EPA-650/2-73-002
June 1973
Environmental Protection Technology Series
-------
EPA-650/2-73-1
CHEMICAL AND PHYSICAL
CHARACTERIZATION
OF AUTOMOTIVE EXHAUST
PARTICULATE MATTER
IN THE ATMOSPHERE
(Year ending June 30, 1973)
by
J. F. Foster, D. A. Trayser,
C, W. Melton, and R. I. Mitchell
Battelle Columbus Laboratories
505 King Avenue, Columbus, Ohio 43201
Contract No. 68-02-0279
Program Element No. 1A1010
EPA Project Officer: Dr. Jack Wagman
Chemistry and 'Physics Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
COORDINATING RESEARCH COUNCIL INC.
30 ROCKEFELLER PLAZA
NEW YORK, NY 10020
APRAC PROJECT CAPE-19-70
and
OFFICE OF RESEARCH AND MDNITORING
U. S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, DC 20460
-------
This report has been reviewed by the Environmental Protection Agency and
approved for publication. Approval does not signify that the contents
necessarily reflect the views and policies of the Agency, nor does
mention of trade names or commercial products constitute endorsement
or recommendation for use.
11
-------
MANAGEMENT SUMMARY
The purpose of the program is to determine the physical and
chemical characteristics of exhaust particulate matter from
internal combustion engines as a function of sampling proce-
dure, engine operating conditions (including emission control
systems), fuel composition, and residence time in the atmos-
phere. Emphasis will be placed on predicting the fate of
exhaust particulate matter in the atmosphere.
The scope of the project work covers the generation of auto-
mobile particulate matter with specially chosen, measured
engines. The auto exhaust is diluted in two stages: first
in a mixing tunnel, then in a large residence chamber with
relatively small wall loss. The properties of exhaust parti-
cles are being studied from their initial formation to their
ultimate remo'val from the chamber atmosphere.
A major experimental study has been necessary to develop test
procedures which give as true as possible values of the
particulate emissions. This has been a consequence of the
very small weights of the particulate matter and the very
large dilutions necessary to simulate the way they exist in
the atmosphere, as well as the fact that the collection of
particulate material is strongly affected by the techniques
used to collect them. Thus the geometry, temperature, and
velocity of the sample stream had to be set to optimums, as
well as the size, weight, and type of filter used, weighing
techniques, and the many car operating variables.
In the study thus far, two identical 1970 Fords with matched
standard 1971 engines (engine modification emission controls)
were conditioned similarly for 5,000 miles on unleaded fuel,
and shown to have equivalent emission characteristics. The
vehicles then were operated for an additional 8,000 miles, one
on unleaded fuel, the other on the same base fuel to which
had been added 2.5 grams lead per gallon measured as TEL Motor
Mix. During this phase, exhaust gas emissions and particulate
matter were measured at preselected intervals. The gaseous
measurements enabled a detailed assessment of vehicle opera-
tion to be made. The particulate matter was sampled from the
dilution tunnel and the residence chamber in each run. The
material collected from the dilution tunnel is considered
representative for fresh auto exhaust as it is emitted into
the atmosphere. The particles sampled from the residence
chamber inidcafe the nature of the air-suspended fraction
after aging for periods up to 24 hours.
In this contract year two major problem areas were encountered
and significant efforts were devoted to solving them. The
-------
first relates to reproducible day-to-day operation of vehicles
for particle generation. Vehicle operation during the cold
start portion of the cycle and the operation of the choke
were found to be critical. These problems have been identi-
fied and the procedures have been standardized for reproduci-
ttlity.
The second problem relates to the measurement by filtration
of the carbonaceous particulate matter in vehicle exhaust.
A filter sampling system and filter media have been selected
that show reproducible weight gain. Work on the filtration
problem continues with the aim of establishing a procedure
for absolute measurements.
The total particulate emission rate from the nonleaded car is
in the range .0.04-0.075 gram per mile for the 1972 FTP (Federal
Test Procedures) Cycle. There is a correlation between the
particulate and the hydrocarbon emission rates. The parti-
culate emissions from leaded fuel were 0.05 to 0.15 g/mile
more than those from unleaded fuel, or from 60 to 100*percent
higher when compared at the same hydrocarbon emission level.
For both fuels the mass median equivalent diameter of the
exhaust particles sus.pended in the dilution tunnel and in the
residence chamber were in the submicron size range. Measure-
ments with the Minnesota Aerosol-Analyzing System indicate
a bimodal size distribution.
Electron microscope studies of the exhaust particles suggest
four possible, mechanisms of particle growth; agglomeration,
condensation, crystal growth, deposition and ablation of exhaust
system deposits.
-------
TABLE OF CONTENTS
MANAGEMENT SUMMARY ......................... i
INTRODUCTION ............................ 1
OBJECTIVE .......... .................. 2
EXPERIMENTAL APPARATUS AND PROCEDURES ............... 2
Fuels ............................. 4
Project Cars . . ....................... 4
Conditioning of Cars ..................... 4
Test Cycles .......................... 5
Test Run Procedure ...................... 5
Preconditioning .............. . ...... 5
Test Run. ........................ 6
Post-Cycle Operation Check ................ 6
Engine Air Flow Measurements ................. 6
Choke Operation ................... ..... 7
Dilution Tunnel ........................ 8
Residence Chamber ....................... 8
Instrumentation ........................ 8
Filtration and Weighing .................... 9
EXPERIMENTAL RESULTS ...... . ................. 12
Exhaust Emission Rates .................... 12
Reproducibility of Mass of Particulate Emissions ....... 14
Correlation of Light Scattering with Mass ........... 18
Addition of Foreign Materials to the Residence' Chamber .... 18
Size Distribution of Auto Exhaust Particles
with the Cascade Impactor .................. 22
iii
-------
TABLE OF CONTENTS
(Continued)
Size Distribution of Particles with the
Minnesota Aerosol-Analyzing System 22
Correlation of Particulate Mass with
Hydrocarbon Emissions 26
Morphological Characteristics of Particles 28
CONCLUSIONS . . 31
Emission Levels .............. . 31
Particulate Characteristics . . . . 34
Morphological Analysis 34
APPENDIX
Fuels A-l
Air Flow Measurements A-4
Car Maintenance f . . . . A-4
Oil Consumption A-7
Development of Acceptable Choke Operation A-7
Apparatus A-7
Characteristic Stages in Choke Schedules A-10
Experimental Studies of Choke Schedules A-ll
Choke Schedules and Starting Sequences A-14
Choke Schedule I (Modified LA-4 Cycle) A-14
Choke Schedule I-S (Standard LA-4 Cycle) A-14
Choke Schedule II (Modified LA-4 Cycle) A-16
Choke Schedule II-S (Standard LA-4 Cycle) A-16
Periods of Choke Schedule Use A-16
IV
-------
TABLE OF CONTENTS
(Continued)
Page
Summary of Test Conditions in Experimental Runs A-18
Dilution Tunnel A-18
Sample Probes » . A-18
Tunnel Sample-Point Pressure A-23
Tunnel-Bag Sample A-23
Calculation of Exhaust Mass Emissions
from Tunnel-Bag Sample. A-26
Lead Deposits in Pipe to Residence Chamber A-26
Test for Uniformity of Mixing in the Tunnel ........ A-27
Gaseous Contaminant Injection System A-27
Solid Contaminant Injection System A-29
Dust Contaminant A-29
Residence Chamber A-32
Chamber Configuration A-32
Dilution Ratio Experiments A-35
Chamber Mixing Experiments A-37
Chamber Humidification. .... A-38
Dilution Ratio Definition and Method of Calculation . . . A-41
Instrumentation. A-42
Gas Analysis . A-42
Speed Controller A-46
Minnesota Aerosol-Analyzing System A-46
Studies of Filtration and Weighing Procedures A-47
Microbalance in Controlled Atmosphere Balance Room. . . . A-47
Repeatability of Weighings with the Microbalance A-47
-------
TABLE OF CONTENTS
(Concluded)
Page
Filter Media . A-49
Pattern of Sampling for Particulate Matter A-51
Experimental Results Supplementary Data A-51
Test Conditions A-rSl
LIST OF TABLES
Table 1. Summary of Test Results . 13
Table 2. Mass Emissions of Particles Collected
From Tunnel by Filtration 15
Table 3. Mass Concentrations of Particles Collected from
Tunnel and Residence Chamber . 16
Table 4. Correlation of Light Scattering with Aerosol
Mass in the Residence Chamber 19
Table 5. Studies of Environmental Variables 21
Table 6. Comparison of Automobile Particulate Data 23
Table A-l. Characteristics of Fuels. ........ A-l
Table A-2. Oil Consumption During Test Series A-8
Table A-3. Test Data Relating to Choke Schedule Development. . . . A-13
Table A-4. Summary of Experimental Studies of Exhaust Emission . . A-19
Table A-5. Chemical Composition of Fine Arizona Dust A-32
Table A-6. Summary of Initial Tunnel-to-Chamber
Dilution Ratio Experiments A-36
Table A-7. Gas Analysis Instruments A-43
Table A-8. Repetitive Weighings of Blank Filters and
Other Materials A-48
Table A-9. Laboratory and Operating Data for Modified and
Standard LA-4 Cycle Runs A-52
vi
-------
LIST OF TABLES
(Continued)
Page
Table A-LO. Air-Fuel Ratios Determined from Steady-State
Exhaust-Gas Analyses After Each Run A-53
Table A-ll. Dilution Ratios of Diluted Exhaust in
Residence Chamber for Modified and Standard
LA-4 Cycle Runs A-55
LIST OF FIGURES
Figure 1. Perspective View of Test Apparatus and
Auxiliary Instrumentation 3
Figure 2. Correlation Between Light Scattering and
Aerosol Mass Concentration . 19
Figure 3. Volume Distribution of Unleaded Particles 24
Figure 4. Volume Distribution of Leaded Particles 25
Figure 5. Correlation Between Particle Emissions and
Hydrocarbon Emissions 27
Figure 6. An Electron Micrograph of Unleaded Exhaust
Particulate Collected at the 0.5 u.m Impactor Stage
After 4 Hours Residence (Run 4-17) 29
Figure 7. An Electron Micrograph of Leaded Exhaust Particulate
Collected at the 0.5 um Impactor Stage After 4
Hours Residence (Run 4-25) 29
Figure 8. Electron Micrographs Showing Nonleaded Exhaust
Particulate From Run 12-15 30
Figure 9. Electron Micrographs Showing Leaded Exhaust
Particulate From Selected Experiments. ........ 32
Figure 10. Electron Micrographs Showing Effects of Dust,
Ammonia, and Humidity from Selected Experiments. ... 33
Figure A-l. Variation of Engine Air Flow and Road Horsepower
with Vehicle Speed Using Unleaded Fuel A-5
Figure A-2. Air-Fuel Ratio Versus Carburetor Air Flow
With Unleaded Fuel A-6
Figure A-3. Device for Controlled Choke Opening on a
Reproducible Time Schedule A-9
VI1
-------
LIST OF FIGURES
(Continued)
Figure A-4. Idealized Choke Opening Schedules for
Modified Cold-Start LA-4 Cycle A-12
Figure A-5. Choke Schedule I A-15
Figure A-6. Choke Schedule I-S A-15
Figure A-7. Choke Schedule II A-17
Figure A-8. Choke Schedule II-S A-17
Figure A-9. Pattern of Probe Inlets in Cross-Section
of Dilution Tunnel at Sampling Point A-21
Figure A-10. Connection at Tunnel of Pipe to Carry Diluted
Exhaust to Residence Chamber A-22
Figure A-11. Recorded Differential Pressure Between Tunnel and
Residence Chamber During a Modified LA-4 Cycle. . . A-24
Figure A-12. Tunnel-Bag Sampling System A-25
Figure A-13. Gaseous Contaminant Injection System A-28
Figure A-14. Schematic of Dust Feeder A-30
Figure A-15. Particle Size Distribution of Classified
Arizona Road Dust A-31
Figure A-16. Layout of Residence Chamber and Purge-
Circulation System A-33
Figure A-17. Tunnel-To-Chamber Sample-Pipe Discharge Nozzle. . . A-34
Figure A-18. Time To Mix Propane in Residence Chamber with
Sample Inlet Nozzle Shown in Figure A-17 A-39
Figure A-19. Schematic of Wick-Type Humidifier in Residence
Chamber Purge Duct A-40
Vlll
-------
FOURTH ANNUAL SUMMARY REPORT
on
CHEMICAL AND PHYSICAL CHARACTERIZATION OF
AUTOMOTIVE EXHAUST PARTICUIATE MATTER.
IN THE ATMOSPHERE
to
COORDINATING RESEARCH COUNCIL
(CRC-APRAC PROJECT NO. CAPE-19-70)
and
ENVIRONMENTAL PROTECTION AGENCY
CONTRACT NO. 68-02-0279)
from
BATTELLE
Columbus Laboratories
July 25, 1974
INTRODUCTION
This report describes a fourth year ' ' of studies directed
toward determining the fate of auto exhaust particles in the atmosphere.
During this year two preconditioned cars have been operated on test cycles
under controlled conditions to generate particulate matter for detailed
(1) C. W. Melton,et al, "Physical-Chemical Characteristics of Particles
Associated with Polynuclear Aromatic Hydrocarbons Present in Automo-
bile Exhaust", Final Summary Technical Report for the period Janu-
ary 24, 1969, to March 31, 1970, to Coordinating Research Council
(APRAC-CAPE-12-68-Neg. 59), January 29, 1970.
(2) J. F. Foster, et al, "Chemical and Physical Characterization of Auto-
motive Exhaust Particulate Matter in the Atmosphere", Final Summary
Report for the period July 1, 1970, to June 24, 1971, to Coordinating
Research Council (CAPE-12-68-Neg. 59 and CAPE-19-70), October 6, 1972.
(3) C. W. Melton, R. I. Mitchell, D. A. Trayser, J. F. Foster, "Chemical
and Physical Characterization of Automotive Exhaust Particulate Matter
in the Atmosphere", Final Summary Report for the period June 25,
1971, to June 30, 1972, to Coordinating Research Council (CAPE-19-70)
and Environmental Protection Agency (Contract No. 68-02-0205),
June 14, 1973.
-------
study. In this report are presented descriptions of the apparatus, includ-
ing cars, exhaust-dilution tunnel, residence chamber for aging particulate
matter in suspension, and instrumentation; and procedures, including pre-
paration and maintenance of the cars, test cycles, and measurements of the
properties of generated particulate matter. Development and standardization
of apparatus and procedures was carried out during the early part of the
year by operation of the cars, as well as by tests on indi-vidual segments
of the associated apparatus.
Subsequently a total of 58 runs was completed in the integrated
test system in which particles were generated from both leaded and unleaded
fuels. Results from the 58 tests are 'tabulated in the experimental section.
OBJECTIVE
The objective of the program is to determine the physical and
chemical characteristics of exhaust particulate matter from internal-
combustion engines as a function of sampling procedure, engine operating
conditions, lead content of the fuel, and the length of residence in the
atmosphere under various conditions.
EXPERIMENTAL APPARATUS AND PROCEDURES
In this project, automobile exhaust particulates are generated
from leaded or unleaded fuels used separately in each of two otherwise
identical automobiles having specially chosen, measured engines selected
for similarity. The auto exhaust is diluted and mixed with air passing
through a 36-foot-long tube (the "dilution tunnel"), then passed into a
2100-cubic-foot residence chamber large enough to minimize wall losses.
Figure 1 is a perspective view of the test apparatus with the test car
positioned on the chassis dynamometer. Exhaust is directed into the dilution
tunnel, from which a portion is taken into the residence chamber.
-------
SAMPLING POINTS:
CHAMBER
TUNNEL-
RESIDENCE CHAMBER
INSTRUMENTATION
u>
FIGURE 1. PERSPECTIVE VIEW OF TEST APPARATUS AND AUXILIARY INSTRUMENTATION
-------
Fuels
Two fuels were purchased from Mobil Research and Development
Corporation in January, 1972. These gasolines were identical except for
lead content. Fuel RE-141B was unleaded and RE-141C had 2.49 g Pb/gal.
Measured characteristics of the two fuels are reported in the Appendix,
Table A-l, pages A-l to A-3.
Project Cars
Two 1970 Fairlane Fords with 1971 engines were prepared last
year for exhaust-gas generation. The engines were 351 CID V-8's with
2-barrel carburetors and automatic transmissions. In order to make
the two automobiles closely comparable, specially matched and measured
1971 engines were made available by the Ford Motor Company and were
installed in the two vehicles by project personnel at Battelle. There
is no significant difference between the two cars, although one is
color coded white (Car W) and the other blue (Car B) for identification.
Only unleaded fuel has been used in the white car. Only leaded fuel
has been used in the blue car, following a 4000-mile conditioning on
unleaded fuel.
Conditioning of Cars
Each car was operated 8,000 miles to stabilize deposits. Be-
ginning at 0 miles, each car was driven on a chassis dynamometer for 4000
miles with unleaded fuel, using a (modified) Durability Driving Schedule .
Periodic measurements of volatile emissions were made during the 4000
miles to compare emission characteristics. Car W continued for another
4000 miles on the modified Durability Driving Schedule with RE-141B unleaded
fuel, but Car B was operated for the second 4000 miles on leaded fuel
RE-141C.
(1) Standard Durability Driving Schedule as described in the Federal Register.
Volume 35, No. 136, Part II, July 15, 1970, without the 70 mph lap.
-------
Test Cycles
A modified Standard Urban Driving Cycle was used during most
of the test runs, in which 35 seconds was added to the standard 20 seconds
of idle after engine start to give a total of 55 seconds idle time before
the first acceleration mode. This cycle is designated the "modified
LA-4" cycle.
In some test runs the standard cycle, with the specified initial
20-second idle, was used. This cycle is designated as the "IA-4" cycle.
Test Run Procedure
In cooperation with the CAPE-19 Project Group the following pre-
conditioning and particulate-emissions test procedure was established:
Prec ond it ioning
(2)
100 miles on modified ' MVMA Durability Schedule
(3)
Two modified EPA (LA-4) Urban Driving Cycles immediately
after the 100-mile durability run
o Record tailpipe CO during EPA cycles
Check tailpipe CO at idle and adjust, if necessary, to
1.0 * 0.2 percent.
a Record tailpipe CO and 0 at idle and 50 mph cruise before
shutdown for overnight soak
Overnight soak period of at least 16 hours at 70 F * 2 F.
(1) Federal Register, Volume 35, No. 136, Part II, July 15, 1970.
(2) Standard Durability Driving Schedule as described in the Federal
Register. Volume 35, No. 136, Part II, July 15, 1970, without the
70 mph lap.
(3) Standard Urban Driving Cycle as described in July 15, 1970 Federal
Register modified as described in Test Cycles above (basically
35 seconds additional time at idle before first acceleration mode).
-------
Test Run
e Actuate throttle, set choke to closed, adjust choke drive
system to starting position
© Simultaneously start particulate-filter sample pump, open
sample-pipe valve to chamber, start tunnel-bag sampling
9 After 15 seconds start car
9 15 seconds after engine start "kick down" off high idle
9 25 seconds after engine start put car in gear
9 35 seconds after engine start, start speed controller tape
(first acceleration mode begins 20 seconds later)
9 5 seconds after full stop at end of last mode of cycle close
sample-pipe valve and stop tunnel-bag and particulate sampling
pumps.
Post-Cycle Operation Check
9 Operate car at idle, 35 mph cruise, and 50 mph cruise.
Record tailpipe CO (Runs 12-1 to 1-15, Appendix, Table A-4) for
evidence of change since preconditioning, and for check on
emissions plus calculation of air/fuel ratio. Supporting
data on tailpipe oxygen, and on C07 and HC in diluted exhaust
of tunnel were found to be less reliable for emissions and
air/fuel ratio determinations, so improved instrumentation
was employed in the next experimental phase.
» Record tailpipe CO, CO 0 , and HC for operation check,
emissions, and air/fuel ratio. (Runs 4-4 to 5-22, Table A-4)
0 Shut off engine.
Engine Air Flow Measurements
A laminar flow meter and a differential pressure transducer were
installed on the unleaded-fuel white car prior to the regular test runs to
measure engine air flow during both the LA-4 cycles and steady cruise
operation under the above operating procedures.
-------
With the laminar flow meter connected to the engine air cleaner
inlet, it was not feasible to preheat the carburetor air as in the normal
configuration. Consequently, carburetor inlet air temperatures were slightly
lower during the air flow measurement runs than during regular test runs.
The effect of this departure from normal engine air flow is negligible.
One run was made to measure engine air flow during a cold-start
LA-4 cycle, and another run to measure engine air flow during a hot-start
IA-4 cycle. The average flow rate during the cold-start run was 28.2 scfm,
and during the hot-start run was 27.2 scfm. These carburetor air rates
are converted to exhaust-gas flow rates, including water vapor, by multi-
plying by a factor of 1.07. Thus, the average exhaust gas flow during the
cold-start modified IA-4 driving cycle was 28.2 x 1.07 = 30.2 scfm. This
value represents a tunnel dilution ratio of 30 for a tunnel flow of 905 cfm.
Measured variations of engine air flow with road speed and of air-fuel ratio
with engine air flow are given in the Appendix, pages A-4 to A-6.
Choke Operation
It was necessary to repeat for each run in a series the same
choke opening schedule because it was found that choke operation influenced
strongly the particle emissions. In addition, slower choke opening than
would normally occur at laboratory ambient temperature was needed in
conjunction with the programmed throttle controller to reduce the chance
of stumble and stall in the first minute or so, which occurred with the
regular vehicle choke. However, decreasing the rate at which the choke
opens increased exhaust emissions, and the extent to which the choke
schedule could be modified was limited by the maximum acceptable HC and
CO composite (tunnel-bag) emissions from the cycle. The tunnel-bag emissions
considered acceptable for the 1971 project cars on the LA-4 driving
schedule were: 4.6 g/mile HC, 47.0 g/mile CO, and 3 g/mile NO. The
small amounts of particulate matter generated in the single modified
or standard LA-4 cycle required that these limits be approached closely
in order to collect weighable amounts. Four different choke schedules
were used in the course of the experimental program to accommodate
operating requirements and other limitations. Details of the experimen-
tal development of 'acceptable choke operation are given in the Appendix,
pages A-10 to A-17.
-------
Dilution Tunnel
A dilution tunnel simulates the real environment by inducing
rapid dilution of the exhaust as it issues from the tailpipe. Mixing with
constant proportions of filtered atmospheric air under reproducible experi-
mental conditions permits accurate sampling of the dilute mixture (usually
diluted 30:1). The dilution tunnel used in these studies was constructed
according to the design developed at the Du Pont Petroleum Laboratory .
The dilution tunnel residence chamber, and sampling apparatus are shown in
Figure 1, page 3. Details of configuration, operation, and arrangement of
auxiliary apparatus are given in the Appendix, pages A-18 to A-32.
Residence Chamber
A rectangular residence chamber with flexible walls of black
polyethylene sheeting on a supporting frame was used to contain a composite
sample of the diluted exhaust taken from the tunnel and further diluted by
filtered air in the chamber. The overall dilution of the exhaust by tunnel
and chamber air was nominally 300:1, and the volume of the twice diluted
sample was about 2100 cu ft. Details of the sampling probes, procedures for
mixing and sampling, and measurement of dilution ratio in each run are
presented in the Appendix, pages A-32 to A-42.
Instrumentation
A critically important part of the facility are the instruments
for indicating and recording the composition of samples of exhaust and
diluted exhaust, the ambient conditions, the operation of the cars, and
the characteristics of the particles suspended in the exhaust gases. All
instruments were selected with the objective of providing reliable and
accurate analyses, checked periodically, given maintenance service as
necessary, calibrated before each run, and then checked after each run to
(1) Habibi, Kamran, "Characterization of Particulate Lead in Vehicle
Exhaust-Experimental Techniques", Environmental Science and Tech-
nology, Vol. 4, No. 3, March 1970, pp 239-253.
-------
insure against unacceptable performance. Measurement techniques were
evaluated and improved when possible throughout the experimental program.
The carbon monoxide meter indicated carbon monoxide concentrations lower
than the true values for Runs 4-11 to 4-26, (Table A-4), so that the cars
Lad CO emissions above the limits specified for the modified LA-4
cycle. Experimental data in this report give the true values for carbon
monoxide concentration, corrected to compensate for the instrument: error. The
series of runs was repeated under conditions to give acceptable CO emissions,
so that two sets of valid data were obtained for interpretation.
Three different oxygen meters of one type were used in the first
half of the test program with recurring difficulties in achieving reliable
operation and rapid response. After investigation a different type was
substituted and used with completely satisfactory results.
In ten runs the size distribution of the aerosol particles in
diluted exhaust in the residence chamber was measured with the Minnesota
Aerosol Analyzing System (MAAS), a group of prototype instruments which
quantitatively characterize aerosol particles in submicron size ranges
and at low concentrations not accessible to other quantitative methods of
measurements. Appropriate calibration procedures to standardize the data
have not yet been worked out, so that the results given in this report are
interpreted in relative terms to indicate the changes that occur in size
distribution during aging of an aerosol by interaction of the particles with
each other and with the walls of the residence chamber.
Two speed controllers were used during the experimental runs for
automatic control of the driving cycle after the car was placed in gear.
The first instrument was programmed by a punched paper tape, which was less
precise than the magnetic tape program used in the later instrument, which
resulted in a marked improvement in reliability and reproducibility.
Specific details of the instruments used and the calibration pro-
cedures are given in the Appendix, pages A-42 to A-47.
Filtration and Weighing
The mass concentration of solid particles in auto exhaust was
measured by passing a measured volume of the diluted exhaust from the tunnel
-------
10
through an absolute filter. If the filter retains only solid particles on
its upstream face and passes all gaseous components, the increase in
weight, AW, of the filter is then the weight of suspended particles collected
from the filtered sample. AW is used to calculate the emission rate, R, of
the car, as follows:
R = AW x V(tunnel)
D(IA-4) V (sample)
where R is the rate of solid emissions, g/mile
AW is the weight gain of the filter when a
constant sample stream is maintained during
one LA-4 cycle
D is the distance driven during one IA-4 cycle, 7.5 miles
V(sample) is the rate of flow through the filter
of a sample stream from the tunnel, cu ft/min
V(tunnel) is the rate of flow of diluted exhaust in
the dilution tunnel, cu ft/min.
Ideally, sampling for particulates should be done isokinetically,
i.e., the flow velocity entering the sample probe should be equal to the
flow velocity in the tunnel. To accomplish this, the ratio rr? ;r-
J v ' V(sample)
should equal the ratio of the cross-sectional areas of the tunnel and
the sample probe. However, when dealing with particulates in the sub-
micron range such as normally found in auto exhaust, large deviations
from isokinetic sampling can be tolerated (such as 10 to 1) without
appreciable effect on the collection efficiency^ ' . The deviation
from isokinetic sampling, expressed as a ratio (Sampling Ratio) is given
by the following equation:
(1) "Subisokinetic Sampling of Particles in an Air Stream", G. A. Sehmel,
AEC Contract No. AT(45-1)-1830, Battelle-Northwest, March 1, 1966.
-------
11
., . . Actual Sample Velocity
Sampling Ratio = Isokinetic Sample Velocity =
V(sample) Area (tunnel)
V(tunnel) Area (sample probe)
The sampling ratios actually used in this program were 1.3 for the
tunnel-to-chamber sampling, and 2.2 and 6.5 for the particu'late sampling
from the tunnel.
Experience in this study and discussions with other investigators
have indicated that the weight change of a filter after filtration of dilute
samples of auto exhaust can be affected by several experimental variables,
such as the size of the sample line, the temperature of the sample stream
at the filter, the face velocity of the sample stream, the size and type of
filter, and the gross weight of filter relative to the weight of solid par-
ticles. Thus the real system used in these studies has required a major
experimental study to develop and specify procedures that give a true value
of AW.
This study has necessarily dealt with very small weights of
collected particulate matter because of relatively large dilution ratios of
about 30:1 in the tunnel and about 300:1 in the residence chamber. The size
of the samples has been limited by controlling the flow rate through the
filter to avoid excessive pressure drop, and by the short sampling time of
23 minutes during one LA-4 cycle. The weights of particles collected have
ranged from a few milligrams down to less than a hundred micrograms. Par-
ticles were collected on two sizes of filters which weighed about 500 rag
in 142-mm diameter, and about 70 mg in 47-mm diameter. A standardized
technique had to be developed to measure weight gains that truly represented
the amounts of solid particulate matter removed from the filtered air.
The following requirements were included in the standard technique after
development studies were completed.
Microbalance with a sensitivity of better than 2 ng
o Control of atmospheric temperature and humidity in
the weighing area
-------
12
» Selection of abrasion resistant filter material
Checks on operator technique by reweighing items of
constant mass
9 Discharge of static electricity from filters before
weighing
e Duplicate simultaneous samples to detect and eliminate
uncontrolled variables
9 Parallel samples with two different membrane filter
materials to detect any effects of filter medium
composition
a A backup filter held in the same filter holder in
contact with the primary filter and weighed after
exposure to the same sample conditions, but without
collection of the solid particles retained by the face
of the primary filter. The weight change of the back-
up filter is applied as a correction to the weight
change of the primary filter.
The use of a backup filter is an important feature of the pro-
cedure used in these CRC studies. The data that support its use are reported
in the experimental section. Details of application of other parts of the
standard technique itemized above are given in the Appendix, pages A-47 to
A-51.
EXPERIMENTAL RESULTS
Exhaust Emission Rates
The results of two series of runs are summarized in Table 1.
These were carried out to examine reproducibility of the emissions with un-
leaded and leaded fuels. In the first series (Runs 4-11 to 4-26) a low
reading caused by a malfunction in the electronic circuit of the CO analyzer
permitted the actual gaseous emission levels of CO to be unintentionally
higher than the limits specified for the LA-4 cycle of 47.0 g/mi CO.
All data taken while the CO analyzer was malfunctioning have been corrected
in Table 1 to give the real values during the tests.
-------
TABLE 1. SUMMARY^F TEST RESULTS
Run
Date
4-4
4-6
4-7
4-9
4-10
4-11
4-12
4-13
4-16
4-17
4-18
4-19
4-24
4-20
4-25
4-26
5-9
5-10
5-11
5-14
5-15
5-16
5-17
5-18
5-19
5-21
5-22
Car
Fuel
Cycle
Test Conditions
Car Fuel
W
W
W
W
W
W
W
W
W
W
B
B
B
B
B
B
B
B
B
W
W
W
W
W
W
B
B
W -
NL -
LA-4
NL
NL
NL
NL
NL
NL
NL
NL
NL
NL
L
L
L
L
L
L
L
L
L
NL
NL
NL
NL
NL
NL
L
L
White
Cycle
LA-4 (M)
IA-4 (M)
IA-4 (M)
IA-4 (M)
IA-4 (M)
IA-4 (M)
IA-4 (M)
IA-4 (M)
IA-4 (M)
IA-4
IA-4 (M)
IA-4 (M)
IA-4 (M)
IA-4 (M)
IA-4 (M)
IA-4 (M)
IA-4 (M)
IA-4 (M)
IA-4 (M)
IA-4 (M)
IA-4 (M)
IA-4 (tf)
IA-4 (M)
IA-4
IA-4
IA-4 (M)
IA-4 (M)
B -
Nonleaded L -
(M)
Schedule
I
I
I
I
I
I
I
I
I
I-S
I
I
I
I
I
I
II
II
II
II
II
II
II
II-S
II-S
II
II
Blue
Leaded
General Comments
Checkout
Checkout
Checkout
Checkout
Checkout
High HC, CO
High HC, CO
High HC, CO
High HC, CO
Std. Cycle
Checkout only
Abort >d
Aborced
High HC, CO
High HC, CO
High HC, CO
Baseline
Baseline
Aborted
Aborted
Baseline
Basel-.ne
Basel-. ne
Std. C/cle
Std. Cycle
Basel .ne
Basel-Lie
KEY
Emissions, e/mi
Part.
--
0.192
0.173
0.195
0.189
0.041
__
--
--
0.319
0.349
0.340
0.135
0.153
__
0.040
0.046
0.044
0.075
0.071
0.211
0.203
HC
--
5.81
5.96
6.13
6.22
3.12
_..
6.07
5.94
5.76
3.91
5.35
.._
3.43
3.56
3.99
4.66
4.12
3.37
3.46
Particular +
Cycle with 55 -second initial idle
CO
--
68.8
73.9
72.5
77.7
58.6
__
--
71.9
68.2
75.8
34.1
47.4
__
35.5
33.2
38.5
52.4
46.8
38.2
44.1
Based on
weights
Sampling
f t%e+ ff»t* a T
NOV
--
2.46
2.53
2.88
2.91
2.72
__
1.99
2.06
1.90
2.60
2.21
_-
2.35
2.48
2.50
2.47
2.49
2.19
1.93
Metric'
only,
Ratio
wo t r» r>
Schedule
LA-4
Cycle with 20-second initial idle
Choke plate opening schedules defined
on Pages A-14 to A-17
isokinetic velocity)"
2.2
NO -H- Not corrected for humidity
-------
14
Rather than use only the data with relatively high gaseous emission
levels as baseline data for particulate emissions, new Choke Schedules II
and II-S described in the Appendix (pages A-16 to A-17) were developed to
lower gaseous emissions to acceptable levels. The baseline runs were
repeated with modified LA-4(M) standard LA-4 cycles (Runs 5-9 through 5-22).
These included a consecutive pair of baseline runs with the modified IA-4 (M)
cycle on leaded fuel, a consecutive series with the modified cycle on un-
leaded fuel, and a consecutive pair using the standard IA-4 cycle with un-
leaded fuel.
Reproducibility of Mass of Particulate Emissions
Samples of diluted exhaust were taken from the downstream end of
the tunnel and from the residence chamber during each of the ten Runs 4-10
to 4-26. Particles filtered from the samples were collected on two types
of membrane filters. All samples used two filters assembled together in
the same filter holder as a primary and a backup disk, according to the
procedure described previously, to distinguish adsorbable vapors in the same
stream. Each type of filter was used in duplicate filter holders and fil-
ter assemblies, so that four sample streams were filtered altogether.
Table 2 shows the mass of particles collected from the tunnel
for each of the valid baseline Runs 4-11 to 4-26, and for each of the filter
types to permit comparison with each other and with the gaseous emissions
in the exhaust. Mean values and standard deviations are given for each
filter type and each fuel. These comparisons show good agreement between
filters and good reproducibility of particle emissions in repeated runs with
the modified cycle,, Emissions from the single run with the standard cycle
are significantly lower than the emissions from the modified cycle.
Table 3 presents emissions data of Table 2 recalculated as mass
concentration of particles, and includes the mass concentrations of par-
ticles sampled from the residence chamber immediately after a tenfold
dilution and mixing of the tunnel sample, and after aging of the mixture
for about 3-1/2 hours. The means and standard deviations, and coefficients
of variation (CV, 7<>) of each group of samples have been calculated for the
modified LA-4 cycle operated with unleaded and lead fuel, using the two
-------
15
TABLE 2. MASS EMISSIONS OF PARTICLES COLLECTED
FROM TUNNEL BY FILTRATION
(Composite samples, dilution ratio, 30,
one cycle, cold start)
Emissions, g/mi
Particles on
Filter Type
Run No. Fuel Cycle
4/11 RE-141B Modified
Unleaded IA-4
4/12 Ditto Ditto
4/13
4/16 " "
Mean
4/17 " Standard
LA-4
4/20 RE-141C Modified
Leaded IA-4
4/25
4/26
Mean
MET(a)
.192
.173
.195
.189
0.188 * .009
.041
.319
.349
.340
0.337 * .019
MIL(b^ CO HC
.204 68.8 5.81
.178 73.9 5.96
.214 72.5 6.13
.208 77.7 6.22
.201 * 0.014
.043 58.6 3.12
.285 71.9 6.07
.360 68.2 5.94
.386 75.8 5.76
.344 * 0.044
NO
2.46
2.53
2.88
2.91
2.72*
1.99
2.06
1.90
(a) Metricel Filter DM450, Part No. 64519, Batch 80557, 142-mm, 0.45//m pore
Flow 4.29± .10 cfm (Isokinetic flow rate 1.95 cfra).
(b) Millipore Filter AAWP 047 00 Lot 66499 15, 47 mm, 0.8 X-'m pore.
Flow 0.70 cfm (Isokinetic flow rate 0.11 cfm). '
-------
TABLE 3. MASS CONCENTRATIONS OF PARTICLES COLLECTED FROM
TUNNEL AND RESIDENCE CHAMBER
Run
4/11
4/12
4/13
4/16
Mean
CV, pc
4/17
4/20
4/25
4/26
Mean
CV, pc
Variables
Fuel Cycle
Clear Modified
141B
Clear Modified
141B
Clear Modified
141B
Clear Modified
141B
Clear Standard
141B
Lead Modified
141C
Lead Modified
141X3
Lead Modified
141C
Net
Tunnel
Composite
MET
2458
2208
2500
2419
2389 ± 145
6.0
526 ± 11
4090
4460
4362
4304 ± 85
2.0
MIL
2600
2297
2753
2680
2582 ± 167
6.5
550 ± 28
3673
4638
4961
4432 ± 586
11.
^\
Particulate Mass, g/m
Residence Chamber
Initial
MET
215
409
302
227
255 ± 113
44.
98 ± 16
390
448
392
415 ± 34
8.
MIL
317
197
394
266
293 ± 71
28.
386
481
368
409 ± 66
16.
Aged 3
MET
356
182
248
224
247 ± 74
30.
79 ± 6
304
348
313
324 ± 24
8.
-1/2 hr
MIL
__
340
336
249
308 ± 64
21.
--
313
367
436
372 i 108
29.
-------
17
types of filter paper. All entries in Table 3 are the mean of the net
weight increase of duplicate simultaneous samples, each of which used two
filters in a single holder. The weight change of the backup filter was
applied as a correction to the particle weight collected on the primary
filter, as described above (page 11).
These experimental data demonstrate the following characteristics
of the specified method of filtration:
9 The coefficient of variation (CV) of the mean
particle mass from the tunnel was about 6 percent
for both unleaded and leaded fuels,
9 There is no significant difference between the
two comparable means of the masses on the two types of
filter paper from both the tunnel and residence
chamber, even though the filter composition, area, face
velocity, and total sample weight and volume for
the two filter systems differed markedly.
The backup filter provides a correction that
eliminates uncontrolled variability related to face
velocity at the primary filter, gaseous composition of
the exhaust, and equilibration of the filter samples
before weighings.
The coefficients of variation of the residence
chamber samples were higher than those of
samples from the tunnel, because of the much
smaller sample from the residence chamber.
« The mean values of the initial residence chamber
samples were each in the expected range of 10
percent of the corresponding tunnel sample,
agreeing with the tenfold dilution of the
tunnel sample determined by gas analyses for
CO and hydrocarbons.
-------
18
Correlation of Light Scattering with Mass
Light scattered by the aerosol in the residence chamber was
measured with an integrating nephelometer initially and after aging of the
particles to determine the correlation of the light scattering coefficient
with the mass of the suspended particles. Table 4 gives the light scattering
coefficients and the mean mass concentrations of samples weighed on two
types of filters for each of Runs 4-11 to 5-22. Figure 2 gives a plot of
3 4
the straight line MASS in u-g/m = 65.5 (10 bt - 0.4) that was calculated
for minimum standard deviation (a = 0.59) of residuals of all points. The
intercept at b = 0.4 is the measured background scattering of filtered
scat
air in the chamber. Correlation coefficient R of MASS with ^scat is 0.91.
There is no apparent systematic influence of particles from either fuel
type or aging -because points in all categories fall around the regressed line,
The correlation of auto exhaust light scatter measured in this study
(1) 4
differs substantially from the Charlson correlation MASS = 38 x 10
b for atmospheric dust, which is also shown in Figure 2. Charlson has
scat
found no deviations from his empirical experimental correlation, but he
points out that it "may or may not hold in all cases", This is an instance
in which either the light scattering of auto exhaust is less than for
atmospheric dust, or the mass measured by filtration of exhaust according
to the methods described above is greater than would be found by filtra-
tion of atmospheric dust with the same scattering coefficient. More
investigation is necessary, but the empirical relation for auto exhaust
may be useful before an explanation is found.
Addition of Foreign Materials to the
Residence Chamber
Following the development of procedures and some shakedown runs,
Runs 12-6 to 1-15 were carried out as summarized in Table 5 to investi-
gate the effects of humidity in the chamber, lead in the fuel, dust and
(1) Charlson, R. J., Ahlquist, N. C., Selvidge, H. and MacCready, P. B., Jr.,
Journal Air Pollution Control Assoc., Vol. 19, pp 937-942 (1969).
-------
19
TABLE 4. CORRELATION OF LIGHT SCATTERING WITH AEROSOL
MASS IN THE RESIDENCE CHAMBER
Run
4-11
4-12
4-13
4-16
4-17
4-20
4-25
4-26
5-15
5-16
5-17
5-18
5-19
5-9
5-10
5-21
5-22
Fuel
Clear
141-B
Ditto
M
ii
"
Lead
141-C
Ditto
11
Clear
141-B
Ditto
ii
it
it
Lead
141-C
Ditto
11
M
LA-4
Cycle
Modified
Ditto
ti
11
Standard
Modified
Ditto
11
Modified
Ditto
it
Standard
Ditto
Modified
Ditto
"
11
h -4 -1
scat (10 m )
Initial 3-1/2 hrs
5.0
4.6
5.5 1.8
5.5
1.5
6.0
7.2 6.9
6.4 6.2
1.4 1.1
1.3 1.4
1.2 1.2
1.9 2.1
2.1 1.7
1.8 2.0
2.8 3.2
2.4 2.2
2.6 2.7
Mass, y,
Initial
266
306
348
247
98
388
465
380
84
70
50
95
120
138
173
197
233
g/m
3-1/2 .hrs
--
--
292
--
--
--
358
375
--
49
19
56
98
133
--
145
135
Nominal sample flow rate from tunnel to chamber = 10.0 cfm, isokinetic flow
rate = 7.5 cfm.
-------
20
o
o
c
0)
'o
o>
o
O
o>
t
o
o
CO
100
200
300
400
500
600
Mass Concentration, /xg/nrr
FIGURE 2. CORRELATION BETWEEN LIGHT SCATTERING
AND AEROSOL MASS CONCENTRATION
-------
TABLE 5. STUDIES OF ENVIRONMENTAL VARIABLES
Run
Tvst Conditions
D«-.{» Car
12-6
12-13
12-15
12-18
12-19
12-20
12-21
12-22
!-]7
1-19
!-23
12-27
12-28
12-29
J-3
1-5
1-10
1-11
1-12
1-15
W
W
W
W
W
W
W
W
W
W
W
B
B
B
0
B
B
B
U
B
Fuel Cycle Sc
XL
Nl.
VI.
Nl.
NL
NL
NL
Nl.
NL
Nl.
NL
1.
L
1.
L
!_
L
L
L
L
I.A-'UM)
I./1 -4 'M)
f.A--; (M)
LA-4 (\'l
I.A-4 (M)
LA--I IM)
LA-.J(M)
I.A-t (M)
I.A-'I(M)
l.A-4 (V.)
I.A--KM)
f.A-4(M)
LA-4(M)
LA-4 (M)
l,A-t(M)
I.A--1 (M)
LA-MM)
LA-! (M)
l.A-1 (M)
LA-4 (M)
Module c:on!ro!lct MAAS
UD DPC
uc :>if:
(JD Dfc:
L'D WC.
\ WO
! IJpiJ
i nn:
I DPC
t Rl
'. IU
1 Kl
t DfC
I I>PC
i DPC
I Rl
1 IU
t m
i m
1 Rl
I Rl
Yes
No
Yes
Ycv
No
No
No
Yes
Wo
No
Yes
Yes
Yes
Yes
No
No
No
No
Yes
Yes
C.enetal Comments Hum.
0
So.ik ;eoip low - Nllj erratic O
O
So.ik remp lew O
CS c'lnnxc.-! - low CO lOle O
O
A
A
O
A
O
Various problems A
R'jn-ln mileage low O
O
Run length low. Idle CO low A
Soak time short A
Soak short, soak temp low A
Idle CO lilg'.i, soak temp low O
Soak temp low O
A
Variables Emissions. >;/mi
Lead
0
O
O
0
O
O
O
O
O
O
0
3
B
B
B
B
B
B
B
B
Dust NU3 Part* IIC
O O 0.011 1.08
O D 0.011 1.70
O O O.OM 2.06
O D 'O.OIR 1.63
O D O.OO'.l 1.73
C O Dust 1.70
O D 0.014 1.89
C O Dust 1.84
O D 0.038 7.69
C 1) Dust 1.0:i
C D Dus: 1.40
C O Uust
C O Dust 5. US
O D 0.255 6.05
C D PuM 4.?2
O O O.iMl 7.'J3
O D 0.2C» 6.33
C D Dust 3.34
O O 0. 1G2 4.04
O O 0. IC4 5.3U
CO
.1.!.
2r».
"2.
27.
2-1.
23.
36.
33.
32.
31.
2f.
--
58.
7).
3-1.
53.
17.
33.
41.
3J.
3
8
I
ft
3
9
n
0
C
7
2
1
6
0
1
7
4
1
0
NO
C.H-
2.7S
2.4P
2.S2
3.23
3. 12
2.25
1.71
:.45
1 . ;i '.
1.30
2.13
2.17
--
C. 53
--
2. i:
l.'-'ii
1.19
1.3S
At this point the cars were leak checked and a small leak waj found In the white car and a very small
Car
Fuel
Cycle
Choke
leak
In
Schedule
NO
the blue car.
W - White
NL - Nonleaded
LA-4(M)
UD
I
Nitrogen oxide
B - Blue
L - Leaded
Cycle with
55 -second
Controller
initial idle Humidity
Partlculate
DPC
Rl
(SI)
+
Under development
Defined In
- Values not
text
corrected
Variable*
for humidity
0
A
B
C
D
Magnetic tape
Paper tape
Steam injection
Based on Metricel weights
Sampling Ratio "2.2
Low or normal values
High humidity
High lead
High dust (solid impurity
In atmosphere)
High ammonia (gaseous
Impurity In atmosphere)
f
-------
22
ammonia in the atmosphere, on the nature and fate of particles from the
two matched cars. Residence time in the chamber was an independent
parameter which was varied from about 1 to 24 hours.
Because the weighing procedure in the early stages was not as
precise, and because of some operating malfunctions, the data are useful
and valid only for interpretations given in following sections; broader
interpretations should not be attempted. Subsequent runs with more
definitive results have already been described in Table 1.
Size Distribution of Auto Exhaust Particles
with the Cascade Impactor
No significant data on cumulative mass as a function of particle
diameter were obtained when particles carried in the exhaust were separated
into aerodynamic equivalent diameter fractions by passing a sample stream
through six stages of the Battelle cascade impactor backed up by an ab-
solute filter. The aerosol suspensions were too dilute and the time avail-
able for sampling was too short to build up weighable amounts of each
fraction on individual impact stages. All stages gained'weights in the
range of a few to tens of micrograms, presumably because there was ad-
sorption of volatile material as well as collection of solids by impact.
Electron micrographs of collections on the impactor surfaces showed the
particles were being classified by size. The morphology of particles in
some of the size fractions is illustrated in a subsequent section.
The absolute filter, which retained unfractionated undersize with
an equivalent diameter less than 0.25 M-m, collected amounts of material
weighing several hundred micrograms. Thus the predominant particles are
in the size range below 0.25 p,m.
Size Distribution of Particles with the
Minnesota Aerosol Analyzing System
The Minnesota Aerosol-Analyzing System (MAAS) was used in ten
tests, five with unleaded fuel and five with leaded fuel, as shown in
Table 6. Size distributions indicated by MAAS for one selected run with
each fuel are plotted in Figures 3 and 4.
-------
TABLE 6. COMPARISON OF AUTOMOBILE PARTICUIATE DATA
Filtration
Mass Concentration,
Test No.
11-29
12-1
12-4
12-6
12-15
12-13
12-18
12-19
1-17
12-21
12-22
1-23
12-20
1-19
1-5
1-15
12-29
1-12
1-10
12-28
1-3
12-27
1-11
Foreign
Material
(a)
(a)
(ae)
(o.)
(o)
(d)
(d)
(d)
(ad)
(ac)
(cd)
(ce)
(acde)
(ab)
(ab)
(bd)
(be)
(abde)
(be)
(abed)
(abce)
(bcde)
Tunnel*
480
230
172
290
176
137
169
121
492
183
3360
1869
1645
1608
3088
2100
3260
2070
3450
6194
4780
20,240
4818
Residence
Initial**
72.5
42
27
37
38
24
10
91
25
152
-
94
233
120
271
186
177
331
246
1066
221
U8/m
Chamber
4 Hours
50
-
18
40
33
6
37
40
14
81
-
85
236
145
213
135
145
222
267
Light Scattering
b -A W1
scat (10 m ),
Initial 4
fm
2.6
2.9
0.82
0.95
0.70
0.65
0.52
0.65
0.65
3.8
3.6
2.5
3.3
5.7
2.4
5.5
1.9
4.7
7.4
5.0
Hours
^
1.7
2.6
0.85
1.0
0.80
1.10
1.15
0.65
0.75
2.6
2.7
2.2
2.3
5.1
2.5
4.6
1.8
4.5
5.4
3.7
16.8 13.6
238
(a) High humidity.
(b) Leaded
(c) Arizona
(d) Ammonia
gasoline.
road dust
4.25
+ Sampling
++ Sampling
3.3
Ratio
Ratio
MASS Instruments
Particulate
Initial
61
386
-
36.4
-
19.7
-
-
-
1152
-
-
-
_
262
697
264
_
1228
-
2786
^
= 2.2
= 1.3
2
Vol. UJP /cc
4 Hours
»
30
24i
-
35.9
-
68.5
-
-
-
356
-
-
-
_
163
422
202
_
476
-
1307
*
contaminant.
gas contaminant.
(e) Long residence time.
(o) Represents all variables at
low level.
NJ
CO
-------
u
o
s
:*.
c
o
.Q
t_
to
<5
O
Test 12/15/72
i i i i i : : i
Unleaded Fuel
10
0
0.001
0.01
10.0
Particle Diameter, p.rr\
FIGURE 3. VOLUME DISTRIBUTION)? UNLEADED PARTICLES
-------
o
o
-------
26
Figure 3 (Run 12-15) shows the results obtained with unleaded
automobile exhaust with no addition of any contaminant or humidity. The
distribution changed little in 3-1/2 hours except for the removal of the
very large particles and some growth in the small particles. Figure 4
(1-12) similarly shows some slight changes in 3-1/2 hours for the size
distribution of leaded exhaust particles. The curves are similar in
shape but differ greatly in the total volume of the two aerosols, since
the ordinate scales differ tenfold. These are two single runs made
with different choke schedules, so that the emissions found are not
comparable as a typical difference in emissions from unleaded and leaded
fuels. Multiple runs in each car under entirely similar controlled
conditions must be made to obtain a significant comparison. The relative
measurements of particles from unleaded and leaded fuels found by filtra-
tion and by MAAS instruments are sufficiently similar to encourage the
use of both methods in detailed studies of the amounts and properties
of particles.
Correlation of Particulate Mass with
Hydrocarbon Emissions
A correlation was found between the mass rate of particle
emissions and total hydrocarbon emissions, which is plotted as the curve
in Figure 5 for unleaded fuel. The plot includes all of the data from
all experiments (except dust injection runs), and one datum from an API
study at Battelle with a so-called "warm start". The correlation is
truly remarkable for unleaded fuel considering the many other parameters
being varied. The leaded data do not correlate as well, particularly
the three high humidity runs marked B, but the trend appears similar to
that for the unleaded curve. The particulate emissions from the leaded
fuel are from 0.05 to 0.15 g/mile above the nonleaded curve, or from
60 to 100 percent higher when compared at the same hydrocarbon emission
level.
The reproducibility of the measurements of emissions in six
sets of consecutive replicate runs was ±6.4 percent for particulate
emissions and ±10.0 percent for gaseous emissions, expressed as the
-------
27
o Modified cycle, unleaded fuel
A Standard cycle, "
a API,
Modified cycle, leaded fuel
I, IE Choke schedules
345
HC Emissions, g/mi
FIGURE 5. CORRELATION BETWEEN PARTICLE EMISSIONS
AND HYDROCARBON EMISSIONS
-------
28
unweighted mean of the percentage standard deviations of each set. These
include all of the data on emissions in Table 1, page 13, except the
single run on 4-17. Reproducibility after a longer period, such as
10 days between the duplicate pairs of Runs 5-9 and 5-10 and Runs 5-21
and 5-22, was not as good. This is the only such set of two pairs
available for comparison, which precludes interpretation and definite
conclusions.
Morphological Characteristics of Particles
Exhaust particulates were collected by means of the cascade
impactor directly onto electron microscope carbon support films on
electron microscope grids. Without further specimen preparation the
particles were examined in the transmission electron microscope.
In general, the same types of particles are seen, independently
of the fuel used and the conditions of generation; agglomerates and
aggregates are comprised of carbon black type material, condensate,
and spherical fly-ash-like particulate matter. Figure 6 shows typical
unleaded exhaust particles; there are several tar droplet particles
which have low enough vapor pressure to persist in the vacuum system of
tine electron microscope under bombardment by the electron beam. Also,
there is usually present a carbon-black-type structure of a chain-like
aggregate, and fairly electron-dense particles which are cemented
together with less dense material, which appears to have condensed
around them.
Figure 7 shows an electron micrograph of leaded auto exhaust.
Carbon black and fly ash are cemented into aggregates by material which
appears to have condensed around them. In some of the small aggregates,
highly electron-dense structures are in evidence and probably are lead-
rich.
Figure 8 is a group of four electron micrographs of unleaded par-
ticles from the residence chamber collected by the impactor in the two
smallest stages, initially and after four hours of aging. The 0.25 um
stage has brushy and chain-like aggregates and many very small individual
particles. After aging many more aggregates of a few particles each are
visible. The initial sample at 0.5 M.m is mostly spherical structures,
several aggregated from three or four units.
-------
29
J22216 10,OOOX
FIGURE 6. AN ELECTRON MICROGRAPH OF UNLEADED EXHAUST
PARTICULATE COI.LKCTED AT TUM 0.5-Um 1MPACTOR
STACF, AFTKK 4 HOURS RFSIDMNCE (Run 4-17)
J22219 10.000X
FIGURE 7, AN ELECTRON MICROGRAPH OF LEADED EXHAUST
PARTICULATE COLLECTED AT THE 0.5-p.m 1MFACTOR
STAGE AFTER 4 HOURS RESIDENCE (Run 4-25)
-------
30
O.Z5>UM COLLECTION
0-HOUK5
COLLECTION
0-HOURS
\.0/U,M
. 1
/£5>UM COLLECTION
JZttT!
4-HOURS
0.5MM COLLECTION
4-MOORS
FIGURE 8. ELECTRON MICROGRAPHS SHOWING NONLEADED
EXHAUST PARTICULATE FROM RUN 12-15
-------
31
Figure 9 is another group of four views of particles from leaded
fuel, not significantly different in shape and degree of aggregation from
the unleaded particles of Figure 8.
Figure 10 shows four individual views of particles from the
residence chamber with foreign materials also present. The dust con-
taminant added in Run 12-20 with unleaded gasoline persists for 20 hours
of aging, and exhaust particles are not distinguishable from the dust.
The presence of alkaline ammonia in the diluted exhaust of Run 12-13 from
unleaded gasoline causes the growth of needles. A condensate layer appears
to enclose aggregates of aged particles in the high humidity chamber at-
mosphere.
CONCLUSIONS
Emission Levels
(1) Particulate emission levels correlated very strongly
with HC (and CO) emissions for the two fuels. For unleaded fuels, the
level of hydrocarbon emissions appears to be the major significant
factor correlating positively'with particulate emissions.
(2) "The same choke schedule resulted in significantly different
HC (and CO) emission levels for the two fuels.
(3) The same choke schedule resulted in significantly
different weights of particulate emissions for the two fuels.
(4) For a given HC emission leaded fuel always gave higher
weights of particulate emissions. For roughly equivalent gaseous
emissions, the particulates from the leaded fuel are about 50 to 100
percent heavier than from the unleaded fuel.
(5) The reproducibility of the emission data on consecutive
run days was a few percent.
(6) The determination of the particulate loading of the auto-
mobile exhaust is heavily dependent upon the characteristics of the
sampling system and, in order to correlate results from different labora-
tories, there should not be any major differences in the filter media or
sampling procedure.
-------
32
0-HOURS
J2I834
COlLECnON
? .:*?
tt-tWURS
*,
. *
r ;
COLLECTION
0-HOOR5
J21644
O.S^UM COLLECT10M
J27000
05XIM COUeCTlON
KEY
RUN
12-28
RUN
1-5
RUN
1-15
RUN
1-15
FIGURE 9, ELECTRON MICROGRAPHS SHOWING LEADED EXHAUST
PARTICULATE FROM SELECTED EXPERIMENTS
-------
LEADED GAGOUNt
tO>UM COLUCTKJM
NON-LEADED GASOUME W HOURS
0.5,14 M COLLECTION DUST
I 0/U.M
NON-LEADED GASOLINE
COLLECTION
.
0-MOUR9
AMMONIA
J21I99
O-HOURS
AMMONIA
AMOURS
HUMIDITY
KEY
RUN
12-20
RUN
12-13
RUN
12-29
RUN
1-15
FIGURE 10. ELECTRON MICROGRAPHS SHOWING EFFECTS OF DUST,
AMMONIA, AND HUMIDITY FROM SELECTED EXPERIMENTS
-------
34
Particulate Characteristics
(1) Both the cascade impactor and MAAS (Minnesota Aerosol
Analyzing System) indicated little difference in the particle size of
the particulates from leaded or nonleaded gasoline. For the most part,
the mass-median diameter of the dilute exhaust measured at the tunnel
(30:1) was less than one-quarter micron. On the basis of the Minnesota
System, the particle size distribution appeared to be bimodal.
(2) High humidity in the chamber produced no apparent effect
on the size of the leaded particles; however, there was a considerable
increase in the size of the unleaded particles.
(3) Ammonia caused no apparent increase in the particle size
of the leaded particulates; however, there was a considerable increase
in size of the unleaded particulates. Ammonia was the only additive to
cause an increase in the particulate mass in the chamber as a function
of time, and this effect occurred only for the unleaded particulates.
(4) The integrating nephelometer data for light scattering by-
the auto exhaust correlates directly with the mass data obtained by
filtration. However, the slope of the correlation curve for auto exhaust
is greater than the slope of the correlation curve proposed by Charlaon
for atmospheric dust.
Morphological Analysis
Four mechanisms of particle growth, listed in order of prevalence,
were identified based on morphological studies of electron micrographs of
the particles from both unleaded and leaded fuels.
First. Agglomeration
Almost every exhaust particle seen in the electron
microscope can be classified as an agglomerate.
Second. Condensation
Many of the particles are surrounded by low density
envelopes of what appears to be a condensate.
-------
35
Third. Crystal Growth
Sometimes crystalline particles are detected, part
of which are associated with oil or tar droplets.
Fourth. Ablation of Deposits
Particles larger than one micron and irregular in
shape are likely to be chunks of material broken
off exhaust system deposits.
The significant differences in particles from the two fuels
were that the leaded particles frequently had an electron-opaque core,
which undoubtedly represented a lead-rich nucleus; that agglomerates
formed as a function of residence time; and that ammonia addition to the
suspension of unleaded particles drastically affected their structure.
The particles produced during runs in which ammonia was added to the
dilution tunnel were highly irregular acicular structures, while in
all other runs the particulates were generally spherical agglomerates.
-------
APPENDIX
-------
A-l
Fuels
TABLE A-l. CHARACTERISTICS OF FUELS
(a)
Properties
Blend Designation
RE -14 IB
unleaded
RE-141C
Leaded
Research Octane No. (RON)
Motor Octane No. (MON)
Vapor Pressure, Reid, Micro (D-2551)
TEL as Lead, ppm (M-1059)
TEL as grams Pb/gal. (M-951)
Sulfur, % wt
Chlorine, ppm (M-600)
Phosphorus, ppm (M-798)
Nitrogen, ppm (M-1042, Col.)
API Gravity (D-287)
ASTM Distillation, °F
Initial Boiling Point
5% Distilled
10% Ditto
20%
30%
40%
50%
60%
70% "
80%
90%
95%
End Point
C, & Lighter, Gas Chroma tographic
D-2887(C)
-16
63
77
142
188
207
231
241
275
292
335
369
Analysis, wt - %
Propylene
Propane
Isobutane
Isobutylene + Butene-1
n-Butane
trans-Butene-2
cis-Butene-2
3-Methylbutene-l
Isopentane
Pentene-1
2-Methylbutene-l
n-Pentane
trans-Pentene-2
cis-Pentene-2
Total Cs & Lighter
93.6
85.4
9.0
0.7
(0.002)
0.01
0.08
0.70
0.07
5.00
0.07
0.09
0.06
10.88
0.16
0.34
2.56
0.41
0.22
(20.7)
100.0
91.7
10.6
2.49
(d)
Present
1.
21
60.3
0.01
0.07
0.69
0.06
5.02
0.07
0.09
0.06
10.99
0.16
0.34
2.59
0.42
0.23
(20.8)
-------
A-2
TABLE A-l. (Continued)
Blend Designation^
Properties
C$ & Heavier, Mass Spec PONA, wt - %
Paraffins
Monoolefins
Cycloolefins & Diolefins
Monocycloparaffins
Dicycloparaffins
Alkylbenzenes
Alkylindanes & "tetralins
Alkylnaphthalenes
Total Cg & Heavier
Approximate Distribution of Alkylbenzenes by Mass
C6
c?
C8
C9
GIO
Cll
C12
Total Alkylbenzenes
Total Sample GC + MS PONA Summary . wt - %
Paraffins
Olefins
Naphthenes
Aromatics
Grand Total CG & MS
Distillation- Chromatoeraphic-UV Analysis
Benz (a )anthracene
Benso(a)pyrene
Re -14 IB
unleaded
/.O Q
*r .3 . O
3.9
0.6
1.6
0.2
27.5
» l # *s
0.8
0.8
(79.3)
Spec, wt - %
1.8
6.9
9.0
6.8
2.2
0.8
0.1
(27.6)
63.1
5.9
1.8
29.2
100.0
3.3 ppm
r * ^^
1.0 ppm
RE -14 1C
Leaded
AO A
H J . w
4.0
0.6
1.8
0.2
27.2
0.9
0.9
(79.2)
1.7
6.5
9.1
6.8
2.2
0.8
0.1
(27.2)
63.0
6.0
2.0
29.0
100.0
(d)
(d)
-------
A-3
Footnotes to Table A-l:
(a) Data supplied by Mobil Research and Development Corporation.
(b) Fuels RE-141B and RE-141C were prepared from a single 6000-gallon-
lot which was composed of five blending components. Those components
which tend to be unstable in long-term storage had antioxidant and
metal deactivator (grades approved for use in military gasoline)
added at the refinery.
Half of the 6000-gallon lot was put into clean drums and labeled
RE-141B. The other half had Tetraethyllead Motor Mix added and,
after mixing, was drummed, then labeled RE-141C. TEL Motor Mix con-
tains ethylene dichlorlde (1.0 theory) and ethylene dibromide (0.5
theory) according to the usual specifications. No other additives
were added during blending of either RE-141B or -C.
(c) Simulated distillation by gas chromatography of RE-141B.
(d) S, BaA, and BaP in RE-141C should be the same as that determined by
analysis of RE-141B.
-------
A-4
Air Flow Measurements
Figure A-l shows measured engine air flow and road horsepower as
functions of vehicle speed from idle (at 0 speed) to 60 mph.
Figure A-2 is a plot of air fuel ratio versus engine air flow
showing the carburetor flow curve supplied by Ford Motor Company (dashed
line) and the data calculated from tailpipe CO, CO , HC, and 0 measurements.
The two curves agree reasonably well up to 50 scfm (about 40 mph). At
higher air flows the carburetor flow curve from Ford shows an air-fuel
ratio higher by almost 1 A/F number. The reason for this is not known.
Car Maintenance
In preparation for each run, performance of the cars was observed
and any necessary adjustment or maintenance was done.
After checkout runs were completed the oil was changed, a new
oil filter and air filters were installed, points checked and cleaned,
timing and dwell checked on both cars, and a series of 17 experimental runs
(Runs 12-18 to 1-15) was completed. Then, the exhaust systems of both
cars were pressure-tested at 1 psi for leaks. On the unleaded-fuel car a
leak was noted at the joint between the Y-pipe and the muffler. A much
smaller leak was noted at the same location on the leaded-fuel car. Both
leaks were stopped by tightening the clamps. The cars were then not used
for about 10 weeks while the data were examined in planning for the next
series.
In preparation for the new series of runs the compression
pressures on each engine were checked under motoring conditions. The
results were as follows in order of cylinder number: unleaded-fuel car,
148, 158, 145, 146, 150, 155, 150, 150; leaded-fuel car, 155, 152, 150, 148,
133, 147, 150, 150. The compression in cylinder No. 5 in the leaded-fuel
car is significantly lower than all other compression pressures. However,
it is not low enough to indicate any major problem such as broken rings,
stuck or warped valve, or burned valve seats.
Services on both project cars were: oil change, new oil and air
filters, new spark plugs, points checked and cleaned, timing and dwell
-------
A-5
90
80
70
60
E
*4
tf>
f*
_g
u_
50
40
30
10
0
0
10
Air flow
f
20 30 40
Road Speed, mph
Hp
20
15
10
0)
£
o
Q.
-------
o
or
o>
I
k-
<
Ford carburetor flow curve
8/6/70
BCL carburetor flow curve
4/14/73
40 50 60
Engine Air Flow, scfm
FIGURE A-2. AIR-FUEL RATIO VERSUS CARBURETOR AIR FLOW WITH UNLEADED FUEL
-------
A-7
checked and set to specifications, and the ignition patterns were observed
with an Ignition Analyzer. No problems or malfunctions were noted after
servicing.
Oil Consumption
When the oil and oil filter in both cars were changed, the new oil
and new filters were weighed before use. Following the test series the oil
was drained and weighed again along with the filter.
Oil consumption by the unleaded-fuel car was 0.58 lb per 1000
miles over a total distance of 2,344 miles, and by the leaded-fuel car was
0.95 lb per 1,000 miles over a total distance of 1,212 miles. The oil
density is 7.35 lb per gallon; thus, the oil consumption rate for the
white unleaded car as shown in Table A-2 was 0.32 quart per 1,000 miles and
for the blue leaded car was 0.52 quart per 1,000 miles.
In appearance, both oils were considerably darker after use than
when fresh. Oil drained from the unleaded-fuel white car was black but not
dirty in looks or feel. Oil from the leaded-fuel blue car was grey-black
and looked and felt dirty.
Accumulated mileage at the beginning and at the end of the oil
economy measurement was 10,583 miles and 12,927 miles, respectively, for
the unleaded fuel car, and 9,217 miles and 10,429 miles, respectively, for
the leaded fuel car.
After six test runs on the leaded-fuel car and ten test runs on
the unleaded-fuel car, a check was made on the oil consumption. Oil
consumption of the unleaded fuel car was 0.31 quart per 1,000 miles, and
of the leaded fuel car was 0.39 quart per 1,000 miles.
Development of Acceptable Choke Operation
Apparatus
A device to move the choke plate automatically according to a
preselected time schedule is illustrated in Figure A-3. A synchronous-type
gearmotor driving a shaft with a tab on its end through an 0-ring belt-and-
-------
A-8
TABLE A-2. OIL CONSUMPTION DURING TEST SERIES
(QUART PER 1,000 MILES)
Run Unleaded- Leaded-
No, fuel car fuel car
11-29 to 1-15 0.32 0.52
4-4 to 4-26 0.31 0.39
-------
Choke plate-
Choke plate
linkages
Vacuum -
break diaphragm'
Drive shaft
Drive rod
Choke lever
Drive pulley-
Idler pulleys'
Driven pulley
FIGURE A-3. DEVICE FOR CONTROLLED CHOKE OPENING ON A
REPRODUCIBLE TIME SCHEDULE
-------
A-10
pulley system was mounted on the choke housing, with the axis of the driven
shaft coincident with the choke-plate shaft. The tab pushed against the
choke lever in the choke housing to move the spring-loaded choke-plate shaft.
The gearmotor shaft speed was 1 rpm and the drive ratio was
approximately 10 to 1. The shaft rotation rate is thus about 34 degrees
per minute. Space limitations prevented using a higher drive ratio. To
accommodate the gearmotor, drive ratio, and mounting housing and plate,
the engine air cleaner housing was raised 1 inch from the carburetor body
by means of a round spacer tube. Gaskets were used at each face of this
spacer tube.
The choke is free to be opened wider than the control lever
requires if intake-air velocity pressure is great enough to overcome the
choke spring force. There was a lack of repeatability of the gaseous mass
emissions during the series of Runs 12-19 to 1-15 inclusive, which may have
been caused in part by too light a choke-plate spring resulting in erratic
action due to the air velocity. Prior to resumption of testing with Runs
4-4 through 5-22 the automatic choke drive system was made more slip-proof
by adding two idle pulleys which increased the 0-ring belt tension and in-
creased the contact length ("wrap") of the belt on both driver and driven
pulleys. A stiffer spring was also installed on the choke-plate shaft to
improve the stability of the choke plate under the influence of engine
vibration and carburetor air flow. The lighter-than-normal spring had
been used in the earlier series to minimize the chance of choke drive-belt
slippage.
Characteristic Stages in Choke Schedules
There are three stages of choke motion. The first stage is a
quick partial opening to about 30 degrees, as soon as sufficient manifold
vacuum is present, v/hich may require only 2 seconds. The second stage is a
hold at partial opening while the thermostatic coil is warmed up by exhaust
heat. The length of time at this stage depends on ambient temperature and
the speed of thermostatic coil warmup. The third stage is under the control of
the thermostatic coil and air velocity pressure on the choke plate. Time at
this stage will depend on the design of the coil and the heat it receives.
-------
A-ll
Figure A-4 illustrates idealized choke-opening schedules for
cold-start IA-4 cycles under various laboratory controlled conditions.
Information used for the three solid curves on this chart was supplied to
Battelle by Dr. W. R. Pierson of the Ford Motor Company and is derived from
special development work conducted on Ford 351 CID engines. Real choke
action is similar to the idealized curves in Figure A-4 but is not as sharply
defined.
The lean and rich curves indicate limits within which -a normal
choke should be operated, for the particular engine and conditions of Ford's
development program, to avoid stumble and stall at the lean end and to meet
emissions standards at the rich end. These limits are only approximate and
may be subject to wide variations with other cars, carburetors, or ambient
conditions. Choke schedules developed in the experimental study of the
"CAPE-19 cars did differ, as shown in the next section.
Experimental Studies of Choke Schedules
Experiments with the choke opening schedule were conducted in con-
junction with development and checkout of other systems, procedures, and
instruments. Proper idle jet setting was also evaluated in these experi-
ments. Initially, during a series of five runs, the idle jets were adjusted
to give lean-side smooth idle at approximately 0.5 percent CO or less in
drive idle. However, during later experiments, the drive-idle CO was
raised to 1 percent. In the test program, the CO in drive idle was
allowed to range between 0.6 and 1.2 percent or was readjusted to
1 percent.
Table A-3 summarizes the results of the test runs which were
conducted in the process of developing the choke opening schedule. Three
runs were made with the initial vacuum break angle of 30 degrees and with
choke movement to full open at 70 degrees. Several tests were run at the
30 degrees vacuum-break angle but at several different time spans to full
open. The vacuum break angle was then changed to 25 degrees for three more
runs with a 3-minute time span to full open. Next, the vacuum break angle
was reduced to 17 degrees. The time span to full open was kept at 3 minutes
-------
A-12
d)
o>
cn
c
601 Full open
50
40
30
Opened by manifold vacuum
I
15 30 45 60 75 90
Time From Start of Engine, seconds
105
120
FIGURE A-4. IDEALIZED CHOKE OPENING SCHEDULES FOR MODIFIED
COLD-START LA-4 CYCLE
-------
A-13
TABLE A-3. TEST DATA RELATING TO CHOKE SCHEDULE DEVELOPMENT
Unleaded RE-14IB Fuel
Tunnel Flow Rate 905 scfm
Sample Point Pressure 1 inch H90
Choke Schedule
Test No.
10-13
10-16
10-18
10-25
10-26
11-15
11-16
11-17
11-29
12-1
12-4
12-6
12-18
CO.
percent
Idle 50
0.3
0.2
0.2
0.2
0.4
1.0
0.7
-
0.6
0.6
0.7
1.1
0.8
0
0
0
0
0
0
0
0
0
mph
.3
.3
.3
-
-
.5
-
-
.4
.5
.4
.5
.4
Time Span
Vacuum to Full
Break(b) Open (c)
Degrees Minutes
30
30
30
25
25
25
17
17
17
17
17
17
17
1
2
2
3
3
3
3
3
3
3
3
3
5
.7
.7
.2
.0
.0
.0
.0
.0
.8
.8
.8
.8
.3
Mass Emissions
R/mile
HC
14.
11.
1.
1.
3.
2.
2.
1.
2.
2.
2.
1.
1.
1<6)
7(e)
34
38
96
09
18
95
02
30
06
90
63
CO
11.
10.
9.
13.
19.
11.
54.
25.
30.
34.
26.
38.
27.
1
1
9
3
6
9
3
6
4
0
3
6
2
(a) Measured at tail pipe.
(b) Full open choke is 70 degrees.
(c) Measured from start of engine.
(d) Computed from tunnel-bag gas composition.
(e) High HC emissions on Tests 10-13 and 10-16 indicated ignition
misfire problem, which was found to be partially shorted spark
(f) Ignition system problem co'rrected.
(d)
NO
3.31
1.36
1.56
2.97
1.55
2.82
2.93
2.89
2.88
2.93
2.63
2.81
2.82
system
plugs.
-------
A-14
then increased to 3.8 minutes0 Finally, a further modification was made
in the choke schedule beginning with Run 12/18 to provide a slightly longer
and also more realistic opening characteristic. The rate of opening was
halved from about 32 degrees per minute to 16 degrees per minute. Also, the
beginning of further opening from the vacuum-break position was advanced to
about 2 minutes after zero time in the starting sequence and the full-open
position was reached at about 5.3 minutes. This sequence was selected for
use in the first tests and designated Schedule I, which completed choke-
schedule experiments. Other schedules designated I-S, II, and II-S were
selected later, to accommodate cycle changes and to maintain emissions with-
in specified limits, as described in the next section.
Choke Schedules and Starting Sequences
Choke Schedule I (Modified IA-4 Cycle)
Figure A-5 shows Choke Schedule I. The sequence-time clock,
exhaust-gas diverter valve, and tunnel bag and dilution air bag sampling
are started simultaneously at time 0. Fifteen seconds later the engine is
started. At 30 seconds the accelerator is "kicked down" and at 40 seconds
the car is put in gear. The choke motor is started at 60 seconds (although
the choke was opened 17 degrees by manifold vacuum when the engine started).
The choke opening rate is 16 degrees/minute so the choke is full open (70
degrees) at 322 seconds (about 5.38 minutes). The first acceleration mode
of the cycle begins at 55 seconds.
Choke Schedule I-S (Standard LA-4 Cycle^
Figure A-6 shows Choke Schedule I-S. The sequence-time clock,
exhaust-gas diverter, valve, tunnel bag and dilution-air bag sampling, choke
motor, and engine are started simultaneously. Ten seconds later the
accelerator is "kicked down", and the car is put in gear at 15 seconds.
The first acceleration mode begins at 20 seconds. The choke opening rate
is 16 degrees per minute so the choke is full open at 262 seconds (4.38
minutes).
-------
A-15
-
0.
e
-o
QJ O>
o> c
Q. <
CO
0)
.
O)
>
o
_c
o
o>
Q.
E
-o" «
0} O>
(U C
o. <
CO
o
_c
o
Choke plate angle
Start engine (15 sec)
Kick down (30 sec)
( In gear (40 sec)
Start choke motor (60 sec)
Elapsed Time From Sequence Start, minutes
FIGURE A-5. CHOKE SCHEDULE I
70
60
50
Choke plate angle
Tr
40
-S? n 30
20
10
Start engine and choke motor (0 sec)
Kick down (10 sec)
In gear (15 sec)
Elapsed Time From Sequence Start, minutes
FIGURE A-6. CHOKE SCHEDULE I-S
-------
A-16
Choke Schedule II (Modified IA-4 Cycle)
Figure A-7 shows Choke Schedule II. The sequence-time clock,
exhaust-gas diverter valve, tunnel bag and dilution-air bag sampling, and
choke motor are started simultaneously. Fifteen seconds later the engine
is started. At 30 seconds the accelerator is "kicked down" and at 40
seconds the car is put in gear. The first acceleration mode begins at 55
seconds. The choke opening rate has been modified to 30 degrees per minute
so the choke is full open in 140 seconds (2-1/3 minutes).
Choke Schedule II-S (Standard IA-4 Cycle)
Figure A-8 shows Choke Schedule II-S. The sequence-time clock,
exhaust-gas diverter valve, tunnel bag and dilution-air bag sampling,
choke motor, and engine are started simultaneously. Ten seconds later the
accelerator is "kicked down", and the car is put in gear at 15 seconds.
The first acceleration mode begins at 20 seconds, and the choke is full
open at 140 seconds (2-1/3 minutes)'.
The dashed lines on Figures A-5, -6, -7, -8 represent the first
and second modes of the driving cycle. Key events such as engine start,
acceleration "kick down", and in-gear are identified.
Periods of Choke Schedule Use
Each of these schedules was developed to meet the requirements of
stable car operation. Choke Schedule I and the corresponding starting
sequence were used in Runs 12-18 to 1-23, 4-4 to 4-16, 4-18 to 4-26.
Choke Schedule I-S was used for the standard IA-4 cycle of Run 4-17.
Choke Schedule II was preferred for Runs 5-9 to 5-17, 5-21 -and 5-22 to
control gaseous emissions closer to the maximum allowable concentrations.
Choke schedule II-S was adapted from II for two standard LA-4 cycle
Runs 5-18 and 5-19. Other operational' and apparatus information on these
runs is summarized in the following section.
-------
A-17
ru
60
Q)
0)
-c CT 50
QL
o 20
0
10
ci
°
~7C\
t(j
60
w
QJ
0)
,c & 50
QL 0)
C7 "U
"g "I. 40
o> c
Q. <
« * *o
o °
£ c1-
> ^
g 20
O
10
7
r,
/
>
/[
!,
0
X^"" Choke plate angle
- Start choke motor (0 sec)
r- Start engine (15 sec) /A/\ / V«\ ^ mph
p Kick down (30 sec) / ^
r~ In aear (40 sec) / v \
// \
1 ^
^*^ %/ V » f \
^^/ * f 1
X/ ' I ' \
i' l ' \
"III 1 \ 1 / 1 1 1 I ,
1 23456
Elapsed Time From Sequence Start, minutes
FIGURE A.-7. CHOKE SCHEDULE II
/*~~ Choke plate angle
- Start engine and choke motor (0 sec) /^
/ A / \
Kick down (10 sec) 'v ^^ / v^ .mph
/ / V \_^^
In gear (15 sec) ^ / v\
/ /"
i/ , \ /, , , \ ,
1 23456
Elapsed Time From Sequence Start, minutes
FIGURE A-8. CHOKE SCHEDULE II-S
-------
A-18
Summary of Test Conditions in
Experimental Runs
Table A-4 summarizes emissions of total particulates, HC,
CO, and NO for the 58 experimental runs completed during tfhe course of
this study.
Dilution Tunnel
The dilution tunnel (Figure 1, page 3) is a steel tube 36 feet
long and 23 inches in diameter constructed from six flanged sections bolted
together. Dilution air is supplied by a blower to a filter section at the
upstream end, and the filtered stream enters the tunnel near the tailpipe
of the car. Exhaust and air both pass through a mixing orifice and thence
through the tunnel to the sampling and outlet sections.
The tunnel outlet system consists of a 3-foot tapered section, a
damper, and the discharge duct. The damper is used to control the pressure
at the sampling point. The tapered transition section ahead of the damper
assures that the tunnel flow patterns at the sample probes will not be
affected by the flow restriction of the damper.
Sample Probes
The sample probes are located 30 feet 8 inches downstream from
the mixing orifice. Figure A-9 is a cross-sectional sketch of the tunnel at
the sampling point showing the location of the various sample probes and
their function. Sampling ratios for the Sinclair-Phoenix smokemeter and
single particle counter were 0.25 and 8.8, respectively, and the cas-
cade impactor was operated at a sampling ratio of 9.2.
Figure A-10 is a detailed sketch of the residence chamber sample
probe. This sample probe is a 2-inch (inside diameter) PVC plastic pipe
enlarged to 2.1 inches at the inlet. The total length of the sample line
is 8 feet with two large-radius bends, one in the tunnel as shown in the
sketch, and one prior to entry into the residence chamber. The sliding-
plate valve permits flow into the residence chamber when open and is used
for precise control of the sampling time.
-------
TABLE A-4. SUMMARY OF EXPERIMENTAL STUDIES OF EXHAUST EMISSION
Test Conditions
Date
10-13
10-16
10-18
10-25
10-26
11-15
11-16
11-17
11-29
12-1
12-4
12-6
12-13
12-15
12-18
12- 17
12-20
12-21
12-22
1-17
1-19
1-23
12-27
12-28
12-29
1-3
1-5
1-10
1-11
1-12
1-15
Car
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
0
0
B
B
li
B
8
B
B
Fuel
NL
.ML
NL
NL
NL
NL
NL
NL
NL
NL
NL
NL
Nl.
NL
NL
NL
NL
NL
NL
NL
Nl.
NL
1.
L
L
L
L
L
L
L
L
Cycle
LA-4(M)
LA-4(V.)
LA-4 (M)
I.A-4 (M)
I.A-4 (M)
LA--l(M)
LA-4(V!)
I.A-4 (M)
LA-4 (M)
I.A-4(M)
LA-4 (M)
LA-4 (M)
l.A-4 (M)
I.A-4 (M)
LA-4IM)
I.A-4(M)
LA-4 fM)
LA-4(M)
I.A-4 (M)
LA-4 (M)
I.A-4(M)
LA--I(M)
I.A-4 (M)
LA-4 (M)
LA-4(M)
I.A-4 (M)
LA-4(M)
I.A-4(M)
LA-4 (M)
LA-4 (M)
LA-4(M)
Schedule
UD-
DO
UD
UD
UD
UD
UD
UD
UD
UD
UD
UD
UD
UD
L'D
1
1
I
1
1
1
I
1
1
I
1
1
I
!
1
1
Controller
RI
RI
RI
RI
DT
Ki
RI
RI
RI
m
DPC:
DPC
DPC
DPC
DPC
DPC
DPC
DPC
RPC
DPC
RI
R!
RI
DPC
DPC
DPC
RI
!U
RI
RI
RI
RI
MAAS
No
No
No
No
No
No
No
No
No
Yes
Yes
Yes
No
Yes
Yes
No
No
No
Yes
No
No
Yes
Yes
Yes
Yes
No
No
No
No
Yes
Yes
General Comments
Ignition problems
Ignition problems
Choke development
Choke development
Choke development
Choke development
Manual drive
Choke development
Low idle CO
Soak temp low - Nllj erratic
Soak temp low - CS changed
CS changed - low CO idle
Various problems
Run-in mileage low
Run length low, idle CO low
Soak time: short
Soak short, soak temp low
Idle CO high, soak temp low
Soak temp low
Hum.
O
A (SI)
.A (SI)
--
A (SI)
--
--
--
A (SI)
A (SI)
A (SI)
0
O
0
O
O
O
A
A
0
A
O
A
O
O
A
A
A
O
O
A
Variables
Lead
O
O
0
O
O
-
--
--
O
O
0
O
O
0
O
O
O
O
0
O
O
O
B
8
B
B
B
B
B
B
B
Dust
O
O
0
O
O
--
--
--
O
O
O
O
O
O
O
0
c
0
c
0
c
c
c
c
O
c
O
O
c
0
O
Nil,
0
O
0
O
O
--
--
--
O
O
O
O
D
O
D
D
O
D
0
D
1)
D
O
O
D
D
O
D
D
O
O
Part*
--
--
--
--
--
--
--
--
0.062'
0.031'
0.015'
o. on
0. nil
0.014
0.018
O.O'.'fl
Dust
0.014
Dust
0.03S
Dust
Dust
Dust
Oust
0.255
Dust
O.C41
0.269
Dust
0. 162
0.164
Emissions. s/nii
HC
14. 1
11.7
1.34
1.33'
3.0C
2.09
2.18
1.95
2.02
2.32
2.05
1.08
1.70
2.06
1.6.1
1.73
1.70
1.89
1.84
1.60
1..V.I
1.40
--
5.95
6.55
4.82
7. '23
6.33
3.34
4.04
5.39
CO
11. 1
10. 1
7.9
13.3
14. C
11.9
54.3
25.6
30.5
34.2
26.2
33.3
25. P
22. 1
27.2
2:1.3
23.9
36.2
33.0
32. C
31.7
2S.2
--
58.1
71.6
3-1.0
53. 1
47.7
33.4
44.1
39.0
NO
3.3!
1.36
1.5C
2.97
1.55
2.82
2.93
2.S9
2.88
2.93
2.63
2.81
2. 78
2.49
2.S2
3.23
3. 12
2.25
1.71
1.46
!.!>!
1.3C
2. i3
2.17
--
1.53
--
2.11
1.98
1.19
1.38
I
t-'
VO
At this point (lie cars were leak checked and a small leak was found in the white car and a very small
leak in the blue car.
-------
TABLE A-4. (CONTINUED)
Kuit
Da le
4-4
4-6
4-7
4-9
4-10
j-n
4-12
4-)3
4-16
4-17
4-18
4-10
4-2-1
4-20
4-25
4-26
5-0
5-10
5-1!
S-!4
5-15
5-1C
5-17
5-13
5-19
5-21
5-22
Test Conditions
Car
W
W
W
W
W
W
W
W
W
V.'
B
&
B
B
3
a
B
B
B
W
W
W
W
W
W
3
B
Car
Fuel
Cycle
Fuel
NL
.NL
NL
NL
NL
NL
NL
NL
NL
NL
L
L
L
L
'L
L
t^
L
L
NL
NL
NL
NL
NL
NL
L
t
Schedule
Instruments
Cycle Schedule
LA -4 (M)
LA-1 CM)
LA-1 (M)
LA -i { M)
U-) (W)
LA -I (M)
LA-» (M)
LA-J(M)
LA -4 (M)
LA~!
LA-1 (M)
LA -4 (M)
LA -4 (M)
LA-» (M)
LA -4 (M)
LA-i (M)
LA -I ( K!)
LA-» (M)
LA-1 (M)
1/.-1 (M)
LA-»(M)
LA-4(M)
U-»(M)
LA-t
LA-»
LA-4(M)
1A-4 (M)
W - White
ML - Nonleaded
LA-4 (M)
LA-4
UD
I and I-S
II and II-S
HAAS
I
I
i
1
1
I
I
I
I
I-t
I
I
I
I
i
I
;t
i;
ii
u
ii
u
u
II-4
II-S
u
ii
Cor,tro!!cr
DPC
DPC
DPC
DPC
i)?C
DPC
DPC
DPC
DPC
DPC
CPC
DPC
DPC
DPC
DPC
DPC
DPC
DPC
npc
DPC
DPC
DPC
DPC
DPC
DPC
DPC
DPC
B - Blue
L- Leaded
Cycle with
Cycle with
MA AS
Mo
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
No
NO
55-second
20-second
Gcr.cr-i Comments
Checkout
C!ie.;kout
C!ie.:kc'jt
Checkout
Chc-.-kc-ut
Hivli KC. CO
High !!C. CO
High JiC. CO
High HC, CO
it.:. Cycle
Checkout only
Atoned
Aborted
High '.1C. CO
High !IC. CO
Kign :iC. CO
Ba'.ciine
llasclinc
Aborted
Aborted
Dateline
Caroline
batcline
Std. Cycle
SfJ. Cycle
Uas.-linc
B«selln«
KEY
Initial idle
Initial idle
Variable.
Hum. Lead
O
O
O
O
O
O
O
O
O
0
O
O
O
0
O
O
O
O
O
O
O
0
O
O
0
0
0
O
O
O
O
O
0
O
O
0
O
B
3
B
E
B
B
B
B
B
O
O
0
O
O
O
B
S
Controller
Humidity
Particulate
Dust
O
O
0
O
O
O
0
0
0
0
O
O
0
O
O
O
O
0
0
O
0
0
O
O
O
O
0
DPC
RI
(SI)
Under development *
Defined In
Defined In
text
text
Variables
Minnesota Aerosol Analysis 'System
0
A
B
C
D
Emissions. R/rriU
Nil., Pan.* HC CO NO
O --
O --
O --
O --
O --
0 0.192 5.81 68.d 2.46
O 0.173 5.96 73.9 2.53
O 0.195 6.13 72.5 2.33
O 0.139 C.22 77.7 2.91
O 0.041 3.12 5S.G 2.72
O --
O
O --
O 0.319 C.07 71.9 1.99
O 0.340 5.04 0:3.2 2.0C
O 0.340 5.70 75.8 1.90
1
O 0.135 3.01 :«.! 2.CO £>
O 0.153 5.35 47.4 2.21
0 --
O --
O 0.040 3.43 3S. 5 2. :IS
O 0.046 3.50 3.-I.2 2.48
<' O.W4 .VJJ W..1" -'. f.O
« 0.07S 4.CC ''-. 4 ;.*7
" 0.071 4. 12 )l'.'< 2.«9
O 0.211 3.37 38.2 2.19
0 0.20J 3.96 «"1 l.»3
Magnetic tape
Paper tape
Steam Injection
Based on Metricel weights, Sampling Ratio " 2.2
Without backup filter
Low or normal values
High humidity
High lead
High dust (solid tnpurlty
in atmosphere)
High ammonia (gaseous
Impurity In atmosphere^
-------
A-21
1/4-in. ID
{CO, C02, NO
analyzers)
l/16-in. ID
(single particle
counter)
1/4-in. ID
(HC analyzer)
1/4-in. ID
(cascade
impactor)
l/4-in. ID
(47mm filters)
l/2-in. ID
(142 mm filters)
l/2-in. ID
(Sinclair -
Phoenix)
2.1 -in. ID
(residence
chamber)
FIGURE A-9. PATTERN OF PROBE INLETS IN CROSS SECTION
OF DILUTION TUNNEL AT SAMPLING POINT
-------
A-22
To
residence
chamber
Tunnel-bag
sampling probe.
Sliding plate valve
Tunnel.
\\\\\\V\\\\\\\\\\\\\\\V\\\\\\\\'
FIGURE A-10.
CONNECTION AT TUNNEL OF PIPE TO CARRY DILUTED
EXHAUST TO RESIDENCE CHAMBER
-------
A-23
Tunnel Sample-Point Pressure
Operation of the tunnel at a positive pressure is necessary to
provide a means for transferring sample from the tunnel to the residence
chamber, which is always at ambient pressure because of its flexible walls.
A further requirement of this sample-transfer operation is that the sample
should always be proportional to the exhaust gases produced throughout the
driving cycle. Since the tunnel flow rate is nearly constant, the sample
flow rate should also be constant to achieve proportional sampling. The
sample flow rate will be constant if there is a constant differential
pressure between tunnel and residence chamber.
The differential pressure between tunnel and residence chamber
was measured during a number of tests using a strain-gage pressure trans-
ducer with a range of 0 to 0.1 psi (2.77 inches H_0). Tunnel sample-
point pressures from about 1/2 inch HO up to about 2 inches H~0 were used.
It was concluded from these tests that a tunnel sample-point pressure of
about 1 inch H20 results in an acceptably constant pressure differential.
Figure A-ll shows a reproduction of the chart record of tunnel-
to-chamber differential pressure during a run in which the sample-point
average pressure was 0.91 inch H.O. The chart shows that AP remains
acceptably constant throughout the cycle. Deviations from a steady pressure
appear to be about ± 2 percent during most of the cycle. In the second
high-speed mode of the cycle the tunnel pressure increased to about 10
percent over the average for a period of about 1 minute.
Tunnel-Bag Sample
A composite sample of the diluted exhaust gas is collected from
the flow going into the residence chamber during the test run. This com-
posite bag sample was analyzed for gaseous components to measure the final
dilution from tunnel to chamber, as well as to monitor gaseous emissions
for assurance of normal engine operation.
Figure A-12 is a sketch of the tunnel-bag sampling system. A
diaphragm pump in conjunction with a rotameter and regulating valve supplies
-------
1.2
1.0
0.8
O
CM
X
en
0.6
o
a.
<3 0.4
I-O
-O
0.2
0
8 12
Elapsed Time, minutes
16
20
24
FIGURE A-11.
RECORDED DIFFERENTIAL PRESSURE BETWEEN TUNNEL AND
RESIDENCE CHAMBER DURING A MODIFIED LA-4 CYCLE
-------
A-25
Tunnel bag
Regulating valve
Diaphragm pump
Bag shut-off valve
Rotameter
I I
Yr_
Filter
Residence chamber
sample pipe
Bag-sample probe
FIGURE A-12. TUNNEL-BAG SAMPLING SYSTEM
-------
A-26
a constant sample flow rate to the bag. Particles are filtered out before
the pump. At the end of the cycle the bag is disconnected from the rota-
meter and connected to the gas analyzers to measure the average gaseous
emissions.
Calculation of Exhaust Mass Emissions
from Tunnel-Bag Sample
A 23-1/2-minute bag sample of diluted exhaust gases from the
tunnel is collected during each test run. This composite sample is
equivalent to a CVS-type bag sample. Mass emissions per mile are computed
as the product of contaminant density (g/cu ft at 68 F and 1 atm pressure)
x contaminant concentration (vol/vol) x V . . The equation in the Federal
(1) mix
Register for calculating V . . the standard volume of the diluted exhaust
° & mix'
gases emitted per mile, is as follows:
V . = total tunnel flow rate (cfm) x run length (minutes)
mix
tunnel absolute pressure (in. Hg)
x
x
29.92
528 1
tunnel temperature (°R) 7.5 miles
Lead Deposits in Pipe to Residence Chamber
Lead deposits were measured in the sample-transfer pipe between
tunnel and residence chamber (Figure A-10, page A-22) after the pipe had
been exposed to diluted exhaust from nine cold-start IA-4 cycles using
leaded fuel. The operating parameters were as follows:
Fuel: 2.5 g Pb/gal; density, 6.25 Ib/gal; rate, 4.25
Ib/cycle Lead in fuel residues from nine cycles:
9 x 2.5 x 4.25/6.25 = 15.3 g
Sample Flow: Tunnel flow, 905 cfm; sample flow through pipe;
10 cfm (average values for all test runs)
Sampling Ratio: 1.3.
(1) Federal Register, Vol. 37, No. 10, January 15, 1972.
-------
A-27
Assuming all lead is discharged in exhaust, then total sample
flow in nine cycles contained 15.3 x 10/905 = 0.17 g Pb.
Found in sample pipe: 0.006 g Pb, or (.006/.17) x 100 =
3.5 percent of Pb in samples.
A particulate sample was collected on filters from the tunnel gas
in Test 1-12 amounting to 0.006 g weight. Sample gas flows were 4.5 cfm
through filters, and 10 cfm through pipe to residence chamber. Gas sample
to chamber contained 0.006 x 10/4.5 = 0.013 g particles. Assuming the
exhaust particles contained a .mean value of 25 percent Pb, and that this
test was typical of nine leaded fuel tests, total lead carried through the
pipe was 0.013 x 9 x .25 = 0.029 g. Sample collected from wall was
(.006/.029) x 100 = 21 percent of Pb exposure, based on the assumptions
made.
It is concluded from these results that lead loss in the transfer
of sample gas to the residence chamber is a significant part of the total
particulate sample, and would affect a quantitative material balance. Never-
theless, losses are not so severe as to preclude study of aging of particles
in the residence chamber.
Test for Uniformity of Mixing in the Tunnel
Simultaneous samples of auto exhaust particles were collected at
the standard sampling location (Page A-18) from a 13-probe assembly to measure
the uniformity of particle distribution across two perpendicular diameters
of the tunnel. The coefficient of variation was 4.5 percent, which is an
acceptably uniform distribution.
Gaseous Contaminant Injection System
Figure A-13 is a sketch of the gaseous contaminant injection
system, which was installed for the study of the effect of a gaseous con-
taminant, ammonia, on aging of auto exhaust particles in the residence
chamber. The gaseous contaminant is introduced into the tunnel for mixing
-------
A-28
Tunnel inlet air flow
Filter-
I U U U
Top Section View
Filter box
Direction of gas injection
"H-Pipe" diffuser
l/32-in. holes on
l-in. centers in
l/4-in. SS tubing
Supply pipe
Rotameter
Regulator
Gas
cylinder
Regulating
valve
FIGURE A-13. GASEOUS CONTAMINANT INJECTION SYSTEM
-------
A-29
with exhaust and dilution air. The H-shaped distributor manifold is in-
stalled immediately downstream of the filters in the dilution tunnel filter
box, and has 1/32-inch-diameter holes drilled in the downstream side.
These holes provide a fairly uniform initial distribution of the contaminant
gas in the dilution air stream before it passes through the tunnel orifice,
where it is mixed with exhaust.
The injection rate of ammonia was measured by bubbling the gas
through a glass tube containing ammonium hydroxide. The flow rate was
determined by counting the bubble rate and estimating the bubble diameter
when operating with ammonia. The same conditions were duplicated with
air, using a calibrated rotameter in the supply line to measure the
quantitative gas flow. Becasue of the corrosive properties of ammonia, a
rotameter could not be used in direct contact with ammonia for the flow
measurement.
Solid Contaminant Injection System
Figure A-14 is a sketch of the dust entrainment apparatus which
was used in some tests to add dust as a solid contaminant to the tunnel
air. The apparatus consists of a motor driven disc containing four V-
shaped grooves. Scraper vanes fill the grooves and level them. The filled
grooves then pass under an aspirator which picks up the dust and dispenses
it. The feed rate was controlled by the disc speed. The resulting aerosol
is fed into the exhaust pipe immediately downstream of the exhaust-gas
diverter valve. Dust injected at this point is quickly entrained and mixed
with the exhaust gas and the mixture is diluted at the tunnel orifice plate.
Dust Contaminant
The dust contaminant was Fine Arizona Road Dust which was screened
before use to remove all particles larger than 3.3 micrometers Stokes
diameter. Figure A-15 is a Rosin-Rammler plot of the particle size dis-
tribution of the sample of dust as received before the larger sizes were
removed and discarded.
-------
A-30
Filtered
air
Vibrator
FIGURE A-14. SCHEMATIC OF DUST FEEDER
-------
A-31
99
98
95
90
80
70
| 60
M 50
& 40
£ 30
u
20
3
<-> 5
0.5
0.2
O.I
0.05
0.01
0.2
0.3 0.4 0.6 0.8 I 2 34 6
Equivalent Particle Diameter, microns
FIGURE A-15. PARTICLE SIZE DISTRIBUTION OF
CLASSIFIED ARIZONA ROAD DUST
8 10
20
-------
A-32
Chemical composition of the classified dust was determined by
optical emission spectrometry, as shown in Table A-5.
TABLE A-5. CHEMICAL COMPOSITION OF FINE ARIZONA DUST
Element
Silicon
Aluminum
Magnesium
Iron
Manganese
Barium
Percent
15-30
5-10
0.7
3.0
0.05
0.05
Element
Chromium
Calcium
Vanadium
Copper
Sodium
Titanium
Percent
0.01
2-4
0.01
0.1
1-2
0.3
Residence Chamber
Chamber Configuration
The residence chamber (Figure 1, page 3) is a rectangular 6-Mnil
black polyethylene bag approximately 9 ft x 12 ft x 20 ft with a filled
volume of about 2160 cu ft.
Figure A-16 is a layout of the residence chamber showing sample
probes, the sample pipe, and the purge circulation system. Diluted exhaust
gases from the dilution tunnel enter the residence chamber through a 2-inch
ID PVC plastic pipe which is located approximately 6 ft from one end of the
chamber and 4 ft down from the top. The sample pipe projects horizontally
into the chamber for a total length inside the chamber of about 5 ft and is
curved slightly so that the sample flow is towards one upper corner of the
chamber. In the final design, the sample pipe has a tapered discharge
nozzle to promote mixing.with dimensions as shown in Figure A-17. This
design was selected after three series of exploratory measurements with other
configurations and locations, described in following sections.
-------
Bleed valve
-!>=
\
Control valve
Chamber walls
Steam supply line
Removable
end cap
v
Removable
end cap
Inlet damper
Particulate filter
Drier bed
Charcoal filter
z
/
Pure
ie duct
i
I
9_in Tn
Humidifier
D
sample probes
(2)
3/8-in. ID
sample probes
(4)
2-in. ID
.sample probe
Outlet damper
Wet/dry bulb-
system
Sample pipe
from tunnel
Purge
blower
OJ
FIGURE A-16. LAYOUT OF RESIDENCE CHAMBER AND PURGE-CIRCULATION SYSTEM
-------
A-34
2-inch ID PVC
sample pipe
\\\\X\\N>
FIGURE A-17. TUNNEL-TO-CHAMBER SAMPLE-PIPE
DISCHARGE NOZZLE
-------
A-35
Chamber samples for gas analysis are withdrawn from the center
of the chamber through 3/8-inch-ID stainless steel probes.. Samples for
particulate and light scattering analysis are withdrawn through two 2-inch-
ID PVC sample probes with intakes near the center of the chamber.
Dilution Ratio Experiments
The dilution tunnel when operated at a total flow rate of 905 cfm
dilutes the auto exhaust by an average ratio of 30 to 1 at 72 F and 1 atm.
To achieve an overall dilution ratio of 300 to 1 in the residence chamber,
a tunnel-to-chamber dilution of 10 to 1 is required. Thus, the tunnel
sample-point pressure, and the sample-pipe and residence chamber design
were selected to achieve this 10 to 1 dilution and to achieve fast and
complete mixing in the chamber.
The 2-inch diameter of the tunnel-to-chamber sample pipe was
selected to achieve approximately isokinetic sampling from the tunnel.
However, the flow through this pipe at 1-inch H^O differential pressure was
considerably greater than required for the 10 to 1 dilution. Consequently,
flow was restricted to the desired rate of about 9 cfm with a discharge
nozzle at the outlet end.
In the actual test runs, the sampling rate averaged approximately
10 cfm. With an average tunnel flow rate of 905 cfm for the test runs,
the isokinetic sample flow rate would be 7.5 cfm for the 2.1-inch-diameter
probe opening, thus, the actual sample flow was about 33 percent greater
than true isokinetic flow.
Table A-6 summarizes the results of the first series of nine
preliminary dilution-ratio experiments, Runs 9-5 to 9-20, at a tunnel flow
of about 1100 cfm. Two locations for flow restrictions were compared, one
at the discharge end of the pipe inside the chamber and the other in the
pipe immediately downstream from the sample-flow control valve. This
configuration was a simple semi-venturi shape with a straight bore between
straight-sided conical converging and diverging sections.
These tests were performed before all of the instrumentation was
available and operational; hence, only HC and CO measurements were made in
the tunnel and chamber to determine dilution ratio. The dilution ratios
-------
A-36
TABLE A-6. SUMMARY OF INITIAL TUNNEL-TO-CHAMBER
DILUTION RATIO EXPERIMENTS
White Unleaded Car
RE-141B Fuel
Sample Point Pressure 1 inch H^O
Tunnel Air Flow, 1100 scfm
Tunnel- to-Chamber
(a\
Dilution Ratio^ '
Run From HC
9-5 4.8
9-6 6.5
9-7 6.2
9-8 12.0
9-12 15.3
9-13 11.0
9-15 11.0
9-18 15.2
9-20 10.0
From CO
11.9
70.0
5.8
14.8
13.4
11.0
14.0
7.4
17.2
Average
4.8
6.5
6.0
13.9
14.4
11.0
12.5
15.2
10.0
Description of
Sample-Pipe Restriction
Open pipe
Open pipe
Open pipe
3/8-inch nozzle
3/8-inch venturi
1/2- inch venturi
with 1" nozzle
it
it
ii
at pipe end
above valve
above valve
at pipe end
(a) Computed from HC and CO measurements in tunnel bag and in residence chamber
before and after run.
(b) Average value taken as value computed from HC data only when HC and CO data
were not in + 10 percent agreement, because of intermittent malfunctions of
CO meter.
-------
A-37
computed from both HC and CO measurements are presented. However, the HC
data are considered more reliable wherever the two numbers do not agree
well, because of intermittent malfunctions of the CO meter.
One configuration, which gave adequate control of flow, had the
1/2-inch primary restriction located downstream of the sliding plate valve
used to cut-off sample flow (Figure A-10, page A-22), plus another 1-inch-
diameter nozzle at the tube end in the chamber. The last four tests, 9-13
through 9-20, were reproducibility tests run with this restriction con-
figuration. It was concluded that flow control was acceptable, but mixing
in the chamber was too slow.
Chamber Mixing Experiments
Beginning with Test 9-15, Table A-6, dilution ratios and mixing
in the chamber were investigated at the same time. It was found initially
that mixing was completed 15 to 30 minutes after the end of the auto-
exhaust-generation run. Particle concentrations were then similar in
various regions in the chamber.
In a second series of tests a more detailed evaluation of chamber
mixing was made with propane tracer, using short sample probes temporarily
installed in four corners of the chamber. These probes projected 1 inch
into the chamber and were each located within about 6 inches of the corner
so that they would be sampling from extreme regions of the chamber. The
regular sample probes in the center of the chamber were also used. The
mixing-study probe locations (listed in the sampling sequence generally used
in the tests) were: side center near chamber middle, side center near
chamber wall, top southeast corner near wall, bottom southwest corner near
wall, top northwest corner near wall, bottom northeast corner near wall.
A full sampling sequence required approximately 12 minutes. This allowed
about 2 minutes at each position to stabilize the reading and to note short-
term changes if they occurred.
A total of 11 experimental runs were conducted in the propane-
mixing study. Propane (LPG from a trailer-type tank) was introduced into
the tunnel filter box with the tunnel air flow at approximately 900 scfra
and with a sample-point pressure of 1 inch H?0. The rate of propane in-
jection was adjusted to yield a constant value around 800 to 1000 ppmC in
-------
A-38
the tunnel-. When stabilized conditions were obtained in the tunnel, the
residence-chamber sample valve was opened for a period of 23 minutes, the
standard LA-4 cycle running time. Immediately following the end of this
sampling period the concentrations of HC in the residence chamber at the
various sample-probe positions wera measured in the predetermined sequence.
The sequence was repeated several times over periods up to almost an hour
a-fter the end of the sampling period.
With the 1-inch-diameter nozzle opening at the end of the sample
pipe in the chamber used in the first series, the velocity of the gas
leaving the pipe was about 28 ft per second. This was insufficient to
promote rapid mixing into corners of the chamber.
Experiments, in the third series were then conducted with one sample-
pipe flow restriction located only at the pipe end. Mixing was substantially
improved, and nozzle dimensions of 7/16-inch diameter by 1 inch long were
found to yield the desired final dilution ratio. With this nozzle (Figure
A-17, page A-34) the velocity of the gas entering the chamber at 9 cfm flow
rate was about 145 ft per second.. A uniform mixture in the residence
chamber was obtained 12 minutes after injection was completed at the end
of the test run, at which time all HC concentrations agreed within 10 per-
cent. Figure A-18 shows the time variation of the HC concentrations in
the residence chamber from the various sampling probes after the simulated
auto-exhaust run with propane, with the final modification of the mixing
nozzle in place.
Chamber Humidification
High residence chamber humidity was initially achieved in Runs 10-
13 to 12-4 (Table A-4, page A-19) by injecting steam directly into the
chamber with the purge-circulation system in operation.
This method worked quite well for bringing the chamber up to 70
percent relative humidity in a short time, but it produced an unacceptable
increase of the background particle level in the chamber.
Other methods were tried and the final technique selected combined
the use of steam to raise the humidity in the whole laboratory, the in-
jection of some steam directly into the purge system makeup-air inlet
(Figure A-16, page A-33) and the use of a wick-type humidifier in the purge
system line as shown in Figure A-19 (also Figure A-16, page A-33). The
-------
A-39
80
o
.a
t»
o
o
o
70
60
50
O
Q.
Q.
0>
"I
o
.c
o
c
o
^ 40
c
CD
O
O
30
20
10
Propane Mixing Study
Test 10-9 =
Tunnel flow rate 960 scfm
Tunnel propane concentration 865 ppmC
10 20 30 40
Elapsed Time From End of Fill Period, minutes
50
FIGURE A-18.
TIME TO MIX PROPANE IN RESIDENCE CHAMBER
WITH SAMPLE INLET NOZZLE SHOWN IN FIGURE 18
-------
A-40
Motor
Clean air
Water
line
FIGURE A-19. SCHEMATIC OF WICK-TYPE HUMIDIFIER
IN RESIDENCE CHAMBER PURGE DUCT
-------
A-41
humidified air passes through the filters before entering the chamber. This
technique requires a longer period of time but does not affect the back-
ground particle concentration in the chamber.
Dilution Ratio Definition and
Method of Calculation
The dilution tunnel dilution ratio is the total tunnel flow divided
by the exhaust flow. The residence chamber dilution ratio is the total
residence-chamber volume (after sample addition) divided by the volume of
diluted exhaust gas added to the residence chamber.
Dilution tunnel dilution ratio is computed directly from measured
tunnel flow and exhaust flow. Residence chamber dilution ratio is computed
indirectly from measurements of CO and HC concentrations in the chamber,
before and after sample addition, and in the tunnel bag sample.
The development of residence-chamber dilution-ratio equations
follows.
Symbol Definitions
Let: Q = residence chamber volume before sample addition
3
Q = volume of diluted exhaust gas added to residence
s
chamber during driving cycle
Q = final volume of residence chamber after sample
addition
C = HC or CO concentration in residence chamber before
a
sample addition
C = HC or CO concentration in diluted exhaust gas
s
C = HC or CO concentration in residence chamber after
sample addition.
Equations
Qt
D.R. (dilution ratio) = r~ (1)
^s
Q + Q = Q (summation of volumes) (2)
Si S C
QC +QC = Q C (summation of emission mass) (3)
a a s s t t
-------
A-42
Solve equations (1), (2), and (3) simultaneously:
Q C Q C Q C
a a , s s _ t t
QC Q C ~ Q C
s a s a s a
C
_t
C
a
C C
(D.R.) - 1 - (D.R.) T = ~ T
C C
s - e
1 P
* *->
t a
D.R. = CS" . . . . . . (4)
Instrumentation
Gas Analysis
Direct reading and recording instruments were used for analyses of
gases from the tailpipe, the dilution tunnel, the residence chamber, the
filtered dilution air, and composite bag samples from the dilution tunnel.
In general, each instrument zero was set with a zero gas and the span with
one or more span gases covering the range of expected concentrations. The
span and zero were checked and adjusted before and after each run.
Manufacturers' guaranteed values for gas composition were accepted for
standard gases to set zero and span at the reported concentration of the
span gas, and the instrument manufacturers' calibration curves were relied
on for interpolation and extrapolation. There was no absolute standardiza-
tion by analysis of the standard gases. On some occasions, zero gas and
two standard gases were used to construct a revised curve through the origin
and these two additional points. Table A-7 shows the record of instruments
used during the year while developing the capability to obtain reliable
analyses of exhaust gas composition. Some instrumental difficulties were
identified and corrected as soon as they were recognized. Specific details
on calibrations and substitutions are described in tabulated comments dis-
cussing the individual gases.
-------
TABLE A-7. GAS ANALYSIS INSTRUMENTS
Gas
CO
HC
Samples
Location
Tailpipe
Tunne 1
Chamber
Background
Tailpipe
Tunnel
Chamber^ ^
Dilution
air
Time
Continu-
ous record
incl.
A/F ratio
modes
Cycle
record
Continuous
11/29-1/23
and 0, 30'
4/11 >
Bag
sample
A/F ratio
modes
Cycle
record
0, 30'
Bag
sample
Approx.
Conc.n
1-5%
2-400
ppm
Low
Low
1000 -
2000 ppm
50 -
100 ppm
Low
Low
Instruments
Description
Olson-Horiba IR
Beckman NDIR
Same
Same
Beckman FID
Same
Same
Same
Standard Gases
Concentra-
tions
0, 4.99%
0, 152 ppm
0, 152, 220
0, 152,1075
0, 76.3 ppm
Use Period
12/6-5/22
12/6-4/25
4/26
5/9-5/22
12/6-5/22
Comments on Techniques and Instruments
Range 0-10%
Instrument malfunction 4/11-4/20;
corrected 4/25. Emissions above 1972
limit 4/11-4/26; adjusted to below
limit, 5/9 " X Bag samples from
tunnel 4/25, 4/26 , instead of con-
tinuous recorded analysis.
>
w
Tailpipe sampled for HC following
cycle during idle, 35 mph, 50 mph
modes. Span calibration before and
after cycle and after 50 mph mode.
Measurement used to detect abnormal
conditions .
Ditto
»
Chamber sample at zero time gives background concentration in chamber; at 30 minutes gives concentration after.
dilution and mixing in the chamber.
-------
£ABLE A.-7.. (CONTINUED)
Gas
°2
Nn
w \j
Samples
Location
Tailpipe
Tunnel
Chamber
Dilution
air
j
Time
A/F ratio
measure-
ment
Bag
sample
0, 30'
Bag
sample
Approx.
Conc.n
1-20
ppm
Instruments
Description
Beckman 715
Beckman 715
Beckman 715
Beckman OM-11
Beckman NDIR
0-250 ppm
range
Standard Gases
Concentra-
tions
Air, 21%
Ditto
0.54,2.2%
Ditto
140 ppm
140,19 ppm
Ditto
Use Period
11/29-1/23
Ditto
4/11-4/26
5/9-5/22
11/29-1/12
1/12-1/23
4/11-5/22
Comments on Techniques and Instruments
First and second instruments gave
unsatisfactory and unreliable
measurements. Identical models.
Third instrument gave acceptable
results with improved techniques,
using long sampling time for slow
response.
Excellent results: Stable reading
with fast response.
>
Doubtful results JL.
-P-
Revised calibration curve to fit two
gases. Improved technique with
drastic dehydration in better cold
trap. Overhauled instrument. Good
results obtained with changes.
Reliable results.
-------
TABLE A-7. (CONTINUED)
Gas
co2
Samples
Location
Tailpipe
Tailpipe
Turme 1
Chamber
Dilution
air
Time
Bag
samples
Direct
reading
0,35,50
mph for
A/F ratio
Cycle
record
0, 30'
Bag
sample
Approx.
Conc.n
Instruments
Description
Beckman NDIR
LIRA IR
0-1%
Ditto
n
Standard Gases
Concentra-
tions
9.94%
12.4%
1100 ppm
7240
Ditto
ii
Use Period
11/29-4/20
4/25-5/22
11/29-1/23
4/11-5/22
Ditto
it
Comments on Techniques and Instruments
Instrument recorded sample from tail-
pipe 11/29-4/20 and 5/18-5/21. Bag
samples carried to instrument 4/25-
5/17.
Two span gases disagreed. Difference
represents 3% of measurement.
1100 ppra gas used routinely; 7240 ppm i
occasionally to check calibration 01
curve. Curve corrected to conform
to two concentrations of span gases.
-------
A-46
Speed Controller
A speed controller was used to drive the test? car automatically
through the LA-4 driving schedule after it was manually started and shifted
into gear. During exploratory tests and choke schedule development* cars
were driven by a Research Incorporated (RI) controller, programmed by a
punched-paper tape (Table A-4, page A-19). Problems were encountered with
faulty control by this instrument in a few tests and it was taken out of
service for repair. While the system was being serviced, a second speed
controller, manufactured by Dynamic Precision Controls (DPC) and using a
magnetic tape program input, was used for Runs 12-1 to 12-22. The RI
punched-tape system was put back in service and used for Runs 1-3 to 1-15.
The DPC magnetic tape speed controller system drives the car
through a more precise cycle than the RI punched tape speed controller,
because the punched tape tends to smooth out the minor speed fluctuations.
Therefore, a new DPC magnetic tape controller was obtained for all sub-
sequent tests.
Minnesota Aerosol Analyzing System
A group of instruments known as The Minnesota Aerosol Analyzing
System (MAAS) was available on loan* for some of the runs (Table A-4, page
A-19) and was used whenever available to characterize the size distribution
of automotive exhaust particles in suspension in samples removed from the
residence chamber.
The MAAS consists of three aerosol particle counters in parallel
operation, which in combination measure the aerosol size distribution
from 0.0032 to 10.0-u.m diameter. The instruments consist of a modified Royco
Model 220 optical counter, a condensation nuclei counter (CNC), and a
Whitby Aerosol Analyzer (WAA). The output of the optical counter is fed
into a Nuclear Data (ND 812) computer, and the outputs of CNC and WAA are
fed into a Hewlett Packard Data Acquisition System (DAS).
* Courtesy of Professor Whitby of the University of Minnesota and Dr.
William Wilson, Jr., of National Environmental Research Center, EPA.
(1) K. T. Whitby, B. Y. H. Liu, R. B. Husar, and N. J. Barsic, J. Colloid
and Interface Science, Vol. 39 No. 1, pp 136-164 (April, 1972)
-------
A-47
The size range of the modified Royco Model 220 sensor is from
0.562 to 10.0 M-m, and this range is classified into 512 channels under
control of the ND 812 computer.
The condensation nuclei counter is a standard Environment-One
counter operated at an under-pressure expansion of 8 inches of Hg vacuum.
The Whitby Aerosol Analyzer is an improved ion mobility analyzer
manufactured by Thermo Systems, Inc. The instrument has feedback-controlled
high-voltage power supplies which permit automatic operation in conjunction
with the DAS. It covers the size range between 0.0032 and 0.562 u.m and the
minimum scan time is 2.0 minutes.
Studies of Filtration and Weighing Procedures
The mass concentration of exhaust particles was determined by
measuring the net weight gain of an absolute filter for filtration of a
measured volume of diluted auto exhaust. A procedure was developed to
check the validity of the weight gains measured. Operations made a part
of this procedure are described in following sections.
Microbalance in Controlled Atmosphere
Balance Room
A new Mettler Model M-5 microbalance xvith a claimed sensitivity of
1 t-tg was set up and used exclusively on this project in an isolated room
with controlled constant humidity and temperature. The microbalance was
also used temporarily in a controlled humidity glove box, and in two other
controlled-atmosphere rooms, with some results that were erratic, before the
satisfactory location was found.
Repeatability of Weighings with the Microbalance
Table A-8 presents typical data on weighings of two types 'of
filters, a stainless steel disk, and a Class M platinum 500 mg weight.
They were weighed every txvo hours one day to check on the operation of the
balance and on the technique of the person performing the weighings. In
-------
TABLE A-8. REPETITIVE WEIGHINGS OF BLANK FILTERS AND OTHER MATERIALS
Weights, g
Oct. 23 1972
Time
8:00
10:00
12:00
2:00
4:00
47-mm Glass
Filter
MSA J7864
0.105173
0.105181
0.105170
0.105165
0.105173
Stainless Steel
Disk
0.213274
0.213274
0.213265
0.213263
0.213272
142-mm Metricel
Filter DM450
0.574450
0.574545
0.574478
0.574509
0.574413
Type M 500 mg
Platinum Wt.
0.499996
0.499996
0.499995
0.499994
0.499995
Temperature,
F
72.3
73.5
74.0
75.0
73.0
Relative
Humidity, %
41.0
46.0
48.8 >
oo
46.5
28.5
-------
A-49
this room the maximum humidity was controlled below 50 percent RH,
but there was no control of minimum relative humidity. The temperature
was controlled between 72.3 and 75.0 F in the weighing area.
The apparent weight gains and losses of the filters in Table
A-8 did not correlate consistently with increases and decreases in humidity.
Therefore, the variations in these weights x^ere suspected to be attributable
to other causes, such as changes in electrostatic charge. A small polonium
source had been used routinely in the balance during these weighings to
dissipate static charges, but the observed variations, if caused by elec-
trostatic charge, meant that the dissipator x^as undersized.
Two new and larger polonium sources were then used to initiate
another series of experimental weighings (data not given) Prior to each
weighing, both sides of each filter were exposed at a distance of one inch
from these polonium sources. After this pretreatment, uniform weights of
± 4 (ig were obtained until there was a change in humidity. When the
humidity remained within ±1.0 percent RH, the weight of a 47-mm Metricel
filter is repeatable within about ± 15 M-g and that of a 142-mm Metricel
filter is repeatable within about ± 150 u.g. The larger size has approxi-
mately ten times the surface area subject to adsorption and desorption of
water vapor.
Filter Media
Two types of polymeric membrane filters were selected for use
in the determination of particulate mass by weighing. Metricel DM 450
membrane filter (Gelman Instrument Company, Ann Arbor, MI, Part No. 64519)
is a copolymer of polyvinyl chloride and acrylonitrile, 0.45-nm pore
size. The filter diameter was 142 mm. The filter was used in a filter
2
holder with an active filtration area of 125 cm . Flow rate, calibrated
for each batch, was about 7 cfm for one filter in the holder and about
4.3 cfm for two in series, with a pressure drop of about 10 cm Hg across
the filter(s). Face velocity at 4.3 cfm was 16,3 cm/s,
Millipore MF AAWP 047 00 membrane filter (Millipore Filter
Corporation, Bedford, MA) is a mixed cellulose ester, reportedly pro-
pionate-butyrate, 0.8 nm pore size. Filter diameter was 47 mm, and active
-------
A-50
2
filtration area was 9.6 cm . Flow rate calibrated for each batch was
about 1.0 cfm for one filter and 0.7 cfm for two in series in the holder,
with a pressure drop of about 10 cm Hg across the filter(s). Face
velocity at 0.7 cfm was 34.4 cm/s.
The membrane filters were chosen because they were expected
to be low in reactivity toward acidic constituents (e.g., 802) i-n tne ex-
haust, and resistant to abrasive losses during insertion in and removal
from the filter holders. Agreement in results from simultaneous parallel
samples taken with both types indicated probably negligible effects
on weight gains because of filter holder configuration, face velocity of
the filtered stream, and chemical composition of the filter medium.
Filter-holder flow rate was controlled by an orifice downstream
of the filter. Orifices for each filter holder were calibrated with the
proper filters in place, using a dry gas meter, and at the same manifold
vacuum as was used during the actual sampling runs. The calibrations were
repeated several times during the program to determine that no substantial
changes in flow rate were taking place.
The SO content of the automobile exhaust was determined, in
order to estimate whether the potential weight change of a reactive filter
would be significant if the sulfur in the fuel (0.036 wt %) were con-
verted to SO- followed by reaction with the filter. It was found that the
S02 in the exhaust diluted in the tunnel (^30:1) contained approximately
0.2 ppm S0?. A 22-liter sample contained by analysis 10.5 micrograms of
S0_. In tests that use a 4.0 cubic meter sample volume, SO could
account for 2,000 micrograms weight increase of the filter if all the S0«
reacted with the filter. This is about the same as the weight of particles
collected from unleaded fuel during one LA-4 cycle so that the effects of
reactions between filter and any substantial part of the SO,, would invali-
date the particle weight data.
On the other hand, these neutral filter media change weight with
humidity changes and by adsorption of exhaust constituents. This disad-
vantage was compensated by using a second filter disc as backup in each fil-
ter holder, and by applying the change in weight of the backup filter as a
correction to the measured weight change of the primary filter.
-------
A-51
Pattern of Sampling for Particulate Matter
Four simultaneous samples were filtered for determination of par-
ticulate mass in almost all experimental tests: two were 142-mm Metrical
filters in stainless steel holders with conical approach sections, and two
were 47-mm Millipore filters in holders of similar configuration.
Two filter discs were placed in each holder, with the primary
filter for particle collection, and the backup filter as a weight-change
control. The weight change in each filter following exposure was determined
by separate weighings.
The efficiency of these filter media is so high that no solid
particulute matter can penetrate to a backup filter in the same holder.
Therefore, the algebraic difference between the weight changes of the primary
and the backup filter after both have been exposed in series to the same
sample stream is a measure of the weight of solid particulate collected on
the primary filter; It is assumed that the gas stream is not significantly
changed in composition by passage through the first filter and that weight
changes from adsorption or desorption of its constituents on the filter
structure are the same in each of the two filters.
Experimental Results Supplementary Data
Test Conditions
Table A-9 summarizes pertinent laboratory conditions and Table A-10
gives calculated air-fuel ratios for the modified and standard LA-4 cycle
test runs. All the runs are included where essentially complete data
were acquired. Problems were encountered on Runs 4-10 and 4-18, which are
noted on the tables. Any significant deviations in the test conditions
from target or normal values are described in footnotes of the tables.
The A/F ratios of Table A-10 were determined from CO, CO , 0
and HC measurements made in the tailpipe at steady-state conditions of 0,
35, and 50 mph immediately following each test run. Each entry is the mean
of three values determined from CO-0 , CO-CO , and 0 -CO , using the
.Eltinge. charts. C0_ and 0_ were corrected for HC. A fuel H/C ratio of
1.90 was assumed.
-------
A-52
TABLE A-9. LABORATORY AND OPERATING DATA FOR MODIFIED AND STANDARD LA-4 CYCLE RUNS
Run No.
Run-in,
Soak Avg. Soak Laboratory Chamber . Run
Tir.c, Temperature, Tfiiporature, Humidity, Humidity , Length
Y
K
Kr/lh dry
Unleaded Fuel Tests
percent:
4-10
4-11
4-12
4-13
4-16,
4-17^
4-18v
4-20
4-25
4-26
5-15
5-16
5-17
5-18
5-19
(f)
42
101
120
139
109
117
60
126
122
111
116
122
120
124
121
19
16
17
16
17
14.2
13.6
18.1
14.7
17.0
(h)
(h)
(h)
69
68
72
70
70
70
71
70
69
68
74
70
17.1
16.7
17.1
16.8
15.7
(h)
Leaded Fuel Tests
70 73
70 77.5
70 71
70 73
Unleaded Fuel Tests
70 70
70 70
70 69
70 71
70 71.5
Leaded Fuel Tests
37
29
35
31
48
48
75
68
37
44
37
40
35
39
46
28, .
33<8)
30,
37
41
(g)
(8)
50
35
30
26
(8)
(8)
30
30
29
27
27
23.41
23.25
23.37
23.28
23.26
22.75
25.13
23.53
23.37
23.35
23.37
23.40
23.40
22.83.
22.78
5-9
5-10
5-21
5-22
121
114
108
118
16.7
17.2. .
15.9
16.7
70
70
70
70
77
72
71
7S
59
67
51
62
31,,,.
42(f)
29
34
23.40
23.42
23.40
23.40
(a) Generally includes about 100 miles on the MV1-1A Durability Driving Schedule and two
consecutive LA-4 cycles.
(b) As measured by wet and dry bulb thermometers just before start of the test run.
(c) Time-sequence clock on to clock off 5 seconds after vehicle stops on last mode, unless
otherwise noted, this time period also represents the interval the exhaust-gas diverter
valve opens the tailpipe to the tunnel and the interval of tunnel-bag sampling.
(d) The tunnel-bag sampling was inadvertently continued for about 40 seconds after the
clock was stopped, sampling dilution air only. The mass gaseous emissions have been
corrected for a 3 percent dilution; the run-in mileage for this test was also low.
(e) The engine killed when put in gear, was restarted and test continued to completion with
about 104 seconds additional time before first mode started. The run-in mileage for
this test was also low.
(f) Tests 4-17, 5-18, and 5-19 arc standard FTP-cycle tests.
(g) The chamber humidity on Tests 4-10, 4-12, 4-16, 4-17, 4-18, 4-20, and 5-10 was above
target of 30 percent.
(h) The soak time on Tests 4-17, 4-18, 4-25, 5-19, and 5-21 was less than the target of
16 hours, but all were greater than the FIT requirement of 12 hours.
-------
A-53
TABLE A-10. AIR-FUEL RATIOS DETERMINED FROM STEADY-STATE
EXHAUST-GAS ANALYSES AFTER EACH RUN
Run No.
4-10
4-11
4-12
4-13
4"16(c)
4-18
5-16 / \
5-17)8(
5"18( )
5-19*1
5_9(g)
5-21
5-22
Idle
Unleaded
14.90
14.29
14.23
14'22(e)
15-01(e
14. 85^
Leaded
15.33
14.49
14.21
13.96
Unleaded
15.10$
14.97
14.66
14.64) ?
14.61(d)
Leaded
S:SS>
14.32
15.24
Air-Fuel Ratio
35 mph
Fuel Tests
15.23
15.46
15.39
15.43
15.26
15.38
FueJL Testa,
15.22
14.91
15.07$
15.02U;
Fuel Tests
15.60
15.57
15.59
15.59
15.48
Fuel Tests
15 31(h)
AJ * J J-/u \
14.79(h)
14.85
14.60
50 mph
tfm
15.15
15.02
15.10
15.03
15,16
14.90
l4'90fM
14.82(h)
15.20
-------
A-54
Table A-11 gives the overall dilution ratio in the residence
chamber of the samples taken for measurement of mass concentrations. The
target for overall dilution ratio was 300. Experimental variations are
caused by differences in the amount of filtered air held in the residence
chamber by partially collapsing the flexible walls. The tunnel sample was
then injected for the final dilution. The variations in ratio shown in
Table A-ll are acceptable..
-------
A-55
TABLE A-11. DILUTION RATIOS OF DILUTED EXHAUST IN RESIDENCE CHAMBER
FOR MODIFIED AND STANDARD LA-4 CYCLE RUNS
Run No.
4-10
4-11
4-12
4-13
4-16
4-17(b)
4-18
4-20
4-25
4-26
Overall
Dilution
Ratio(a)
__
223
284
278
268
293
278
262
271
252
Run No.
5-15
5-16
5-17
5-18
5-9
5-10
5-21
5-22
Overall
Dilution
Ratio
265
275
325
325
325
268
297
332
(a) Calculated using average dilution ratio in tunnel of 31.9 and
measured dilution ratio from tunnel to chamber using HC values.
(b) Tests 4-17, 5-18, and 5-19 are standard LA-4 cycle tests.
-------
ENVIRONMENTAL PROTECTION AGENCY
Technical Publications Branch
Off tc« of Administration
TriangU Park. N C 27711
POSTAGE AND FEES PAID
ENVIRONMENTAL PROTECTION AGENCY
EPA - 335
OFFICIAL BUSINESS
AN EQUAL OPPORTUNITY EMPLOYER
Return this sheet if you do NOT wish to receive this material EH,
or if change of address is needed CD. (Indicate change, including
ZIP code.)
------- |