Characterization of Emissions from Malfunctioning Vehicles
Fueled with Oxygenated Gasoline-MTBE Fuel — Part I

Fred Stump, Silvestre Tejada, and David Dropkin
National Exposure Research Laboratory
U. S. Environmental Protection Agency
Research Triangle Park, NC 27711

Colleen Loomis
Clean Air Vehicle Technology Center, Inc.
Research Triangle Park, NC 27709


-------
Table of Contents

Abstract	ii

List of Tables	iii

List of Figures	iv

Abbreviations and Symbols	v

INTRODUCTION	1

EXPERIMENTAL METHODS AND MATERIALS	2

Test Fuels	2

Test Schedule/Conditions	2

Test Vehicles	3

Test Facilities	3

Test Procedures	4

Tailpipe Emissions	5

RESULTS AND DISCUSSION	5

Regulated Emissions	8

Toxic Emissions	12

Particulate Emissions	13

SUMMARY AND CONCLUSIONS	17

ACKNOWLEDGMENTS	19

DISCLAIMER	20

REFERENCES	20

l


-------
ABSTRACT

Two vehicles, a 1993 4-cyUnder Chevrolet Cavalier and a 1993 6-cylinder Ford Taurus,
were tested using three different fuels - a winter grade fuel containing 11.3% methyl tertiary
butyl ether (MTBE), a winter grade fuel (base fuel) without MTBE, and a summer grade fuel
without MTBE. Vehicle tests were conducted at ambient temperatures of 75°( with summer
grade fuel only), 20°, 0°, and -20°F. The vehicles were first tested under a normal mode
(vehicles were tuned to manufacturers specifications) and then tested under two simulated
malfunction modes - 1) the oxygen (02) sensor was disconnected and 2) the exhaust gas
recirculating valve (EGR) was disconnected and plugged. The malfunction modes were not
tested simultaneously. The vehicles were tested on the Urban Dynamometer Driving Schedule
(IJDDS) of the Federal T est Procedure (FTP). Two high speed REP05 test cycles were run after
each of the IJDDS tests. The exhaust emissions determined were particulate matter (PM2.5 and
PM10), gaseous total hydrocarbons (HC), carbon monoxide (CO), and oxides of nitrogen (NOx),
speciated (individual) hydrocarbons, MTBE, and speciated aldehydes.

The IJDDS hydrocarbon emissions from both vehicles increased as test temperatures
decreased. Under normal mode, the HC emissions at 20°F tripled, the CO emissions more than
doubled, while the NOx changed only slightly when the vehicles were tested at -20°F. HC
emissions with the MTBE fuel were on the average 9-14% less than those with the base fuel. The
Cavalier emitted on the average 12% less CO with the MTBE fuel than with the base fuel while
the Taurus, 5% more CO with the MTBE fuel. The Cavalier HC, CO, and NOx emissions on the
average increased 235%, 489% and -8%, respectively with the base fuel and 226%, 483%) and
-27%, respectively with the MTBE fuel over the normal mode emissions when the oxygen sensor
was disconnected at the winter test temperatures. The Taurus HC, CO, and NOx emissions on the
average increased 216%>, 347%) and 55% with the base fuel and 91%>, 138%> and 52%
respectively with the MTBE fuel. Disconnecting the EGR had a minimal impact on the HC and
CO emissions but a significant one on NOx emissions. The Cavalier NOx emissions increased
81% and 148% with the base and MTBE fuel respectively over those of the normal mode, while
those of the Taurus increased 31 % and 41%> respectively.

The emissions of such toxic compounds benzene and 1,3-butadiene tended to increase as
the testing temperature decreased. Disconnecting the 02 sensor generally increased the
emissions of the toxic compounds, while the fuel containing MTBE reduced some toxic
compound emissions. The measured emissions of formaldehyde and acetaldehyde did not show
trend associated with changes of test temperatures, modes or fuels used in this study.

Particulate emissions were mostly PM2.5 particles. Particulate emissions correlated very
well (R2 > 0 .920) with HC emissions for every vehicle-fuel-malfunction mode combinations.
Particle emissions increased 4-5 times when the test temperature was decreased from 20°F to
-20°F. Both vehicles emitted more particles with the base fuel than with the MTBE at all test
conditions. Taurus particle emissions ranged 13-73 mg/mi with the base fuel and 6-34 mg/mi
with the MTBE fuel. The Taurus particulate emissions were 2-3 times greater than those of the
Cavalier. Maximum particulate emissions were obtained when the oxygen sensor was
disconnected. Particle emissions with a disconnected EGR differed only slightly from normal
mode emissions for both vehicles.

u


-------
List of Tables

Table 1 Characteristics of test fuels	2

Table 2 Test Schedule/Conditions for 1993 Chevrolet Cavalier and Ford Taurus vehicles . . 3

Table 2a Vehicles tested	3

Table 3 Vehicle tailpipe emissions rates at 75°F with summer grade fuel	5

Table 4 Vehicle UDDS tailpipe emissions at 20°F	6

Table 5 Vehicle UDDS tailpipe emission rates at 20°F	6

Table 6 Vehicle UDDS tailpipe emission rates at -20°F	7

Table 7 Vehicle REP05 tailpipe emission rates	7

in


-------
List of Figures

Figure 1 Correlation plots of benzene vs HC as a function of fuel	12

Figure 2 Correlation plots of PM10 and PM2.5 particulate emissions from all

vehicle-fuel-mode combinations at all test temperatures	13

Figure 3 Correlation plots of PM2.5 vs HC as a function of fuel and oxygen sensor malfunction
	14

Figure 4 Correlation plots of PM2.5 vs HC as function a function of fuel and EGR malfunction
	15

Figure 5 Correlation plots of REP05 PM2.5 vs HC a function of fuel and oxygen sensor

malfunction	16

Figure 6 Correlation plots of PM2.5 vs HC as a function of fuel and EGR malfunction . . . 17

iv


-------
Abbreviations and Symbols

°F

Degree Fahrenheit

AME

Archived Mobile Source Emissions Data Base

CFR

Code of Federal Regulations

CO

Carbon monoxide

CVS

Constant Volume Sampler

EGR

Exhaust Gas Recirculation valve

FI

Fuel Injection

FTP

Federal Test Procedure

g/cm3

gram per cubic centimeter

g/mi

gram per mile

GC-MS

Gas chromatography - mass spectrometry

HC

Hydrocarbon

IBP

Initial Boiling Point

LA-4

Los Angeles cycle, corresponding to stabilized cold cycle of UDDS

mg

microgram

mg/mi

milligram per mile

mph/sec

Miles per hour per second

MTBE

Methyltertiarybutylether

NC

North Carolina

NOx

Oxide of nitrogen

02

oxygen

02 S

Oxygen sensor

R2

Correlation coefficient

REP05

A high speed, high acceleration driving cycle

RVP

Reid Vapor Pressure

TWC

Three Way Catalyst

v


-------
INTRODUCTION

Motor vehicles emit large quantities of hydrocarbons (HCs), carbon monoxide (CO), and
oxides of nitrogen (NOx). These emissions participate in the atmospheric photochemical
processes to form ozone and other oxidants. Motor vehicles also emit toxic compounds such as
formaldehyde, acetaldehyde, benzene, 1,3-butadiene and particulate matter (PM2.5 and PM 10).
The 1990 Clean Air Act Amendment specify a reduction in these toxic emissions.1 This study
characterized the emissions from two recent model light duty gasoline-powered vehicles tested at
ambient temperatures of 75°, 20°, 0°, and -20°F with the vehicles normally tuned to
manufacturers' specifications using three commercially available fuels - a winter grade fuel
without MTBE, a winter grade fuel with MTBE, and a summer grade fuel without MTBE. In
addition, the two vehicle were tested using the same fuel and temperature combinations as the
normal mode under two distinct simulated malfunction modes - 1) the 02 sensor was
disconnected and 2) the EGR valve was disconnected and plugged.

Particulate emissions are becoming of major concern due to their adverse effects on
materials, visibility reduction, atmospheric reactivity, and human health. Recent epidemiological
studies have indicated health concerns for particulate matter emissions with an aerodynamic
diameter of 10 //m or less.2'3 In this study, individual PM2.5 and PM 10 filters were taken from
each of the UDDS tests (duplicate tests at each condition). After the completion of each UDDS
test, the particulate emissions from two REP05 driving cycles were collected on individual
PM2.5 and PM 10 filters. Currently there is little research information regarding particulate
emission rates from recent model in-use light-duty gasoline vehicles4 and very few studies have
been reported on vehicles tested at malfunction conditions at various ambient temperatures.

1


-------
EXPERIMENTAL METHODS AND MATERIALS

The test fuels, vehicles, facilities, and procedures are described in this section.

Test Fuels

The fuels used in this study were a summer grade fuel, a winter grade fuel containing no
methyl tertiary-butyl ether (MTBE), and a winter grade fuel containing 11.3% (volume) MTBE.
These fuels were purchased locally and are representative of the fuel in-use. The test fuels are
listed in Table 1.

Table 1. Characteristics of test fuels.

Fuel Property

Summer

Winter

Winter-MTBE

Specific gravity, g/cm

RVPa

Distillation, °F
IBP

10%

50%

90%

End Point
Paraffins, %

Olefins, %

Aromatics, %
Benzene, %

MTBE, %

a Reid vapor pressure

0.75
6.87

102
133
219
351
421
45.43
14.08
40.23
1.23
0.38

0.75
9.72

101
131

219
355
425
55.33
15.89
28.61
1.23
0

0.72

14.78

77
100
173
314
432
62.09
12.85

24.79
0.63

11.3

Test Schedule/Conditions

A single UDDS test was performed with each vehicle at 75 °F with the summer grade fuel
at each test mode. Duplicate UDDS tests at 20, 0 and -20°F at each test mode were performed
with each of the winter grade fuels (Table 2). Each UDDS test was followed by two consecutive
REP05 tests. The vehicles were conditioned with each test fuel at each test condition before
actually testing the vehicle for data collection. The vehicles were conditioned each day by
running an LA-4 driving cycle after completion of the high speed REP05 tests.

2


-------
Table 2. Test Schedule/Conditions for 1993 Chevrolet Cavalier and 1993 Ford Taurus
vehicles.

Description

Summer Test Conditions

Winter Test Conditions

Driving Cycles
Test Temperature
Duplicate Runs
Fuel Type
MTBE

Malfunction Modes
Emissions Measured:

IJDDS

75°F

Single test only
Summer Fuel
Without
02 and EGR
Gaseous Emissions
No Particles

IJDDS + 2 REP05
20, 0, -20°F.
Duplicate runs
Winter Fuel
With and Without
02 and EGR
Gaseous Emissions
Particle Measured:
PM2.5 and PM10

Test V ehicles

The vehicles used in this study are described in Table 2a.

Table 2a. Vehicles tested.

Vehicle	Cyl. Vehicle Displaced Fuel	Emission

	Miles	Liters	System	System1

1993 Chevrolet Cavalier 4	9545	2.2	FI* EGR/TWC/02S

1993 Ford Taurus	6	5320	3	FI EGR/TWC/02S

^ EGR _ Exhaust Gas Recirculation, TWC ~~ Three Way Catalyst, 02S ~~ Oxygen Sensor
* FI _ Fuel Injeetion

Test Facilities

Vehicle road simulations were conducted on a Horiba Model CDC800/DM A915
computerized DC electric chassis dynamometer. The dynamometer was housed in a temperature
controlled chamber capable of maintaining vehicle test temperatures from -20° to 110°F.
Vehicle emissions were transferred from the vehicle tailpipe to a constant volume sampling

3


-------
(CVS) system through a 7.62-cm i.d. (3 inch) section of flexible stainless steel tubing heated to
230°F. The CVS system, which dilutes the tailpipe emissions with charcoal-filtered room air, has
been described previously.5 A heater has been added behind the dilution air filter to raise the
tunnel dilution air temperature to 150°F to prevent formaldehyde and other compound losses in
the system.6

Test Procedures

Emission tests were conducted in close accord with the Code of Federal Regulations
(CFR) Title 407 using the Urban Dynamometer Driving Schedule (UDDS) and a high speed,
high acceleration rate "off-cycle" sequence (REP05). The UDDS cycle has a maximum speed of
about 57 mph, a maximum acceleration rate of about 3.3 mph/sec, and an average speed of about
19.6 mph. The cycle includes three phases: a cold start transient phase (505 sec), a stabilized
phase (808 sec), and a hot start transient phase (505) with a 10 minute engine-off soak between
phases two and three. The REP05 cycle has a maximum speed of about 80 mph, a maximum
acceleration rate of about 8.5 mph/sec, and an average speed of about 51.5 mph. It includes two
test phases: a high speed driving mode (1195 sec) and a final high acceleration rate mode (205
sec).

At the beginning of the test week, the vehicle to be tested was pre-conditioned by driving

the vehicle over the UDDS cycle with the test fuel, in the test mode, and at the test temperature
to be used for data collection. The daily UDDS test served (as duplicates were run at all
temperatures except 75 °F) as the pre-conditioning for the actual test. During the test week when
the vehicle, fuel, temperature, or mode were to be changed, the vehicle was conditioned in the
afternoon after the daily test was over, by running an LA-4 cycle (phase one and phase two of
the UDDS cycle) at the test conditions to be used the following day. Each UDDS test cycle was
followed by two REP05s. Particulate PM2.5 and PM 10 emissions from a single UDDS cycle
were each collected on single filters using a PM2.5 and a PM 10 cyclone system (University
Research Glassware, Carrboro, NC). Single filters were used to collect the PM2.5 and PM 10
emissions from two consecutive REP05 cycles. Vehicle emissions were measured at 75°, 20°,
0°, and -20 °F for each of the UDDS driving cycles and the REP05 high speed cycles. No
REP05 particulate filter data were obtained at 75°F due to equipment problems.

Integrated emission samples were collected from the CVS system in Tedlar bags for each
UDDS test phase including the initial 124 seconds of phase one. Also, a background air sample
was taken after the charcoal dilution air filter. The regulated emissions (HC, CO, and NOx)
were measured in real time with instruments connected to the CVS system.

Vehicle emissions were characterized to determine speciated (individual) hydrocarbons,
speciated aldehydes, regulated gaseous emissions, and MTBE. Identification or structural
formulas for more than 95% of the emitted HCs have been determined by established gas
chromatographic-mass spectrometric (GC-MS) techniques. Supplementary emission data of
regulated, unregulated, speciated HCs, aldehydes, and MTBE are available from the authors.

4


-------
Tailpipe Emissions

The vehicle UDDS bag samples were analyzed for speciated hydrocarbons and MTBE by
gas chromatography. Data were manually transferred to a PC where peak assignments were
made using a Lotus 1-2-3 program developed for identifying compounds in complex
chromatograms8. The data were finally transferred to the Archived Mobile Source Emissions
Data Base (AME). The concentrations of the regulated emissions were reported as g/mi and
mg/mi for individual HCs, aldehydes, and MTBE.

Aldehyde emissions were sampled from the CVS system through a heated (212°F)
stainless steel line and collected on silica gel cartridges coated with acidified
2.4-dinitrophenylhydrazine. A mass flow controller was used to regulate the sampling rate of the
aldehydes in the exhaust stream at 1 L/min. The aldehydes were analyses previously described
liquid chromatographic procedures.9

Particle mass was determined by weighing the filter before and after sample collection in
an electronic balance with mass resolution of ±1 jig. All filter samples were equilibrated for 24
hours in a constant humidity room prior to weighing.

RESULTS AND DISCUSSION

Single UDDS test was performed with each vehicle at 75°F with the summer grade fuel
at each test mode. Duplicate UDDS tests were performed at 20°F, 0°F, and -20°F with each of
the winter grade fuel at each test mode. The individual UDDS test with the winter grade fuel was
followed by two consecutive REP05 tests. The data were not statistically evaluated since only-
single and duplicate tests were made. Any comparison should be viewed from this perspective.
The THC emission results reported were the values determined by the analytical bench and total
hydrocarbon analyzer (using the flame ionization detector) rather than a summation of the

Table 3. Vehicle tailpipe emission rates at 75°F with summer grade fuel.

Vehicle	 	Cavalier	 	Taurus	

Mode	No	02 Sensor	EGR	No	02 Sensor	EGR

Malfunction Disconnected Disconnected Malfunction Disconnected Disconnected

Fuel, Summer Grade

THC, g/mi

0.19

1.71

0.19

0.3

0.67

0.24

CO, g/mi

4.51

48.71

3.27

4.58

11.61

3.95

NOx, g/mi

0.69

1.16

2

0.59

1.69

0.82

MTBE, mg/mi

NA+

NA

NA

NA

NA

NA

Formaldehyde, mg/mi

0.9

1.17

1.36

0.76

1.33

0.81

Acetaldehyde, mg/mi

1.37

2.84

1.29

0.57

1.18

0.61

Total Aldehydes, mg/mi

3.42

5.64

3.93

2.63

4.27

2.4

Benzene, mg/mi

16.3

106.94

11.07

8.07

45.5

7.69

1,3-Butadiene, mg/mi

0.64

2.82

0.03

1.21

0.20

0.77

^ Data not available

5


-------
Table 4. Vehicle UDDS tailpipe emissions at 20°F.

Vehicle		Cavalier	 	Taurus	

Mode	No	02 Sensor	EGR	No	02 Sensor	EGR

Malfunction	Disconnected	Disconnected	Malfunction	Disconnected	Disconnected

Fuel

Base*

MTBE*

Base

MTBE

Base

MTBE

Base

MTBE

Base

MTBE

Base

MTBE

THC, g/mi

0.49

0.46

2.43

2.08

0.66

0.49

0.62

0.46

1.93

0.8

0.64

0.44

CO, g/mi

10.35

10.63

94.39

77.38

12.75

11.14

6.94

6.1

28.45

11.42

7.64

6.86

NOx, g/mi

0.55

0.64

0.52

0.7

1.61

1.83

0.73

0.45

0.98

1.49

0.91

0.67

MTBE, mg/mi

NA

5.12

NA

8.68

NA

5.42

NA

15.49

NA

12.05

NA

11.63

Formaldehyde, mg/mi

7.6

2.7

3.79

5.94

3.25

7.27

1.57

2.2

2.45

2.19

2.37

2.67

Acetaldehyde, mg/mi

4.55

1.27

7.25

6.94

1.45

3.64

1.1

1.12

1.97

1.22

1.37

1.38

Total Aldehydes, mg/mi

18.93

5.25

20.15

17.52

7.25

10.18

4.24

4.77

7.41

6.4

6.19

7.13

Benzene, mg/mi

23.08

16.12

86.64

62.84

30.71

17.3

25.72

8.56

64.2

28.15

35.12

8.69

1,3-Butadiene, mg/mi

0.7

3.36

7.48

5.82

1.69

1.48

2.78

2.12

9.91

2.48

4.32

1.94

PM2.5, mg/mi

3.58

3.99

7.38

4.72

7.31

4.61

12.7

6.22

15.97

6.74

17.37

7.05

PM10, mg/mi

3.99

4.56

7.21

5.44

8.04

4.46

12.5

5.7

15.35

7.31

17.52

7.05

^ Winter grade fuel without MTBE, ^ Winter grade fuel with MTBE

Table 5. Vehicle UDDS tailpipe emission rates at 0°F.

Vehicle	 	Cavalier	 	Taurus	

Mode	No	02 Sensor	EGR	No	02 Sensor	EGR

Malfunction	Disconnected	Disconnected	Malfunction	Disconnected	Disconnected

Fuel

Base*

MTBE*

Base

MTBE

Base

MTBE

Base

MTBE

Base

MTBE

Base

MTBE

THC, g/mi

1.05

0.92

3.03

2.65

1.15

0.94

0.91

0.81

2.71

1.73

1.17

0.86

CO, g/mi

24.58

19.71

120.9

100.3

25.17

18.95

8.91

9.96

39.96

24.73

10.09

9.92

NOx, g/mi

0.61

1.23

0.47

0.54

0.86

1.86

0.56

0.47

0.94

0.97

0.87

0.73

MTBE, mg/mi

NA

20.46

NA

27.42

NA

19.92

NA

34.58

NA

39.01

NA

34.33

Formaldehyde, mg/mi

4.08

4.16

3.66

7.42

4.01

3.91

3.23

3.72

4.72

4.22

4.36

4.2

Acetaldehyde, mg/mi

2.09

1.99

7.42

4.88

2.2

2.05

1.44

1.55

3.25

2.3

1.3

1.59

Total Aldehydes, mg/mi

8.43

9.99

20.4

14.89

8.87

8.74

6.89

9.21

12.69

10.76

8.35

8.63

Benzene, mg/mi

49.05

22.19

109.5

74.12

50.55

22.06

29.81

21.1

73.14

60.22

24.82

22.19

1,3-Butadiene, mg/mi

3.29

2.7

15.07

6.92

3.89

2.7

4.93

4.34

12.74

9.88

4.51

4.14

PM2.5, mg/mi

14.1

9.12

16.59

10.52

13.79

8.4

28.62

16.8

31.63

18.1

40.7

19.86

PM10, mg/mi

12.34

9.96

16.28

11.61

12.44

9.12

27.74

17.6

32.72

19.86

36.66

19.08

^ Winter grade fuel without MTBE, * Winter grade fuel with MTBE

6


-------
Table 6. Vehicle UDDS tailpipe emission rates at -20°F.

Vehicle	Cavalier	Taurus

Mode	No	02 Sensor	EGR	No	02 Sensor	EGR

Malfunction	Disconnected	Disconnected	Malfunction	Disconnected	Disconnected

Fuel

Base*

MTBE*

Base

MTBE

Base

MTBE

Base

MTBE

Base

MTBE

Base

MTBE

THC, g/mi

1.63

1.5

3.6

3.58

1.74

1.65

1.62

1.53

3.92

2.85

1.67

1.44

CO, g/mi

31.8

25.71

115.95

132.5

32.61

29.14

11.59

13.65

44.69

38.17

10.83

12.26

NOx, g/mi

0.73

0.67

0.85

0.47

0.8

0.54

0.63

0.53

1.05

0.81

0.73

0.64

MTBE, mg/mi

<0.01

56.14

<0.01

45.36

<0.01

59.63

NA

73.65

NA

62.82

NA

57.05

Formaldehyde, mg/mi

3.14

2.71

2.85

7

2.41

3.83

2.36

4.16

3.03

1.33

1.97

3.64

Acetaldehyde, mg/mi

2.68

1.1

7.6

6.47

2.66

2.34

2.03

2.63

3.86

1.18

1.77

2.56

Total Aldehydes, mg/mi

7.96

6.3

22.39

21.21

8.7

10.84

8.83

10.24

15.59

4.27

7.09

10.21

Benzene, mg/mi

72.69

35.47

132.14

91.33

82.63

39.76

51.23

39.63

164.8

90.41

61.03

36.94

1,3-Butadiene, mg/mi

5.39

4.52

25.1

9.85

6.58

5.15

8.85

7.99

29.08

16.79

9.87

7.3

PM2.5, mg/mi

25.55

17.3

37.65

22.4

31.9

19.05

60.46

38.06

73.26

33.18

63.15

33.86

PM10, mg/mi

25.05

21.9

36.25

24.8

32.8

21.8

59

37.54

71.86

33.5

63.26

34.58

t	t

Winter grade fuel without MTBE, Winter grade fuel with MTBE

Table 7. Vehicle REP05 tailpipe emission rates.

Vehicle	 	Cavalier	 	Taurus	

Mode	No	02 Sensor	EGR	No	02 Sensor	EGR

Malfunction	Disconnected	Disconnected	Malfunction	Disconnected	Disconnected

Fuel	Base* MTBE* Base MTBE Base MTBE Base MTBE Base MTBE Base MTBE

Temperature,

THC, g/m

20 F

0.19

0.19

1.36

0.99

0.21

0.15

0.27

0.18

0.68

0.19

0.16

0.15

CO, g/mi



8.38

9.49

70.17

44.04

9.5

8.61

6.48

7.22

16.67

5.71

6.27

5.18

NOx, g/mi



0.59

0.59

0.97

1.35

1.17

0.72

0.57

0.35

1.13

2.06

0.73

0.51

PM2.5, mg/mi



2.0

1.36

2.33

2.0

2.41

1.14

1.9

1.69

2.29

1.35

1.82

1.33

PM10, mg/mi



2.37

1.38

2.61

1.96

2.75

1.31

2.43

1.75

2.26

1.65

2.2

1.46

Temperature,

THC, g/m

OF

0.32

0.18

1.47

1.22

0.25

0.17

0.28

0.15

0.98

0.29

0.2

0.17

CO, g/mi



13.43

9.09

73.99

69.41

10.82

9.5

8.24

5.41

19.2

8.37

5.98

7.6

NOx, g/mi



0.79

0.74

0.94

0.85

0.99

0.97

0.48

0.41

1.07

1.42

0.63

0.61

PM2.5, mg/mi



1.86

1.68

2.53

1.9

1.82

1.82

3.02

1.22

2.27

1.06

2.44

1.97

PM10, mg/mi



2.23

2.19

2.8

2.36

1.82

2.41

3.22

1.82

3.07

1.36

2.71

2.58

Temperature,

THC, g/m

-20 F

0.32

0.15

1.51

1.33

0.27

0.15

0.32

0.17

1.28

0.58

0.28

0.15

CO, g/mi



12.28

8.75

76.65

74.07

11.24

10.45

8.51

6.89

23.52

13.93

8.28

5.96

NOx, g/mi



0.71

0.49

0.99

0.8

1.03

1.04

0.5

0.37

1.11

1.1

0.57

0.47

PM2.5, mg/mi



2.0

2.19

3.75

2.98

2.59

2.54

3.68

1.96

4.1

4.77

3.1

2.27

PM10, mg/mi



1.9

2.64

3.66

4.83

3.04

3.08

4.06

1.59

4.57

5.17

3.52

2.95

t	i

Winter grade fuel without MTBE, Winter grade fuel with MTBE

7


-------
individual hydrocarbons as determined by gas chromatographic analysis. Tables 3-6
summarize the UDDS regulated, toxic, and particulate emission data and Table 7, the REP05
gas phase and particulate data. Emission rates in Table 7 are averages of two tests.

Regulated Emissions

TotalHydrocarbons-UDDS Cycle — Under normal mode, the Cavalier emitted 6-12 %
less hydrocarbons when using the MTBE fuel compared to the base fuel at ambient temperatures
in the 20° to -20°F range. The Taurus emissions were 6-26% less under the same test conditions.
The hydrocarbon emissions increased with decreasing ambient temperatures for all vehicle-fuel
combinations. With the base fuel, the Cavalier emission of 0.49 g/mi at 20°F more than doubled
to 1.06 g/mi at 0°F and tripled tol .63 g/mi at -20°F. Corresponding emissions with the MTBE
fuel were 0.46 g/mi, 0.62 g/mi, and 1.5 g/mi. The Taurus emission of 0.62 g/mi at 20°F with the
base fuel increased to 0.91 g/mi at 0°F and to 1.62 g/mi at -20°F. With MTBE, the Taurus
emission of 0.46 g/mi at 20°F almost doubled to 0.81 g/mi at 0°F and more than tripled to 1.53
g/mi at -20°F. Hydrocarbon emissions with the summer grade fuel at 75 °F were 0.19 g/mi and
0.3 g/mi for the Cavalier and the Taurus respectively.

The vehicle's engine air/fuel mixture is controlled by a feedback loop via the oxygen
sensor. When the sensor is disconnected, the vehicle operates in a slightly rich fuel condition
throughout the test cycle. In this malfunction mode, all vehicle-fuel combinations exhibited
dramatic increases (86 - 800%) in hydrocarbon emissions compared to the normal mode at all
test temperatures. This malfunction mode had a significantly greater impact on the emission of
the Cavalier than for the Taurus. The malfunction emissions of the Cavalier which has a smaller
engine than the Taurus, were significantly greater for most winter grade fuel-temperature
combinations. In only one instance, at -20°F with the base fuel, was the Taurus malfunction
emission greater than that of the Cavalier. At 75 °F the Cavalier malfunction mode emission of
1.71 g/mi was nine times those of the normal mode and the Taurus emission of 0.67 g/mi, 2.2
times that of the normal mode. With both winter fuels, the malfunction mode emissions were
about 5, 3 and 2 times those of the normal mode at 20°, 0°, and -20 °F respectively for the
Cavalier. For the Taurus using the base fuel, the malfunction mode emissions were 3.1,3, and
2.4 times those of the normal mode at 20°, 0° and -20 °F respectively. With the MTBE fuel, the
Taurus malfunction emissions were 1.7, 2.2, and 1.9 times those of the normal mode.
Malfunction emissions for all vehicle-fuel combinations increased as temperature decreased. The
Cavalier malfunction emissions with the base fuel at 20°F increased 25% at 0°F and 48% at
-20°F; with the MTBE fuel, the corresponding increases were 27% and 72% respectively. For
the Taurus, the corresponding increases were 40% and 103%) respectively with the base fuel, and
76%) and 156% respectively with the MTBE fuel.

Disconnecting the EGR resulted in no change in emission of 0.19 g/mi over the normal
mode at 75°F for the Cavalier. However, the Taurus emission of 0.24 g/mi was 20% lower than
that under normal mode. The malfunction emissions at 20° to -20°F for all vehicle-fuel
combinations differ slightly (~ 10%, except the Cavalier with the base fuel at 20°F) from the
corresponding normal mode emissions. Malfunction emissions with the base fuel were also
slightly higher than with the MTBE fuel. It was surprising to note, however, that at the same test
temperature, the Cavalier malfunction emission, in absolute amount, was almost the same as that

8


-------
of the Taurus malfunction emission despite the difference in engine size of the vehicles. It may
not be unreasonable to explain this observation as due to the difference in efliciency/size of the
catalytic converters of the vehicles. For all vehicle-fuel combinations, the malfunction emissions
approximately doubled at 0°F and approximately tripled at -20°F from their initial values at

20°F.

Total Hydrocarbons-REP05 Cycle — The REP05 tests in this study were preceded
either by a UDDS or another REP05 test. The tests started with an already warm engine and
with the catalytic converter performing at or near peak efficiency. REP05 THC emissions for all
vehicle-fuel-temperature-mode combinations were therefore expected to be lower than the
corresponding UDDS emissions. This has been found to be the case in this study.

Under normal mode, the Cavalier THC emissions ranged from 0.19 to 0.32 g/mi with the
base fuel and 0.15 to 0.19 g/mi with the MTBE fuel in the temperature range 20° to -20°F. For
the Taurus, the emission ranged from 0.27 to 0.32 g/mi with the base and from 0.15 to 0.18 g/mi
with the MTBE fuel in the same temperature range. Both the emissions of the Cavalier and the
Taurus with the base fuel showed a discernible upward trend with decreasing temperatures. With
the MTBE fuel, however, the Cavalier emissions showed a discernable downward trend with
decreasing temperatures while with the Taurus, no definitive trend with decreasing temperature
was observed. It must be noted that with the MTBE fuel, the emissions at 20°, 0°, and -20°F
differed from each other by no more than 0.04 g/mi for the Taurus. For all practical purposes
these emissions can be considered almost equivalent. We believe that the predominant factors
influencing REP05 hydrocarbon emissions under the normal mode are the temperatures of the
engine and the catalyst and not the ambient test temperatures.

With the oxygen sensor disconnected, all vehicle-fuel combinations showed dramatic
increases in hydrocarbon emissions compared to the normal mode at all test temperatures. With
the base fuel, the increases ranged from 359-616% for the Cavalier and 152-300% for the
Taurus. With the MTBE fuel, the increases ranged from 421-787% for the Cavalier and 6-241%
for the Taurus. As with the UDDS tests, the malfunction emissions of the Cavalier were
significantly greater than those of the Taurus at all test temperatures. All vehicle-fuel
combinations showed increasing emissions with decreasing test temperatures. From their initial
values at 20°F, the Cavalier malfunction emissions at -20°F increased 11% and 34% with the
base and MTBE fuels respectively. The corresponding increases for the Taurus were 88% and
205% respectively.

Disconnecting the EGR resulted in slight decrease or no change in the malfunction
emission relative to the normal mode for all vehicle-fuel combinations, except in case of the
Cavalier with the base fuel at 20 °F and of the Taurus with the MTBE fuel at 0°F where the
malfunction mode emissions were slightly greater than the normal mode emissions. With the
base fuel, the malfunction emissions of the Cavalier increased 29% and those of the Taurus
increased 75% when the temperature was changed from 20° to -20°F. With the MTBE fuel, the
malfunction emissions of both practically remained the same when the temperature was changed
from 20° to -20°F.

Carbon M on oxide- UDDS Cycle — The Cavalier normal mode CO emission of 4.51 g/mi
at 75 °F was about the same as that of the Taurus. However, from 20 °F to -20 °F, the Cavalier
emissions of 10.4-31.8 g/mi with the base fuel were significantly (49-74%) higher than the
corresponding emissions from the Taurus; with the MTBE fuel, the Cavalier emissions of
10.6-25.7 g/mi were likewise significantly (74-98%) higher. Emissions from all vehicle-fuel
combinations increased with decreasing temperatures. Except at 20°F, emissions from the

9


-------
Cavalier were higher with the base fuel than with MTBE fuel while emissions from the Taurus
were higher with the MTBE fuel than with the base fuel. The Cavalier emissions at 20°F
increased 207% with the base fuel and 104 % with the MTBE fuel at -20°F. The Taurus
emissions at 20°F increased 67% with the base fuel and 124% with the MTBE fuel at -20°F.

Disconnecting the oxygen sensor increased dramatically (100-900%) the CO emissions
relative to the normal mode emissions for all vehicle-fuel combinations at all test temperatures.
At 75°F, the Cavalier emission of 4.51 g/mi increased ninefold and the Taurus emission of 4.58
g/mi more than doubled. Malfunction emissions increased with decreasing temperatures. The
Cavalier emission of 94.4 g/mi with the base fuel at 20°F tripled at -20°F; the emission of 77.4
g/mi with the MTBE fuel more than doubled at -20°F. The Taurus emission of 28.4 g/mi with
the base fuel increased 57% at -20°F; the emission of 11.4 g/mi with the MTBE fuel more than
tripled at -20°F. Except in one instance (Cavalier at -20°F), the malfunction emissions from both
vehicles with MTBE fuel were generally lower than the corresponding vehicle emissions with
the base fuel. With the base fuel, the Cavalier emissions were about three times those of the
Taurus. With the MTBE fuel, the Cavalier emissions were 3.5-6.8 times as much as those of the
Taurus.

The CO emissions with the winter fuels from both vehicles with entirely malfunctioning
EGRs, differed from the normal mode emissions by no more than 23% at all low test
temperatures. The Cavalier emissions at this temperature range were 1.6-3 times greater than
those of the Taurus. The Cavalier consistently emitted more CO with the base fuel than with the
MTBE fuel. The Taurus emissions with the base fuel were slightly greater at 20°F and 0°F than
with the MTBE fuel but slightly less at -20°F. The CO emissions of both vehicles increased as
the temperature decreased. With the Cavalier, the CO emissions increased by about 160% at
-20°F for both the base and the MTBE fuels from their values at 20°F. With the Taurus, the CO
emission increases were 42% and 79% for the base and MTBE fuels respectively.

Carbon M on oxide-RE P05 Cycle — U nder normal mode, the Cavalier CO emissions
(8.3-13.4 g/mi, with base fuel; 8.8-9.5 g/mi, with MTBE fuel) were consistently higher than
those of the Taurus (6.0-8.5 g/mi, with base fuel; 6.41-7.2 g/mi, with MTBE fuel) at all test
temperatures. Except at 20°F, emissions from both vehicles with the base fuel were greater than
those with the MTBE fuel. Emissions from both vehicles with the base fuel showed apparent
upward trend with decreasing temperatures while with the MTBE fuel, there were slight
indications of downward trends with decreasing temperatures.

Disconnecting the oxygen sensor increased the Cavalier emissions with the base fuel
from 4.7 to 8.4 times the corresponding normal mode emissions; with the MTBE fuel, 4.6-8.5
times. The Taurus malfunction emissions with the base fuel increased by about 2.5 times the
corresponding normal mode emissions; with the MTBE fuel, the emission increase range from
-20% to 50% of the normal mode emissions. The Cavalier emissions were consistently higher
(100-680%) than those of the Taurus at all temperatures. The CO emissions for all vehicle fuel
combinations exhibited increasing trends with decreasing temperatures.

Depending on the fuel, the CO emissions of the Cavalier with the EGR disabled ranged
from -20% to 20% of the normal mode; with the Taurus, the CO emissions ranged from -28% to
40% of the normal mode. The Cavalier emissions were consistently higher (25-81%) than those
of the Taurus at all temperatures. With one exception (Taurus at 0°F), emissions with the base
fuel were higher than with the MTBE fuel for all vehicle fuel combinations. The CO emissions
for all vehicle fuel combinations exhibited increasing trends with decreasing temperatures.

10


-------
Oxides of Nitrogen - UDDS — Under normal mode at ambient temperature range 20 °F to
-20°F, the Cavalier NOx emissions ranged 0.55-0.73 g/mi with the base fuel and 0.64-0.67 g/mi
with the MTBE fuel. For the Taurus, the emissions ranged 0.56-0.73 g/mi with the base fuel and
0.45-0.53 g/mi with the MTBE fuel. There were slight indications of upward trends with the
Cavalier emissions with both fuels and the Taurus emissions with the MTBE fuel with
decreasing temperatures. Emissions of the Taurus with the base fuel, however, showed a slight
downward trend with decreasing temperatures. The Cavalier emission at 75°F was 0.69 g/mi; the
Taurus, 0.59 g/mi.

Disconnecting the oxygen sensor almost doubled (1.16 g/mi) the Cavalier emission and
almost tripled (1.69 g/mi) the Taurus emission at 75°F relative to their normal mode values. At
the lower test temperatures, both vehicle emissions with the base fuel showed a slight upward
trend with increasing temperatures; vehicle emissions with the MTBE fuel showed a downward
(especially the Taurus with MTBE fuel) trend. A disabled oxygen sensor had a greater impact on
the Taurus NOx emissions than those of the Cavalier. Depending on the temperature, the Taurus
NOx emissions were 19-53% greater than those of the Cavalier.

Disabling the EGR had a greater impact on the NOx emissions of the Cavalier than those
of the Taurus. The impact was more pronounced with the MTBE fuel. At temperature range 0°F
to 75°F, the Cavalier NOx emissions increased 40-190% relative to the normal mode emissions.
Over this temperature range, the Taurus emissions increased only 25-55% relative to the normal
emission values. Depending on the fuel and temperature, the absolute NOx emissions of the
Cavalier were 1-1.8 times those of the Taurus. At -20°F, the Cavalier emissions were 10%
higher and -16% lower than the Taurus emissions for the base and MTBE fuels respectively. All
vehicle-fuel combinations showed decreasing emissions with decreasing winter test
temperatures.

Oxides ofNitrogen-REPOS — Under normal mode, the Cavalier emitted more NOx
(0.49-0.79 g/mi) than the Taurus (0.41-0.58 g/mi) for all vehicle-fuel combinations at all test
temperatures. The Cavalier emission with the base fuel showed a slight upward trend with
decreasing temperatures while the Taurus showed a downward trend. With the MTBE fuel, the
NOx emissions from both vehicles showed a definite downward trend with decreasing
temperatures. With one exception (Cavalier at 20°F), NOx emissions with the MTBE fuel were
consistently lower than with the base fuel for both vehicles.

With a completely disabled oxygen sensor, the Taurus NOx emissions (1.07-2.06 g/mi)
were consistently higher than those the Cavalier emissions (0.8-1.35 g/mi) for all vehicle-fuel
combinations. With the base fuel, the Taurus NOx emissions increased 98-133% over the normal
mode values; the Cavalier, 18-64% over the normal mode values. With the MTBE fuel, the NOx
emissions of both vehicles decreased with decreasing test temperatures. With the base fuel, the
Cavalier NOx emissions were about the same over the winter test temperature range. Similar
variations were observed with the Taurus emissions.

The NOx emissions of the Cavalier with a completely disabled EGR increased 22-112%
over the normal mode emissions. The Taurus NOx emissions increased 14-45% under the same
test conditions. The Cavalier emissions were consistently higher (41-121%) than the Taurus
emissions at all test temperatures. The Cavalier and the Taurus emissions with the base fuel
showed a downward trend with decreasing temperatures while the Cavalier emissions with the
MTBE fuel showed an upward trend with decreasing temperatures.

11


-------
Toxic Emissions

Benzene — Besides being present in the fuel itself (see Table 1), benzene is emitted from
the tailpipe as a result of its formation during the combustion process involving other fuel
components, such as cyclohexane and the alkylaromatics.1011 In one study from 2° to 7% of the
benzene was determined to be the result of the rearrangement of these molecules during
combustion.12 Benzene emissions showed a general increase as test temperature decreased. A
regression plot of HC vs benzene had a R2 of 0.880 and a slope of 32.099 for the Cavalier and a
R2 of 0.906 and a slope of 37.37 for the Taurus indicating good correlation with the HC
emissions for both vehicles at all test conditions. The correlation of benzene with HC emission
as a function of the individual fuel is shown in Figure 1. The MTBE fuel reduced benzene
emissions, as compared to the base fuel, for both vehicles and all test modes. Benzene
emissions from the Cavalier ranged from a low of 11.07 mg/mi, at 75 °F (EGR disconnected), to
a high of 132.14 mg/mi at -20°F (02 sensor disconnected and base fuel), and for the Taurus a
low of 7.69 mg/mi, at 75°F (EGR disconnected) to a high of 164.80 mg/mi at -20°F (02
disconnected).

140 -
130 -
120 -
110 -
100 -
90 -
60 -
70 -
60 -
50 -
40 -
30 -
20 -
10 -

o -

Figure 1. Correlation plots of benzene vs HC as a function of fuel.

1,3-Butadiene — 1,3-Butadiene is not a gasoline component but a by-product of the
combustion process. This compound was emitted primarily in the initial 2 minutes (present in the
124 second sample bag) of vehicle start-up, when the air-to-fuel mixture was rich and the
vehicle's emission control system was warming up. 1,3-Butadiene emissions from the Taurus
increased as test temperature decreased for both fuels and all test modes. A regression plot of HC
vs 1,3-butadiene had a R2 of 0.720 and a slope of 5.067 for the Cavalier and a R2 of 0.919 and a
slope of 6.742 for the Taurus indicating good correlation with the HC emissions for both vehicles
at all test conditions. Emissions from the Cavalier appeared to be test temperature and fuel
dependent. The MTBE fuel reduced emissions from the Taurus for all test conditions and in
general, also for the Cavalier. The Cavalier emissions ranged from a low of 0.64 mg/mi, at

12


-------
75°F, to a high of 25.10 mg/mi, at -20°F, and the Taurus emissions ranged from 0.77 mg/mi, at
75 °F, to a high of 29.08 mg/mi at -20°F.

Formaldehyde and Acetaldehyde — Formaldehyde and acetaldehyde were not present in
the fuel but are by-products of the incomplete combustion of the fuel. These two aldehydes are
usually the major aldehydes emitted, with formaldehyde the major emission product. A
regression plot of HC vs formaldehyde had a R2 of 0.012 and a slope of 0.174 for the Cavalier
and a R2 of 0.081 and a slope of 0.360 for the Taurus indicating no correlation with the HC
emissions for either vehicle. Further regressions when temperatures or modes were deleted
showed no correlation improvement for either vehicle. The Cavalier formaldehyde emissions
ranged from a low of 0.90 mg/mi, at 75°F, to a high of 7.42 mg/mi, at 0°F, and the Taurus
formaldehyde emissions ranged from 0.76 mg/mi, at 75 °F, to 4.72 mg/mi at 0°F. A regression
plot of HC vs acetaldehyde had a R2 of 0.650 and a slope of 1.835 for the Cavalier and a R2 of
0.679 and a slope of 0.745 for the Taurus indicating fair correlation with the HC emissions for
both vehicles at all test conditions. A regression plot of the data without the -20 °F values
increased the R2 of the Taurus (0.841) but did not make a significant improvement in the R2 of
the Cavalier (0.681). The Cavalier acetaldehyde emissions ranged from a low of 1.10 mg/mi, at
-20°F, to a high of 7.60 mg/mi at -20 °F. The aldehyde emissions were greatest when the oxygen
sensor was disconnected (vehicle running rich).

Particulate Emissions

The mass emission rates of the PM2.5 and PM10 particles were about the same for all
vehicle-fuel combinations at all test modes and temperatures. Figure 2 shows a very good
correlation of the PM10 and PM2.5 particle data. The regression parameters (R2 = 0.995, slope

PM2.5, mg/mi

Figure 2. Correlation of PM10 and PM2.5 particulate emissions from all
vehicle-fuel-mode combinations at all test temperatures.

13


-------
= 0.976, intercept = 0.437) strongly indicate that the particles collected on the cyclone filters
were mostly PM2.5 fine particles.

PM2.5 and PM10 - UDDS Cycle — The particulate emissions (PM2.5 and PM10) from
the UDDS cycle followed the HC and CO trends, and increased as the temperatures decreased.
The Taurus, which had a larger engine than the Cavalier, emitted significantly (25-250%) more
particles than the Cavalier under normal mode at all test temperatures. Except in one instance
(Cavalier at 20°F), both vehicles emitted more particles when using the base fuel than when
using the MTBE fuel. The Taurus emitted about twice as much particles with the base fuel than
with the MTBE fuel while the Cavalier emitted about 10-35% more with the base fuel than with
the MTBE fuel. The Taurus PM2.5 emission of 12.7 mg/mi at 20°F, with the base fuel,
quadrupled at -20°F. With the MTBE fuel, the emission (6.2 mg/mi) at 20°F increased sixfold at
-20°F. Under the same sets of conditions, the Cavalier PM2.5 emission with the base fuel (3.58
mg/mi) and with the MTBE fuel (3.99 mg/mi) increased 7 and 4.3 times respectively at -20°F.
Maximum particulate emission observed for the Taurus was 60 mg/mi with the base fuel and 28
mg/mi with the MTBE fuel, both at -20°F. Corresponding maxima for the Cavalier were 25
mg/mi and 21 mg/mi, respectively.

As was noted earlier, disabling the oxygen sensor resulted in dramatic increase of both
the HC and CO emissions at all test temperatures compared to the normal mode and that this
malfunction mode impacted the emissions of the Cavalier more than those of the Taurus.
Particulate emissions did increase with the disconnected oxygen sensor, but the increases were
more modest than those of the CO and the hydrocarbons. The Taurus emissions were
consistently higher (34-116%) than those of the Cavalier for both fuels at all test temperatures.
Over the temperature range 20 to -20°F, the Cavalier particulate malfunction emissions with the
base fuel increased on the average by about 54% over the normal mode emissions; with the
MTBE fuel, the average increase was about 18%. For the Taurus, the average increase was about
19% with the base fuel and 5% with the MTBE fuel. The malfunction particulate emissions of
both vehicles at 20°F increased about 5 times at -20°F. Maximum particulate emissions for the
Taurus were 73 mg/mi and 38 mg/mi for the base and MTBE fuel respectively at -20°F. For the
Cavalier, maximum particulate emissions were 38 mg/mi and 25 mg/mi for the base and MTBE

Figure 3. Correlation plots of PM2.5 vs HC as a function of fuel and oxygen sensor
malfunction.

14


-------
fuel respectively at the same temperature. Figure 3 shows the impact of the fuel and the disabled
oxygen sensor on PM2.5 emissions. Under the normal mode, the PM2.5 particle generation rates
with HC (slope of the regression line) of the Taurus with the base and the MTBE fuels were
about twice those of the Cavalier with the same fuel. For the Taurus, the particle generation rate
was slightly higher with the base than with the MTBE fuel; for the Cavalier the particle
generation rates were about the same for both fuels. However, under the oxygen malfunction
mode, the particle generation rates of both vehicles are comparable.

With the EGR disconnected, the Cavalier particulate emissions with the base fuel
increased on the average by about 43% over the normal mode but hardly no increase (~1%) at all
with the MTBE fuel. The Taurus emissions with the base fuel, on the average increased by about
27% over the normal mode; with the MTBE fuel, the increase was about 7%. The Taurus
emissions were also consistently higher (53-200%) than the corresponding Cavalier emissions.
For both fuels, the Cavalier emissions at 20°F more than quadrupled at -20°F, while the Taurus
emissions almost quadrupled with the base fuel and quintupled with the MTBE fuel. Maximum
particulate emissions for the Taurus were 63 mg/mi and 35 mg/mi at -20°F with the base and
MTBE fuel respectively. Corresponding values for the Cavalier were 32 mg/mi and 22 mg/mi
respectively. Figure 4 shows for the same fuel, the EGR malfunction particle generation rates
with hydrocarbon emissions are comparable to those under normal mode.

Figure 4. Correlation plots of PM2.5 vs HC as a function of fuel and EGR malfunction.

PM2.5 and PM10 - REP05 Cycle — The REP05 hydrocarbon emissions were generally
low and several factors less than the corresponding values observed with the UDDS cycle. The
same low emissions were observed with the PM 2.5 and PM10 particles. Since as indicated
earlier, the particles collected on the PM2.5 and the PM10 cyclone filters were mostly PM2.5
particles, it is perhaps more appropriate, especially at these low levels, to present the particulate
emission as an average, i.e., (PM2.5 + PM10)/2. Average particulate emissions, in general,
increased as the test temperature decreased.

Under normal mode, the average particulate emissions with the base fuel were about
twice those with the MTBE fuels for both vehicles except for the Cavalier at -20°F. Over the
temperature range 20°F to -20°F, the average Cavalier particulate emissions ranged 2.0-2.4
mg/mi and 1.3-2.4 mg/mi with the base and MTBE fuel respectively while those of the Taurus

15


-------
ranged 2.7-3.9 mg/mi and 1.5-1.8 mg/mi for the base and MTBE fuel respectively. The average
Cavalier particulate emission at 20 °F with the MTBE fuel about doubled at -20 °F; with the base
fuel, the emission decreased by about 10%. The average Taurus particulate emission of 2.16
mg/mi at 20°F with the base fuel doubled at -20°F while emissions with the MTBE fuel
remained about the same.

Disconnecting the oxygen sensor increased the average particulate emissions of both
vehicles. Over the temperature range 20°F to -20°F, the average Cavalier particulate emissions
ranged 2.7-3.7 mg/mi with the base fuel and 2.0-2.9 mg/mi with the MTBE fuel. The average
Taurus particulate emissions ranged 2.4-4.3 mg/mi with the base fuel and 1.4-4.8 mg/mi with the
MTBE fuel. Both vehicle emissions showed upward trends with decreasing temperature. Figure
5 shows regression plots of PM2.5 vs HC emissions of the Cavalier and the Taurus running on
base and MTBE fuels under normal and oxygen sensor malfunction mode. No meaningful
correlation was observed between the PM2.5 and the HC emissions of the Cavalier with either of
the winter grade fuels nor of the Taurus with the MTBE fuel under normal mode. Fair to good
correlations were obtained with the Cavalier and the Taurus under malfunction mode and the
Taurus with the base fuel under normal mode. The malfunction PM2.5 particle generation rate
with HC emissions (slope of the regression line) of the Cavalier was almost three times faster
with the base fuel than with the MTBE fuel. With the Taurus, the PM2.5 particle generation rate
with HC emissions was in the reverse direction of the fuel; the generation rate was three times
faster with the MTBE than with the base fuel.

3,50 ¦
3.00 1

0,50 ¦
0.00 •

I ~ ~

~ nD

Base Fuel, normal mode, R2=0.000, Slope=0.0
MTBE fuel, normal mode, R2=0.021, Slope=-1.7
Base fuel, 02 sensor disabled, R2=0.622, Slope=7.5
MTBE fuel, 02 sensor disabled, R2=0.592, Slope=2.6

TAURUS
REP05

Base Fuel, normal mode, R2=0.801, Slope=20.8
MTBE fuel, normal mode, R2=0.039, Slope=3.1
Base fuel, 02 sensor disabled, R2=0.705, Slope=3.0
MTBE fuel, 02 sensor disabled, R2=0.834, Slope=9.2

0.60

0.80

1,00

1.20

1.40

0,60

0.80

1.00

1.20

Figure 5. Correlation plots of REP05 PM2.5 vs HC as a function of fuel and oxygen
sensor malfunction.

Average particulate emissions of the Cavalier under the EGR malfunction mode ranged
1.8-2.8 mg/mi with the base fuel and 1.5-2.8 mg/mi with the MTBE fuel. The Taurus
malfunction emission ranged 2.0-3.3 mg/mi and 1.4-2.6 mg/mi with the base and the MTBE fuel
respectively. The malfunction emissions trended upwards with decreasing test temperatures. The
correlation of malfunction PM2.5 particles with HC emissions of the Cavalier was poor (R2 <
0.1) for both fuels. The same was true for the Taurus with the MTBE fuel. Correlation of the
normal PM2.5 and the malfunction PM2.5 emissions with HC emissions for the Taurus operating
on the base fuel was fairly good (R2 > 0.80). Figure 6 shows regression plots of the normal and

16


-------
CAVALIER
REP05

Base Fuel, normal mode, R2=0.000, Slope=0.0
MTBE fuel, normal mode, R2=0.021, Slope=-1.7
Base fuel, EGR disabled, R2=0.074, Slope=6.0
MTBE fuel, EGR disabled, R2=0.012, Slope=14

Base Fuel, normal mode, R2=0.801, Slope=20.8
MTBE fuel, normal mode, R2=0.039, Slope=3.2
Base fuel, EGR disabled, R2=0.849, Slope=10.5
MTBE fuel, EGR disabled, R2=0.009, Slope=4.4

0.10

0,20

0,30

0,00 -

0.40	0,00

TAURUS
REP05

Figure 6. Correlation plots of PM2.5 vs HC as a function of fuel and EGR malfunction.

malfunction PM2.5 emissions vs HC. The malfunction PM2.5 particulate generation rate as a
function of HC emission for the Taurus is about twice as fast as the generation rate under normal
mode.

SUMMARY AND CONCLUSIONS

The regulated (HC, CO, andNOx), the toxic (benzene, 1,3-butadiene, formaldehyde, and
acetaldehyde), and the particulate (PM2.5 and PM10) emissions data were obtained from two
vehicles operated on three different fuels. These test vehicles could or could not be
representative of the on-road fleet. The malfunction conditions introduced are extreme conditions
in which the vehicle's oxygen sensor and the EGR valve were rendered completely inoperable.
In actuality, the condition of these simulated malfunctions (oxygen sensor and EGR valve
disconnected) on-road vehicles could be anywhere in the range of being completely operable and
inoperable. Limited resources restricted our testing to two vehicles and to duplicate runs (one at
75°F) at 20°, 0° and -20°F. Also, due to resource limitations we were only able to take only
single PM2.5 and PM10 particulate filter from each of the UDDS test cycles (all three phases
combined) rather than individual PM2.5 and PM10 particle filter from each of the UDDS's three
phases.

Hvdrocarbons-UDDS cycle — The MTBE fuel produced less HC emissions than the base
fuel for both vehicles. Hydrocarbon emissions increased as test temperature decreased.
Maximum THC was emitted when the oxygen sensor was disconnected. Hydrocarbon emissions
with the EGR disconnected differed very slightly from the normal mode emissions. The Taurus,
which had a bigger engine than the Cavalier, emitted more hydrocarbons at 75 °F than the
Cavalier. However, at the winter test temperatures under the normal mode and the EGR
malfunction mode, the Taurus emissions were comparable to those of the Cavalier. At the same
winter test temperatures when the oxygen sensor was disconnected, the Cavalier emissions were
significantly greater than those of the Taurus at all temperatures. We suggest that these

17


-------
significant emission differences can be reasonably attributed to the capacity/efficiency
differences of the catalytic converters of the vehicles.

Hydrocarbons-REPOS cycle — Hydrocarbon emissions were a few factor lower than the
corresponding UDDS emissions. The MTBE fuel produced less emissions than the base fuel for
both vehicles. Emissions increased as test temperature decreased. Maximum HC was emitted
when the oxygen sensor was disconnected. Under this conditions, the Cavalier emissions were
significantly greater than those of the Taurus for both fuels at all test temperatures. The EGR
malfunction emissions differ only slightly from the corresponding normal mode emissions.

Carbon M on oxide- UDDS cycle — Carbon monoxide emissions with the summer grade
fuel were about he same for the Cavalier and the Taurus. At the winter test temperatures, the
Cavalier emissions were significantly (49-170%) greater than those of the Taurus under normal
mode. The Cavalier emissions with the base fuel were generally greater than those with the
MTBE fuel at all test conditions and temperatures. The Taurus emissions with the base fuel were
comparable to those with the MTBE fuel under the normal and the EGR malfunction mode.
Maximum CO was emitted at each test temperature when the oxygen sensor was disconnected.
Under this condition, the Cavalier emissions with the base fuel were about three times higher
than those of the Taurus with the base fuel and the Cavalier emission with the MTBE fuel were
three to seven times higher than those of the Taurus with the MTBE fuel.

Carbon M on oxide-RE P05 cycle — The REP05 CO emissions were decidedly lower
than the UDDS emissions under all modes and test temperatures. Emissions were generally
higher with the base fuel than with the MTBE fuel for both vehicles. Emissions generally
increased as the temperature decreased. EGR malfunction emissions were comparable to the
normal mode emissions for each vehicle. Maximum CO was emitted at each test temperature
when the oxygen sensor was disconnected. Under this mode, the Cavalier CO emissions with the
base fuel were 3-4 times higher than those of the Taurus; with the MTBE fuel, the Cavalier
emissions were 5-8 times higher than those of the Taurus with the MTBE fuel.

Oxides of Nitrogen - UDDS cycle — With the summer grade fuel at 75 °F, the Cavalier
normal mode NOx emission (0.69 mg/mi) increased 68% when the oxygen sensor was
disconnected and 190% when the EGR was disconnected. The Taurus normal mode emission
(0.59 mg/mi) increased 186%) when the oxygen sensor was disconnected but only 38 % when the
EGR was disconnected. With the winter fuels at winter test temperatures, the NOx emissions of
both vehicles with the base fuel showed discernable increasing trends with decreasing
temperatures under the normal or the oxygen sensor malfunction mode. With MTBE, NOx
emissions showed a discernable downward trend. Maximum NOx was emitted at each
temperature when the EGR was disconnected for all vehicle-fuel combinations. Emissions
trended downward with decreasing temperature under the EGR malfunction mode.

Oxides ofNitrogen-REPOS cycle — Normal mode emissions of the Cavalier were about
the same at 20° and 0°F for both the base and the MTBE fuel. At -20°F, emission with the base
fuel was about 44% higher. NOx emissions of the Cavalier were significantly (4-121%) higher at
all temperatures than those of the Taurus under the normal or EGR malfunction mode. However,
when the oxygen sensor was disconnected, the Taurus emissions were significantly (12-79%)
higher than those of the Cavalier. Except for the Cavalier with the MTBE fuel, NOx emissions
trended downward with decreasing temperatures.

Benzene — Benzene emissions generally increased as test temperature decreased.
Disconnecting the oxygen sensor produced the highest benzene emission rate and the MTBE fuel

18


-------
reduced benzene emissions. Regression analysis shows good correlation between HCs and
benzene with both vehicles at all test conditions.

1,3-Butadiene — The 1,3-butadiene emissions increased as test temperature decreased.
Disconnecting the oxygen sensor generally produced greater emissions and the MTBE fuel
generally reduced emissions. Regression analysis of HCs and 1,3-butadiene indicated fair
correlation (R2 = 0.720) for the Cavalier and very good correlation (R2 = 0.919) for the Taurus.

Aldehydes — Formaldehyde and acetaldehyde did not show any trends with test
temperature, test mode or test fuel. Regression analysis of HCs and formaldehyde showed no
correlation (R2 less than 0.1 for both vehicles) but a regression of HCs and acetaldehyde showed
a fair correlation (R2 0.650 and 0.679).

PM2.5 and PM 10particulate-UDDS cycle — Particulate emissions were mostly PM2.5
particles. Particulate emissions correlated very well (R2 > 0.920) with HC emissions for every
vehicle-fuel-malfunction mode combinations. Particle emissions increased 4-5 times when the
test temperature was decreased from 20°F to-20°F. Both vehicles emitted more particles with the
base fuel than with the MTBE at all test conditions. Taurus particle emissions ranged 13-73
mg/mi with the base fuel and 6-34 mg/mi with the MTBE fuel. The Taurus which had a bigger
engine than the Cavalier emitted 2-3 times as much particles as the Cavalier. Maximum particles
were emitted when the oxygen sensor was disconnected. Particle emissions with a disconnected
EGR differed only slightly from normal mode emissions.

PM2.5 and PM If) particulate-REPOS cycle — The REP05 particle emissions (2-5
mg/mi) were several factors lower than the UDDS particle emissions for all
vehicle-fuel-malfunction mode combinations. Particle emissions trended upward with decreasing
temperatures. The normal mode and the EGR malfunction mode particle emissions of the
Cavalier for both the base and the MTBE fuels were about the same. The EGR malfunction
mode emissions of the Taurus with the MTBE fuel were slightly higher than those of the normal
mode. Maximum particles were emitted when the oxygen sensor was disconnected.

ACKNOWLEDGMENTS

The authors acknowledge and express gratitude to William Crews and Richard Snow of
Clean Air Vehicle Technology Center, Inc. for providing analytical assistance. We also
acknowledge Jerry Faircloth, Versal Mason, and Phil Carter of ManTech Environmental for
vehicle testing and data processing assistance. The authors would also like to express their
appreciation to Jim Braddock of EPA for his technical and editorial assistance in preparing this
manuscript.

19


-------
DISCLAIMER

The U. S. Environmental Protection Agency through its Office of Research and
Development funded and managed the research described here under Contract 68-D5-0156 to
Clean Air Vehicle Technology Center, Inc. It has been subjected to Agency review and approved
for publication. Mention of trade names or commercial products does not constitute an
endorsement or recommendation for use.

REFERENCES

1.	Health Effects Institute Communications, Research Priorities for Mobile Air Toxics, No. 2

(1993)

2.	D. W. Dockery, C. A. Pope., III. Annu. Rev. Public Health, 15, 107-132 (1997)

3.	Particulate air pollution and daily mortality: Replication and validation of selected
studies; Health Effects Institute: Boston, 1995

4.	S. Cadle, P. A. Mulawa, J. Ball, C. Donase. A. Weibel. J. C. Sagebiel, K. T. Knapp, R.
Snow, Environ. Sci. Technol., 31, 3405-3412 (1997)

5.	R. Snow, L. Baker, W. Crews, C. O. Davis, J. Duncan, N. Perry, P. Siudak, F. Stump, W.
Ray, J. Braddock, "Characterization of emissions from a methanol fueled vehicle". J. Air
Pollu. Control Assoc. 39, 48 (1989)

6.	F. Stump, S. Tejada, F. Black, W. Ray, W. Crews, R. Davis, "Compound injection to
assure the performance of motor vehicle emissions sampling systems". SAE Paper
961118, Society of Automotive Engineers, Warrendale, PA, 1996

7.	Code of Federal Regulations, Title 40, Part 86, U. S. Government Printing Office,
Washington, DC, 1983

8.	S. Tejada, J. Sigsby, "Identification of chromatographic peaks using Lotus 1 -2-3", J.
Chromatogr. Sci., 26, 292 (1988)

9.	S. Tejada, "Evaluation of silica gel cartridges coated in situ with acidified
2,4-Dinitrophenylhydrazine for sampling aldehydes and ketones in air", Intern. J.

Environ. Anal. Chem. 26, 167 (1986)

10.	E. W. Kaiser, W. O. Siegl, D. F. Cotton, R. W. Anderson, "Effects of fuel structure on
emissions from a spark-ignited engine II. Naphthalene and aromatic fuels". Environ. Sci.
Technol., 26, 1581-1586 (1992)

11.	E. W. Kaiser, W. O. Siegl, Y. I. Henig, R. W. Anderson, F. H. Trinker , "Effects of fuel
structure on emissions from a spark- ignited engine". Environ. Sci. Technol., 25, 2005
(1991)

12.	F. D. Stump. K. T. Knapp, W. D. Ray, C. Burton, R. Snow, "The seasonal impact of
blending oxygenated organics with gasoline in motor vehicle tailpipe and evaporative
emissions-Part II", SAE Paper 902129, Society of Automotive Engineers, Warrendale, PA, 1990

20


-------