Content uploaded by Norbert E. Ligterink
Author content
All content in this area was uploaded by Norbert E. Ligterink on Nov 06, 2016
Content may be subject to copyright.
Earth, Life & Social Sciences
Van Mourik Broekmanweg 6
2628 XE Delft
P.O. Box 49
2600 AA Delft
The Netherlands
www.tno.nl
TNO report
TNO 2016 R11247
Emissions of three common GDI vehicles
Date
6 October 2016
Author(s)
Norbert E. Ligterink
Copy no
2016-TL-RAP-0100300044
Number of pages
16
Number of
appendices
-
Sponsor
Dutch Ministry of Infrastructure and the Environment
PO Box 20901
2500 EX THE HAGUE
The Netherlands
Project name
LD In-use Compliance project 2012-2016
Project number
060.14432
All rights reserved.
No part of this publication may be reproduced and/or published by print, photoprint,
microfilm or any other means without the previous written consent of TNO.
In case this report was drafted on instructions, the rights and obligations of contracting
parties are subject to either the General Terms and Conditions for commissions to TNO, or
the relevant agreement concluded between the contracting parties. Submitting the report for
inspection to parties who have a direct interest is permitted.
© 2016 TNO
TNO report | TNO 2016 R11247 | 6 October 2016
2 / 17
Samenvatting
De laatste jaren is het aandeel benzinevoertuigen met directe injectie (Gasoline
Direct Injection, ofwel GDI) sterk toegenomen. In 2015 betrof dit in Nederland bijna
de helft van de verkoop van alle nieuwe benzineauto’s. Het is daarom belangrijk om
vast te stellen of deze voertuigen vergelijkbare emissies hebben als andere
benzineauto’s, of dat deze groep bijzondere aandacht behoeft in de
emissieregistratie en luchtkwaliteitsmodellen. In het bijzonder kunnen de emissies
van deeltjesaantallen, fijnstofmassa (PM10) en elementair koolstof afwijken van
andere benzineauto’s ten gevolge van de specifieke GDI technologie.
Van de vijf meest verkochte GDI voertuigmodellen van de laatste jaren zijn er drie
geselecteerd van verschillende fabrikanten. Deze drie voertuigen zijn op de
rollenbank in het laboratorium bij een temperatuur van 23-25 °C onderworpen aan
testen op een voor de praktijk representatieve ritcyclus. De instellingen en de testen
komen overeen met een belading en rijweerstand die aan de hoge kant liggen van
normaal gebruik. Dit is gedaan om er voor te zorgen dat de emissies niet
onderschat worden met deze testen. Desondanks presteerden alle voertuigen
goed, met emissies die vergelijkbaar zijn met de emissielimieten op de
typekeuringstest.
De fractie van het elementair koolstof in de fijnstof-emissies is hoger dan voor
andere benzinevoertuigen. Maar de absolute fijnstof-emissies in milligrammen per
kilometer zijn dusdanig laag dat de uitstoot van elementair koolstof door deze GDI’s
niet substantieel bijdraagt aan de fijnstof-concentraties in de lucht. Ook is de
uitstoot van de andere schadelijke emissies laag en vergelijkbaar met de uitstoot
van andere moderne benzinevoertuigen.
Vanaf Euro-6 moeten GDI’s voldoen aan een deeltjesaantallen-eis van 6 x 1012
deeltjes per kilometer in de typekeuring. Deze eis is minder streng dan voor
vergelijkbare dieselvoertuigen. De deeltjesaantallen-emissies van GDI’s in de
praktijktesten zijn inderdaad hoger dan van dieselvoertuigen met een gesloten
roetfilter, maar de gemeten waarden zijn consistent met de eisen die aan deze
voertuigen gesteld worden. De GDI’s vertonen in de praktijk geen afwijkend gedrag
ten opzichte van de typekeuringstest op de rollenbank. Vanaf 1 september 2017
moeten nieuwe GDI modellen aan de strengere eis voor deeltjesaantallen van
6 x 1011 deeltjes per kilometer voldoen. Naar verwachting zullen de praktijkemissies
van deze voertuigen daarmee nog verder verbeteren.
TNO report | TNO 2016 R11247 | 6 October 2016
3 / 17
Summary
The share of petrol passenger cars with Gasoline Direct Injection (GDI) technology
quickly increases in the total sales of new petrol passenger cars. In The
Netherlands almost half of the total sales of petrol vehicles in 2015 are GDI’s.
Therefore, it is important to establish the emissions of these vehicles and to
investigate whether a distinction as a special vehicle category in the emission
inventories and air quality models would be appropriate. Special attention is given
to the emissions of particulate mass and particulate numbers, and the elemental
carbon fraction in the particulate mass. On these pollutants GDI technology is
expected to deviate the most from other petrol cars.
Three vehicles with common GDI engines from different manufacturers in the top
five of GDI sales of the last years have been selected. The vehicles have been
tested in a laboratory on a chassis dynamometer with a test cell temperature of 23-
25 °C with a real-world test cycle and normal to high engine loads, to ensure that
the effects of real-world emissions are not underestimated. Still, all the vehicles
perform well. The emission results in the real-world tests are more or less
consistent with the vehicles’ performance on the type-approval test.
The fraction of elemental carbon (EC) in the PM10 emissions of GDI vehicles is
higher than for other petrol vehicles. Given the low absolute PM10 emission levels
there is no specific concern regarding EC emissions of GDI’s, nor regarding the
other pollutant emissions considered.
The current particulate number emission limit for Euro-6 GDI’s is less strict than for
diesel vehicles. The tests indicate that GDI vehicles perform worse on particulate
numbers than their diesel counterparts. However, the particulate number emissions
in real-world tests are generally consistent with the type-approval limits. The type-
approval PN emission limit for GDI will become more stringent from 1 September
2017 onwards. The current limit value of 6 x 1012 particulates per kilometre will
reduce to 6 x 1011 particulates per kilometre. The real-world emissions are
expected to decrease accordingly.
TNO report | TNO 2016 R11247 | 6 October 2016
4 / 17
Contents
Samenvatting ........................................................................................................... 2
Summary .................................................................................................................. 3
1 Introduction .............................................................................................................. 5
2 Emergence of GDI’s................................................................................................. 7
3 Test specification and vehicle selection ............................................................... 8
3.1 High load chassis dynamometer emission tests ....................................................... 8
3.2 Vehicle selection ........................................................................................................ 9
4 Emission test results............................................................................................. 12
5 Determination of the fraction of elemental carbon ............................................ 15
6 Conclusions ........................................................................................................... 16
7 Signature ................................................................................................................ 17
TNO report | TNO 2016 R11247 | 6 October 2016
5 / 17
1 Introduction
Particulates from the vehicle exhaust are known to cause health problems.
Therefore, particulate matter emissions are regulated for relevant vehicles, which
can emit substantial amounts of particulate matter. For a long time this concerned
only diesel vehicles, but with the introduction of modern direct fuel injection
technology (GDI: Gasoline Direct Injection) also petrol vehicles are reported to emit
particulate matter. These emerging group of petrol vehicles are the subject of the
study reported here. The study concerns the performance of such vehicles under
real-world driving conditions, not the legal results under type-approval test
conditions., and aims to establish the actual contribution of these vehicles to air-
quality problems.
However, particulate matter emissions cannot yet be tested on-road. At the moment
no official mobile PEMS equipment is available to measure these emissions.
Therefore, the tests are performed in the laboratory mimicking real-world conditions
under which particulate emissions may be substantial, such as high velocities and
high vehicle loads
1
. In the future, accurate PEMS-PN equipment should enable on-
road testing and enable the assessment of the real-world performance of such
vehicles within the limits.
Direct fuel injection technology with petrol passenger cars was for a long time
applied only in a small fraction of the petrol passenger cars sold in the Netherlands.
Recently, petrol direct injection technology has become more common. Soon more
than half of the petrol vehicles sold will have direct injection. Given the fact that
these vehicles also emit particulate matter (PM) they are of some concern for air
quality and will be considered as a separate vehicle category in emission
inventories.
Table 1 The different emission limits for PM and PN associated with the different Euro-classes for diesel
passenger cars and GDI's.
As shown in Table 1 with the coming into force of Euro-5 (starting from 2009) the
type-approval test for particulate matter was compulsory for GDI’s, with the same
emission limit of 5 mg/km as for diesel vehicles. The more strict particulates number
emission limit of 6 x 1011 #/km, from Euro-5b (2011) for diesel vehicles, was applied
with a lower limit of 6 x 1012 #/km from Euro-6 (2014) for GDI vehicles. In type-
approval GDI’s clearly have higher particulate number emissions than diesel
vehicles (see figure 1). Moreover, there is some correlation between the particulate
number and the particulate mass emissions on the type-approval test. This situation
1
For more information on TNO test methods for laboratory and on-road testing see:
TNO 2016 R11178, “Assessment of road vehicle emissions: methodology of the Dutch in-service
testing programme”, V.A.M. Heijne et al., 2016
new models all models diesel GDI diesel GDI
Euro-4 1-Jan-2005 1-Jan-2006 25 - - -
Euro-5a 1-Sep-2009 1-Sep-2010 5 5 - -
Euro-5b 1-Sep-2011 1-Sep-2012 4.5 4.5 6.0E+11 -
Euro-6 1-Sep-2014 1-Sep-2015 4.5 4.5 6.0E+11 6.0E+12
Euro-6c 1-Sep-2017 1-Sep-2018 4.5 4.5 6.0E+11 6.0E+11
introduction dates
PM [mg/km]
PN [#/km]
TNO report | TNO 2016 R11247 | 6 October 2016
6 / 17
is meant to be temporary until three years after the initial date of the introduction of
Euro-6. With the coming into force of Euro-6c the particulate mass and number
emissions of GDI’s should satisfy, in all aspects, the same limits and standards as
those of diesel vehicles. This also means that RDE legislation should incorporate a
particulate number test and set an appropriate limit.
Figure 1 The particulate mass and particulate number measurement results in the NEDC type approval test
of diesel vehicle models (green) and GDI's (red) which are sold in the last two years. The
particulate number limit of GDI's is 6 1012 #/km, while the diesel vehicles have a limit of 6 1011
#/km.
The tests reported here were also designed and executed to obtain elemental
carbon (EC) emission factors for GDI’s. For the GDI vehicles limited information on
EC emissions is available. From the technology it is to be expected that the fraction
of elemental carbon in the particulate mass is higher than for other petrol vehicles.
This test program fills an important gap in the knowledge underlying the EC
emission factors used in the Netherlands.
TNO report | TNO 2016 R11247 | 6 October 2016
7 / 17
2 Emergence of GDI’s
In the past GDI’s were a small group among all petrol vehicles sold. With the
introduction of European CO2 targets, GDI technology became a way to increase
the efficiency of a petrol engine. Hence in the last few years GDI’s were sold in
larger numbers. In 2010 the GDI’s were only 1% of the total petrol vehicle fleet.
Currently, summer 2016, the GDI’s form 9% of the total Dutch petrol fleet. In the last
two years GDI’s make up 44% of the total sales. In particular, small engines (0.9-
1.2 litres) with substantial power of 60–100 kW are sold in large numbers.
Given the current sales of GDI’s rapidly approaching 50%, the replacement of the
other, older petrol technology will be at a rate of about 4% a year. Therefore, it is
decided to take GDI’s into consideration as a separate vehicle category in the
national emission inventories and air quality models for which separate emission
factors may be needed.
Given the fact that particulate mass emissions of GDI’s are regulated from early on,
the particulates mass emissions of GDI’s may be low. The more lenient
intermediate particulate number emission limit of GDI’s from Euro-6 onwards,
however, gives concern that GDI technology may not keep pace with the more
stringent particle number standards for diesel vehicles.
Particulate matter emissions are not the only concern related to GDI’s, but also the
sophisticated engine control, for which more complex and specific optimizations can
be used. These optimizations may lead to an increase in NOx or CO emissions in
real-world driving conditions. The most quoted reason to use adapted control
strategies for GDI’s is the reduction of real-world fuel consumption. If the engine is
operated under partial lean-burn conditions, fuel consumption may decrease, but a
deviation from near stoichiometric operation associated with the optimal functioning
of the three-way catalyst, might increase NOx, CO, and HC exhaust gas emissions.
TNO report | TNO 2016 R11247 | 6 October 2016
8 / 17
3 Test specification and vehicle selection
3.1 High load chassis dynamometer emission tests
High particulates emissions are associated with high engine load and high fuel
consumption. As the test program is set up to assess the risks of high PM and PN
emissions, the tests are conducted with engine loads that are at the higher end of
the spectrum of normal vehicle use in the Netherlands. This means that the engine
loads in the test are higher than what is common for the type-approval tests. This is
achieved by setting the driving resistance and test mass on the high side of normal
use. The applied values are comparable to carrying two passengers and, for
example, the use of C or D-label tyres.
Table 2 The settings of the chassis dynamometer for the test mass and driving resistance.
For the cold urban tests the vehicles were soaked at 23 °C. All chassis
dynamometer tests are carried out with a test cell temperature of 23 - 25 °C.
Vehicles are tested on the CADC-150 driving cycle
2
, which is considered
representative of more aggressive driving within the spectrum of normal driving.
The variant of the CADC-cycle used has a maximum velocity of 150 km/h.
Therefore, all in all the power demand on the vehicles is on the high side, and the
emissions may be somewhat higher than can be expected from average driving.
The need of this high demand lies in the fact that emissions may increase rapidly
with engine demand and therefore average driving does not necessarily result in
average emissions.
Table 3 The parameters of the driving cycle.
Since particulate matter is collected on a filter, the tests have been repeated three
times to collect sufficient particulate matter for an accurate filter particulate mass
determination and subsequent chemical analyses. Rather than collecting the
emissions of a full CADC cycle on a single filter, the urban, rural and motorway part
2
For more information on TNO test methods for laboratory and on-road testing see:
TNO 2016 R11178, “Assessment of road vehicle emissions: methodology of the Dutch in-service
testing programme”, V.A.M. Heijne et al., 2016
road-load settings
mass 1450 kg
F0 130 N
F1 0 N/[km/h]
F2 0.04 N/[km2/h2]
CADC-150 distance [km]
urban 4.8
rural 17.2
motorway 29.5
total 51.5
TNO report | TNO 2016 R11247 | 6 October 2016
9 / 17
are collected on separate filters to have representative particulate mass results for
each of these traffic situations.
Moreover, cold starts, when the engine itself is at the ambient temperature, may
lead to additional particulates emissions. Cold starts occur mainly in urban driving
situations. In the tests cold starts are included in the result of the urban test. For two
vehicles the cold start was in each of the three repetitions of the urban test. For the
third vehicle only a single cold start was used in the three tests. It is a rather
cumbersome procedure to ensure a cold start, which was therefore abandoned later
in the test program.
The filters were of pure quartz. These filters can be used for elemental carbon
determination using the SUNSET
3
method from the EUSAAR. In this method of
determination the filters are heated up to 800 °C. Other types of filters cannot
withstand such temperatures, and the filter material itself might contaminate the
results.
3.2 Vehicle selection
The vehicle were selected from the top five of high sales of GDI engines in vehicles
from different manufacturers. In analysing the sales database engines were
characterized by engine volume and rated power. The GDI classification was
induced from the fact that the petrol engine was type-approved for particulate
matter.
The selected vehicles each represent about 4% of current sales of GDI engines.
This small fraction is due to the recent introduction of many more GDI engines. The
specifications of the selected vehicles are given in Table 4 to Table 6. They are
common models with even more common engines for these manufacturers. In the
past Volkswagen had the largest share of GDI vehicles in the total fleet. The
Volkswagen Golf was selected from that period. But more recent other
manufacturers have increased their share.
3
F. Cavalli et al., Toward a standardised thermal-optical protocol for measuring
atmospheric organic and elemental carbon: the EUSAAR protocol, Atmos. Meas. Tech. (2010) 3
p.79-89.
TNO report | TNO 2016 R11247 | 6 October 2016
10 / 17
Table 4 Vehicle specifications of the Ford Focus
Trade Mark
[-]
Ford
Type
[-]
Focus
Body
[-]
Hatchback
Vehicle Category
[-]
M
Fuel
[-]
Petrol
Engine Code
[-]
Swept Volume
[cm^3]
999
Max. Power
[kW]
74
Euro Class
[-]
Euro 6
Vehicle Empty Mass
[kg]
1176
Vehicle Test Mass
[kg]
1450
Odometer
[km]
11628
Registration Date
[dd-mm-yy]
17-03-15
Table 5 Vehicle specifications of the Peugeot 308
Trade Mark
[-]
Peugeot
Type
[-]
308
Body
[-]
Hatchback
Vehicle Category
[-]
M
Fuel
[-]
Petrol
Engine Code
[-]
Swept Volume
[cm^3]
1199
Max. Power
[kW]
81
Euro Class
[-]
Euro 6
Vehicle Empty Mass
[kg]
1055
Vehicle Test Mass
[kg]
1450
Odometer
[km]
14268
Registration Date
[dd-mm-yy]
11-12-15
TNO report | TNO 2016 R11247 | 6 October 2016
11 / 17
Table 6 Vehicle specifications of the Volkswagen Golf
Trade Mark
[-]
Volkswagen
Type
[-]
Golf
Body
[-]
Station wagon
Vehicle Category
[-]
M
Fuel
[-]
Petrol
Engine Code
[-]
CJZA
Swept Volume
[cm^3]
1197
Max. Power
[kW]
77
Euro Class
[-]
Euro 5
Vehicle Empty Mass
[kg]
1135
Vehicle Test Mass
[kg]
1450
Odometer
[km]
32331
Registration Date
[dd-mm-yy]
03-12-14
TNO report | TNO 2016 R11247 | 6 October 2016
12 / 17
4 Emission test results
In Table 7 the results of the three tested vehicles are reported: the urban, rural and
motorway parts are listed separately. The emission tests give results for all the
regulated pollutants, which are generally in line with the vehicle performance of the
vehicles on the type-approval test. Moreover, the test results show no peculiar or
systematic deviations.
Table 7 The measurement results of the emission tests of the three vehicles with GDI.
Ford Focus HC CO CO2 NOx NMHC CH4 PN FC PM (sum)
mg/km mg/km g/km mg/km mg/km mg/km #/km l/100km mg/km urban test
urban 165 931 228.7 116 154 12 5.95E+12 9.974 cold
114 1178 225.8 112 102 13 8.37E+12 9.860 cold
108 1076 226.1 132 96 14 8.20E+12 9.865 6.1 cold
average 129 1062 226.9 120 117 13 7.51E+12 10 6.1
rural 15 220 129.9 76 13 2 1.98E+12 5.632
7192 127.2 54 6 1 1.92E+12 5.514
17 348 127.8 69 15 2 1.79E+12 5.551 0.8
average 13 253 128.3 66 11 2 1.90E+12 6 0.8
Motorway 10 301 169.0 55 9 1 2.56E+12 7.331
21 528 171.6 95 19 3 2.81E+12 7.457
14 419 172.1 63 12 2 2.56E+12 7.471 2.4
average 15 416 170.9 71 13 2 2.64E+12 7 2.4
Peugeot 308 HC CO CO2 NOx NMHC CH4 PN FC PM (sum)
mg/km mg/km g/km mg/km mg/km mg/km #/km l/100km mg/km
urban 75 475 230.9 96 68 7 7.49E+12 10.025 cold
16 117 219.5 330 12 4 2.18E+12 9.503 warm
11 40 221.3 406 8 4 2.58E+12 9.572 1.8 warm
average 34 210 223.9 277 29 5 4.08E+12 10 1.8
rural 374 122.4 130 2 1 8.21E+10 5.300
4107 125.2 107 3 1 9.98E+10 5.420
474 125.0 146 3 1 1.08E+11 5.410 0.1
average 485 124.2 128 3 1 9.67E+10 5 0.1
Motorway 2224 165.1 66 2 1 3.02E+11 7.156
3180 164.4 78 2 1 3.22E+11 7.121
3210 164.2 87 2 1 2.89E+11 7.117 0.3
average 3205 164.6 77 2 1 3.04E+11 7 0.3
VW Golf HC CO CO2 NOx NMHC CH4 PN FC PM (sum)
mg/km mg/km g/km mg/km mg/km mg/km #/km l/100km mg/km
urban 55 563 224.7 137 48 8 1.92E+12 9.760 cold
110 1039 232.1 170 100 11 3.02E+12 10.123 cold
72 937 239.1 121 64 9 1.26E+12 10.413 0.6 cold
average 79 846 232.0 143 71 9 2.07E+12 10 0.6
rural 18 77 134.7 98 16 1 9.30E+11 5.834
10 65 134.5 83 8 1 7.33E+11 5.820
245 134.6 76 2 1 5.75E+11 5.825 0.1
average 10 62 134.6 86 9 1 7.46E+11 6 0.1
Motorway 14 435 171.3 54 10 4 2.26E+12 7.440
16 583 171.2 68 12 4 2.53E+12 7.447
17 684 170.5 68 13 5 2.27E+12 7.421 0.9
average 16 567 171.0 63 12 4 2.35E+12 7 0.9
TNO report | TNO 2016 R11247 | 6 October 2016
13 / 17
In all cases the NOx emissions exceeded the NEDC emission limits, but taking into
account that the tests were performed with high load settings of the chassis
dynamometer and a normal though demanding cycle and that the separate urban
test had a short distance of only 4.8 km, the measured emissions do not raise
special concern. In the case of an NEDC type-approval test the additional emission
due to a cold start would have been spread over double the distance, yielding lower
g/km results for the urban part of the test.
Table 8 The average test results of the three vehicles, based on equal weighing of vehicles and tests.
Per three tests one single PM filter result is available, which is the sum result of the
three repetitions of the same test, collected on a single filter. The PM values are
very low, close to the measurement accuracy, such that this repetition is needed to
arrive at a reasonable accuracy of the measured particulate mass.
One notable aspect is the NOx emission of the Peugeot which are higher in the
warm urban tests (330 and 406 mg/km) than in the cold tests (96 mg/km). Looking
at the second-by-second emission rate over these tests, the two warm tests show
similar incidental occurrences of high emissions (see Figure 2). This indicates a
failure of the three-way catalyst control strategy, rather than a systematic deviation.
The robustness of the three-way catalyst control seems somewhat limited for this
vehicle. The introduction of RDE legislation, with on-road testing may help to reduce
such incidents of high emissions.
In general emissions of all vehicles and all pollutants are higher under cold start
conditions. In this effect the engine soak temperature plays an important role. The
cold start tests were conducted at 23 °C laboratory temperature. At lower
temperatures
4
the cold start effect is typically larger. The effect is expected to
increase gradually as the engine soak temperature decreases. For example, for
pollutants that are reduced by the three-way catalyst, which needs to reach
operation temperatures, at 0 °C ambient temperature the additional emissions of
the cold start are expected to be two or three times higher than at 23 °C. For PM
and PN the temperature sensitivity of the cold start effect is smaller.
All vehicles have higher particulate number emissions than comparable diesel
vehicles equipped with a DPF (Diesel Particulates Filter) which have emissions
below 6 x 1011 #/km. Especially, the cold start urban tests have high particulate
number emissions. The values are of minor concern, given the fact that a Euro-4
diesel vehicle without DPF produces typically 1013 to 1014 #/km. This is a factor 10
to 100 higher than the particulate number emissions of these GDI’s.
4
The average ambient temperature in The Netherlands is 11 o C.
HC CO CO2 NOx NMHC CH4 PN FC PM (sum)
averages mg/km mg/km g/km mg/km mg/km mg/km #/km l/100km mg/km
urban 80 706 227.6 180 72 9 4.55E+12 9.8993 2.8
rural 9134 129.0 93 8 1 9.13E+11 5.5895 0.3
Motorway 11 396 168.8 71 9 2 1.77E+12 7.3289 1.2
total 33 412 175.1 115 30 4 2.41E+12 7.6059 1.4
TNO report | TNO 2016 R11247 | 6 October 2016
14 / 17
Figure 2 The NOx emission rates of the Peugeot 308 on the three urban tests show that the higher
emissions are incidental, with some correlation between the warm tests.
Also the particulate mass emission is low. Current emission factors for all petrol
vehicles are about 5 mg/km. The measured values, despite the demanding tests,
are below this current estimate of 5 mg/km. Hence, in terms of absolute emission
levels GDI’s are of little concern. Moreover, the differences between the vehicles
may appear large, but at current levels, below the emission limits, these differences
should not be exaggerated. In particular, at these low levels the variation in PM
emissions between tests with the same vehicle can be large.
TNO report | TNO 2016 R11247 | 6 October 2016
15 / 17
5 Determination of the fraction of elemental carbon
The quartz filters were analysed in a chemical laboratory for the determination of
the elemental carbon (EC) fraction compared to the remaining organic carbon
fraction. The elemental carbon and the organic carbon together make up the total
carbon in PM10 emissions. Generally, there is only a small mass fraction from other
elements and the majority of the total mass is carbon. The SUNSET method used is
considered to be best suited to determine EC fractions. Other methods, because of
their lower end temperatures, may not collect all the carbon from the filters.
Table 9 The fraction of EC, as opposed to the fraction of carbon as part of organic compounds.
Surprisingly, the fractions of EC in particulate matter are constant across the tested
vehicles for a given road type, but vary widely between the road types. The EC
fraction is higher than for petrol vehicles without direct injection, which are generally
estimated at 30% and less, depending on technology and driving behaviour. It
should be noted that organic carbon also contains hydrogen and oxygen, and the
total mass fraction of the organic matter is somewhat higher than the mass fraction
of carbon alone with about 15% non-carbon components in organic material. In
some cases hydrates, metal and ash is collected on the filter, increasing the fraction
of carbon-free components. However, since these vehicles are relatively new and in
good order, it is not expected that the carbon-free fraction in PM is more than 15%.
Table 10 Typical composition of PM10 emissions of GDI vehicles
Fraction of PM10
Of which carbon
EC
>60%
100%
Organic matter
<30%
85%
carbon-free fraction
<15%
--
Although the fraction of elemental carbon is substantial, especially for an urban
tests with cold start, the absolute levels are very low, ranging from 0.3 to 3 mg/km.
Consequently, GDI’s have a limited contribution to the total EC emissions of traffic.
Only by the time that the large majority of diesel vehicles is equipped with well-
functioning diesel particulate filters, the EC emissions of GDI’s will become a
significant remainder.
Ford Focus elemental carbon fraction
urban 82.6%
rural 33.4%
motorway 59.8%
Peugeot 308
urban 80.6%
rural 28.0%
motorway 54.3%
Volkswagen Golf
urban 89.5%
rural 37.7%
motorway 54.8%
TNO report | TNO 2016 R11247 | 6 October 2016
16 / 17
6 Conclusions
Given the growing group of vehicles with GDI, specific attention for the emissions of
these vehicles is justified. There are suggestions that these vehicles have
significantly higher PM emissions in normal use than other petrol vehicles.
In order to gain more insight in PM, PN and EC emissions of GDI vehicles CADC
tests were conducted in a chassis dynamometer test programme. To establish the
EC fraction, the PM emissions were collected on quartz sample filters, which were
analysed in detail later in a chemical laboratory.
For the three tested GDI’s, which are common makes and models, the PM10
emissions are in the range of 0.1 to 6.1 mg/km. In general the emissions are
consistent with the type approval emission limits of 5.0 mg/km. Deviations can be
explained by the more severe test conditions in this test programme, compared to
the type-approval test, and multiple inclusion of cold starts.
The ambient air contains particles from many different emission sources, including
sand, dust, and sea salt. The particulates from vehicle exhaust are only a small
fraction in the total particulate matter. Zooming in on the ambient concentration of
specific components of PM10, in particular EC or black carbon, a substantially
higher fraction is directly related to tailpipe emissions from vehicles. For most diesel
vehicle categories specific EC emissions have been determined in previous test
programmes. Based on that the relative contribution of particular vehicle categories
to the ambient EC concentration could be established. The intention is that a
complete picture for EC emissions of all road vehicles will be available for Dutch air-
quality assessments. By adding specific information on EC emissions of GDI’s, this
study fills an important gap in the current understanding needed for the full
assessment.
The particulate number emissions of the three vehicles are in line with the current
emission limit value of 6 x 1012 #/km. But unlike the real-world emissions of modern
diesel vehicles, the real-world emissions of GDI’s are close to the limit value. When
the type-approval limit for GDI vehicles will be lowered to 6 x 1011 #/km – as of 1
September 2017 for new models –it is expected that the real-world particulate
number emissions will decrease accordingly. Only if particulate filters will be the
common technology for GDI’s, real-world particulate number emissions might be
substantially lower than the limit values, as is already the case for diesel vehicles
equipped with particulate filters.
TNO report | TNO 2016 R11247 | 6 October 2016
17 / 17
7 Signature
Delft, 6 October 2016 TNO
Peter van der Mark Norbert E. Ligterink
Project leader Author
