Fuel A Fuel Fuel B Fuel Fuel Fuel Fuel A Fuel Fuel
Fuel A Fuel B Fuel Fuel Fuel Fuel A Fuel Fuel Fuel Fuel Fuel Fuel Fuel A Fuel B Fuel Fuel Fuel
A5S B5S B5S B5S A5E B5E A5S B5S
A5S B5S A5S B5S B5S A5E B5E B10E A10E A10S B10S B5S

Gain in canister weight Gain in canister weight

VEHICLE 5 VEHICLE 6
Initial canister weight Initial canister weight
Final canister weight Final canister weight
25 720 45 740

40 720
700
20
35 700
680
30 680
Canister weight, g

Canister weight, g
Weight gain, g

15
Weight gain, g

660
25 660

640 20 640
10

15 620
620

5 10 600

600
5 580

0 580 0 560
Fuel A Fuel B Fuel Fuel Fuel A Fuel Fuel Fuel A Fuel Fuel Fuel A Fuel Fuel Fuel B Fuel A Fuel A Fuel B Fuel A5S Fuel A5S Fuel B5S Fuel A Fuel A5S Fuel Fuel Fuel Fuel A5E Fuel B5E Fuel A
A5S B5S A5E B5E B10S B10E A5S B5S A10S A10S B10S

Gain in canister weight

VEHICLE 7 Initial canister weight
Final canister weight
60 2740

2720
50

2700
40
Canister weight, g
Weight gain, g

2680

30

2660

20
2640

10
2620

0 2600
Fuel A Fuel A Fuel B Fuel B Fuel A5S Fuel B5S Fuel A5S Fuel B10S Fuel A

63
In case of a vehicle equipped with a fuel system featuring fuel return, the temperature of
gasoline in the tank could be well higher than the ambient temperature due to heated gasoline
returning to the tank. Two vehicles of the test fleet used in this programme had a fuel return
(vehicles 1 and 7). It was possible to monitor the fuel temperature only in the case of vehicle
1; as shown in Fig. 16, in this case the fuel return did not influence significantly the
temperature of the gasoline in the tank.

Figure 16: Typical gasoline and engine lubricant temperature profiles during the Hot
Soak test (not available for vehicle 7)
HOT SOAK TEST
Gasoline in Tank: Temperature Profile

35

30

25

Vehicle 1
Temperature (C)

20 Vehicle 2
Vehicle 3
Vehicle 4
15 Vehicle 5
Vehicle 6

10

0
0 10 20 30 40 50 60 70
Time (min)

HOT SOAK TEST

Lubricant Temperature Profile

120

100

80
Vehicle 1
Temperature (C)

Vehicle 2
Vehicle 3
60
Vehicle 4
Vehicle 5
Vehicle 6
40

20

0
0 10 20 30 40 50 60 70
Time (min)

64
The bar charts showing the time series of Hot Soak evaporative emissions results are
presented in Figure 17. Hot Soak emissions were in general low and exhibited a high degree
of variability; consequently it is not possible to identify any pattern of Hot Soak evaporative
emissions or effect of fuel quality

For all test vehicles Hot Soak emission levels were also quite low compared to diurnal
emissions. This is reflected in the small contribution of the hot soak test to the total
evaporative emissions. The average Hot Soak contribution, as a percentage of the total
evaporative emissions (Diurnal +Hot Soak) ranged from 3 % to 13%, with similar results for
all the vehicles, apart perhaps for vehicle 1. This was true even for vehicles having very
different evaporative emissions (e.g. vehicle 5 and vehicle 7). This suggests that, comparing
different vehicles, hot soak and diurnal emissions mainly come from the same sources.

Table A4.1: Hot Soak contribution to total evaporative emissions (%)

Hot Soak contribution to total Total emissions – Diurnal+Hot

evaporative emissions (%) Soak (grams/test)
Min - Max Average Min - Max Average
Vehicle 1 1–6% 3% 1.17 – 3.92 1.90
Vehicle 2 3 – 17 % 7% 0.48 – 1.36 1.01
Vehicle 3 5 – 10 % 8% 1.58 – 5.44 2.59
Vehicle 4 3–9% 7% 1.81 – 8.01 3.36
Vehicle 5 11 – 16 % 13 % 1.05 – 7.81 2.36
Vehicle 6 3 – 21 % 10 % 0.42 – 1.37 0.71
Vehicle 7 5 -16 % 11% 0.74 – 1.35 0.91

65
Figure 17: Hot Soak evaporative emissions

Evaporative - Hot Soak

Vehicle 1 Vehicle 2

0.10 0.12
0.09
0.08 0.10
0.07 0.08
0.06

g/test
g/test

0.05 0.06
0.04
0.04
0.03
0.02 0.02
0.01
0.00 0.00
A B A5S B5S A A10SB10S A A10EB10EB10E A A5E B5E B B A B B A5S B5S A A A10S B10S A A10S B10S A

Fuel Fuel

Vehicle 3 Vehicle 4
0.50 0.25
0.45
0.40 0.20
0.35
0.30 0.15
g/test

0.25 g/test
0.20 0.10
0.15
0.10 0.05
0.05
0.00 0.00
A B A5S B5S A A B5S B B5S A5E B5E A A5S B5S A B A5S B5S A5S A B5S B5S A5E B5E B10EA10E A B A10SB10S B5S
Ext pg Ext pg Ext pg
Fuel Ext pg Ext pg
Fuel

Vehicle 5 Vehicle 6

0.45 0.16
0.40 0.14
0.35
0.12
0.30
0.10
g/test

0.25
g/test

0.20 0.08
0.15 0.06
0.10 0.04
0.05
0.02
0.00
0.00
A

A5S

B5S

A5E

B5E

B10S

B10E

A5S

B5S

A B A5S A5S B5S A A5S A10S A10S B10S A5E B5E A

FID
Aborted
Fuel Fuel

Vehicle 7

0.16
0.14
Key:
0.12
0.10 ExtPg: Extra purging carried out prior to test
g/test

0.08 FID: FID analyzer saturated, result invalid

0.06
0.04 Aborted: Test aborted due to technical problems

0.02
0.00
A A B B A5S B5S A5S B10S A
Fuel

66
 Diurnal Test

During Diurnal tests VT SHED temperature is varied between 20 and 35 °C to simulate

ambient temperature fluctuations occurring during a typical summer day. The Diurnal test has
to be started at least 6 hours and maximum 36 hours after the end of Hot Soak test to allow
cooling down of test vehicle. In this programme, after the Hot Soak test the vehicles were left
inside the VT SHED at a temperature of 20 °C for 6-7 hours. This period of time was always
sufficient to cool the vehicles down to 20 °C that was, in all cases, the initial temperature of
the lubricant and of the gasoline in the tank at the start of the Diurnal test.

The temperature profile of the Diurnal test is specified in the Directive 98/69/EC; the
maximum temperature of 35 °C is reached after 12 hours, then the temperature starts
decreasing. The temperature profile of Diurnal test is plotted in Figure 18.
Gasoline temperature does not follow the same temperature profile of the air contained in the
VT SHED. In fact, the tank is filled with a volume of gasoline corresponding to 40% of the
tank capacity that means typically about 20-30 litres of gasoline. Gasoline in the tank has a
certain thermal inertia and therefore its temperature increases more slowly compared to
ambient temperature. This also means that the maximum temperature reached by gasoline is
lower than 35 °C (usually 1-2 °C lower). This behaviour is shown in Fig. 18 where the
temperature profile of the air in the VT SHED and of the gasoline in the vehicle tank is
plotted over the Diurnal test duration. The gasoline temperature profile over the diurnal test
resulted very similar for all the vehicles.

Figure 18: Typical temperature profiles of gasoline vs. Diurnal test temperature profile
(not available for vehicle 7)
Diurnal Test
Gasoline Temperature vs Shed Temperature
40

35

30

25
Temperature (C)

20
Vehicle 1
15 Vehicle 2
Vehicle 3
Vehicle 4
10 Vehicle 5
Vehicle 6
5 Temperature Profile

0
0 200 400 600 800 1000 1200 1400
Time (min)

The VOC concentration profiles over time for all the test vehicles and for the extreme fuels in
terms of DVPE (fuel A and Fuel B5S) are shown in Fig. 19. The plots show that the Diurnal

67
test can be divided into three phases on the basis of the temperature profile and on the
consequent VOC concentration increase rate:

- Phase 1 (0 – 200 min): temperature in the VT SHED increases so slowly that after 200
minutes it is still below 22 °C. During this phase, the VOC concentration increased in most
of the cases at a very low rate. Due to the little increase of temperature, evaporation of
gasoline is limited and vapour pressure in the tank does not increase significantly.
Furthermore, at this stage the activated carbon in the canister normally has still a large
adsorbing capacity and therefore the canister is able to trap any gasoline vapours coming from
the tank. For the above mentioned reasons, VOC concentration increase observed in this
phase is likely to be due mainly to fuel permeation and bleed emissions.

- Phase 2 (201 – 800 min): temperature starts being raised more quickly and reaches its
maximum value (35 °C) after 660 min. Also gasoline temperature increases (see Fig. 18) and
reaches its maximum after about 800 min. The increase of fuel temperature obviously leads to
an increase of the internal pressure of the tank and to compensate this increase some air has to
flow from the tank towards the external ambient. It is during this phase that the canister plays
a very important role in avoiding a release of hydrocarbons into the atmosphere.
The VOC concentration profile recorded during this part of Diurnal test was dependent on the
vehicle (each test vehicle exhibited a different behaviour) and also on the fuel tested. In most
of the cases and especially with the high DVPE fuels a sharp change in the slope of the VOC
concentration profile was noticed in the neighbourhood of the maximum temperature in the
VT SHED. As a consequence the major part of evaporative emissions turned out to be emitted
by the vehicle during this phase.
It is not easy to identify what sources contribute more to evaporative emissions during this
phase; in theory it is in this part of the Diurnal test that breathing losses can occur as there is
an outgoing flow from the tank towards the external environment. However, a properly
working canister should be able to trap all the hydrocarbons vapours generated in the tank; on
the contrary, vapours can freely escape from the tank if the canister has reached the
breakthrough condition but this occurrence can be usually identified.
On the other hand bleed emissions and fuel permeation through hoses and plastic materials
are still present; fuel permeation rate is known to be influenced by temperature (typically
doubles for a 10º C rise in temperature [3]) while it is not clear to what extent bleed emissions
are affected by a temperature change.

- Phase 3 (801 – 1440 min): the VT SHED is cooled down back to the initial
temperature of 20 °C. In most of the tests performed the VOC concentration did not increase
further in this phase or increased at a very reduced rate. To explain this behaviour two factors
have to be taken into consideration: first of all, the falling temperature in the VT SHED, and
consequently of the gasoline in the tank, results in a decrease of pressure in the tank itself; as
a consequence, air starts flowing from the atmosphere into the tank to compensate for the
pressure decrease.
Similarly, due to the volume compensation system, air starts flowing into the VT SHED to
compensate the decreasing pressure in the measuring chamber. Air is taken from the
atmosphere outside the VT SHED and usually contains only few ppm of hydrocarbons
corresponding to the VOC background levels; as a results, the internal atmosphere of the VT
SHED is diluted and a decrease of VOC concentration can be even noticed.
At this stage, only fuel permeation and non fuel-related sources can contribute to evaporative
emissions.

68
 Figure 19: Typical VOC concentration profile over Diurnal test with fuel A and B5S

DIURNAL TEST DIURNAL TEST

VOC CONCENTRATION VOC CONCENTRATION
Vehicle 1 Vehicle 2

140 40
35 40

35 30 35
120
30

CONCENTRATION (ppm)
30 25

Temperature Profile (C)

CONCENTRATION (ppm)

100

Temperature Profile (C)

25
25
20
80
20
20
15
60 15
15
10
40 10
10
5 5
20 5

0 0
0 0
0 200 400 600 800 1000 1200 1400
0 200 400 600 800 1000 1200 1400 TIME (Minutes)
TIME (Minutes)

Fue A Fuel B5S Temperature Profile Fuel A Fuel B5S Temperature

DIURNAL TEST DIURNAL TEST

SHED VOC CONCENTRATION VOC CONCENTRATION
Vehicle 3 Vehicle 4

200 40
250 40
180 35
35
160
30 200
CONCENTRATION (ppm)

140 30
Temperature Profile (C)

CONCENTRATION (ppm)

Temperature Profile (C)

120 25
150 25
100 20 20
80 15 100 15
60
10 10
40 50
20 5 5
0 0 0 0
0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400
TIME (Minutes)
TIME (Minutes)

Fuel A Fuel B5S Temperature Profile Fuel A Fuel B5S Temperature Profile

DIURNAL TEST DIURNAL TEST

SHED VOC CONCENTRATION VOC CONCENTRATION
Vehicle 5 Vehicle 6

90 40 50 40
80 35 45 35
70 40
30 30
CONCENTRATION (ppm)

Temperature Profile (C)

CONCENTRATION (ppm)

35

Temperature Profile (C)

60
25 25
50 30
20 25 20
40
15 20
30 15
10 15
20 10
10
10 5
5 5
0 0
0 0
0 200 400 600 800 1000 1200 1400
0 200 400 600 800 1000 1200 1400
TIME (Minutes)
TIME (Minutes)

Fuel A Fuel B5S Temperature Fuel A Fuel B5S Temperature Profile

DIURNAL TEST
VOC CONCENTRATION
Vehicle 7

40 40

35 35

30 30
CONCENTRATION (ppm)

Temperature Profile (C)

25 25

20 20

15 15

10 10

5 5

0 0
0 200 400 600 800 1000 1200 1400
TIME (Minutes)

Fuel A Fuel B5S Temperature Profile

69
APPENDIX 5: EVAPORATIVE EMISSION VOC SPECIATION AND ETHANOL
CONTENT: ANALYTICAL METHOD

VOC Speciation
Ten liters of air sampled from the VT SHED were collected in Tedlar™ bags. A 2D–Gas
Chromatograph (2D-GC) equipped with dual flame ionisation detector (FID) is used for the
determination of the C2-C9 ozone precursor hydrocarbons specified in the Ozone Directive
2002/3/EC.

Figura – thermal desorption gas chromatograph

Chromatography analysis (2D-GC)

The Tedlar™ bags containing the exhausts are connected to the analyzer inlet. Only a limited
sample volume of 80 ml (20 ml/min for 4 minutes) was collected and analyzed for each
driving test, allowing replicated analysis if necessary. A thermal desorption unit (UNITY™)
and an auxiliary sampling device (Air Server™) from Markes International (Pontyclun, UK)
are used to collect samples from the Tedlar bags of exhaust gases.
Samples are transferred to a multilayer adsorbent trap (Markes) that operates at -15°C. The
trap is then heated up to 320°C and a whole-sample thermo-desorption / injection step
follows. The released compounds are injected into the gas chromatograph via a heated
transfer line (130°C) lined with uncoated, deactivated fused silica.
Analysis were performed with a gas chromatograph (6890, Agilent, Wilmington, DE, USA)
equipped with a “Deans switch” followed by a two column system with a dual flame
ionisation detector (FID). [1]

[1] Latella, A., Stani G., Cobelli, L., Duane M., Junninen, H., Astorga, C. Larsen, B. R., Semicontinuous
GC analysis and Receptor Modeling For Source Apportionment of ozone precursor hydrocarbons in Bresso,
Milan 2003. J. Chromatography A, 1071 (2005)29-39.

70
 UNITY™
Thermal
Desorber

UNITY
Air Server™
module

Pump

LT: connection tube

1-4 : solenoid valves
P : pressure transducer
MFC : mass flow controller
NV : needle valve

Thermal desorption unit

All the effluent, transferred to the gas chromatograph, is directed into the first column
(“PLOT” 50m×0.32mm, coated with Al2O3 / KCl, 8μm). For the first 15 minutes the C2-C5
compounds elute from this column and are sent, via the “Deans switch”, towards FID 2.
Afterwards the “Deans switch” activates directing the C6-C9 fraction into the second column
(“WCOT” 50m×0.32mm, coated with polidimethyl-siloxane, 1.2 μm) and then to FID 1.
From this point on two chromatograms are produced in parallel, one on each FID.
The two columns work with the same program of temperatures: isotherm at 40°C for 2
minutes, ramp at 7 °C/min, isotherm at 120°C for 3 minutes, ramp at 7 °C/min, isotherm at
200°C for 35 minutes. The carrier gas is Helium.

71
 Figure – 2D gas chromatograph configuration

Finally, the VOCs emissions (g/km) are calculated as follows:

Cvi ⋅ PMi ⋅ Vtot

Mvi =
D ⋅ d ⋅ 1000
Mvi = mass of VOCi emitted expressed as g/Km
Cvi = concentration of VOCi expressed as ppm
Pmi = molecular weight of VOCi (g/mol)
Vtot = Vmix + VPM + VLPI (m3 )
D = 24 mol/l
d = distance in km

ETHANOL Analysis

The polarity of ethanol is too high for chromatography with polar column/adsorbent material,
such as the ones necessary for the VOC analysis. Hence, a method is used by which ethanol is
sample-enriched from the Tedlar bags containing the evaporative emissions onto a mild
adsorbent (Tenax TA) and successively thermally desorbed and analysed by GC-MS. This
method has been adapted from a previously described method for analysis of polar organic
compounds (49).
The procedure for analysis is as follows: 1000 mL of air is sampling onto a Tenax TA
cartridge with a sampling rate 200 mL/min. The adsorption trap is thermally desorbed using

72
an automatic desorption system (Perkin Elmer ATD 400) at 200 °C for 20 min and re-trapped
on a Tenax-packed cold trap maintained at -20 °C. The trapped compound is transferred by
ballistic heating of the trap to 200 °C (1 min hold time) into a HP 5890 GC with a MEGA
OV-1, 30 m x 0.32 I.D., 0.1-0.15 μm film thickness. The GC is interfaced to a HP 5970 MSD
(Mass Selective Detector) operating in EI (70 eV) scan mode, scanning from 30 to 200 m/z.
The column temperature is held at 30 °C for 4 min, raised from 30 to 240 °C at 8 °C/min,
with a final holding time of 6 min. This produces a complete run time of 20 min. The carrier
gas for thermal desorption as well as gas chromatography is helium. Ethanol is quantified
from the integrated signal of the single ion extraction chromatogram (m/z 46) in comparison
with external standards produced by spiking blank Tenax TA cartridges with known amounts
of ethanol.

73
VOC speciation detailed results
All emissions g/test
Vehicle 1 A B A5S A A10S B10S A A10S B10E B5E A5S
C4- Alkanes 0.371 1.13 0.698 0.228 0.357 1.50 0.285 0.310 0.565 0.443 0.973
C4- Unsaturates 0.158 0.135 0.101 0.0583 0.079 0.225 0.048 0.058 0.103 0.064 0.116
C5+ Alkanes 0.677 1.012 0.585 0.548 0.833 1.63 0.710 0.901 0.825 0.636 0.502
C5+ Unsaturates 0.0607 0.0286 0.0000 0.0552 0.0736 0.1177 0.0635 0.0843 0.0876 0.0583 0.184
Aromatics 0.00657 0.225 0.134 0.304 0.247 0.338 0.232 0.317 0.263 0.227 0.196
Benzene 0.00563 0.00675 0.00394 0.00896 0.00926 0.01121 0.00754 0.0138 0.0190 0.0153 0.0227
Ethanol 0 0 0 0.0171 0.0208 0.0375 0.0143 0.0254 0.0263 0.0230 0.0131
Total 1.28 2.54 1.52 1.22 1.62 3.86 1.36 1.71 1.89 1.47 2.01

Vehicle 2
Fuel A B A5S B5S A A A10S B10S A A10S A
C4- Alkanes 0.0252 0.137 0.0631 0.0817 0.0557 0.0400 0.0457 0.0897 0.0467 0.0565 0.0501
C4- Unsaturates 0.00325 0.0253 0.0137 0.0386 0.0216 0.0223 0.0348 0.0319 0.0285 0.0279 0.0377
C5+ Alkanes 0.305 0.574 0.275 0.337 0.511 0.382 0.480 0.388 0.526 0.729 0.502
C5+ Unsaturates 0 0 0.0274 0.0332 0.0429 0.0460 0.0196 0.0471 0.0523 0.0275 0.0480
Aromatics 0.0569 0.217 0.272 0.215 0.293 0.235 0.372 0.416 0.461 0.258 0.439
Benzene 0.00848 0.00700 0.00617 0.00722 0.01252 0.00886 0.00989 0.0104 0.0106 0.0123 0.0104
Ethanol 0 0 0.00656 0.0256 0.0223 0.0161 0.0320 0.0643 0.0349 0.0472 0.0354
Total 0.40 0.96 0.66 0.74 0.96 0.75 0.99 1.05 1.16 1.16 1.12

Vehicle 3
Fuel A B A5S A A B5S B B5S A5E B5E A A5S B5S
C4- Alkanes 0.176 0.239 0.355 0.273 0.119 0.091 0.133 0.293 0.114 0.218 0.0794 0.0716 0.783
C4- Unsaturates 0.0372 0.0612 0.118 0.125 0.0581 0.0343 0.0458 0.0721 0.0498 0.0572 0.0453 0.0460 0.227
C5+ Alkanes 0.962 1.19 1.50 2.24 1.39 1.10 1.12 1.30 1.39 1.04 1.11 1.13 3.62
C5+ Unsaturates 0.0370 0.0634 0.0715 0.0443 0.0367 0.0264 0.0133 0.0194 0.0163 0.0412 0.0309 0.0241 0
Aromatics 0.197 0.255 0.283 0.417 0.285 0.244 0.463 0.401 0.396 0.555 0.303 0.338 0.267
Benzene 0.0204 0.0206 0.0237 0.0217 0.0122 0.0105 0.0129 0.0160 0.0160 0.0304 0.0190 0.0161 0
Ethanol 0 0 0 0.03 0.01 0.01 0.02 0.04 0.03 0.05 0.02 0.02 0
Total 1.43 1.83 2.35 3.15 1.91 1.51 1.81 2.14 2.01 1.99 1.61 1.65 4.90

74
Vehicle 4 Polo
Fuel A B A5S A B5S B5S A5E B5E B10E A10E A B
C4- Alkanes 0.283 0.266 0.083 0.290 1.250 2.431 0.117 0.359 0.231 0.135 0.193 0.298
C4- Unsaturates 0.0444 0.0541 0.0349 0.0585 0.0899 0.131 0.0523 0.0576 0.0451 0.0230 0.0493 0.0326
C5+ Alkanes 0.728 1.10 1.23 1.55 1.79 2.95 1.57 1.33 1.29 1.76 1.84 1.67
C5+ Unsaturates 0.106 0.0837 0.181 0 0 0.251 0.104 0.137 0.0143 0.0496 0.0671 0.113
Aromatics 0.481 0.480 0.550 0.678 0.855 0.855 0.412 0.510 0.668 0.353 0.471 0.391
Benzene 0.0253 0.0156 0.0153 0.0232 0.0278 0.0242 0.0164 0.0271 0.0141 0.0241 0.0253 0.0162
Ethanol 0 0 0.0140 0.00400 0.0222 0.00789 0.0419 0.0130 0.0140 0.0131 0.0135 0.00457
Total 1.67 2.00 2.11 2.60 4.04 6.65 2.32 2.44 2.27 2.35 2.66 2.53

Vehicle 5
Fuel B10S A A5S B5S B A A A
C4- Alkanes 0.482 0.122 0.133 0.610 2.435 0.254 0.0262 0.0325
C4- Unsaturates 0.596 0.0598 0.0571 0.0949 0.348 0.0636 0.0362 0.0255
C5+ Alkanes 1.74 1.39 1.20 1.55 6.53 0.985 0.771 0.837
C5+ Unsaturates 0.0788 0.116 0.0585 0.0920 0.0294 0.0647 0.0167 0.0244
Aromatics 0.234 0.151 0.753 0.814 0.969 0.429 0.353 0.391
Benzene 0.0127 0.0280 0.0221 0.0244 0.0254 0.0352 0.0103 0.00105
Ethanol 0.0606 0.171 0.123 0.171 0.0760 0.0396 0.0122 0.00898
Total 3.20 2.04 2.34 3.35 10.41 1.87 1.23 1.32

Vehicle 6
Fuel A10S B10S A5E B5E
C4- Alkanes - 0.475 0.0194 0.241
C4- Unsaturates - 0.0385 0.0140 0.0239
C5+ Alkanes - 0.480 0.327 0.369
C5+ Unsaturates - 0.0147 0.0630 0.0157
Aromatics - 0.223 0.142 0.205
Benzene - 0.00554 0.00368 0.00761
Ethanol 0.0325 0.0524 0.0414 0.0491
Total 0.66 1.29 0.61 0.91

75
Vehicle 7
Fuel A B B5S B10S
C4- Alkanes 0.0216 0.0285 0.0523 0.212
C4- Unsaturates 0.0111 0.0069 0.0136 0.0249
C5+ Alkanes 0.577 0.514 0.616 0.846
C5+ Unsaturates 0.00452 0.00294 0.00623 0.0258
Aromatics 0.136 0.121 0.0950 0.119
Benzene 0.00470 0.00484 0.00268 0.00592
Ethanol 0 0 0.00819 0.0113
Total 0.76 0.68 0.79 1.25

76
APPENDIX 6: EXHAUST EMISSION RESULTS
(Where blanks appears, no valid measurement was taken on that test)

NEDC - HC

Vehicle 1 Vehicle 2

0.070 0.050
0.045
0.060
0.040
0.050 0.035
0.030

g/km
0.040
g/km

0.025
0.030 0.020
0.020 0.015
0.010
0.010 0.005
0.000 0.000
A B B A5S B5S A A A10S B10S A A10S B10S A
A B A5S B5S A A10SB10S A A10EB10EB10E A A5E B5E B B
Fuel Fuel

Vehicle 3 Vehicle 4

0.090 0.070
0.080 0.060
0.070
0.050
0.060
g/km

g/km
0.050 0.040
0.040 0.030
0.030
0.020
0.020
0.010 0.010
0.000 0.000
A B A5S B5S A A B5S B B5S A5E B5E A A5S B5S A B A5S B5S A5S A B5S B5S A5E B5E B10EA10E A B A10SB10S B5S

Fuel Fuel

Vehicle 5 Vehicle 6

0.080 0.040
0.070 0.035
0.060 0.030
0.050 0.025
g/km

g/km

0.040 0.020
0.030
0.015
0.020
0.010
0.010
0.005
0.000
0.000
A
B
A5S
B5S
A
A5E
B5E
A
B10S
B10E
A
A5S
B5S
B
A
A
A
A
A
B
A
A
A5E
A5E
A5E
B5E
A5S
A5S
A5S
A
A

A B A5S A5S B5S A A5S A10S A10S B10S A5E B5E A

Fuel Fuel

Vehicle 7

0.080
0.070
0.060
0.050
g/km

0.040
0.030
0.020
0.010
0.000
A A B B A5S B5S A5S B10S A

Fuel

77
 NEDC - CO

Vehicle 1 Vehicle 2

0.40 0.35
0.35 0.30
0.30 0.25
0.25
0.20

g/km
g/km

0.20
0.15
0.15
0.10
0.10
0.05 0.05

0.00 0.00
A B A5S B5S A A10S B10S A A10E B10E B10E A A5E B5E B B A B B A5S B5S A A A10S B10S A A10S B10S A

Fuel Fuel

Vehicle 3 Vehicle 4

0.40 0.80
0.35 0.70
0.30 0.60
0.25 0.50
g/km

g/km
0.20 0.40
0.15 0.30
0.10 0.20
0.05 0.10
0.00 0.00
A B A5S B5S A A B5S B B5S A5E B5E A A5S B5S
A B A5S B5S A5S A B5S B5S A5E B5E B10EA10E A B A10SB10S B5S
Ext pg Ext pg Ext pg Ext pg Ext pg
Fuel Fuel

Vehicle 5 Vehicle 6

0.60 0.60

0.50 0.50
0.40 0.40
g/km

g/km

0.30 0.30
0.20
0.20
0.10
0.10
0.00
0.00
A
B
A5S
B5S
A
A5E
B5E
A
B10S
B10E
A
A5S
B5S
B
A
A
A
A
A
B
A
A
A5E
A5E
A5E
B5E
A5S
A5S
A5S
A
A

A B A5S A5S B5S A A5S A10S A10S B10S A5E B5E A

Fuel Fuel

Vehicle 7

0.25

0.20

0.15
g/km

0.10

0.05

0.00
A A B B A5S B5S A5S B10S A
Fuel

78
 NEDC - NOx

Vehicle 1 Vehicle 2

0.040 0.050
0.035 0.045
0.040
0.030
0.035
0.025 0.030

g/km
g/km

0.020 0.025
0.015 0.020
0.015
0.010
0.010
0.005 0.005
0.000 0.000
A B A5S B5S A A10S B10S A A10E B10E B10E A A5E B5E B B A B B A5S B5S A A A10S B10S A A10S B10S A

Fuel Fuel

Vehicle 3 Vehicle 4

0.120 0.120

0.100 0.100

0.080 0.080
g/km

g/km
0.060 0.060

0.040 0.040

0.020 0.020
0.000 0.000
A B A5S B5S A A B5S B B5S A5E B5E A A5S B5S
A B A5S B5S A5S A B5S B5S A5E B5E B10EA10E A B A10SB10S B5S
Fuel Fuel

Vehicle 5 Vehicle 6

0.045 0.080
0.040 0.070
0.035
0.060
0.030
0.050
g/km

0.025
g/km

0.020 0.040
0.015 0.030
0.010 0.020
0.005
0.010
0.000
0.000
A
B
A5S
B5S
A
A5E
B5E
A
B10S
B10E
A
A5S
B5S
B
A
A
A
A
A
B
A
A
A5E
A5E
A5E
B5E
A5S
A5S
A5S
A
A

A B A5S A5S B5S A A5S A10S A10S B10S A5E B5E A

Fuel Fuel

Vehicle 7

0.080
0.070
0.060
0.050
g/km

0.040
0.030
0.020
0.010
0.000
A A B B A5S B5S A5S B10S A
Fuel

79
 NEDC - CO2

Vehicle 1 Vehicle 2

200 180
180 160
160 140
140 120
120
100

g/km
g/km

100
80
80
60
60
40 40
20 20
0 0
A B A5S B5S A A10S B10S A A10E B10E B10E A A5E B5E B B A B B A5S B5S A A A10S B10S A A10S B10S A

Fuel Fuel

Vehicle 3 Vehicle 4

200 180
180 160
160 140
140 120
120
g/km

g/km
100
100
80 80
60 60
40 40
20 20
0
0
A B A5S B5S A A B5S B B5S A5E B5E A A5S B5S
A B A5S B5S A5S A B5S B5S A5E B5E B10EA10E A B A10SB10S B5S
Fuel Fuel

Vehicle 5 Vehicle 6

250 250

200 200

150 150
g/km

g/km

100
100
50
50
0
0
A
B
A5S
B5S
A
A5E
B5E
A
B10S
B10E
A
A5S
B5S
B
A
A
A
A
A
B
A
A
A5E
A5E
A5E
B5E
A5S
A5S
A5S
A
A

A B A5S A5S B5S A A5S A10S A10S B10S A5E B5E A

Fuel Fuel

Vehicle 7

300

250

200
g/km

150

100

50

0
A A B B A5S B5S A5S B10S A
Fuel

80
 NEDC - FC - l/100km
Vehicle 1 Vehicle 2

9 8
8 7
7 6
6
5

l/100km
l/100km

5
4
4
3
3
2 2

1 1
0 0
A B A5S B5S A A10S B10S A A10E B10E B10E A A5E B5E B B A B B A5S B5S A A A10S B10S A A10S B10S A

Fuel Fuel

Vehicle 3 Vehicle 4

8 8
7 7
6 6
5 5
l/100km

l/100km
4 4
3 3
2 2
1 1
0
0
A B A5S B5S A A B5S B B5S A5E B5E A A5S B5S
A B A5S B5S A5S A B5S B5S A5E B5E B10EA10E A B A10SB10S B5S
Fuel Fuel

Vehicle 5 Vehicle 6

10 9
9 8
8 7
7
6
l/100km

6
l/100km

5 5
4 4
3 3
2 2
1
1
0
0
A
B
A5S
B5S
A
A5E
B5E
A
B10S
B10E
A
A5S
B5S
B
A
A
A
A
A
B
A
A
A5E
A5E
A5E
B5E
A5S
A5S
A5S
A
A

A B A5S A5S B5S A A5S A10S A10S B10S A5E B5E A

Fuel Fuel

Vehicle 7

12

10

8
l/100km

0
A A B B A5S B5S A5S B10S A
Fuel

81
 NEDC - Energy Consumed

Vehicle 1 Vehicle 2

300 250

250 200
200

MJ/100km
MJ/100km

150
150
100
100
50
50

0 0
A B A5S B5S A A10S B10S A A10E B10E B10E A A5E B5E B B A B B A5S B5S A A A10S B10S A A10S B10S A

Fuel Fuel

Vehicle 3 Vehicle 4

250 250

200 200
MJ/100km

MJ/100km
150 150

100 100

50 50

0
0
A B A5S B5S A A B5S B B5S A5E B5E A A5S B5S
A B A5S B5S A5S A B5S B5S A5E B5E B10EA10E A B A10SB10S B5S
Fuel Fuel

Vehicle 5 Vehicle 6

300 300

250 250
200
MJ/100km

200
MJ/100km

150 150
100
100
50
50
0
0
A
B
A5S
B5S
A
A5E
B5E
A
B10S
B10E
A
A5S
B5S
B
A
A
A
A
A
B
A
A
A5E
A5E
A5E
B5E
A5S
A5S
A5S
A
A

A B A5S A5S B5S A A5S A10S A10S B10S A5E B5E A

Fuel Fuel

Vehicle 7

400
350
300
MJ/100km

250
200
150
100
50
0
A A B B A5S B5S A5S B10S A
Fuel

82
 NEDC - Formaldehyde
Vehicle 1 Vehicle 2

0 0

μg/km
μg/km

0 0
A B A5S B5S A A10S B10S A A10E B10E B10E A A5E B5E B B A B B A5S B5S A A A10S B10S A A10S B10S A

Fuel Fuel

Vehicle 3 Vehicle 4

120 300

100 250

80 200
μg/km

μg/km
60 150
40 100
20 50
0 0
A B A5S B5S A A B5S B B5S A5E B5E A A5S B5S
A B A5S B5S A5S A B5S B5S A5E B5E B10EA10E A B A10SB10S B5S
Ext pg Ext pg Ext pg Ext pg Ext pg
Fuel Fuel

Vehicle 5 Vehicle 6

450 120
400
100
350
300 80
μg/km

250
μg/km

200 60
150
40
100
50 20
0
0
A
B
A5S
B5S
A
A5E
B5E
A
B10S
B10E
A
A5S
B5S
B
A
A
A
A
A
B
A
A
A5E
A5E
A5E
B5E
A5S
A5S
A5S
A
A

A B A5S A5S B5S A A5S A10S A10S B10S A5E B5E A

Outlier Fuel Fuel

Vehicle 7

180
160
140
120
μg/km

100
80
60
40
20
0
A A B B A5S B5S A5S B10S A
Fuel

83
 NEDC - Acetaldehyde

Vehicle 1 Vehicle 2

0 0

μg/km
μg/km

0 0
A B A5S B5S A A10S B10S A A10E B10E B10E A A5E B5E B B A B B A5S B5S A A A10S B10S A A10S B10S A

Fuel Fuel

Vehicle 3 Vehicle 4

160 450
140 400

120 350
300
100
μg/km

μg/km
250
80
200
60
150
40
100
20
50
0 0
A B A5S B5S A A B5S B B5S A5E B5E A A5S B5S
A B A5S B5S A5S A B5S B5S A5E B5E B10EA10E A B A10SB10S B5S
Fuel Fuel

Vehicle 5 Vehicle 6

400 300
350
250
300
250 200
μg/km

μg/km

200 150
150
100 100

50 50
0
0
A
B
A5S
B5S
A
A5E
B5E
A
B10S
B10E
A
A5S
B5S
B
A
A
A
A
A
B
A
A
A5E
A5E
A5E
B5E
A5S
A5S
A5S
A
A

A B A5S A5S B5S A A5S A10S A10S B10S A5E B5E A

Fuel Fuel

Vehicle 7

500
450
400
350
300
μg/km

250
200
150
100
50
0
A A B B A5S B5S A5S B10S A
Fuel

84
APPENDIX 7 – Aldehyde Emissions: analitycal method and statistical analysis details

Tab. A7.1, - Aldehyde emissions: table of means

Car Fuel Aldehydes (mg/km) Car Fuel Aldehydes (mg/km)
Acetald Acetald
Formald. . Formald. .

Vehicle 3 A 56.63 106.74 Vehicle 6 A 70.08 109.36

Vehicle 3 A5E 106.81 134.95 Vehicle 6 A5E 92.28 138.78
Vehicle 3 A5S 27.62 36.44 Vehicle 6 A5S 57.55 90.86
Vehicle 3 B 63.13 58.00 Vehicle 6 A10S 83.01 211.54
Vehicle 3 B5E 60.54 83.24 Vehicle 6 B 39.65 94.40
Vehicle 3 B5S 66.81 94.57 Vehicle 6 B5E . .
Vehicle 6 B5S 62.85 98.98
Vehicle 4 A 81.88 93.70 Vehicle 6 B10S 62.15 84.04
Vehicle 4 A5E 53.38 30.91
Vehicle 4 A10E 199.60 313.10 Vehicle 7 A 106.65 186.01
Vehicle 4 A5S 99.37 85.89 Vehicle 7 A5S 140.81 378.64
Vehicle 4 A10S 77.56 89.74 Vehicle 7 B 118.08 167.67
Vehicle 4 B 128.77 117.43 Vehicle 7 B5S 165.55 404.05
Vehicle 4 B5E 98.06 131.18 Vehicle 7 B10S 92.48 305.02
Vehicle 4 B10E 135.33 418.68
Vehicle 4 B5S 123.48 162.13
Vehicle 4 B10S 88.96 113.26

Vehicle 5 A 109.29 135.97

Vehicle 5 A5E 99.28 202.81
Vehicle 5 A5S 119.75 210.98
Vehicle 5 B 104.14 221.47
Vehicle 5 B5E 113.89 214.29
Vehicle 5 B10E 136.55 359.54
Vehicle 5 B5S 155.20 229.96
Vehicle 5 B10S 127.03 214.09

Analytical method

Carbonyl compounds are measured taking a sample of the diluted exhaust gas from the
dilution tunnel. The diluted exhaust sample was drawn through 2,4-dinitrophenyl-hydrazine
coated silica cartridges (WatersTM Sep-Pak® DNPH-cartridges). Aldehydes and ketones
present in the exhaust flow are collected as their non-volatile 2,4-dinitrophenylhydrazone
derivatives. The cartridges are then eluted with 2.5mL acetonitrile and 2.5 mL water and
analyzed by HPLC-UV (λ=365nm).
The cartridges are then eluted with 2.5mL of acetonitrile, diluted with 2.5 mL of water and
stored at 5oC until analysis. The samples were analyzed by HPLC-UV (λ=365nm)

85
Statistical analysis details

Fuel effects on formaldehyde and acetaldehyde emissions were analysed on a vehicle-by-

vehicle basis for the five cars for which data was available (vehicles 3, 4, 5, 6 and 7).
Arithmetic mean values for each vehicle × fuel combination may be found in Table A7.1.
(Note: the mean aldehyde values in this table have not been adjusted for the significant time
trend seen in the vehicle 4)

Outliers

A formaldehyde emission result (vehicle 5, fuel B, 407.7μg/km) was excluded as a statistical

outlier as it was over 2.5 times higher than emissions in any other test.

Statistical techniques for evaluating fuel effects

(In the EPEFE gasoline project [27] and recent CONCAWE emission studies [25] [26], the
variability in emission measurements has typically been found to follow the lognormal
distribution with the degree of scatter increasing as the emission level increases. Variations
between repeat measurements are a combination of measurement error and genuine variations
between tests.

The aldehyde emissions data behave similarly and therefore the data need to be analysed
using weighted analysis of variance and/or weighted regression techniques (see [27], annex
5). Unfortunately, due to the lack of repeats, the arithmetic mean emissions for each vehicle
and fuel are too inaccurate to calculate weights reliably via the usual equation

weight = 1/(mean emission for that vehicle and and fuel ) 2 .

This problem has been overcome by using iteratively reweighted least squares. The first step
in IRLS regression is to perform an unweighted regression analysis and calculate the
predicted value for each observation. A second regression is then performed with each
observation given a weight of

weight = 1/( predicted value) 2

The predicted values and weights are then recomputed and a subsequent weighted regression
is conducted. This process is continued until the regression coefficients have converged.

86
European Commission

EUR 22713 EN – DG Joint Research Centre, Institute for the Environment and Sustainability
Title: Effects of gasoline vapour pressure and ethanol content on evaporative emissions from modern cars
Authors: Edited by G. Martini
Luxembourg: Office for Official Publications of the European Communities
2007 – 87 pp. – 21x 29.7 cm
EUR - Scientific and Technical Research series; ISSN 1018-5593
ISBN 978-92-79-05249-1

Abstract
A test programme designed to investigate the influence of gasoline vapour pressure and ethanol content
on evaporative emissions from modern passenger cars has been carried out by the Joint Research Centre
of the European Commission jointly with CONCAWE and EUCAR. Seven gasoline passenger cars
representative of current EURO 3/4 emissions technology were tested for evaporative emissions with ten
different test fuels. The test fuel matrix comprised 60 and 70 kPa hydrocarbon base fuels with 5 and 10%
ethanol splash blends and 5 and 10% ethanol matched volatility blends. The evaporative emission tests
were carried out according to a test protocol based on the European homologation test procedure, with no
additional vehicle conditioning. Although this test protocol turned out to have a considerable influence on
the results, the programme has provided valuable information and several clear conclusions can be
drawn. The programme confirmed that vapour pressure (DVPE) is a key fuel variable for evaporative
emissions. However the effect of vapour pressure is strongly non-linear; the ethanol blends with final
DVPE around 75 kPa gave considerably higher evaporative emissions than the lower volatility fuels in
most of the vehicles. Differences between fuels with DVPE in the range 60-70 kPa were small. Additional
tests on two vehicles performed after the main programme have raised some questions about possible
effects of ethanol on carbon canister working capacity and on the role of permeation in determining
evaporative emissions.
 LB-NA-22713-EN-C
The mission of the JRC is to provide customer-driven scientific and technical support for the
conception, development, implementation and monitoring of EU policies. As a service of the European
Commission, the JRC functions as a reference centre of science and technology for the Union. Close
to the policy-making process, it serves the common interest of the Member States, while being
independent of special interests, whether private or national.