EPA 600/R-18/299 | September 2019 | www.epa.gov/research
United States
Environmental Protection
Agency
v>EPA
Cold Temperature Effects on Speciated VOC
Emissions from Modern GDI Light-Duty Vehicles
Office of Research and Development

-------
Cold Temperature Effects on Speciated VOC Emissions from Modern GDI
Light-Duty Vehicles
Ingrid George, Michael Hays, Richard Snow, James Faircloth,
Thomas Long, Richard Baldauf
Office of Research and Development
United States Environmental Protection Agency
109 T.W. Alexander Drive
Research Triangle Park, NC 27711
george.ingrid@epa.gov
Disclaimer: This product has been reviewed in accordance with U.S. Environmental Protection
Agency policy and approved for publication.
ABSTRACT
In this study, speciated volatile organic carbon (VOC) emissions were characterized from three
modern gasoline direct injection (GDI) light-duty vehicles. The vehicles were tested on a chassis
dynamometer housed in a climate-controlled chamber at two temperatures (20 and 72 °F) over
the EPA Federal Test Procedure (FTP) and a portion of the Supplemental FTP (i.e. US06) that
represents more aggressive driving conditions. The vehicles operated on gasoline blended with
10% ethanol. VOC emissions from diluted vehicle exhaust were sampled with SUMMA
canisters for EPA Method TO-15 analysis and with 2,4-Dinitrophenylhydrazine (DNPH)
cartridges for carbonyl analysis by EPA Method TO-11 A. This presentation will report the
impact of ambient cold temperature, driving cycle, and GDI technology on speciated VOC
emissions.
INTRODUCTION
The transportation sector contributes approximately 20% of total non-biogenic volatile organic
compound (VOC) emissions in the United States. Therefore, it is imperative to obtain a detailed
understanding of speciated VOC emissions from mobile sources to accurately assess the air
quality and health impacts of the transportation sector. To address the need for mobile source
speciated VOC emissions data, one major goal of the U.S. EPA's Office of Research and
Development vehicle emissions research is to comprehensively characterize speciated emissions
from modern vehicles and assess the impacts of various fuels, temperatures, newer engine, and
aftertreatment technologies. This study focused on measurements of speciated exhaust emissions
from gasoline direct injection (GDI) vehicles representing newer engine technologies. GDI
engines were introduced into the vehicle market in the U.S. in 2007, and the technology has
since quickly risen in popularity. Currently, GDI vehicles represent nearly a half of the U.S.
light-duty vehicle market share. However, the speciated emissions data on these types of vehicles
is extremely sparse and the effect of cold temperature on GDI vehicle emissions is unknown.
Therefore, the objective of this study was to characterize speciated emissions from three GDI
vehicles at two ambient temperatures.

-------
METHODS
Three light-duty GDI vehicles were tested during this vehicle emissions study. All vehicles were
in the U.S. EPA Tier 2 Bin 5 emission standard class and each vehicle represented different types
of GDI technologies. The model years (MY), odometer readings at the study start (ODO), engine
displacements (ED), and GDI technology types are given below for the three vehicles (VI, V2,
V3).
1)	VI: MY 2014 (Tier 2, Bin 5), ODO=12,700 miles, ED = 2.4 liter, GDI technology:
Naturally aspirated, wall-guided GDI engine
2)	V2: MY 2015 (Tier 2, Bin 5), ODO=10,500 miles, ED = 1.5 liter, GDI technology:
Spray-guided, turbocharged GDI engine
3)	V3: MY 2014 (Tier 2, Bin 5), ODO=9,200 miles, ED = 1.8 liter, GDI technology: Wall
and air guided, turbocharged GDI engine
Vehicles were operated on 10% ethanol blended with gasoline (E10) sourced from a local
distributor. Vehicle testing was conducted on a 48 inch roll chassis dynamometer housed inside a
climate controlled chamber. Vehicles were tested at two ambient temperatures (20 and 72 °F).
Each vehicle was driven on the Federal Test Procedure (FTP) followed by a portion of the
Supplemental Federal Test Procedure (also called US06) and vehicle testing for each test
condition was conducted in triplicate.
The vehicle exhaust was diluted by approximately 20-30 times in a constant volume sampling
dilution tunnel, where the flow was controlled by a critical flow venturi. Real-time emissions
were characterized by continuous emissions monitors for carbon dioxide (CO2), carbon
monoxide (CO), methane (CH4), total hydrocarbon (THC) and nitrogen oxides (NOx). Time-
integrated samples for VOCs were taken from the dilution tunnel for the three phases of the FTP
(FTP1, FTP2, FTP3) and US06. VOC samples taken included SUMMA canisters for EPA
Method TO-15 analysis and 2,4-Dinitrophenylhydrazine (DNPH) cartridges for gas-phase
carbonyls by TO-11A analysis. Canister samples were analyzed by gas chromatography-mass
spectrometry (GC/MS). DNPH cartridges were extracted with acetonitrile and extracts were
analyzed by high-performance liquid chromatography (HPLC).
RESULTS
Real-time Hydrocarbon Emissions
Real-time non-methane hydrocarbon (NMHC) emissions determined from the difference
between the THC and CH4 measurements from a representative test for V2 were evaluated to
better understand how the VOC emissions varied over the driving cycles. The NHMC trace over
the FTP driving cycle for V2 showed that there was an extremely large spike early in the cold

-------
start FTP1 phase. The majority of the total NMHC emissions over the FTP were represented by
this early cold start peak within the first 200 s of the FTP. After this initial peak, a few minor
peaks in NHMC emissions were observed during the FTP during driving accelerations. The
US06 driving cycle represents substantially more aggressive driving compared to the FTP. As a
result, the NMHC trace over the US06 for the same test showed numerous spikes in NHMC
emissions that coincided with intensive accelerations during the driving cycle.
Time-integrated VOC emissions
Time-integrated emission rates were calculated for 134 individual VOCs over each driving cycle
and test condition for VI, V2 and V3. Figure 1 shows the sum of all speciated VOC emission
rates (XVOCs) that were averaged over three replicate tests for each condition. It was observed
that cold start FTP1 had substantially higher EVOC emissions compared to other phases of the
FTP and US06, and were between 4 to 400 times higher than warm start FTP3 tests. EVOC
emission rates during FTP1 20 °F tests, as represented by the striped bars in Figure 1, were
strongly enhanced compared to FTP1 72 °F tests, where the cold temperature enhancements
varied for each vehicle. However, cold temperature enhancements in VOC emissions were
modest for other driving phases and US06. It was also found that VOC emission rates for V2
were substantially higher than for the other two vehicles with the exception of FTP 1 V3 testing
at 20 °F.
Figure 1. Sum of speciated VOC emission rates for each test condition and vehicle.

1.20

1.00 -


0)
0.80 -
E

cn

—
0.60 -
ui

C)
/
o
>
0.10 -



0.05 -

0.00 -
i a
v\
Vi
!
FTP1
mamf2	
FTP2
iEI
V1 72F
F7Z7I V1 20F
V2 72F
rzzz] V2 20F
V3 72F
V3 20F
FTP3
US06
VOC emission profiles for the three vehicles during FTP1 cold start of the 25 highest emitted
VOCs are shown in Figure 2 for 72 °F (top panel) and 20 °F (bottom panel). The highest emitted
VOCs during the FTP1 include a number of hazardous air pollutants, such as benzene and
toluene, as well as a number of hydrocarbons found in E10 gasoline. During the warm
temperature tests, VOC emission rates from V2 were substantially higher than for the other two
vehicles across all VOCs measured. However, during cold temperature tests V3 emissions were
higher for most of the major VOCs shown in Figure 2. It is currently unclear what the underlying

-------
mechanism for these vehicle specific temperature effects might be. Real-time measurements and
other ancillary data acquired during the study will be further examined to better understand these
observed trends.
Figure 2. VOC profiles for FTP1 at 72 °F (top) and 20 °F (bottom)
CONCLUSIONS
In this study, detailed speciated VOC emissions from three GDI vehicles at two ambient
temperatures (20 and 72 °F) were measured. We observed substantial differences in VOC
emissions between vehicles that were temperature dependent and VOC emissions were higher
during the cold start FTP1 compared to other driving phases and cold temperature FTP1 tests
compared to warm temperature test conditions. This work significantly increases the available
emissions data for modern light-duty GDI vehicles that will improve emission inventories and air
quality model predictions.

-------
Key Words
Volatile organic compounds
Vehicle exhaust
Gasoline direct injection
Chassis dynamometer

-------