SECOND ANNUAL REPORT
ON
GASOLINE COMPOSITION AND
VEHICLE EXHAUST GAS
POLYNUCLEAR AROMATIC CONTENT
Period Ending April 15, 1971
CRC-APRAC Project No. CAPE-6-68
APCO/EPA Contract CPA-70-104
Submitted to
The Coordinating Research Council, Inc.
and
Environmental Protection Agency
Esso Research and Engineering Company
Products Research Division
Linden, New Jersey 07036
-------
c_
SECOND ANNUAL REPORT
ON
GASOLINE COMPOSITION AND VEHICLE
EXHAUST GAS POLYNUCLEAR AROMATIC CONTENT
CRC-APRAC Project No. CAPE-6-68
APCO/EPA Contract CPA-70-104
Period Ending April 15, 1971
Submitted to
The Coordinating Research Council, Inc.
c/o
Mr. A. E. ^engel, Project Manager
Coordinating Research Council, Inc.
30 Rockefeller Plaza
New York, New York 10020
Esso Research and Engineering Company
Products Research Division
Linden, New Jersey
07036
Report By:
G. P. Gross
-------
CRC-APRAC PROJECT CAPE-6-68
SECOND ANNUAL REPORT
TABLE OF CONTENTS
LIST OF TABLES iii
PERSONNEL v
1. SUMMARY 1
2. INTRODUCTION AND OBJECTIVES 2
3. EXPERIMENTAL 7
3.1 Test Fuels 7
3.2 Production of Samples 8
3.3 Analysis of Samples 9
3.4 Test Vehicles 10
4. RESULTS 12
4.1 Phenol Emission 12
4.2 Statistical Treatment of PNA-Emission Data 18
4.3 Effects of Vehicle Emission-Control Systems on 23
PNA Emission
4.4 Immediate Effects of Fuel Composition on 27
PNA Emission
4.4.1 Fuel Aromatics Content 27
4.4.2 Fuel PNA Content 27
4.4.3 High-Boiling Naphtha 29
4.4.4 Antiknock Lead 29
4.5 Long-Term (Deposit) Effects of Fuel Composition 33
on PNA Emission
5. DISCUSSION 38
5.1 Effect of Carburetion 38
5.2 Phenol Emission 39
5.3 Emission-Control-System Effects on PNA Emission 41
5.4 Immediate Fuel Effects on PNA Emission 41
5.4.1 Fuel Aromatics Content 42
5.4.2 Fuel PNA Content 42
5.4.3 High-Boiling Naphthas 43
5.4.4 Antiknock Lead 44
5.5 Deposit Effects on PNA Emission 44
5.6 Areas for Further Study 46
-------
TABLE OF CONTENTS (CONTINUED)
APPENDIXES
A. REFERENCES 47
B. TEST FUELS 49
C. TEST VEHICLES AND TEST-OPERATIONS DETAILS 56
D. CORRELATION AND APPLICATION OF THE GC/UV PNA METHOD 85
E. ANALYTICAL PROCEDURES 90
ii
-------
LIST OF TABLES
Table No. Title Page
1 Summary Data Sheet 13
2 Key to Test Numbers 14
3 Summary of Phenol Emission Data 15
4 Corrected PNA Data for 1966 Plymouth 19
and 1968 Chevrolet for Two Contract Years
5 Summary of Benzo(a)pyrene Data 20
6 Summary of Benz(a)anthracene Data 21
7 Repeatability of PNA Emission Tests 24
8 Emission Control System Effects on PNA 25
Emission
9 Emissions from Low- and High-Aromatic Fuels 26
10 Effects of Fuel Aromatics on PNA Emission 26
11 PNA Emissions from Low- and High-PNA Fuels 28
12 Effect of Fuel PNA on PNA Emission 28
13 PNA Emissions with a High-Boiling Naphtha 30
Present
14 Effect of High-Boiling Naphtha on PNA 30
Emission
15 PNA Emissions with Varying Lead Concentra- 31
tion
16 Effect of Lead on PNA Emission 32
17 Sequence of Engine-Deposit Fuels 34
18 PNA Emissions with Deposits from Base A Fuels 35
19 Effect of Deposit-Fuel Composition on PNA 37
Emission
20 Phenol Emission Rates 40
B-l Fuel Compositions 52
B-2 Test Fuel Inspections 53
B-3 Mass Spectrometer Analyses of Test Fuels 54
B-4 GC Analyses of Aromatic Fractions from 55
Silica-Gel Separations of Test Fuels and
Blend Components
C-l Tests of Dubious Validity as to PNA 60
C-2 NDIR Hydrocarbon Emissions 61
C-3 through
C-43 Exhaust Analysis Data D/-OZ
C-44 Muffler-Exit and Exchanger-Inlet 83
Temperatures
C-45 Muffler Retention of PNA Species 84
iii
-------
Table No. Title Page
D-l PNA Analyses by TLC and GC Methods 87
D-2 Multiple-Species Analyses of Exhaust 88
Samples
D-3 Ratios of PNA Species in Exhaust Samples 89
E-l Observed Retention Times 100
E-2 UV Calibration Data for some PNA Hydrocarbons 102
iv
-------
PERSONNEL
(Second Year)
Supervisory Project Group
Mr. Charles R. Begeman, Chairman
Research Laboratories
General Motors Corp.
Mr. Harold F. Elkin
Sun Oil Company
Mr. Paul L. Gerard
Research and Development Dept.
Mobil Research and Development Corp.
Mr. D. S. Gray
Research and Development Dept.
American Oil Co.
Dr. A. J. Pahnke
Petroleum Laboratory
E. I. duPont de Nemours & Co., Inc.
Mr. T. W. Stanley (Project Officer for CPA-70-104)
Environmental Protection Agency
Project Manager for Coordinating Research Council
Mr. A. E. Zengel
Coordinating Research Council, Inc.
Project Manager for Esso Research and Engineering Co.
Dr. George P. Gross
Products Research Division
Esso Research and Engineering Company
Note;
The following personnel changes will be in effect for the start of
the third-year program of this Project:
1) Mr. J. S. Ninomiya, Ford Motor Co., new member of the
Project Group.
2) Dr. J. H. Somers, Environmental Protection Agency, replaces
Mr. T. W. Stanley as Project Officer.
v
-------
- 1 -
1. SUMMARY
Fuel composition, engine deposits and vehicle emission-control
systems have been studied as variables affecting the exhaust emissions
of phenols and of polynuclear aromatic hydrocarbons (PNA, measured as benzo-
(a)pyrene (BaP) and benz(a)anthracene (BaA)). This program was the second-
year continuation of research supported jointly by the Coordinating Research
Council, Inc., and the Environmental Protection Agency.
The exhaust from uncontrolled (1966) and engine-modification (EM)
(1968 and 1970) vehicles in 7-mode cyclic tests with cold starts was cooled
and filtered to collect PNA and phenols. Emission rates, per gallon of
fuel, ranged from 74 to 776 milligrams of phenol, 3 to 165 micrograms of
BaP, and 13 to 276 micrograms of BaA. Standard deviations for single
determinations were estimated to be 38 milligrams of phenol, and 25% and 11%
for BaP and BaA emissions, respectively. Standard calculated emission
results for CO, HC, and NO were obtained concurrently with the PNA tests.
Based on first-year results, PNA emissions from the 1966 and 1968 vehicles
were corrected to a uniform CO level for each vehicle. For the 1970 vehicle,
smaller variations in CO emission made PNA corrections unnecessary.
Fuel composition variables affecting emissions were found to
have both direct effects, which were observed in short-term tests over
existing engine deposits, and indirect effects, which appeared as emission
differences between different engine deposits. The fuel-composition vari-
ables that were studied with typical-volatility fuels in direct-effect
tests included aromatics content at 11-12%, 28%, and 46%; fuel PNA (as BaP,
0 and 3 ppm); fuel lead (0, 0.5, and 3 grams/gal.); and the content of a
high-boiling naphtha (0 and 16% at 12% aromatics, 0 and 2% at 46% aromatics).
Engine deposits were formed with commercial gasoline which provided a high
lead level with phosphorus and, using a different base fuel and in only
two vehicles, low-lead and no-lead levels without phosphorus. Comparable
tests in the three vehicles provided data on the performance of emission
control systems in controlling PNA and phenol emissions.
PHENOL EMISSION
Phenol emission from each vehicle correlated linearly with fuel
aromatics, approaching zero at very low aromatics; other fuel variables
had no effect. At typical aromatics levels, phenols from the 1970 vehicle
were down 30% from the nearly-equal 1966 and 1968 vehicles. Engine-deposit
type had no effect on phenol emission.
PNA EMISSION
PNA emissions from the 1968 and 1970 vehicles (EM) were similar
at an average of only 30% of the PNA from the 1966 uncontrolled vehicle.
The CO emissions were also similar at 0.6-0.8%, against 1.5% in the 1966
-------
- 2 -
vehicle; emissions of CO and PNA correlated. This implies that more-nearly
typical (higher-CO) 1966 and 1968 vehicles would emit more PNA than was
observed here.
In the 1966 and 1968 test vehicles, increasing fuel aromatics
from 117o to 467> caused PNA emissions to increase by an average of 1347,,
with BaP emission increasing more rapidly than BaA emission. In the 1970
vehicle, the effect of increased aromatics on PNA emission was not statis-
tically significant.
Increasing fuel PNA content (as BaP, from 0 to 3 ppm, i.e., the
field maximum) caused a 327o increase in BaA emission and a small, non-
significant increase in BaP emission.
A high-boiling naphtha (catalytically cracked naphtha + polymer,
FBP 419°F by ASTM D-86) present at 27, in a 467,-aromatic fuel did not signi-
ficantly affect the emission of BaP or of BaA. When present at 167o in a
1270-aromatic fuel, this naphtha apparently caused substantial PNA-emission
changes. However, increases in two vehicles were offset by decreases in
the third, with no conclusive over-all effect for the naphtha.
The presence of antiknock lead in fuel had variable effects
on PNA emission; both increases and decreases were observed, compared with
tests using unleaded fuels. Limited data at 0.5 grams Pb/gallon indicated
average decreases of 26% (BaA emission) to 43% (BaP emission). More
extensive data at 3 grams/gallon indicated no significant BaP change and a
17% BaA increase.
While tests in the 1966 and 1968 vehicles showed that increases
in fuel aromatics caused substantial increases in PNA emission, tests in
the same vehicles showed that changes in fuel-derived engine deposits also
affected PNA emission by a factor of more than two. As a part of these
tests, high-aromatic fuels produced PNA emissions which were not higher
than the PNA emissions from low-aromatic fuels under different, higher-
emission, deposit conditions. The essential properties of fuels for
forming low-emission or high-emission deposits are not clearly identified
at this time.
-------
- 3 -
2. INTRODUCTION AND OBJECTIVES
The first year of research on vehicular emission of polynuclear
aromatics (PNA) and phenols under CRC-APRAC Project CAPE-6-68 resulted
in a number of conclusions (references are given in Appendix A) which
influenced the definition of objectives 'for the second-year research pro-
gram. Second-year objectives also reflected an increased interest in
the use of low-lead and lead-free gasolines. In reflecting these in-
fluences, the second-year objectives unavoidably continued deferring action
on some of the original objectives in areas such as vehicle operating
modes, oil consumption, warm-up, spark timing, etc., which had been out-
lined in the original program plans but were not undertaken in the first year,
As a background for the second-year objectives, it is appropriate
to review briefly the highlights of the first-year conclusions as to the
effects of different controlled variables on the emission of PNA
[benzo(a)pyrene (or BaP) and benz (a) anthracene (or BaA) ] and phenols.
Generally, these conclusions were based on comparing results obtained
in test pairs which differed in respect to the variable under study, and
each variable was usually evaluated by several available test pairs. The
concurrently measured carbon monoxide (CO) emissions in the two tests of
a pair commonly differed, and this difference was shown to affect the PNA
emissions of the tests and thereby to influence the apparent effect of the
variable being studied. A formula was developed to take this CO effect
into account within a given vehicle and the conclusions reached are believed
to be reasonably independent of this source of error. However, where the
variable was "emission control system" and the "effect" is the difference
between the PNA emissions of two vehicles which differ as to the presence
of a control system for CO and HC emission, it was recognized that the
difference in CO between the two vehicles was, in fact, part of the variable
itself, rather than a source of error. Therefore, the CO-correction approach
was not applied to the between-vehicle studies. (In the present report,
the observed CO emission for every test in both years is used, together
with the first-year formula, to convert all observed PNA emissions from a
vehicle (1966 or 1968 only) to a standard CO level for that vehicle,
and even between-vehicle comparisons are then on a more uniform basis.)
The highlights of the first-year studies, particularly as they
influenced second-year plans, were:
1) Increases in fuel aromatics increased PNA emission, but to an
apparently lesser degree in a 1968 Chevrolet (with EM control)
than in a 1966 Plymouth without emission control (NC) ,
2) A high-aromatic fuel emitted less PNA, other things being equal,
when it contained more C-j aromatics, at the expense of C
aromatics; i.e., volatility or molecular weight played a9part .
-------
4 -
3) The presence, in a low-aromatic test fuel, of a small amount
of a high-boiling naphtha (a polymer/heavy catalytic naphtha
mixture) was suspected as the cause of a relatively high PNA
emission from the low-aromatic fuel and a consequent appearance
of relatively low sensitivity to increased aromaticity, per se.
This component was not rich in PNA content (i.e., BaP), however.
4) The presence, in a fuel, of PNA (measured as BaP) at the
maximum field level found in gasolines resulted in PNA emission
about 25% greater than if the fuel BaP content were near zero,
and this relative effect was nearly constant over different
fuels in two different vehicles (NC and EM), i.e., over a wide
range of emission levels. Very little of even a large amount
of BaP present in the fuel actually survived and was emitted,
or caused the emission of other BaP that would appear to have
survived from the fuel.
5) The presence of lead (3 ccs TEL/gallon) in fuel apparently
caused little or no change in BaP emission and an increase of
about 30% in BaA emission.
6) Accumulation of lead-free deposits in previously-cleaned com-
bustion chambers caused no consistent effect on PNA emission.
The deposits were accumulated in programmed city/suburb/highway
operation on a dynamometer, with an average speed of 30 mph.
7) Conversion of these lead-free deposits to leaded deposits (with
lead and phosphorus present in a different base fuel) slightly
more than doubled the emission of PNA. These were single-vehicle,
three-fuel results, and could be due to base-fuel composition, and/
or lead, and/or phosphorus.
8) Emission of PNA from leaded fuel used in the presence of leaded
(+ phosphorus) deposits was about double that from lead-free fuel
and lead-free deposits; the effect was apparently due largely to
the deposit change, in line with the preceding item, Tests of
only two fuels in one vehicle were available for data on this
point, and the result could derive either from lead, from phos-
phorus, or from the use of two different base fuels for deposit
formation.
9) The PNA emission from a 1968 (EM system) Chevrolet was about
30% as great as from a 1966 (NC) Plymouth, with comparable
deposits present in both and with five fuels tested in both
vehicles. The CO emissions of both vehicles were relatively
low, for the model years, so that PNA may have been relatively
low as well, but the relative between-vehicle comparisons of
PNA should remain valid.
10) An increase of 0.5% CO in the 7-mode cyclic test result from a given
vehicle was associated with an apparent 45% increase in PNA emission.
11) Phenol emission more than tripled on increasing fuel aromatics
from 12% to 46%; lead appeared to have a small suppressing
-------
- 5 -
action on phenols, while deposit accumulation increased phenols.
The 1968 emission control system decreased phenols only slightly,
if at all, compared to the 1966 uncontrolled case. PNA in fuel,
and fuel volatility at constant aromatics, did not affect phenol
emission.
Most of these results, particularly as they relate to PNA . ..
emission, are in at least directional agreement with findings by Begeman, ' '
by Griffing(5) and by others. The most unexpected, and significant, results
were those relating to a direct (but small) effect of lead, to an effect of
deposits, and to a suspected effect of high-boiling materials (aromatics?) in
fuels, lying between the single-ring aromatics and the 4-ring or 5-ring poly-
nuclear aromatics. The apparent reduction in PNA (but not phenol) by the 1968
emission-control system, while not expected, was judged important enough to
justify tests in an additional controlled vehicle and to encourage testing in
advanced experimental low-emission vehicles.
The second-year objectives developed from these considerations
included the study, with respect to emissions of PNA and phenols, of
the following factors:
1) The effects of lead antiknock (a) instantaneously, as a result
of its presence in the test fuel, with a low, non-zero, level
included, and (b) long-term, as lead residues appear in deposits,
or disappear by extended use of reduced-lead and zero-lead
gasolines.
2) The effects of fuel aromatics content, including an intermediate
level around 30%.
3) The effects of fuel PNA content.
4) The effects of high-boiling aromatics, as distinct from total
aromatics, such as might occur in heavy catalytic naphthas or
heavy catalytic reformates and which are not PNA but may act
as PNA precursors.
5) The effects of current exhaust emission control systems, extending
the study of the above factors beyond the 1966 (NC) and 1968 (EM)
vehicles to a 1970 EM vehicle.
6) The effects of advanced low-emission experimental systems (thermal
and catalytic) as means of providing control on PNA and phenol
emission; this heading also includes the effect, on PNA and phenols,
of systems, either the same or separate, for controlling nitrogen
Oxide emiss-iona..
-------
- 6 -
It was foreseen that these objectives would be approached by the
use of new test-fuel blends prepared without the high-boiling naphtha that
was previously used, and including a cross-blend at an intermediate aroma-
tics level. It was further foreseen that the lead-level experiments re-
quired should use this cross-blend as a base fuel, with "no-lead," "low-
lead," and "high-lead" levels represented. Deposit studies would employ
similar lead levels but in necessarily different base fuels. As far as
possible, the second-year work was to be comparable with results already
obtained in the first year. In the first year, however, deposits had been
changed from no-lead to leaded in one car, while for the second year the
change was to be in the opposite direction (leaded to no-lead), and in two
or more cars.
Most of these objectives have been attained in, at least, a limited
way. In few cases can it be assumed that work is complete, however, and
no vehicles with experimental low-emission control systems were available for
use.
The results obtained with the 1970 vehicle present a number of
inconsistencies compared to earlier results, so that certain earlier con-
clusions from the first two vehicles are now less clear than they appeared
to be initially. Specifically, the 1970 vehicle, with relatively leaner
carburetion and with a retarded spark, emits less phenol than the earlier
vehicles, and its PNA emisssions are less sensitive to some fuel composi-
tional variables. Only one deposit condition was studied for the 1970
vehicle.
Two additional objectives have been achieved, although they were
not formally specified for the second year: (1) The PNA accumulated in
a vehicle muffler (three different vheicles) during a test has been evaluated
(see Appendix C) in relation to the PNA emitted at the tailpiper (2) The
analytical methods for PNA have been converted to the GC/UV method developed
by CAPE-12-68 (Appendix E), so that results are available on a larger variety
of PNA species in each test. Data interpretation is still based on the
emission of BaP and BaA, however, since these are the only PNA results for
earlier tests.
-------
- 7 -
3. EXPERIMENTAL
3.1 Test Fuels
The considerations involved in preparation of the test fuels
used for first-year work are given at length in Section 4 and Appendix B
of the First-Year Report(1). Prominent among these considerations was
the requirement for volatility (Engler D86 distillation) properties,
including the 90%-FBP range, that were commercially typical.
For the second year (see Appendix B of this report) these con-
siderations continued to apply but presented new problems because the
second-year objectives required that blends be prepared which were as
similar as possible to the first-year blends but which did not contain
the high-boiling naphtha (heavy catalytic naphtha/polymer/pentanes)
that was used in the first-year blends. Appendix B of this report
gives the compositions used, with identically the same components other
than the heavy naphtha, to prepare the new (1970) basic blends for com-
parison with the earlier (1969) blends. Appendix B also contrasts the
distillation and other properties of the two sets of blends, as summarized
below:
Aromatics:
Year:
RVP
IBP °F
10%, °F
50%
90%
FBP, °F
% Arom.
% Olef.
FUEL PROPERTIES
Low
1969
10.6
82
112
218
320
400
12
6
1970
10.5
92
116
215
296
384
11
1
Mid
1970
10.2
91
116
217
310
383
28
2
High
1969
10.7
84
112
220
332
387
46
5
1970
10.5
92
116
222
322
381
46
1
The second-year (1970) blends match the 1969 blends reasonably well
except around the 90% - FBP range of the low-aromatic fuel. Here the
exclusion of the heavy naphtha was only partly compensated by the increases
made in certain aromatic fractions (see Appendix B). The 1970 base gaso-
lines (11% and 46% aromatics) were low in PNA content, as were the 1969
fuels. PNA addition, when required, was done in both years by adding a
384+° FVT reformate bottoms, at 22 grams per gallon, thus adding 8000 yg
of BaP per gallon and substantial amounts of other PNA species. (The
8000 yg BaP level was approximately the maximum found (First-Year Report)
by Esso Research in a field sampling study, not at CAPE-6 expense, and
-------
- 8 -
disclosed to CAPE-6 for guidance in preparing the blends.) Lead contents
of 0, 0.5, and 3 gms./gallon were used in the 1970 fuels; the 0.5 level
was used only in the mid-aromatic fuel, and the low-aromatic fuel was
tested only with lead at the 3-gm. level.
The exclusion of the high-boiling naphtha (HBN) in formulating
the second-year (1970 set) test fuels was prompted by the suspicion that
its presence at 16% in the first-year (1969) low-aromatic (12%) fuels was
causing misleadingly high PNA emissions. These, in turn, caused mis-
leadingly small apparent effects for increasing fuel aromatics from 12% to
46%, since only 2% of the HBN was present in the 1969 46%-aromatic test
fuel. When this suspicion was confirmed (Section 4) by the finding of
lower emissions for the 1970 low aromatic (no HBN) fuels than for the 1969
fuels, the 1970 set became the basic reference blends for evaluating fuel
variables in the present report, and data from the 1969 fuel set provided
a measure of the effect of the presence of the high-boiling naphtha.
Different fuels, in larger quantities, were used for deposit
accumulation or modification by operation of the test vehicles for periods
of 1000-4000 miles on a Mileage-Accumulation Dynamometer (see Appendix C.)
Where "leaded" deposits were specified in the first-year report (see foot-
note on page 9 of that report) the fuel used contained about 2.5 gms. of
Pb per gallon, along with 0.2 theories of phosphorus, and was a commercial
leaded premium fuel, or a prototype blend of that type. In this report,
this fuel is coded as C3P. At the beginning of the second year, the 1966
and 1968 test vehicles contained such lead + phosphorus deposits, as did the
1970 vehicle when it was first used during the year. During this year, low-
lead and lead-free fuels were also used for deposit changes in the 1966 and
1968 vehicles. These fuels, which were free of phosphorus, were based on
a commercially-available, high-aromatic, (35-407=,) lead-free premium
gasoline, coded as fuel A in this report. Motor Mix TEL was added, at 0.5
gms. of Pb/gallon, to a portion of this fuel to prepare the "low-lead"
blend with the code symbol A.5. Thus the high-lead deposits were invariably
associated with one type of base gasoline and had phosphorus present. Con-
versely, the low-lead and no-lead deposits were invariably formed from a
different type of base fuel and did not contain phosphorus. These dis-
tinctions as to phosphorus and as to base fuel in addition to lead level
may be important in interpreting the deposit-effect results in this report,
and exploration of their significance is one part of the third-year program.
At the same time, any deposit effects from either base fuel, lead, or
phosphorus can be detected, so that concurrent changes in all the variables
is the most efficient procedure for initial studies.
3.2 Production of Samples
The procedure and equipment for sample production were unchanged
from the first year and are fully described in the first annual report.
The test vehicle is operated on a Clayton chassis dynamometer in three
cold-start blocks of twelve 137-second, 7-mode cycles each. The vehicle
drives about 29 miles and uses about 2-2.4 gallons (weighed) of fuel.
-------
- 9 -
The total exhaust, except a sidestream for infrared gas analysis, is
cooled and filtered and the condensed water is recovered. The samples
for analysis include the water condensate (90+% of all phenols), the
filter media, and the solvent washings of all surfaces from the inlet
end of the vehicle muffler to the outlet side of the filter; most of
this surface is on the walls of the 84-tube heat-exchanger. The side-
stream scrubber tower on the filtered exit gas was not used in the
second year, because first-year results showed it to be of doubtful
value for PNA or phenols. As shown in Appendix C of the present report
and on the basis of 18 tests, the muffler sample might often be omitted
with little loss, but the need to clean the muffler anyway means the
sample may as well be included and this is normally done.
3.3 Analysis Of Samples
Analytical techniques for phenols were unchanged in the
second year from the descriptions in the previous report. The aqueous
condensate ususally contains over 90% of all the phenols, while the
Soxhlet extract of the fiber glass filter media contains no phenols.
Although many different phenols are present, the data are reported in
terms of phenol, i.e., the parent compound. CRC-APRAG-CAPE 12-68 at
Esso Research is currently studying techniques for estimating individual
phenols in the sample.
Analyses for PNA in exhaust samples in the early part of the
second year (through Test 32) continued to be done with a TLC (thin-layer
chromatography) finishing step, as described in the first-year report.
Concurrently, CAPE-12-68 analyzed samples from twelve CAPE-6 tests by
the GC/UV method which they had developed. The data by the two methods
are compared in Appendix D, and Appendix E, prepared by CAPE-12, is a
detailed description of the procedure for the GC/UV method. The
regression equations for the data in Appendix D are
BaP by GC = 8.61 + 0.997 (BaP by TLC), with R = 0.997,
and BaA by GC = 15.36 + 0.830 (BaA by TLC), with R = 0.990.
Viewed differently, averaging the ratios of GC/TLC leads to
BaP by GC = 1.16 (TLC), and
BaA by GC = 0.94 (TLC).
These comparisons indicate reasonable agreement between the methods.
(The BaA by GC is probably more accurate than by TLC, according to
CAPE-12 studies, because the TLC band for BaA, when analyzed by mass
spectroscopy, includes the methyl derivatives.) However it appears
preferable to compare results within a given method, where this is
possible, in examining the effect of a controlled variable.
-------
- 10 -
Tests earlier than number 26 include five tests covering a
range of yields, for which both TLC and GC analyses were done (see
Appendix D); for 26-32, both methods were used, and for tests 33-59
only GC analyses are available. Test 26 marked the beginning of a
study of lead levels, using the 28%-aromatic fuel. In order to avoid
ambiguity in the data that might arise over the choice of analytical
results, this report uses GC/UV results for PNA yields on test 26 and
all tests thereafter. However, previously-issued quarterly reports
used TLC data through test 31, and GC results thereafter. Thus
certain values here are slightly different from those previously
reported. Another, smaller, variation from the quarterly report data
also occurs because the present data, in some cases, reflect higher-
precision radiotracer counting data than were used in previously
issued reports.
The GC/UV method of exhaust-sample PNA analysis makes
available the yield data on up to eleven different PNA species. In the
earliest tests (26-31), only BaP and BaA were determined, but six
species are known for most tests through number 54, with eleven species
assayed thereafter, in some of the earlier tests, some GC peaks were
not successfully trapped and yield results are missing in these cases.
The available results are included in Appendix D, along with the
calculated ratios between certain species. These ratios may suggest
some differences in sample composition that is related to other
variables, but no clear-cut conclusions are now available in this area.
3.4 Test Vehicles
A detailed discussion of the test vehicles appears in Appendix C,
At this point, it is appropriate to compare the three vehicles briefly as
to exhaust-gas composition (CO and oxygen) and temperature (at the muffler
exit). These factors appear likely to influence the survival of PNA in
the exhaust system and its ultimate emission. The importance of CO (expres-
sed as the cyclic-test results) was demonstrated in the first-year report
and is emphasized again in a later section of this report, fundamentally,
the presence of CO is to be taken as a symptom of an oxygen-deficient con-
dition that should favor the formation and survival of PNA. The CO levels
found for the 1966 and 1968 vehicles were only about half of the typical
values for these model years, so their PNA may also be relatively low.
Exhaust oxygen (as %) was estimated for each vehicle from average values of
the computer-recorded correction factors used in cycle-test calculations to
force % CO + % C02 to 15.00. The correction factors were taken as measures
of air dilution and the oxygen contents calculated accordingly. Exhaust
temperatures (muffler exit, Table C-44) were taken as the average of hot-
cycle maximum values at the peak of the 15-50 MPH acceleration mode. The
table on the next page compares the vehicles with respect to CO, 02, and the
exhaust temperature; reference to the differences will be made at a later
point in this report.
-------
- 11 -
VEHICLE EXHAUST CONDITIONS
Avg. % CO Est. Avg. Avg. Max.°F.
Vehicle (Fed. Test) % 02 Muffler Exit
1966 Plymouth 1.49 1.2 568
1968 Chevrolet 0.74 1.8 599
1970 Chevrolet 0.65 2.6 706
Oil consumption rates of the vehicles were observed (Appendix C)
in each operation for mileage accumulation in changing engine deposits. The
consumption rates remained low, at around 0.5 quarts/1000 miles; high oil
consumption could be expected to influence PNA emission, according to the
results reported by Begeman (4).
-------
- 12 -
4. RESULTS
A summary of data on all tests for the second contract year
is presented in Table 1. This table includes data on observed emission
rates of benzo(a)pyrene (BaP), benz(a)anthracene (BaA), and phenols
(as phenol), all per gallon of test fuel consumed. Fuel consumption,
aqueous condensate recovery, ambient weather, and computed Federal
cycle emissions results (1968-1969 model-year method) are also included.(a)
Tables C-3 through C-43 in Appendix C present average results from one to
three test blocks for each mode in six of the twelve cycles that form a
block in a sample-production operation. The extreme right-hand columns of
these Appendix C tables present the Federal cycle results that appear in
Table 1.
Table 2 provides a test-number key to all the PNA/Phenol
emissions tests since the start of Project CAPE-6-68, with the exception
of a few tests specified in a footnote on the table. Reference to
Table 2 will, immediately indicate whether a given vehicle/deposit/test
fuel case has been tested and specify the number of the test. All test
pairs whose comparison can serve to evaluate a given controlled variable
can be readily selected from this table. Similarly, sets of tests can
be selected whose data can form a matrix for statistical analysis.
Finally, this same format, with yields/gallon of fuel inserted in place
of test numbers, can provide useful tabulations (in three tables) of
emission rates for phenol, BaP, and BaA, within which the same selections
of test pairs, etc., can be made. In the following, the effects of
variables will be discussed in terms of data from both contract years,
rather than from only the second year.
4.1 Phenol Emission
The format of Table 2 is used in Table 3 to present the phenol
emission data for both contract years. Inspection of the data immediately
indicates a dependence on fuel aromatics content, as was shown in the
first year, and it also appears that the 1970 vehicle emits phenol at a
lower level, for a given fuel, than the 1966 or 1968 vehicles. The early
data in the 1966 Plymouth show a deposit effect which was interpreted, in
the first year, as meaning that leaded deposits (C3P) caused higher emission
than the (preceding) unleaded deposits (A). Newer data, in both the 1966 and
1968 vehicles, examine the phenol level after returning from C3P deposits,
through low-lead (A.5), to the no-lead (A) condition, and it is clear that
this change did not reduce phenol emission in either vehicle. It now appears
more likely that the deposits for the early no-lead (A) tests, which
(a) For consistency, results from all three vehicles (1966, 1968, and 1970)
are presented on the basis of concentrations (ppm or %). For the 1970
vehicle (4000-Ib. class), the factors of 23.8 and 0.0127, respectively,
will convert the tabulated % CO and ppm HC to the basis of grams/mile.
-------
- 13 -
p -,, , 0.-0- , ,_, ._», ._ _ , .._ _
u . - . . ... . . . . . ...
^ ,-t ^-i r-i r-t r-t ^-i^-i^-i^-i^-i,-! o OOOOO O OOOOO OOO OO OO O OOOOO OOO
t-i !-!,-< CM ,-1 ^-i ^-i n-( ^-( ^ ^-1 CM ,-1 CMO OOO OO O>O> O\ OCT>OOO OOO
imoom o omooo
-------
- 14 -
Table 2
CAPE -6-68 -SECOND YEAR
REPORT
KEY TO TEST NUMBERS
(Covering Tests for Two Years fa))
Number of PNA/Phenol Test for Indicated
Test Vehicle and Engine Deposits (b)
Fuel
Set
1969
1970
(a)
(b)
Test Fuel
% BaP,
Arom. yg./g.
12 0
8K
46 0
8K
11 0
8K
28 0
46 0
8K
1966 NC
TEL No
cc./g. Deposit A
3
0 39
3 4 12
0 18
3 2 11
3
3
3
0
0.5
3
0
3
3
Plymouth
C3P A. 5
17
16
20
19
18
22
23
27 34
28 32
26 37
1968 EM Chev.
A C3P A. 5 A
13
7
42 21 39
5,6
14
24
25
41 30 35 38
31 33
29 36
43 40
Test 10 (special fuel) and tests of doubtful validity (see Table 1) are
Engine deposits are specified by the fuel used for deposit formation.
1970 EM
Chev.
C3P
47,58
46
48
51B
44,52
57
56
55
53
49
59
omitted.
A com-
mercial leaded premium fuel with phosphorus (CodeC3P ), a commercial unleaded
premium (A), and the latter with 0.5 grams Pb/gallon (A.5) are included.
-------
- 15 -
Table 3
CAPE -6-68 -SECOND YEAR REPORT
SUMMARY OF PHENOL EMISSION DATA
(Covering Tests for Two Years(a))
Phenol Emission, mg./gal., for
Indicated Test Vehicle and Engine Deposits(b)
Fuel
Set
1969
1970
Test Fuel
% BaP,
Arom. yg./g.
12 0
8K
46 0
8K
11 0
8K
28 0
46 0
8K
TEL
cc./g.
3
0
3
0
3
3
3
3
0
0.5
3
0
3
3
1966 MC Plymouth
No
Deposit A C3P A. 5 A
165
117 154
117 148 179
498 565 732 772
358 487 689
663
113
110
350 279 435
424 389
379 334
776
1968 EM Chev.
C3P A. 5 A
116
165
685 691
663,667
663
78
88
380 346 314
406 370
348 401
607
1970 EM
Chev.
C3P
107,119
106
434
78
123, 74
287
259
256
420
460
485
(a) Test 10 (special fuel) and tests of doubtful validity (see Table 1) are omitted.
(b) Engine deposits are specified by the fuel used for deposit formation. A commercial
leaded premium fuel with phosphorus (Code C3P) , a commercial unleaded premium (A),
and the latter with 0.5 grams Pb/gallon (A.5) are included.
-------
- 16 -
were formed from 4,000 miles of operation after the clean-chamber tests,
were not at equilibrium; i.e., that additional miles on the no-lead (A) fuel
would have caused greater emission of phenol. Evidently, deposit
accumulation to an equilibrium level does increase phenol emission, but
the equilibrium phenol emission level is not strongly affected by the
composition of the deposits. On this basis, linear regressions of all
the data (phenol yield vs. % aromatics) for each vehicle except the first
two columns of data from the 1966 vehicle have been calculated, making
the implicit additional assumption that other variables, such as TEL and
PNA, are not important.
The regression equations take the form
Phenol = a + b (% Aromatics),
where a and b are constants and phenol is expressed in mg./gallon of
fuel burned. The results of the analyses for the three vehicles are
summarized below and are presented graphically in Figure 1.
PHENOL REGRESSIONS
Test Vehicle
1966 NC 1968 EM 1970 EM
Number of points 16 17 13
Intercept, a -82.0 -74.4 -15.2
Slope, b 17.2 15.9 10.1
Std. error in b 1.03 0.58 0.39
Std. error of estimate 53.3 31.4 20.8
Correlation coefficient 0.976 0.990 0.992
Variance due to regression, % 95.2 98.0 98.4
The negative intercepts predicted by linear regression suggest that the
true relationships are not linear. It is more likely that phenol emis-
sion approaches zero asymptotically as aromatics content decreases
toward zero. Aromatics levels below 11% were not tested. The regres-
sions against aromatic content explain over 95% of the total variance
in the data from each car, so it appears that no other variable (TEL,
PNA, etc.) is important, as long as equilibrium deposits are present.
The 1966 and 1968 vehicles are apparently equivalent in terms of their
phenol emissions, since their regressions are not significantly differ-
ent. The 1970 vehicle, however, emits only about 70-75% as much
phenol as the 1968 vehicle when typical fuels are considered:
Phenol, mg./gal.
1968
Fuel Aromatics
10%
20
30
40
1968
Estimate
85
244
403
562
Est.
86
187
288
389
1970
% of 1968
101
77
71
69
-------
- 17 -
700
< 600
o
500
o
CO
o
s
QJ
31
Q.
400
300
200
100
0
I
I
10 20 30 40
% AROMATICS IN FUEL
50
Figure 1
Effects Of Fuel Aromatics And Emission
Control Systems On Phenol Emission
-------
- 18 -
According to the regressions, phenol emissions from all three vehicles
are equal at about 10% aromatics in the fuel, i.e., just below the lower
limit of the data. The 30% reduction in phenol that is found for the
1970 vehicle at higher aromatics contents, where the phenol levels of
all vehicles are substantially higher, is clearly the more important
conclusion, however.
4.2 Statistical Treatment Of PNA-Emission Data
In the first-year report for Project CAPE-6-68, the emission
of PNA was shown to be correlated with the level of carbon monoxide (CO)
emission, and the interpretation of first-year data was substantially
improved by taking this effect into account. The Federal cycle % CO
result, as determined at least once in each PNA emission test, was
apparently a suitable measure of CO for the purpose. It was observed
that if, between two PNA tests which differed as to some controlled
variable, there was also an increase in CO, then the apparent relative
PNA emission affect of the variable was unduly increased along with
the CO change; the converse was also true. Where several measures of
the relative PNA effect of a variable were available, along with as-
sociated between-test differences in CO emission, the results could be
correlated to define a best estimate of the effect of the variable as
that effect which appeared likely to occur if there were no attendant
change in CO emission. By this means, an estimate of the change in
PNA emission for any change in CO emission was obtained for use within
a given vehicle (1966 Plymouth or 1968 Chevrolet) and with other factors
constant. The relationship found (page 38 of reference 1) was
Y = Y /(I + 0.9(C. - C )),
S u t S
where Ys and Yt are respectively the PNA yield rates (yg./gal.) at the
standard CO level (Cs, in %) and at the emission-test CO level (Ct) ,
This relation was applicable to both BaP and BaA yield rates, but presumably
not beyond differences of about + 0.5% CO from a selected standard.
Tfyis relationship can be used to calculate correction factors for con-
verting PNA yield rates to a standard CO level, as typified in the fol-
lowing table.
PNA CORRECTION FACTOR FOR
CO-LEVEL VARIATION
Lean Mixture Rich Mixture
Ct - Cs, %: -0.6 -0.4 -0.2 +0.2 +0.4 +0.6
Corr. Factor: 2.17 1.56 1.22 0.85 0.74 0.65
The use of a correction factor of this type agrees directionally with the
reports by Begeman (4), Griffing (5), and Gofmekler (6) that rich mixtures
result in increased PNA emission.
-------
'66 Plym.
'68 Chev.
- 19 -
Table 4
CAPE-6-68 SECOND-YEAR REPORT
CORRECTED (a^ PNA
AND 1968 CHEVROLET
Test No.
1
2
3
4
8
9
10
11
12
16
17
18
19
20
22
23
26
27
28
32
34
37
41
42
43
5
6
7
13
14
21
24
25
29
30
31
33
35
36
38
39
40
% CO
2.08
1.92
1.67
1.78
1.52
1.56
1.47
1.62
1.95
1.43
1.39
1.27
1.70
1.31
1.40
1.45
1.40
1.42
1.57
1.66
1.54
1.57
1.45
1.38
1.20
0.68
0.64
0.75
0.96
0.85
0.66
0.72
0.79
0.85
0.60
0.91
0.77
0.80
0.60
0.76
0.70
0.62
DATA FOR 1966 PLYMOUTH
FOR TWO CONTRACT YEARS
PNA
BaP
Observed
54
118
49
91
55
59
38
60
79
85
70
131
149
104
45
35
54
72
67
20
37
31
37
51
45
29
27
26
28
38
26
9.7
13
38
8.8
9.8
3.4
14
2.8
14
10.5
9.9
Emission,
Corr .
36
86
43
73
54
56
39
54
56
91
78
165
127
125
50
37
59
78
63
18
36
29
39
57
62
32
32
27
25
36
30
10.5
13
36
11
8.9
3.5
14
3.4
15
12
12
yg./gal.
BaA
Observed
142
264
168
214
96
133
80
138
219
198
164
219
303
163
99
128
143
120
92
34
51
49
49
73
76
67
54
81
59
89
48
37
75
78
23
22
13
15
12
16
23
19
Corr
94
193
146
171
94
126
82
124
155
212
182
276
258
196
109
134
157
130
86
30
49
46
51
82
104
75
63
85
52
85
55
40
75
75
28
20
13
15
15
17
25
23
Corrected to 1.5% CO for 1966 Plymouth and 0.8% CO for
1968 Chevrolet. Data through Test 25 by TLC analysis,
then GC/UV analysis.
-------
- 20 -
Table 5
CAPE -6-68 -SECOND YEAR REPORT
SUMMARY OF BENZO(a)PYRENE DATA
(Covering Tests for Two Years(a))
BaP Emission, yg./gal., for
Indicated Test Vehicle and Engine Deposits(c)
Fuel
Set
1969
1970
Test Fuel
% BaP,
Arom. yg./g.
12 0
8K
46 0
8K
11 0
8K
28 0
46 0
8K
TEL
cc./g.
3
0
3
0
3
3
3
3
0
0.5
3
0
3
3
1966 NC Plymouth(b)
No
Deposit A C3P A. 5 A
78
43 56
73 56 91
36 54 125 57
86 54 127
165
50
37
78 36 39
63 18
59 29
62
1968 EM ChevP>)
C3P A. 5 A
25
27
30 12
32,32
36
10.5
13
11 14 15
8.9 3.5
36 3.4
12
1970 EM
Chev.
C3P
12,20
10.5
29
19
20,34
48
36
26
26
17
28
(a) Test 10 (special fuel) and tests of doubtful validity (see Table 1) are omitted.
(b) PNA yields from the 1966 NC and 1968 EM vehicles are corrected to standard levels
of CO emission, by equations based on first-year data.
(c) Engine deposits are specified by the fuel used for deposit formation. A commercial
leaded premium fuel with phosphorus (Code C3P), a commercial unleaded premium (A),
and the latter with 0.5 grams Pb/gallon (A.5) are included.
-------
- 21 -
Table 6
CAPE -6-68 -SECOND YEAR REPORT
SUMMARY OF BENZ(a)ANTHRACENE DATA
(Covering Tests for Two Years(a))
BaA Emission, yg./gal., for
Indicated Test Vehicle and Engine Deposits(c)
Fuel
Set
1969
1970
Test Fuel
% BaP,
Arom. yg./g.
12 0
8K
46 0
8K
11 0
8K
28 0
46 0
8K
TEL
cc./g.
3
0
3
0
3
3
3
3
0
0.5
3
0
3
3
1966 NC Plymouth (b)
No
Deposit A C3P A. 5 A
182
146 126
171 155 212
94 94 196 82
193 124 258
276
109
134
130 49 51
86 30
157 46
104
1968 EM Chev. (t>l
C3P A. 5 A
52
85
55 25
75,63
85
40
75
28 15 17
20 13
75 15
23
1970 EM
Chev.
C3P
28,31
38
52
57
70,72
70
61
51
60
42
63
(a) Test 10 (special fuel) and tests of doubtful validity (see Table 1) are omitted.
(b) PNA yields from the 1966 NC and 1968 EM vehicle are corrected to standard levels
of CO emission, by equations based on first-year data.
(c) Engine deposits are specified by the fuel used for deposit formation. A commercial
leaded premium fuel with phosphorus (Code C3P), a commercial unleaded premium (A),
and the latter with 0.5 grams Pb/gallon (A.5) are included.
-------
- 22 -
However, it is doubtful that such a relationship will be useful
for vehicles carburetted near, or leaner than, stoichiometric. For this
reason, and lacking data suitable for demonstrating its validity, this
relation is not regarded as applicable to the recently-obtained PNA data
from the 1970 Chevrolet test vehicle. Instead, the PNA emitted by the
1970 vehicle may reflect exhaust oxygen content and temperature. As pointed
out earlier in Section 3.4, the average oxygen content of the 1970 vehicle's
exhaust was estimated at 2.6%, compared with 1.8% for the 1968 vehicle, even
though the CO contents were similar. The higher oxygen content in the 1970
vehicle exhaust resulting from leaner carburetion, together with a hotter
exhaust, resulting from retarded spark timing, suggests that some PNA formed
in the engine may be destroyed before being emitted from the exhaust system;
this could, in turn, introduce an added source of variability in PNA results
from the 1970 vehicle.
All PNA emission-rate data, from both contract years, for the
1966 and 1968 vehicles have now been "corrected," by the above equation,
to the expected rates at standard (near-average) CO levels of 1.5% and
0.8% for the 1966 and 1968 test vehicles, respectively. The CO levels
used are the composite computed results from Federal 7-mode tests, averaged
over from one to three of the three 12-cycle blocks used to form a PNA
sample. The resulting "corrected" PNA yields can be compared directly,
without the need to quality the comparisons by noting concurrent change in
% CO. The corrected PNA data are listed in Table 4. Tests 15 (see first
year report) and 54 (see Appendix C, this report) are omitted, as they do
not provide useful data.
These corrected PNA yield data, for both years, are presented in
Tables 5 (BaP) and 6 (BaA) in the same format that is used for Tables 2
(Test Numbers) and 3 (Phenols). The tables also include PNA data from the
1970 Chevrolet test vehicle. These latter data are not corrected for CO
variation, for reasons given above. Tables 5 and 6 provide the data base
for statistical analysis and for the presentation of the effects of dif-
ferent variables on PNA emission. Data from both years are used where
applicable. Where replicate tests are listed, the average result is used.
Statistical analyses were used to identify and quantify the
significant variables in PNA emission; BaP and BaA emissions were analyzed
separately. The variables were found to have fairly constant relative
(%-change) effects, so that the effect of a variable could be appropriately
stated as the geometric mean of ratios between the results from a number
of test pairs with the variable at two levels. Statistical analyses using
the data in logarithmic form for all pairs related to each variable then
lead directly to the geometric mean ratios as best estimates of the effect
of each variable. All mean ratios (R^) reported here are geometric means
of ratios observed in from two to fifteen pairs of tests. Where a need was
apparent, three-factor analyses of variance were used to detect interactions,
such as the effectiveness of a variable in two vehicles but not in a third
vehicle; parts of the data were then analyzed in smaller two-factor analyses
to redefine the geometric mean ratios.
-------
- 23 -
In presenting the results for different variables, each geometric
mean ratio is accompanied by a statement of the 95%-conf idence interval.
The interval assists in determining what effects are significant and whether
two mean ratios are meaningfully different. Each interval results from
multiplying, and dividing, the geometric mean ratio by a factor based on the
number of pairs in the mean, and on the estimated standard deviation in the
determination of the logarithm of a single BaP (or BaA) yield, in yg/gallon.
These estimates of standard deviations in log^Q of the yield were
initially based on four replicate-test pairs. Three other pairs were added
as replicates after being shown to be internally equivalent. Table 7 lists
the replicate pairs used for the estimates. Reference to Table 2 indicates
that designating Tests 8 and 42 as replicates, in the first line of the
table of replicates, implies that emissions for deposits from a given fuel are
reproducible even if other deposits have intervened. The next three lines
represent true replicates, while the last three lines include three test pairs
that were classified as replicates after data-analysis showed no difference
between the PNA from 46%-aromatic fuels with, or without, 2% of a high-boiling
naphtha (see Section 3.1). The estimates of standard deviation are for
single determinations of the log^o of the yield in ug/gallon. These corres-
pond, approximately, to standard deviations of 25% and 11% for one determina-
tion of the BaP and BaA yields, respectively. The relatively poor repeati-
bility for BaP is due to three test pairs from the 1970 vehicle.
4.3 Effects Of Vehicle
Emission-Control Systems On PNA Emission
The effects on PNA emission of using emission-control-equipped
vehicles were measured by comparing results from the 1966 vehicle (no control)
with fifteen comparable results from the 1968 vehicle and with ten comparable
results from the 1970 vehicle. Relevant data are given in Table 8. Within
the 95% confidence limits, the two control-system vehicles were equivalent
and emitted 30% as much of both PNA species as was emitted by the 1966 uncon-
trolled vehicle; i.e., a 70% reduction in PNA emission is indicated for the
production-vehicle control systems. It should be recalled (Section 3.4)
that the average CO emissions of the 1966 and 1968 vehicles were about half
of typical or permitted levels. It follows that more typical vehicles for
these years might show somewhat higher emissions of both CO and PNA, but the
relative rates (for 1968/1966) should remain similar to those found here.
-------
- 24 -
Table 7
REPEATABILITY OF PNA EMISSION TESTS
Emission
Test Nos. In Rates In Repeat
Vehicle Replicate Tests, yg/gallon
Year Pair BaP BaA
Fully-replicate pairs:
1966
1968
1970
1970
Est. of s.,
Pairs shown
1966
1968
1970
Est. of s,
8, 42(a) 54, 57 94, 82
5, 6 32, 32 75, 63
44, 52 20, 34 70, 72
47, 58 12, 20 28, 31
Iog10 of yg/gal. 0.1134 0.0377
to be internally equivalent:
42, 43 57, 62
39, 40 12, 12
48, 49 29, 17
login of yg/gal. 0.1062
82, 104
25, 23
52, 42
0.0478
based on seven pairs
(a) Deposits present were from the same deposit
fuel, with other deposits intervening.
-------
- 25 -
Table 8
CAPE- 6- 68 SECOND-YEAR REPORT
EMISSION CONTROL SYSTEM EFFECTS ON PNA EMISSION v
Benzo(a)pyrene
yg./gallon
Uncontrolled
1966
Vehicle
125
127
165
78
63
59
50
37
78
91
36
18
29
54, 57, 62
39
EM System
1968
Vehicle
30
32, 32
36
11
8.9
36
10.5
13
25
27
14
3.5
3.4
12, 12
15
1968/1966
Ratio, RM(b)
95% Interval
f°rRM
0.25
0.21 to
0.30
1970
Vehicle
26
17, 29
28
48
36
26
19
20, 34
12, 20
10.5
1970/1966
0.30
0.24 to
0.38
Benz(a)anthracene
./gallon
Uncontrolled
1966
Vehicle
196
258
276
130
86
157
109
134
182
212
49
30
46
94,82, 104
51
EM System
1968
Vehicle
55
63, 75
85
28
20
75
40
75
52
85
15
13
15
25, 23
17
1968/1966
0.32
0.30 to
0.35
1970
Vehicle
60
42, 52
63
70
61
51
57
70, 72
28, 31
38
1970/1966
0.32
0.29 to
0.36
(a) Data on each line relate to a common fuel and engine-deposit
condition.
(b) Geometric mean ratio of PNA emissions, emission-controlled/
uncontrolled.
-------
- 26 -
Table 9
CAPE-6-68-SECOND-YEAR REPORT
EMISSIONS FROM LOW-AND HIGH-AROMATIC FUELS
Benzo(a)pyrene
Benz(a)anthracene
Test
Vehicle
1966
1968
1970
rueo.
PNA, as
ppm BaP
0
3
0
3
0
3
Emission,
11% Ar.
50
37
10.5
13
19
20, 34
Ug./gal.
46% Ar.
127
165
32, 32
36
17, 29
28
Ratio Ja)
2.54
4.46
3.05
2.77
1.21
1.04
Emission,
11% Ar.
109
134
40
75
57
70, 72
yg./gal.
46% Ar.
258
276
75, 63
85
52, 42
63
Ratio (a
2.37
2.06
1.72
1.13
0.82
0.89
(a) Replicate data were averaged for ratios.
Table 10
CAPE-6-68-SECOND-YEAR-REQUEST
EFFECTS OF FUEL AROMATICS ON PNA EMISSIONS
Test
Vehicles
No. of
Ratios
in Mean
Geometric Means of Ratios of Emission at
46% Aroma tics to Emission at 11% Aromatics
Benzo(a)pyrene Benz(a)anthracene
Mean Ratio 95% Interval Mean Ratio 95% Interval
1966
1968
1970
2
2
2
3.37
2.91
1.12
2.07-5.49
1.79-4.75
0.69-1.82
2.21
1.39
0.86
1.77-2.76
1.11-1.74
0.69-1.08
1966, 1968
3.12
2.63-3.71
1.76
1.63-1.90
-------
- 27
4.4 Immediate Effects of Fuel Composition on PNA Emission
The reduction in PNA emission that was observed in the two control-
system vehicles does not completely eliminate the value of studies on the
effects of fuel variables, because the PNA from vehicles in current use does
vary with fuel changes. The present section relates to immediately-observ-
able PNA-emission effects where different test fuels were used; a later
section discusses indirect effects that occurred after extended use of de-
posit-forming fuels.
4.4.1 Fuel Aromatics Content
Test fuels with 117>, 28%, and 467o aromatics were used in PNA emission
tests in the 1966, 1968, and 1970 test vehicles. In order to simplify the
presentation, tests with 11% and 46%, aromatics will be reviewed here; the 287o-
aromatic data are available in Tables 5 and 6. Two test pairs are available
for each vehicle to illustrate the effect of increasing aromatics from 11% to
467o. The data are presented in Table 9, with the second data line for each
vehicle representing tests in which PNA (BaP at 3 ppm) had been added to the
test fuels.
The ratios of emission at 46% aromatics to emission at 11% aroma-
tics, as given in Table 9, range from 0.82 to 4.46. The geometric mean
ratios for each vehicle are given in Table 10, and, in view of the 957= con-
fidence intervals, it is clear that the PNA emission from the 1970 vehicle
was not sensitive to fuel aromatics. Excluding this vehicle, the last line
of Table 10 indicates the geometric mean ratios, with confidence intervals,
for the 1966 and 1968 vehicles. The confidence intervals are based on the
statistical discussions given in a previous section.
The mean ratios show that, in the 1966 and 1968 vehicles increas-
ing fuel aromatics from 11% to 46%, caused average increases in emissions of
BaP and of BaA amounting to about 212% and 76%, respectively. BaP emission
was apparently more than twice as sensitive as BaA emission, although more
data on this point appear to be needed. The mean ratios for each vehicle
also show a trend, particularly in BaA emission: the sensitivity of the
emission rate to increases in fuel aromatics decreased between vehicles in
the order of 1966>1968>1970 i.e., the more recent vehicles were less sen-
sitive to aromatics.
4.4.2 Fuel PNA Content
At a given level of fuel aromatics (11%, or 46%,), the presence of
PNA (BaP, BaA, and other species) already in the fuel can affect the emission
of PNA. The emission-test fuels were prepared with near-zero PNA contents,
and PNA was added, as reformate still-bottoms, to increase BaP content to
the field-maximum level of 8000 yg. of BaP/gallon, or about 3 ppm. Compari-
sons of PNA emissions from fuels with and without PNA addition are avail-
able in Table 11. Two low-aromatic fuels and one high-aromatic fuel were
used to provide three test pairs in each of three vehicles, with fuel PNA
content as the only within-pair variable. The ratios do not differ signi-
ficantly between base fuels or between vehicles. The geometric means of
two sets of nine ratios each are given, with the 95% confidence intervals,
in Table 12. An increase of 327o in BaA emission is demonstrated by the
data but the 87o apparent increase in BaP emission is not statistically
significant.
-------
- 28 -
Table 11
CAPE-6-68 SECOND-YEAR REPORT
PNA EMISSIONS FROM LOW- AND HIGH-PNA FUELS
(a)
Benzo(a)pyrene
Benz(a)anthracene
Test
Vehicle
1966
1968
1970
Emission,
0 BaP 3
78
50
127
25
10.5
32, 32
12, 20
19
17, 29
US/Ral.
jppm BaP
91
37
165
27
13
36
10.5
20, 34
28
Ratio ,
3/0
1.17
0.74
1.30
1.08
1.24
1.12
0.66
1.42
1.22
Emission
0 BaP
182
109
258
52
40
75, 63
28, 31
57
52, 42
> US -/gal.
3 ppm BaP
212
134
276
85
75
85
38
70, 72
63
Ratio ,
3/0
1.16
1.23
1.07
1.63
1.88
1.23
1.29
1.24
1.34
(a) Fuel PNA content is expressed as ppm BaP; replicates
are averaged for ratios.
Table 12
CAPE 6-68 SECOND-YEAR REPORT
EFFECT OF FUEL PNA ON PNA EMISSION
Means of Nine Ratios
95% Interval
Significant Effect?
Geometric Means of Ratios of Emissions
at 3 ppm BaP to Emission at 0 ppm
BaP Emission
1.08
0.85 to 1.35
No.
BaA Emission
1.32
1.19 to 1.48
Yes.
-------
- 29 -
4.4.3 High-Boiling Naphtha
Tests with and without a high-boiling naphtha component in the
test fuels showed that its presence caused changes in PNA emission which
varied with the amount present and with test-vehicle identity. The component
used was a mixture of pentanes, heavy catalytic naphtha and polymer gaso-
line. Its presence in test fuels did not significantly affect the content
of aromatics or of PNA. The effect of the naphtha on PNA emission was tested
in all three vehicles, using two concentration levels: 2% in a 467,-aromatic
fuel, and 167, in an 117o - aromatic fuel. Olefin content was maintained at
about 17= without the naphtha, and at 67o when it was present at either con-
centration level. The PNA emissions data, with and without the component
present, are summarized in Table 13. Parallel statistical analyses of the
logarithms of the yield data for BaP and for BaA in the first three lines
(27o naphtha) of the table showed no significant effects on either BaP or
BaA emission at even 907o confidence for 27° of the naphtha, either as a
main effect or as an effect showing interaction with vehicle identity. The
error mean squares used in these analyses were the variances calculated from
four replicate test pairs. The three test pairs relating to 27, of the naphtha
in the above table were thereafter treated as additional replicate pairs and
thus increased the available replicates to seven.
Statistical analyses of the data on 167, naphtha showed an inter-
action effect: the naphtha effect depends on the vehicle identity, being
reversed in the 1970 vehicle from its effect in the 1966 and 1968 vehicles.
Analysis of the data for the 1966 and 1968 vehicles only yielded the results
in Table 14. The ratios imply increases of about 1097, in BaP emission and
407» in BaA emission for the presence of 167, high-boiling naphtha. These
results are clearly restricted to the particular tests included in the analy-
sis. No explanation is available for the reversed effect of the naphtha in
the 1970 vehicle, and the overall picture is not clear.
4.4.4 Antiknock Lead
The effect of lead antiknock (tetraethyllead as Motor Mix) on PNA
emission can be examined by comparing emissions from ten tests using lead-
free fuels with the emissions from ten tests with lead present at 3 grams/
gallon and with five tests with lead at 0.5 grams/gallon. The 3-gram tests
included fuels with 117,, 287o, and 46% aromatics, but the 0.5-gram tests were
all in the 287,-aromatic fuel; all three test vehicles were used in the tests
and gave similar results. The emissions data are summarized in Table 15.
The geometric mean ratios for the effect of lead, together with the 957o con-
fidence intervals, are given in Table 16. The effects for the presence of
3 grams of lead included a non-significant small decrease in BaP emission
and a small, but significant, increase in BaA. With 0.5 gram of lead, the
decreases observed are statistically significant but are based on a smaller
number of tests using only one base fuel. Further testing appears necessary
in order to clarify and establish this complex set of effects for the use of
lead.
-------
- 30 -
Table 13
Vehicle
1966
1968
1970
1966
1968
1970
CAPE-6-68 SECOND-YEAR REPORT
PNA EMISSIONS WITH A HIGH- BOILING NAPHTHA PRESENT
% Ar.
in
Fuel
46
46
46
11
11
11
11
11
11
Benzo
Emission
0% HBN
62
12
17
0% HBN
50
37
10.5
13
19
20, 34
(a)pyrene
, yg/gal.
2% HBN
57
12
29
16% HBN
78
91
25
27
12, 20
10.5
Benz (a) anthracene
Ratio
2%/0%
0.92
1.00
1.71
1.56
2.46
2.38
2.08
0.84
0.39
Emission,
0% HBN
104
23
42
0% HBN
109
134
40
75
57
70, 72
yg/gal.
2% HBN
82
25
52
16% HBN
182
212
52
85
28, 31
38
Ratio
2%/0%
0.79
1.09
1.24
1.67
1.58
1.30
1.13
0.52
0.54
Table 14
CAPE-6-68 SECOND YEAR REPORT
EFFECT OF HIGH-BOILING NAPHTHA ON PNA EMISSION
(Low-aromatic Fuels in Two Vehicles)
Geometric Means of Ratios of Emission
for 16% Naphtha to Emission for 0%
BaA Emission
Means of Four Ratios
95% Confidence Intervals
BaP Emission
2.09
1.48 to 2.95
1.40
1.20 to 1.64
-------
- 31 -
Table 15
CAPE-6-68 SECOND-YEAR REPORT
PNA EMISSIONS WITH VARYING LEAD CONCENTRATION
Test
Vehicle
1966
1968
1970
% Arom
in
Fuel
11
28
28
46
46
28
28
46
28
46
Emission, yg./gal.,
Lead in grams/gal.
0
Emission of
56
78
36
125
54, 57, 62
11
14
30
48
26
0.5
Benzo
63
18
8.9
3.5
36
3.0
(a)pyrene
56
59
29
127
54
36
3.4
32, 32
26
17, 29
Ratios
0.5/0
0.81
0.50
0.81
0.25
0.75
3/0
1.00
0.76
0.81
1.02
0.93
3.27
0.24
1.07
0.54
0.88
Emission of Benz (a) anthracene
1966
1968
1970
11
28
28
46
46
28
28
46
28
46
126
130
49
196
94, 82, 104
28
15
55
70
60
__
86
30
20
13
61
155
157
46
258
124
75
15
63, 75
51
52, 42
0.66
0.61
0.71
0.87
0.87
1.23
1.21
0.94
1.32
1.33
2.68
1.00
1.25
0.73
0.78
-------
- 32 -
Table 16
CAPE-6-68 SECOND-YEAR REPORT
EFFECT OF LEAD ON PNA EMISSIONS
Geometric Means of Ratios of Emission
with Lead to Emission without Lead
BaP Emission
Lead at
0.5
Lead at
3
BaA Emission
Lead at
0.5
Lead at
3
Geometric Mean
95% Interval
0.57
0.42
to
0.78
0.86
0.70
to
1.08
0.74
0.64
to
0.85
1.17
1.06
to
1.29
Effect Significant? Yes,
No.
Yes.
Yes.
-------
- 33 -
4.5 Long-Term (Deposit) Effects of Fuel Composition on PNA Emission
The comparisons of tests, or groups of tests, that have been
made up to this point have been between results from different fuels
or from different vehicles. In every comparison, the engine deposits
have been constant, i.e., they were deposits formed under programmed
cyclic city-suburb-highway conditions with essentially the same deposit-
formation fuel. Deposits, as characterized by the fuel from which they
were formed, are the only remaining test variable on which data are
available.
The effect of deposits on PNA emission was studied initially
by tests in the 1966 vehicle; the studies were later expanded to
include the 1968 vehicle, but no deposit studies have been done in the
1970 vehicle. After cleaning the combustion chambers in the
1966 vehicle, lead-free deposits (fuel A) were accumulated as the first
step in the deposit sequence shown in Table 17. The choice of deposit
fuels for steps 1 and 2 of the 1966 vehicle was intended to allow a de-
cision on whether lead-free and leaded deposits give different PNA emis-
sions. However, when a difference was seen between deposits from the
two marketed gasolines, a conclusion relating specifically to lead could
not be reached confidently because, in addition to lead, the two deposit
fuels differed as to phosphorus and base-stock composition. The next
deposit steps, using A.5 and A fuels, were intended to overcome this ob-
jection, by using a single base fuel, but the effect of lead, via de-
posits, was tested at the new low-lead level of 0.5 g./gal. rather than
at 3 grams; two vehicles were used. The program also allowed a test of
the reversibility of deposit effects in the 1966 vehicle, i.e., could
the conditions of step 1 be reached again in step 4 after two other de-
posit conditions?
The test results obtained with deposits from fuels A and A.5,
in two vehicles, are summarized in Table 18. Two conclusions may be
drawn from the data, although each conclusion is based on emission tests
with only one fuel: (1) a deposit condition, as measured by the PNA
emission level, was re-established after other deposits and emission
tests had intervened, i.e., the PNA effects of deposits were reversible;
and (2) this specific deposit fuel formed deposits with the same PNA
emission levels, whether it was lead-free or leaded at 0.5 g. Pb/gallon.
Additional tests of lead effects, using more lead and other base fuels
for deposit formation, appear necessary in order to generalize any con-
clusion as to the effect of deposit lead on PNA emission. Within the
present data, however, data from A and A.5 deposits can be combined as
representing results characteristic of A-based fuel deposits.
-------
- 34 -
Table 17
CAPE-6-68 SECOND-YEAR REPORT
SEQUENCE OF ENGINE-DEPOSIT FUELS
Step No.
for Vehicle
1966 1968
1
2 1
Code
A
C3P
Deposit Fuel
Description
Commercial lead-free premium.
Commercial leaded (2-3 g/gal.
) prei
with phosphorus additive.
3 2 A.5 Commercial lead-free as above, + 0.5 g.
Pb/gal.
43 A Commercial lead-free as at start.
-------
- 35 -
Table 18
CAPE-6-68 SECOND-YEAR REPORT
PNA EMISSIONS WITH DEPOSITS FROM BASE A FUELS
Emissions, yg./gal., with deposits
from Base A leaded, as shown, in g./gal.
Benzo(a)pyrene Benz(a)anthracene
Vehicle % Ar.-PNA-Pb 0 0.5 0 0.5
1966 46-0-0 54 -- 94
46 - 0 - 0^' 57 82
46 - 0 - 0(b) 62 104
28-0-0 39 36 51 49
1968 28-0-0 15 14 17 15
(a) This test is a repeat of the test on the line above and occurred
after two intervening deposit conditions in the 1966 vehicle.
(b) This fuel, and the 28%-aromatic fuels, contain no high-boiling
naphtha.
-------
36 -
Comparisons (seven in the 1966 vehicle and five in the 1968
vehicle) of results from A-base deposits with results from C3P deposits
are listed in Table 19 along with ratios, A/C3P of emissions from each
test pair. The geometric mean ratios given in the table make it clear
that deposit-fuel identity caused greater than two-fold differences in
PNA emission from a variety of emission-test fuels.
Fuel C3P differed from A and A.5 in respect to lead content,
phosphorus content, and hydrocarbon base fuel. The broad fact that the
base fuels (hydrocarbons) were different implies the possibility of a
variety of more specific differences, such as aromaticity, olefinicity,
traces of high-boiling species present, etc. At the present time, it
is not clear which of these factors, or a combination of the factors,
is important in controlling the PNA-emission character of deposits.
Information on this point may be a matter of some practical
interest, because in the course of the deposit studies, PNA emissions
were observed from high-aromatic fuels, used with A-fuel deposits,
which were similar to, or lower than, the emissions from fuels of appre-
ciably lower aromatic content used in the presence of C3P deposits.
-------
- 37 -
Table 19
CAPE-6-68 SECOND-YEAR REPORT
EFFECT OF DEPOSIT-FUEL COMPOSITION ON PNA EMISSION
Test
Vehicle
1966
1968
Fjnission
Test Benzo
Fuel Deposit
% Ar. A
46 54, 57, 62
28 39
12 (a) 56
46 54
28 36
28 18
28 29
46 12, 12
28 15
28 14
28 3.5
28 3.4
lean Ratios
nee Intervals
(a)pyrene
Deposit
C3P
125
78
91
127
78
63
59
30
11
11
8.9
36
0.
Ratio
A/C3P
0.46
0.50
0.62
0.43
0.46
0.29
0.49
0.40
1.36
1.27
0.39
0.09
0.46
38 to 0
, yg./gal.
Benz(a)
, Deposit
A
94, 82, 104
51
155
124
49
30
46
23, 25
17
15
13
15
.57
anthracene
Deposit
C3P
196
130
212
258
130
86
157
55
28
28
20
75
0.40
Ratio,
A/C3P
0.47
0.39
0.73
0.48
0.38
0.35
0.29
0.44
0.61
0.54
0.65
0.20
0.44
to 0.'
(a) With high-boiling naphtha, TEL, and high PNA content.
-------
- 38 -
5. DISCUSSION
The origin and emission of polynuclear aromatic hydrocarbons
in vehicle exhaust has been studied in numerous laboratories since atten-
tion was directed to their presence and possible medical significance in
the 1950's, and many of the factors governing the exhaust concentration of
PNA have been examined. A smaller number of similar studies on phenol
emission has also appeared.
5.1 Effect of Carburetion
Several investigators (4, 5, 6, 7) have shown that vehicles with
progressively higher air/fuel ratios and lower emissions of CO will also
emit less PNA, and measures that control CO are apparently one of the most
effective single approaches to minimizing PNA. Intuitively, however, the
presence of CO in exhaust gas is much less likely to be a direct cause of
PNA formation and emission than it is to be a symptom of an oxygen-deficient
environment, in the engine, that is favorable to the synthesis and survival
of PNA and of the free radicals that must combine for PNA synthesis. Radio-
tracer BaP experiments (5, 7,) have, in fact, shown that PNA are destroyed
by oxidation or pyrolysis to an appreciable degree even in an exhaust with
substantial CO content. Other experiments (4, 7) have shown the value of
exhaust system oxidation with air-injection systems to reduce the amount of
PNA that is emitted, as well as the value of engine-modification systems
to combine the benefits of reduced CO and of whatever oxidation of PNA may
occur from available exhaust oxygen. Also, reports have appeared recently
of PNA control with an "advanced" emission control system (7) using a thermal
reactor, exhaust recirculation, and particulate traps, and of PNA destruction
in a prototype exhaust catalytic system (10). Paralleling or accompanying
these reports on opportunities for control of PNA through engine design, ad-
justments, and the use of devices, there have been reports (2, 4, 7, 10, 11)
of the effects of fuel compositional variables on PNA emission. In general,
it appears that these reports of fuel effects are based on data obtained in
tests characterized by the presence, in the exhaust, of appreciable CO;
emissions of CO at a level of less than 1% (i.e., within the standard for a
4000-lb. 1970 vehicle) are not common in the reports on the effects of fuel
factors on PNA.
The situation with respect to publications on exhaust phenols is
somewhat different. Early tests (2, 12, 13) had implicated fuel aromatics
as potent sources of exhaust phenols, using tests conducted under almost
assuredly fuel-rich conditions. Now, a recent report (14) presents a con-
vincing correlation of fuel aromaticity and exhaust phenol yields from three
vehicles, of which two were 1970 models and must be assumed to have emitted
less than 1% CO in the standard test.
The studies presented in the present report may be viewed as com-
plementing the existing literature for both polynuclear aromatics and phenols.
The three test vehicles that were used include one (1966) without emission
controls and two (1968 and 1970), with engine-modification control. The
vehicles were adjusted for relatively lean mixtures (Section 3.4), and
operated in the near-stoichiometric range of air/fuel ratio, where relatively
-------
- 39 -
little published information on the effects of fuel factors on PNA has
been obtained.
With respect to phenol emission, we confirm the recently pub-
lished results on fuel factors, as obtained with lean carburetion (1970)
vehicles), and we extend the study of fuel factors toward a slightly
higher CO level.
The existence of carbon monoxide emissions from an engine was
referred to above as a symptom of an environment in which PNA or PNA pre-
cursors might form and survive, rather than being destroyed by oxidation
or pyrolysis. Two other exhaust-gas factors, oxygen content and tempera-
ture, will obviously also play roles in PNA survival and emission. Phenols,
as species capable of being oxidized, should also be sensitive to the
same variables. It is thus appropriate, before reviewing our emission re-
sults, to note that both CO content, oxygen content, and temperature were
different between the three vehicles (Section 3.4). We feel that these
tabulated characteristics typify a progressively changing environment,
between the three vehicles, that influenced the emission of PNA and phenols
from the vehicles and also influenced the sensitivity of the vehicles, in
terms of PNA and phenol emission, to changes in fuels. The influence on
PNA emissions of even small changes in CO, between tests in a given vehicle,
was sufficiently evident to provide a basis for a correlation to correct
PNA data for CO variation, as discussed in a preceding part of this paper.
5.2 Phenol Emission
The three regression equations, developed from the data on each
vehicle, for the emission of phenols as a function of fuel aromatics,
explained over 95% of the variance in phenol rates (mg./gallon of fuel);
as a result, no other within-vehicle variables (fuel composition, A/F ratio,
etc.) are likely to be found significant, either statistically or practically.
The equations do indicate substantial differences between the two nearly
equivalent earlier vehicles (1966 and 1968) and the 1970 vehicle. The
differences are typified by the phenol rates in Table 20, which were calcu-
lated from the equations that describe each vehicle. Our results
for the 1970 vehicle are in good agreement with the recent report by Hinkamp
(14) on phenol emission with fuels of varying aromaticity, as observed
in two 1970 vehicles and one 1968 vehicle. The fact that phenol emission
depended strongly on fuel aromatics in all three of our test-vehicles, but
was lower in our 1970 vehicle than in the nearly-equal 1966 and 1968 vehicles,
suggests that phenol formation depends on a reaction involving aromatic
molecules that is independent of CO level, while phenol destruction or
oxidation depends on an adequate oxygen supply and sufficiently high tempera-
tures .
If the 30% decrease in phenol emission from our 1970 test vehicle,
relative to the 1966 and 1968 vehicles, is assumed to be typical, it suggests
-------
- 40 -
Table 20
CAPE-6-68 SECOND-YEAR REPORT
PHENOL EMISSION RATES
Test Vehicle
1966 1968 1970
NC EM EM
Emission Rate, rag./gal.
28% aromatic fuel 400 371 268
46% aromatic fuel 709 657 449
Slope, mg./gal./% aromatics 17.2 15.9 10.1
-------
- 41 -
that phenol emission has not been reduced as effectively, in reaching
the 1968 and 1970 stages of emission control, as have the emissions of
other pollutants.
5.3 Emission-Control System Effects on PNA Emissions
The 1968 and 1970 vehicles, both equipped with engine-modifica-
tion control systems, emitted an average of about 30% as much PNA (BaP
or BaA) as was emitted by the 1966 uncontrolled vehicle. The 70% reduction
(range, 29% to 88%) agrees generally with results published by Begeman (4)
and by Hoffman (7). The range of reductions in BaP emissions reported by
Begeman was 66% to 96%, relative to the same engines with the control
system inoperative or replaced by a no-control configuration. Hoffman
reported on eleven 1964-1966 vehicles and seven 1968 (controlled) vehicles,
with two fuels tested in each vehicle. The average BaP emissions of the
seven controlled cars were reduced by 79%, for one fuel, and 67%, for
the other, from the average BaP emissions, on the same fuels, in the uncon-
trolled group. For both of these reports, the % CO (emission-test result)
was generally 2% to 4% in the "uncontrolled" tests and was still, usually,
a little greater than 1% in the "controlled" tests. By contrast, our
1966 vehicle and 1968 vehicle average CO levels were 1.49% and 0.75%, i.e.,
each was about half to two-thirds as great as in the other published
studies, or as is typical for those years. We interpret this as meaning
that our 1966 and 1968 vehicles perform correctly relative to each other,
but that the PNA from both may be on the order of 50% low. The other
published studies do not include comparable data in 1970 vehicles, which,
as a class, were required to meet lower emission standards for CO (typical-
ly, 1% vs. the prior 1.5%). Our 1970 vehicle shows no difference in PNA
emission from our 1968 vehicle, but we attribute this to the fact, indicated
above, that both the 1966 and 1968 test vehicles had depressed emissions be-
cause their carburetion was abnormally lean.
In the preceding discussion of phenol emissions from our test
vehicles, we proposed, in effect, that about 30% of the phenols formed in
the 1970 vehicle engine were oxidized in the exhaust system so that it
emitted less phenol than did the 1966 and 1968 vehicles. It is reasonable,
if this is true, to expect that PNA from this engine would also be oxidized
to some extent. Yet we find that the 1970 and 1968 vehicles have very
similar emissions of PNA. We conclude either that the expected oxidation
of PNA did not occur in the 1970 vehicle exhaust, or that it did occur
but that relatively more PNA was initially formed, for unknown reasons.
5.4 Immediate Fuel Effects on PNA Emission
Changes in PNA emission rates with changes in fuel composition
can occur in two ways. An immediate effect may occur as soon as a change
is made in fuel composition, or the extended use of any one fuel will
result in engine deposits which may influence the emission of PNA, just
as deposit accumulation is known to affect the emission of unburned
hydrocarbons. In this program, the two types of change have been studied
independently by examining several fuels in each of several successive
-------
- 42 -
deposit conditions, and the two types of change will be discussed separately.
In no case were the deposit fuels and the emission-test fuels identical.
5.4.1 Fuel Aromatics Content
The effects of increasing fuel aromatics from 11% to 46% were
observed in two test pairs (which differed in PNA content) in each of
three vehicles, and both BaP and BaA emission data were obtained. The
summary of geometric mean ratios in Table 9 showed trends with vehicles
and with the species of PNA detected. The 1970 vehicle could not signifi-
cantly detect the effect of increased aromaticity, while the 1966 vehicle
was the most sensitive. Sensitivity to fuel aromatics was also greater
for BaP emission than for BaA. The overall geometric mean ratio for two
PNA species in only the two earlier vehicles is 2.34, indicating a 134%
PNA emission increase for increasing aromatics from 11% to 46%.
The finding that increased fuel aromatics increases PNA emission
is in agreement with numerous other studies (2, 7, 11, 12). Only rarely [Griffing
(5), and Padrta (10)] have cases been reported in which there was not a clear
PNA'relation to aromatics content. Hoffman (7) has reported on uncontrolled and
1968-model controlled vehicles, with each car in both groups testing low-
and high-aromatic fuels. In contrast to our finding, the PNA emissions
of the controlled vehicles displayed fully as much relative response to
increased fuel aromatics as did the emissions of the uncontrolled vehicles.
Hoffman also reports experiments in which two different control systems
were added to a 1966 uncontrolled car. Again the response of PNA emission
to aromatics changes did not decrease with the use of controls, nor was
BaA emission less sensitive to aromatics than BaP emission. These tests
were conducted at generally higher CO levels than were present in our
work, however.
It would appear to be very desirable to know from a larger
sample whether or not the PNA emissions of vehicles with lean carburetion
and retarded spark, such as the 1970 test vehicle used here, are sensitive
to fuel aromatics. However, the question is one of differences in PNA
emissions which are already reduced substantially from earlier levels
because of the emission control systems.
5.4.2 Fuel PNA Content
The presence of PNA in fuel, commonly expressed in terms of
BaP content in yg./gallon, or ppm, where 1 ppm is about 2800 yg./gallon,
has been shown to cause increased emission of exhaust PNA. Distillation to
remove the high-boiling material from a gasoline that initially contained
0.7 ppm of BaP was reported by Hoffman (7) to have caused an 87% decrease
in BaP emission. Begeman (4) found that Indolene, with about 4 ppm BaP,
gave four to five times as much PNA emission as a regular-grade gasoline
with 1.1 ppm.
The present studies find much smaller effects for PNA addition
to fuels of initially low PNA content. The PNA added was the still-bottoms
from careful distillation of a large quantity of catalytic reformate.
This approach seemed preferable, in studying fuel PNA, to either distilling
-------
- 43 -
a finished fuel, with the chance of removing materials other than PNA,
or to comparing fuels that may differ in other ways than their analyzed
PNA contents. The distillate from the reformate was used for test-fuel
blends, which had about 0.015 ppm BaP, and the bottoms were added as
required to introduce 2.9 ppm BaP. This BaP level was near the field
maximum found in a small-scale survey (1). A composite of nationwide
premium and regular gasoline samples obtained separately would have about
one-tenth of this maximum BaP level (1). Hoffman (7) suggests that
most fuel BaP derives from catalytic reformate.
The effect of adding PNA to introduce 2.9 ppm BaP was observed
in two low-aromatic fuels and one high-aromatic fuel, with all three
fuels used in three vehicles, and data were obtained on both BaP and
BaA emissions. Neither vehicle identity nor fuel aromatics influenced
the observed effects of fuel PNA, which were unexpectedly small: an 8%
increase in BaP emission was not significant, and the increase in BaA
emission was only 32%. These relatively modest effects for fuel PNA,
compared to those reported in the other papers, are believed to be due
to differences in details of the experiments that were carried out. The
earlier studies were conducted in relatively rich, fully warmed-up labora-
tory engines with appreciably longer tests than were used here. Griffing
(11) has reported that high-PNA fuels seem to deposit PNA that can persist
into a later test with a low-PNA fuel, suggesting that longer tests might
give greater emphasis to fuel PNA. Pending further data, our findings
of a modest effect for fuel PNA appear correct for short-term effects
in low-CO vehicles, but long-term effects and other engine conditions
might change this picture.
5.4.3 High-Boiling Naphthas
The fuel PNA studied in the foregoing experiments was derived
from the higher-boiling portion of a catalytic reformate and was rich
in aromatic hydrocarbons. Other high-boiling materials can also be pres-
ent in gasoline, for example, from the blending of heavy catalytic naphthas;
this component is relatively common in regular-grade gasolines. In the
studies described here, low-aromatic (11%-12%) fuels were compared, with
or without 16% of a high-boiling naphtha (419°F end-point, ASTM-D86) that
included heavy catalytic naphtha as one component. Fuel compositions
and inspections in Appendix B show that blending the naphtha caused an
increase of 16 degrees (F) in end-point; the fuel PNA level was very low
and was not increased significantly by the naphtha. The data obtained
were not entirely clear as to the effect of the naphtha on PNA emission,
since the emissions were increased by the naphtha in the first two vehicles
used, but were reduced in a third vehicle.
The data from the first two vehicles do suggest that a high-boiling
naphtha may promote PNA emission under at least some circumstances. Griffing
(5) has also noted that a low-aromatic, low-PNA fuel with a high end-point
gave unexpectedly high BaP emission. Hoffman's data (7) indicate that the
heavy catalytic naphthas which he inspected contained no BaP or BaA, and
the blends we used with the naphtha present were low in PNA content. It
appears that any effects of heavy catalytic naphthas on PNA emission may
involve smaller molecules, or different species, than PNA. Further study
of high-boiling materials in gasolines appears desirable.
-------
- 44 -
In reviewing the fuel-composition studies on the PNA-emission
effects of fuel aromatics, fuel PNA, and a high-boiling naphtha, it is
noteworthy that these three variables are actually independent, and
that each of them, under at least some circumstances, caused significant
increases in PNA emission. Under these circumstances, experiments to
evaluate the effect of any one variable must be carefully controlled as
to the other variables. Furthermore, while they were not investigated
here, it is reasonable to assume that within-class differences in potency
as PNA precursors will exist. For example, C7, CR, Cq, and C1f) aromatics
are likely to differ, as are 2-ring, 3-ring, and 4-ring polynuclears,
in potency as precursors for, say BaA, or BaP. Evidently, much more
research is necessary before the different possible descriptions of
fuel composition can be definitively related to PNA emission.
5.4.4 Antiknock Lead
The immediate effect of the presence of lead in gasoline on PNA
emission has been examined in several laboratories. Begeman and Colucci
(4) found both small increases and small decreases in PNA emission
for the presence of lead in Indolene fuel. Hoffman (7) presents data
showing more increases than decreases in PNA as a result of lead addition
when testing several fuels in a 1966 engine (about 3% CO), but consistent
substantial decreases in BaP emission, when lead was added, for tests
conducted in two 1970, presumably fuel-lean, vehicles. Griffing (11),
employing two different 1967 vehicles, found no effect for lead on BaP
emissions. All of these studies compared results from fuels with 0 and
3 grams of lead (or ccs. of TEL/gallon).
Our tests included the 1966, 1968, and 1970 test vehicles and
permit comparisons of PNA emissions obtained with lead at both 0.5 and
3 grams/gallon with corresponding lead-free tests. The effects observed
for lead were not statistically different between the three vehicles,
but there was a difference between the two lead levels in their effects
on PNA emission. The results for 0.5 grams, which were fewer in number
and limited to one fuel, showed decreases in PNA emissions due to lead,
of 26% (BaA) and 43% (BaP). For 3 grams, with more tests and in three
fuels, we found a non-significant 14% BaP decrease and a statistically
significant 17% increase in BaA emission.
Our lead-effect results indicate that there may be complex
effects of lead on PNA emission, varying with specific situations, but
that these effects are relatively small. No large overall effect of lead
on PNA emission is apparent in terms of the lead itself. As a material
which may affect the presence of aromatics, and the composition of gaso-
lines, in general, lead probably can have indirect effects.
5.5 Deposit Effects on PNA Emission
The possibility that fuel composition could influence PNA
emission slowly by changes in engine deposits has received much less
attention in the published literature than has been applied to the
immediate PNA-emission effects of fuel changes. The Griffing (11)
-------
- 45 -
observation of a carryover effect of fuel PNA on PNA emission has been
previously noted; the implication of possible deposit effects was not
apparently pursued, but steps were taken to avoid the problem pre-
sented. Hoffman (7) reports tests in two 1970 vehicles on the PNA-
emission effect of lead in deposits. Four measurements of BaP emis-
sion in each deposit condition yielded data showing three increases
and one decrease in BaP associated with lead in the deposits. The
PNA emissions were described as "essentially the same" for leaded and
unleaded deposits. It is not clear whether the lead-free and leaded
deposit fuels were prepared from a single base fuel in this study.
Our approach to studying deposits was also directed, initially,
to the question of "leaded" vs. "unleaded" deposits, and two commercial
premium gasolines, lead-free and leaded (with phosphorus) were used for
deposits. Thus, the base gasolines were not identical in this compari-
son. The PNA emissions from the leaded (+P) deposits were substantially
higher than from the lead-free deposits, according to tests with three
emission-test fuels in each deposit condition. This initial study of
deposit effects, which was limited to the 1966 vehicle, was followed by
additional deposit tests in both the 1966 and 1968 vehicles. In these
tests, deposits from the commercial leaded (+P) premium gasoline were
converted to low-lead, and then lead-free, deposits by the successive
use of the same lead-free commercial premium gasoline, first with lead
added to 0.5 grams/gallon, and then without lead.
For both vehicles, the PNA-emission levels from the low-lead
and no-lead deposits were equivalent and these levels were only about
one-half of the level that had been obtained with deposits from the
leaded (+P) gasoline. The PNA emissions of the 1966 vehicle had returned
to the same level that was observed in the first tests of the lead-free
premium gasoline, before going to leaded deposits. Two distinct levels
of emission, related to deposits, were thus established, but the evidence
remains unclear, at this time, whether the effect observed was due to
base-fuel composition, phosphorus, or lead level, or to interactions be-
tween these or other factors.
Lead, phosphorus, and fuel composition have all been examined
in relation to some aspects of deposits by various investigators in the
past. Although these studies did not include PNA-emission differences
between deposits, the results that were obtained, and the fuel composition
factors that were considered, may be relevant here. In studies of lead,
at several levels from 0 to 3 grams/gallon, and of phosphorus with lead,
Gagliardi (8) and Gagliardi and Ghannam (9) found that the emission ot
volatile hydrocarbons was elevated equally by any non-zero lead level,
relative to the emission from lead-free deposits, while the use of phos-
phorus, with lead, did not affect hydrocarbon emission. Studies on the
base-fuel composition aspects of deposits have been concerned primarily
with combustion abnormalities, such as knock and surface-ignition, ra-
ther than with the emission of air pollutants. Differences in combus-
tion related to deposits, have been ascribed to increased overall fuel
aromatics content at a constant lead level (15); thus overall deposit-
fuel aromaticity should be considered as a possible factor in the ef-
fect of deposits on PNA emission. In addition to overall aromatics,
-------
- 46 -
attention has been directed toward high-boiling components (15, 16) whose
presence may be apparent from ASTM D-86 distillation data, and to high-
boiling aromatics specifically, such as naphthalene (17). The latter,
extending in principle to include trace amounts of PNA in fuels, may not
appear in terms of distillation data, but are found in PNA analyses,
such as our small-scale field survey (1) on the BaP content of several
gasolines.
Thus the fuel composition factors that may affect PNA emission
indirectly because of their effects on deposits may be somewhat the same
fuel factors that appeared to influence PNA emission directly in the experi-
ments described earlier. These factors were aromatics, high-boiling
naphthas, and polynuclear aromatics (PNA, as BaP); within these broad
classes there are likely to be hydrocarbons with widely differing potency
for both directly, and indirectly, influencing PNA emission.
5.6 Areas For Further Study
Our studies are currently directed toward getting a better under-
standing of the fuel composition factors that influence PNA emission in
current vehicles, with particular reference to higher-boiling materials.
However, on the basis of the tests reported here, it appears that future
low-emission vehicles will minimize the significance of any of these factors.
-------
- 47 -
APPENDIX A
REFERENCES
(1) Gross, G.P., "First Annual Report on Gasoline Composition and
Vehicle Exhaust Gas Polynuclear Aromatic Content" (CRC-APRAC
Project CAPE-6-68), Period Ending February 17, 1970.
(2) Begeman, C.R., "Carcinogenic Aromatic Hydrocarbons In Automobile
Effluents, "SAE Technical Progress Series, Vol. 6, McMillan,
New York (1964). (Presented to the SAE, January, 1962, Detroit.)
(3) Begeman, C.R., and Colucci, J.M., "Benzo (a) pyrene In Gasoline
Partially Persists in Automobile Exhaust," Science, 161.
271 (1968).
(4) Begeman, C.R., and Colucci, J.M., "Polynuclear Aromatic Hydro-
carbon Emissions from Automotive Engines," SAE Transactions, 79.
1682-1698 (1970) .
(5) Griffing, M.E., Maler, A.R., and Cobb, D.G., "A New Tracer
Technique for Sampling and Analysis of Exhaust Gas for Benzo (a)
pyrene, Using Carbon 14," presented to Division of Petroleum
Chemistry, Inc., American Chemical Society, New York City, Sept.
7-12, 1969, (Preprints, page B-162, Vol. 14, No. 3).
(6) Gofmekler, V.A., et al, "Correlation Between The Concentrations
of 3,4-Benzopyrene and Carbon Monoxide In The Exhaust Gases of
Automobiles," GJRiena i Sanit., 28 (8), 3-8 (1963). (In Russian,
see CA 69_, 8546b (1964) .
(7) Hoffman, Jr., C.S., Willis, R.L., Patterson, G.H., and Jacobs,
E.S., "Polynuclear Aromatic Hydrocarbon Emissions from Vehicles,"
presented to Division of Petroleum Chemistry, Inc., American
Chemical Society, Los Angeles, March 28-April 2, 1971, (Preprints,
Vol. 16, No. 2, page E-36) . Note: A privately-distributed preprint
contains additional data.
(8) Gagliardi, J.C., "The Effect of Fuel Antiknock Compounds and Deposits
on Exhaust Emissions," Paper No. 670128 at the Automotive Engineering
Congress of the Society of Automotive Engineers, Detroit, January
9-13, 1967.
(9) Gagliardi, J.C., and Ghannam, F.E., "Effects of Tetraethyl Lead Con-
centration on Exhaust Emissions in Customer Type Vehicle Operation,"
Paper No. 690015 at the International Automotive Engineering Congress
of the Society of Automotive Engineers, Detroit, January 13-17, 1969.
-------
- 48 -
(10) Padrta, F.G., Samson, P.C., Donohue, J.J., and Skala, H., "Polynuclear
Aromatics in Automobile Exhaust," presented to Division of Petroleum
Chemistry, Inc., American Chemical Society, Los Angeles, March 28 -
April 2, 1971 (Preprints, Vol. 16, No. 2, page E-13).
(11) Griffing, M.E., Maler, A.R., Borland, J.E., and Decker, R.R., "Apply-
ing a New Method for Measuring Benzo(a)pyrene in Vehicle Exhaust to
the Study of Fuel Factors," presented to Division of Petroleum Chemistry,
Inc., American Chemical Society, Los Angeles, March 28 - April 2, 1971
(Preprints, Vol. 16, No. 2, page E-24).
(12) Hoffmann, D., and Wynder, E.L., "Studies on Gasoline Engine Exhaust,"
Journal Air Poll. Cont. Assn., 13., 322-327 (1963) .
(13) Hoffmann, D., Theisz, E., and Wynder, E.L., "Studies On The Carcino-
genicity of Gasoline Exhaust," Journal Air Poll. Cont. Assn., 15 (4),
162 (1965).
(14) Hinkamp, J.B., Griffing, M.E., and Zutaut, D.W., "Aromatic Aldehydes
and Phenols in the Exhaust from Leaded and Unleaded Fuels," presented
to Division of Petroleum Chemistry, Inc., American Chemical Society,
Los Angeles, March 28 - April 2, 1971 (Preprints, Vol. 16, No. 2,
page E-5) .
(15) Wiese, W.M., "if You Squeeze Them, Must They Scream?" Paper No. 61J
presented at the SAE Summer Meeting, Atlantic City, June 8-13, 1958.
(16) Fleming, C.L., Jr., Hakala, N.V., Moody, L.E., Scott, R.W., and Tongberg,
C.O., "Can Better Fuels Curb Engine Knock?" Oil and Gas Journal, 53_
(43), 100-104 (February 28, 1955), based on "Control of Engine Knock
Through Gasoline and Oil Composition," presented at tne SAE Golden
Anniversary Meeting, Detroit, January 10-14, 1955.
(17) Shore, L.B., and Ockert, K.F., "Combustion Chamber Deposits - A Radio-
tracer Study," Paper No. 145 presented at the SAE Summer Meeting,
Atlantic City, June 2-7, 1957.
-------
- 49 -
Appendix B
TEST FUELS
The emission-test fuels used in the second-year program of
CAPE-6-68 were blended from the same catalytic-reformate distillation
fractions and other blending stocks used for the first-year blends. These
components are described in detail in Appendix B of the first-year report;
along with compositions and inspections of the first-year test fuels. In
the following, selected first-year fuel information is repeated to permit
comparison with second-year fuels. First-year fuels are designated as
"1969" fuels and those for the second-year as "1970" fuels.
Section 6.3 of the first-year report indicated the intention
to prepare, for the second year, new blends that would not contain a high-
boiling stock (a mixture of heavy catalytic naphtha, polymer gasoline,
and pentanes) that was present in the 1969 set; total fuel aromatics was
to remain constant at each extreme. The purpose of the change was the
elimination, particularly from the low-aromatic test fuel, of suspected
high-boiling PNA-precursor materials that were thought to be responsible
for relatively high PNA emission from the low-aromatic fuel and a resulting
decrease in the apparent effect on PNA emission from increasing the over-
all fuel aromatics content from low (12%) to high (46%). A consequence of
the change in fuel composition would be a decrease in final boiling-point,
as measured by the ASTM-D86 method, as well as decreases in the distillation
temperatures for the entire higher-boiling parts of the fuels. These
effects could be partly compensated by using relatively more heavier
aromatics, particularly in the low-aromatic fuel where the high-boiling
naphtha represented 16% of the 1969 blend, against only 2% in the high-
aromatic fuel.
The blend compositions used in both years are given in Table B-l
of this Appendix. The components used in the two years are identical
materials except for the C,Q-C;Q aromatic solvent, where different batches
manufactured to the same specification were used. As indicated in the
table, essentially the same total aromatics were provided in both years,
but the olefin content decreased from about 6% to 1% when the high-boiling
naphtha was deleted. This change, which was unavoidable in the low-aromatic
fuel, was matched in the high-aromatic fuel by decreasing the amount of
light catalytic naphtha from 13% to 4% and then compensating with other
components.
Routine inspections of the fuel blends for both years are com-
pared in Table B-2. For the 1970 set, an equal-volume cross-blend of
the high- and low-aromatic fuels is included, and the table also lists
typical inspections for two fuels used for accumulating combustion-chamber
deposits in test vehicles during both the first and second program years.
The fuel codes for both years, as used in Table B-2 and at other points
in this report, are listed as follows:
-------
- 50 -
FUEL CODE DESIGNATION
Cross
11-12% Arom. Blend 46% Arom.
PNA TEL
1969
100
108C
1970
1LOO
1L03
1970
3000
3000.5
3003
1969
5HOC(a)
5HO
1970
SHOO
5H03
5H80
Low 0
0.5
3
High 0
3 108 1L83 5H8 5H83
(a)
A similar fuel, 5LOC, is more volatile in the mid-fill region, and
was used only in the first year.
This code includes certain fuels that were prepared but have not yet been
tested. It also includes the cross-blend of the basic 1970 low-aromatic
(1LOO) and high-aromatic (5HOO) fuels and indicates that this cross-blend
was tested at low lead content (0.5 ccs. TEL, as Motor Mix, per gallon),
as well as at zero and 3 ccs./gallon. The "low" and "high" PNA levels
indicated in the code chart refer to the as-made fuels with less than 50yg. of
benzo(a)pyrene (BaP) per gallon (low PNA) and to the fuels with PNA added,
in the form of reformate still bottoms, in an amount sufficient to add
8000 yg. of BaP/gallon to the fuel. Further details on this PNA-additive
material are in Appendix B of the first-year report.
The major differences in the inspections given in Table B-2
between the 1969 and 1970 fuels are in the distillation temperatures and
the sulfur contents. It is clear that the expected decreases in back-end
distillation temperatures have occurred, and that the low-aromatic fuel
now distills, throughout its range up to the end-point, at lower tempera-
tures than the high-aromatic fuel. It appears reasonable to assume from
this that high-boiling PNA precursors have been minimized in the new
fuels, although other more definitive inspections of the fuels would be
more informative. The decrease in sulfur is also a predicted result of
excluding the high-boiling naphtha. Its significance, if any, for PNA
emission is unknown.
Table B-3 presents mass-spectrometer analyses of the two fuel
sets, including also an analysis showing the changes observed when the
reformate still-bottoms (PNA) is added. These analyses indicate appreciable
decreases, between the 1969 and 1970 sets, in the amounts of GH and C^2
aromatics and indans, with only small changes in naphthalenes. As
expected, "naphthenes + olefins" also decreased substantially because of
the olefin decrease. The changes from 1969 to 1970 fuels appear reasonable,
except that the decrease in Ci2 aromatics from 0.18-0.22% in the two 1969
fuels to zero in both of the 1970 fuels is not consistent with the quite
modest change in composition between the 1969 and 1970 high-aromatic fuels
-------
- 51 -
as against a large change between the low-aromatic fuels (see Table B-l,
higher-boiling components).
Table B-4 presents GC analyses on the aromatic hydrocarbons
present in the same basic fuels, in the two batches (essentially equal)
of ClO-ll aromatic solvent (see above), in the lead-free deposit-accumulation
gasoline, and in the high-boiling naphtha used in the 1969 fuel set. These
GC data apparently do not reach to as high a boiling-point as do the MS
analyses; identification is essentially complete at CIQ> so that C]_i and
Ci2 cannot be discussed adequately. Like the MS, the GC indicates
decreased indan in 1970 fuels. However, the GC data are necessarily suspect
when it is noted that the 1970 low-aromatic and high-aromatic fuels appear
to have almost equal amounts of a variety of high-boiling aromatics, whereas
the compositions (Table B-l) show the high-aromatic fuel has from 2 to 18
times as much of any of the reasonable sources of these aromatics as are
present in the low-aromatic fuel. The GC analysis of the aromatic part
of the high-boiling naphtha does not reveal any uniquely abundant species
that would markedly diminish if this component were omitted from gasoline,
yet the lower Engler (D86) final boiling points (Table B-2) for 1970 show
that some high-boiling materials have indeed been excluded.
It appears from these analyses, both MS and GC, that techniques
for getting better data on the higher-boiling materials in gasoline would
be a distinct advantage in a study of PNA emission as a function of fuel
composition.
-------
- 52 -
TABLE B-l
FUEL
COMPOS 11 IONS
Low Aromatic
Stock
n-Butane
Isopentane
Motor Alkylate
Lt. Cat. Naphtha
Toluene
155-220 Reformate
220-241 Reformate
241-298 Reformate
298-384 Reformate
^10-11 Aromatic Solvent
H. Cat. N. + Polymer + C5's
7o Aroma tics
% Olefins
1969
4
16
54
2
1
1
1
1
1
3
16
11.9
5.8
1970
4
17
65
2
1
1
1
2
3
4
None
11.0
0.8
High Aromatic
1969
8
15
2
13
3
15
15
11
9
7
2
46.3
5.2
1970
7
19
5
4
3
17
18
11
9
7
None
46.2
1.4
-------
- 53 -
0 C
PEH -r
CO Ct
rH r-
CU ^
3 E
E
r-H C.
Cfl <
H
O 4-
01 Cf
S C
B C
o a
u o
0
rH
Cfl
S5
s
H
<^ W
pq 5:
w S
03
01 m m
ONrHCN
r^ 1 1 II f^ n-t r*1 rr\ ^^ ^^ ^^ ^^ ^^ ^^ ^^
1-1
'sO r^* oo O^ iO 'O co
*
invOvoc\ij" LO OO ^-^ CO vO 00 vD -t- i^\ /T\ rr\ i~*\ /TS
/v-v *^s. i i i i *^i *Ci fM «.i /«rt ^ *X ZX "^ " Ol OO tJ OA
OO Oi i~H i~H r~H iH CM tN O\l CN ") CO CO CO ,
i~i
O CM 00 cy> in ^- u-, CTtoOOO
CO O^ il i ( i t i 1 Csl CSI CN| CN CN CO CO CO ^^
&\ OO ON r-H LTj CJN CO
* * * * * * *
O CN ^3" 00 CN 00 i 1
CU
o
4J
T)
(11
(U
CU
rl
H_
*
cu co
J2 3
CU O
CO P.
CO
TJ a
CU rCi
u ex
CO
rl CO
i 1 CU
rl
0) S-l
CO O
O 0)
,C J3
4-1 4-1
> CN
r4 CO
^ fl
C
>» cfl
6 C
f-\
W
CO rH
CU i 1
,C CO
o so
4J ~-~«
CO rQ
4-> W
C &
CU Cfl
CU 60
m in
H
13 CN
MH ^
O 4->
H
co &
C
o ^*»
H H
U rH
0 CO
CU O
P. -H
CO CX
PJ t>>
M H
Cfl rP
Cfl
-------
- 54 -
CO
1
M
CU
cfl
H
^^
cfl
CO
W
fe
H
CO
W
H
Pn
0
CO
W
00 CO
^O ^H
I J
vD <
pq ^
PH
U W
H
W
S
O
H
0
W
PH
CO
C/D
C/3
^o !**» t^» CNJ csj o ^^ i r^* ^d"
vo D r*^ vO oo r^^ *^~ ^^ c^j to ^ co c*^ c*^ i~H
rH COrHCOOOVDCMOOrH CO r-» CM CM
CM rH LO CO
^-x
t*i
**-*
-IJ-IJ-lrGCO J3MH <4H (3
CU CU M V-l ^ . C
(UOOOCTvrHrHrHCOC O Cfl CO O
PPHUCJOUOSM H S P-i CJ
-d
c
cd
oo C
o- -HO
*H
CTi 0) 4-1
O^i ^O CO
B
Pi O
CO >-i
g cfl
1
co c
a o
H a
1 i
CO ^J
00 g 4-1
O^ O O
t_l tfi
W KM
ON Cfl
as
rH
Cfl CO
4-1 CU
O rH fi
4J ft O
g rH
CU Cfl rH
,TJ CO Cfl
4-1 bO
CU ---
4-J »C PH
d 4-J Cfl
00 vQ PQ
O> fj
CTi CO O
CT\ -H ID
CO fi
CO cfl tot)
o -I i ^-j
d
Q) ,J3 O
> o
HO 0
4J a oo
cfl
CU CU T3
rl g Cfl
oo d
O1! CO W O
CO 41
CT> C3 Cfl
CT* O T3 tfl
TJ -i
co o
as C i-i c MH
as co cu o cu
O 4-1 -H )-i
CO ft cfl H
CU g rH >
co o d FM
>~, 0 O
rH rH 0
CO 0) Cfl +
C js o
1 *-l rH Cfl
CO O Cfl
W M 4J
-------
- 55 -
TABLE B-4
GC ANALYSES OF AROMATIC FRACTIONS FROM SILICA-GEL
SEPARATIONS OF TEST FUELS AND BLEND COMPONENTS
Compound
Benzene
Toluene
Ethyl Benzene
m- & p-Xylene
o-Xylene
Benzene, i-C,,
n-CJ
1 M, 3 E
1,3,5 TM
1 M, 2 E
1,2,4 TM
1,273 TM
1,3 DE
Indan
Benzene, n-C
1,3 DM, 5 E
1 M, 2 n-C,
1,4 DM, 2 E
1,3 DM, 4 E
1,2 DM, 4 E
Unknown A
B
Benzene 1,3 DM, 2 E
1,2 DM, 3 E
1,2,4,5 TM
1,2,3,5 TM
Unknown C
D
E
F
G
H
I
J
Benzene, 1,2,3,4 TM
Unknown K
L
M
N
0
Naphthalene
Unknown P
Q
R
Methylnaphthalene
Total, Wt. Vo Aroma tics
Wt.
Low Aromatic
1969
0.31
3.40
0.49
1.79
0.68
<0.01
0.15
0.94
0.35
0.19
1.08
--
--
0.21
0.27
0.25
0.15
0.25
0.10
0.22
0.25
0.50
0.07
0.03
0.21
0.10
0.40
0.56
0.06
0.04^
0.10
0.06J
0.09
0.25
0.02
0.07
0.20
0.24
0.03
0.11
0.07
0.08
0.37
0.03
-.
--
--
14.8
1970
0.28
2.75
0.56
1.67
0.69
0.04
0.19
0.85
0.34
0.18
0.90
~-
--
0.18
0.19
0.09
0.18
0.19
0.12
0.21
0.29
0.51
0.07
0.08
0.09
0.19
0.44
0.64
0.06
> 0.13
0.06
0.15
0.05
0.35
0.25
0.10
0.03
0.10
0.06
0.34
0.06
--
--
0.05
13.7
% of Total
Gasoline
High Aromatic
1969
3.50
21.96
2.86
7.94
2.74
0.12
0.62
2.86
1.06
0.56
2.74
--
0.48
0.38
0.34
0.22
0.38
0.17
0.36
0.44
0.73
0.02
--
0.23
0.19
0.67
1.06
__
0.05
0.08
0.25
--
--
0.32
0.22
0.02
0.05
0.06
--
0.49
__
--
--
54.2
1970
3.55
24.88
3.21
6.89
1.98
0.10
0.45
1.87
0.75
0.42
1.92
--
0.36
0.29
0.12
0.25
0.29
0.17
0.29
0.64
0.37
0.08
0.11
0.10
0.22
0.56
0.82
0.07
0.14
0.10
0.19
--
0.07
0.41
0.28
0.11
0.06
0.13
0.07
0.57
0.06
-.
--
0.06
53.0
Lead-Free
Comm. Pre .
0.35
20.40
1.53
6.10
2.23
0.10
0.53
2.63
1.18
0.66
3.24
0.10
0.08
0.61
0.48
0.10
0.35
0.53
0.10
0.23
0.25
0.41
__
--
0.04
0.07
0.20
0.29
0.04
0.12
--
0.11
--
0.11
0.14
0.05
--
0.04
--
0.39
__
--
--
43.8
Wt.
% of Aromatic
Cm- 11 Arom. Solv.
1969
-_
--
--
--
--
._
--
0.3
--
0.9
4.2
2.7
2.5
3.7
2.8
4.4
5.4
11.1
__
3.0
3.8
8.3
12.2
1.4
0.2
2.2
0.8
1.1
3.5
0.5
0.5
5.8
4.4
0.3
2.9
0.7
2.1
6.5
-.
0.4
0.5
0.9
100.0
1970
..
--
--
--
--
__
--
--
0.3
--
0.6
3.2
2.2
2.1
2.6
2.5
4.1
5.3
10.6
__
--
3.1
4.3
8.5
12.5
1.5
0.2
2.3
0.7
1.3
3.6
1.2
0.6
7.1
5.2
0.3
2.9
0.7
2.1
6.7
--
0.4
0.4
0.8
100.0
Fraction
HCN + Poly.
0.9
7.4
2.9
15.2
5.7
0.1
1.3
9.0
4.1
2.1
12.8
--
2.4
2.6
3.6
--
2.6
0.6
1.4
1.5
2.4
0.6
0.2
1.6
0.5
1.4
1.9
0.5
0.1
0.9
0.5
0.7
2.1
0.6
0.8
--
2.0
0.2
1.9
1.0
0.9
1.6
0.8
0.3
100.0
-------
- 56 -
Appendix C
TEST VEHICLES AND TEST-OPERATIONS DETAILS
Appendices C and D of the First Annual Report described the
essentials of test-vehicle preparation and operation, the test-stand
facilities, the design and functioning of the PNA/phenol sample collector,
and the recovery of the samples from the collector for analysis. Sample
production continues to involve three cold-start blocks of 12 7-mode,
137-second test cycles, with substantially all the exhaust processed.
In the second year of CAPE-6-68, there have been no major changes in any
of these areas, except that a third vehicle has been incorporated in the
test fleet. This vehicle, a 1970 350 CID Chevrolet V-8, has thus far
yielded PNA data that appear less consistent than that from the earlier
vehicles, and certain tests have had to be discounted on the basis of
observed abnormalities.
Test Vehicles
The 1966 318 CID Plymouth without emission control (NC) and the
1968 307 CID Chevrolet (EM) that were used in the first-year work also
supplied the majority of the second-year data. Both have undergone com-
bustion-chamber deposit changes as part of their test use. Oil consump-
tion (10W-40 commerical in all cases) apparently rose slowly in the Plymouth
according to data obtained during vehicle operation on a mileage-accumulation
dynamometer for deposit modification. However, recent third year data show
this trend was due to an oil leak and that true oil consumption is near 0.5
qts./lOOO miles. Oil consumption rates in the 1968 and 1970 Chevrolets
have remained at low levels:
MAD Operation
C ity/Suburb/Highway
30 MPH Average
1000 miles, lead + P
4000 miles, no lead
2000 miles, lead + P
5000 miles, low TEL
5000 miles, no lead
Deposit
Fuel
Code
C3P
A
C3P
A.5
A
Quarts/1000 miles
Plymouth
1966
(b)
0.67
0.46
0.53
0.65
0.70
Chevrolets
1968
307 CID
l.ll(c)
0.56
0.46
1970
350 CID
0.32
(a)
(b)
(c)
All vehicles had additional prior deposits from commercial leaded
premium gasoline with phosphorus (Code C3P).
Before initial tests (first year), deposits were removed from the
Plymouth, and lead-free deposits were later introduced with a
commercial lead-free premium.
Affected by an oil leak.
-------
- 57 -
The fuel indicated as "low TEL" was a commercially available lead-free
premium, with 0.5 ccs. of TEL added per gallon, as Motor Mix, and without
phosphorus. Without TEL or P, the same gasoline was used where "no lead"
is indicated in the table.
The 1970 Chevrolet, in common with the other two vehicles, was
loaned by Esso Research from its test fleet. Prior to CAPE-6 use it had
been used in a program in which its exhaust emissions were within the 1970
standards. Vehicle preparation, similar to that described on p. 61 of
the First Annual Report, was followed by a 1000-mile MAD operation (see
above) and a regular emissions test for CO, HC and NO. However, repeated
emissions tests (cold-start 7-mode) were required, interspersed with
engine tuning and testing and, in some cases, replacement of parts
related to the vacuum spark advance controls, intake-air heater controls,
distributor spark advance schedule, ignition points, etc., in order to
pass the emissions standards for a 1970 vehicle. The vehicle was
ultimately accepted for PNA use when its composite cycle results were
0.58% CO (= 13.8 grams/mile), 178 ppm HC (= 2.2 grams/mile), and 853 ppm
NO. The precise cause or causes of the initial difficulty in meeting
1970 emission standards are not known. The problem centered on HC emis-
sion in the deceleration modes, in most cases. In the course of subsequent
PNA emissions testing, the vehicle malfunctioned on other occasions, with
the loss or invalidation of certain PNA tests as a result. In one case,
clean-up solvent was trapped in the laminated walls of the muffler and
boiled out during the next test, causing very high apparent HC emissions
and other abnormalities; the muffler had to be reconstructed with welded
joints. In another case, the vacuum hose for vacuum spark advance became
detached during a test, resulting in unusually high exhaust temperatures
at speeds above about 23 mph. Other difficulties included an apparent
binding or sticking in the automatic choke linkage, an apparent
deviation in hot-idle RPM which resulted in high HC emissions in decelera-
tion modes, and evidence of questionable performance of two resistor-type
spark-plug cables. The net effects of such difficulties, together with
other irregularities such as variations in choke action at start-up, were
that three fuels had to be tested twice and one fuel three times in the
course of the work reported here; single tests are normally used. The
rejection or questioning of PNA results from certain of these tests is
discussed further in Table C-l. Also listed in this table of rejected
tests is a Test (No. 54) from the Plymouth vehicle that was carried out
as a base-point test prior to a deposit change needed for third-year work. As
noted in the table, this test had evident defects from the outset, and
these are believed to have caused an abnormally high. PNA-emission result.
In this particular case, there appear to be factons above and beyond
explanation in terms of the observed CO-emission level in the test. Accord-
ing to the first-year conclusion as to the effect of CO differences, the
PNA yield in this test would be increased only about 15% above an earlier
-------
- 58 -
equivalent test because of the observed CO-level difference between the
first blocks of the two tests, with the later two blocks of the tests
equal. In fact, the PNA difference shown in the table is much larger,
but the cause is obscure. The long (4-month) idle period between tests,
and the heavy dynamometer load at startup, are both potential explanations
and represent situations to be avoided. The earlier test, which was
consistent with other data, will have to be used as the base case for
evaluating future results, and the results of Test 54 are rejected.
Test Operations Details
Table C-2 presents data on the average unburned hydrocarbon
emissions (NDIR, C(,~) from all three vehicles and in different deposit
conditions over both contract years. Two facts are clear: (1) Depending
on fuel aromaticity, the apparent HC emission can vary widely; this is
at least partly an instrument sensitivity effect. (2) Deposit condition
has a marked effect on HC emission, with the "lead + phosphorus" case,
at least for the fuels used here, clearly giving the highest HC emission,
and with the "lead-free" deposits giving less. It should be noted,
however, that the "lead + P" and "lead-free" fuels used different base
stocks, as well as differing in additive content. Evidently, the
emissions of the 1970 Chevrolet (see above) would have been acceptable
at the outset if a different choice of test fuel and deposits had been
made. The corresponding vehicle emissions of CO and NO, as given in the
data tables in the main text are not tabulated here in the same way,
because, as shown in the first year (for the Plymouth), CO and NO were
related (by regression equations) to ambient humidity rather than to
fuel or deposits. Second-year Plymouth CO data fit the same humidity
correlation until test 41, at which point carburetor service established
a new relationship for subsequent tests.
Tables C-3 through C-43 of this Appendix show the average
emission of CO, IRHC, and NO for each mode of each cycle in cycles 1-4
and 6-7 in each test of the second contract year. The values are based
on "corrected" (to C02 + CO = 15) values from computer printout sheets,
averaged over all available printouts for 1 to 3 of the 3 cold-start
blocks that form a PNA emission test. As in the first-year work, the
choke action of all three vehicles at startup was variable, so that start-
ing mixture ratio and engine speed ranged over fairly wide limits in
different tests or in different blocks of the same test.
Table C-44 records the maximum and minimum muffler-exit and heat-
exchanger inlet gas temperatures observed in hot test cycles for all second-
year tests. The 1968 Chevrolet exhaust is only slightly warmer than that
from the 1966 Plymouth, while the 1970 Chevrolet is over 100 Fahrenheit
degrees hotter. This higher muffler-exit temperature, for the 1970 car,
combined with a slightly higher gas flow rate (measured at the PNA collector
outlet) also results in higher exchanger-inlet temperatures; even the
exchanger-outlet temperatures are about 5° higher for the 1970 vehicle.
Examination of temperatures for individual tests in Table C-44 reveals
some conspicuous departures from typical values. No explanation can be
-------
- 59 -
offered for these variations; the likely causes, such as dynamometer loading
and spark timing do not appear to be generally useful. In some, but not all
cases, higher temperatures appear associated with higher fuel consumption,
but the variations in fuel consumption are, in themselves, unexplained. In
any case, it is likely that the higher average temperatures and fuel consump-
tion for the 1970 Chevrolet, relative to the 1968 Chevrolet, are at least
partly due to the lack of vacuum spark advance below 23 mph because of
the action of the TCS (transmission-controlled spark) system in the 1970
vehicle.
Between-vehicle differences in exhaust-gas temperatures at the
muffler might be expected to cause differences in the tendency of the mufflers
of the different cars to retain PNA that is emitted from the engines in
the course of a test. The data in Table C-45 suggest that this is indeed
the case, and that at the temperature level of the 1970 Chevrolet it is
not really necessary to include the muffler washings in a PNA-sample
workup, even though cleaning the muffler between tests does appear to be
in order. The data in this table were obtained by separate GC/UV analyses
of muffler washings from 18 tests, covering a variety of fuels in three
vehicles. The data are admittedly unbalanced in terms of vehicle identity,
fuel TEL, engine deposits, etc., but no further separate analyses are planned.
Muffler washings will continue to be added to the main samples for work-up.
Muffler clean-out was improved, during the year, by drilling a small hole
near the side of one end-wall of each muffler and extending its entire
length. This hole promoted drainage of clean-up solvent from the muffler;
it was closed by a screw plug when neither filling nor drainage of solvent
was needed. With cover plates also blocking the gas inlet and outlet,
each muffler could be charged with a small amount of solvent and mechanically
rotated to ensure thorough washing. One problem encountered here was that
cleanout solvent could penetrate into an insulation layer between the inner
and outer shells of the 1970 Chevrolet muffler; the consequences are mentioned
in Table C-l; the remedy was rebuilding the muffler and adopting, as the
last step of clean-up, a procedure in which filtered air is drawn through
the muffler while it is warmed with electric heating tape.
-------
- 60 -
Table C-l
CAPE-6-68
TESTS OF DUBIOUS VALIDITY AS TO PNA
(Tests 45, 50, 51 in 1970 Chevrolet, 54 in 1966 Plymouth)
Test
No. Comment
45 Solvent from prior test trapped in laminated muffler wall, boiled
out in early part of first block, caused off-scale HC results
with other analyses normal. Aborted Block 1 after Cycle 3; choke
was then open but oil and coolant not at equilibrium temps. Block
2 was run for 21 cycles, with oil at «v225°F for 9 cycles (20
minutes), where normally oil only reaches 220°F. at end of 12-cycle
block. Sample from muffler dark yellow (abnormal), high in
BghiP and BaP. Yields of phenol and BaA "normal," of BaP, high.
Note: Similar HC retention in muffler wall in Test 46, but with
the 3 standard 12-cycle blocks, showed low emission of all PNA
and low muffler hold-up.
51 Same fuel as Test 45, no muffler abnormality. Incorrect idle
speed setting is believed to be cause of abnormally high HC
emissions in deceleration modes (modes 4 and 7). All PNA high
except pyrene; samples dark colored, abnormal.
50 In third block, lost vacuum spark advance hose connection; suppressed
NO, elevated peak exhaust temperatures~80°F. at muffler exit;
strongly oxidized exhaust HC, particularly in Mode 7. PNA yields
are very high; muffler sample and main sample abnormally dark
and dirty. BaP exceeds BaA (abnormal).
54 Vehicle was idle for 3.7 months. Dynamometer was pre-loaded
for 16 BIIP at 50 MPH, i.e., 2 x normal; reduced to near normal by
approximately 12th minute of test. The entire first block of
this test was rich, relative to the earlier test 41 which was
expected to be matched:
_%_C_0_;_ Cycle Results
Block: 1 2 _3 BaP/gal. BaA/gal. Fuel, gal.
Test 41 1.42 1.49 1.45 37 49 1.93
Test 54 1.89 1.50 1.44 128 195 2.04
-------
- 61 -
Table C-2
NDIR HYDROCARBON EMISSIONS - CAPE-6-68 TESTS
(a)
(Data of Both Contract Years)
Average Hydrocarbons, Cfi, ppm (No. of Tests)
Vehicle, Fuel Aromatics:
Deposits Fuel Set(b)
'66 NC
Clean
Lead-free (A)
Lead + P (C3P)
Low-lead (A. 5)
Lead-free(A)
'68 EM
Lead + P (C3P)
Low-lead (A.5)
Lead-free (A)
11% - 12%
1969 1970
677(2)
776(2)
817(2) 740(2)
458(2) 417(2)
28%
1970
641(3)
570(3)
511(2)
298(3)
293(3)
273(1)
46%
1969
590(2)
600(2)
624(4)
496(1)
285(4)
193(1)
1970
491 (
185 (
'70 EM
Lead +
(C3P)
205(3) 232(4) 178(3) 135(1) 134(3)
(a)
(b)
(c)
Each value is the average of 1 to 4 test results, each of which is the
average from results on 1 to 3 test blocks out of 3 blocks comprising
a test. Deposit conditions listed in order of occurrence.
Fuel sets (by year) are identified in Appendix B.
Tests 50 and 51 (abnormal) omitted from averages. First blocks of
certain other tests, in which muffler clean-out solvents affected
apparent IRHC at startup, are also excluded; see Table C-l.
-------
- 62 -
-------
-------
- 64 -
i\
U 13
U £{
-J CM
- , CM 00
>S
-------
- 65 -
I
I
-------
- 66 -
-------
- 67 -
i M
< o
I
o i
I
OH
<
S
-------
- 68 -
-------
- 69 -
i i
-------
- 70 -
-------
- 71 -
-------
- 72 -
^i u
-------
- 73 -
a
oi 3
12
Oj .
o i o
-------
- 74 -
PH
<
3
I
3 ^
< o
CMf ,
o I.
g o.
-------
- 75 -
f oo
\
e
o.
a
i
|
-------
- 76 -
PLI M M-l
I
o
fX
§
o
-------
- 77 -
ol
4J JJ -rl 0)
4-* £ O 01
(U W
1 J=
4-J (0
^ t) 4J
-------
- 78 -
PL. M 4-)
-------
- 79 -
o s.
-------
- 80 -
bl to
0. M
< CO
3 t-
i
Pu
<
3
S
K E sO
OiO.0
M a^-'
-------
- 81 -
o t.
ta a.
I
o I.
g O.
-------
- 82 -
TABLE C-4
CAPE-6-68
EXHAUST ANALYSIS DATA - TEST NO.
59
Data are averages from first 7 cycles of ^ 12-cycle blocks.
Avg., Cone, in Sees. Read, Corr. to CO + CO = 15
Mode No.
Wtg. Factors
% CO
IRHC,
ppm
(C6)
NO,
ppm
Cycle 1
1
0.042
3.78
2 j -84
3 ! .21
4 -19
Avg. 1-4' L26
6 .32
7
.34
Avg. 6-7; -33
2
0.244
4.42
.73
.64
.16
1.60
.57
.55
.56
3
0.118
2.30
.16
.14
.21
.69
.14
.16
.15
4
0.062
2.70
.47
.21
.14
.90
.28
.28
.28
5
o.qsc)
1.06
.18
.15
.52
.39
.15
.16
.16
6
0.455
1.46
.52
.48
.52
.75
.53
.56
.55
7
0.029
2.77
.56
.48
.48
Corr. &
Wtd.
V
1.07 .95
.55
.65
.60
X
.43
Composite Test Result, % CO 60
1
2
3
277
134
131
4 125
Avg. 1-4 167
6 121
7
110
Avg. 6-7 j H6
251 1
116
122
107
149
103
91
98
"
257
172
167
173
193
156
153
155
225
167
177
191
190
164
188
177
128
83
97
74
96
71
67
69
169
116
126
105
129
111
107
109
331
412
479
455
419
499
519
509
_ <
/ \
151
-w .>-
124
Composite Test Result, ppm HC ' l }?
1
2
3
4
Avg. 1-4
6
7
Avg. 6-7
409
222
287
313
308
311
314
313
365
771
885
811
709
745
720
733
977
1667
1666
1615
1481
1675
1746
1711
445
730
893
774
711
801
799
800
269 992
301
298
299
292
310
315
313
1335
1441
1457
1307
1473
1408
1441
575
_M6_
__8_68_
825
\ ^'^
A
769 1033
839
814
^^^
827 H37
Composite Test Result, ppm NO
1100
-------
Vehicle
'66 Plym.
'68 Chev.
'70 Chev.
- 83 -
Table C-44
MUFFLER EXIT AND EXCHANGER-INLET TEMPERATURES
(Hot Cycles)
Test
No.
20
22
23
26
27
28
32
34
37
41
42
43
54
21
24
25
30
31
33
35
36
38
39
40
44
45 (b)
46
47
48
49
50^°'
51 W)
51B
52
53
55
56
57
58
59
Muffler
Block 1
427/557
417/550
410/550
416/549
423/560
/
456/590
447/565
457/569-
452/575
457/587
450/577
457/575
475/588
485/595
467/572
480/590
485/607
499/615
485/606
485/602
477/597
475/590
477/595
562/685
/
577/701
578/705
570/708
583/714
568/710
535/680
583/703
585/710
583/708
607/735
615/743
575/700
565/690
565/698
Exit, °F.,
Block 2
425/560
417/547
421/550
419/550
418/560
451/586
458/591
454/565
449/571
452/575
457/583
450/577
457/578
495/605
485/593
467/573
496/610
487/605
497/615
483/603
490/606
494/602
485/600
480/595
570/690
570/693
577/698
578/705
570/710
582/711
570/712
542/680
588/712
590/702
583/707
607/735
603/735
567/685
561/682
563/698
min . /max .
Block 3
425/555
417/547
420/550
422/550
420/555
452/585
455/585
450/570
452/572
457/577
455/585
452/580
460/575
493/600
478/585
470/575
492/610
487/605
489/612
483/603
487/605
492/605
485/599
484/599
575/689
580/697
574/704
573/704
572/705
585/715
643/800
538/685
589/710
587/700
585/709
607/734
612/737
567/685
565/685
568/706
Exchanger
Block 1
212/293
217/292
212/287
207/279
220/297
225/310
230/308
/
/
/
235/315
220/300
230/305
245/
259/322
243/302
250/305
250/312
265/330
255/313
254/311
254/315
256/319
251/307
302/369
/
303/372
305/374
307/373
312/380
303/375
278/357
322/385
317/380
326/390
/393
327/404
300/370
298/370
298/365
Inlet, °F.,
Block 2
215/294
217/292
216/288
207/280
220/295
228/311
226/305
234/303
227/305
/
237/317
222/300
228/305
263/325
260/320
243/302
260/320
250/312
263/325
254/311
258/317
261/326
260/325
__/__
307/370
318/383
307/373
310/375
310/375
312/
/376
286/363
327/391
323/385
330/393
322/395
324/398
293/364
295/365
298/368
min . /max .
Block 3
212/293
217/291
217/292
210/280
220/296
230/310
228/306
230/300
227/305
222/300
235/315
225/304
227/303
264/324
257/315
244/302
257/318
253/315
258/322
252/309
256/317
266/326
262/325
__/__
305/370
318/383
307/373
307/374
310/375
316/385
347/432
285/363
327/391
327/385
332/395
324/397
325/403
293/360
298/369
301/372
(a)
Defect in temperature recorder.
First block aborted at Cycle 3; Block 2 was 21 cycles.
(c)
Lost vacuum spark advance (above 23 MPH only) in Block 3.
(d)
All temperatures low, reason unknown. Deceleration hydrocarbon high.
-------
- 84 -
Table C-45
CAPE-6-68
MUFFLER RETENTION OF PNA SPECIES
Test Test
No. Vehicle
34 '66 Plym.
37
41
42
43
38 '68 Chev.
39
40
44 '70 Chev.
45 (a)
46
47
48
49
50(a)
51(a)
51B
52
Fuel
TEL
0
3
0
0
0
0
0
0
3
3
3
3
3
3
3
3
3
3
Muffler PNA
BaP
18
10
8
5
12
17
3
5
2
10
2
6
1
2
4
2
4
4
BeP
_
3
3
3
6
2
10
1
1
5
0.8
0.8
0.8
0.8
1
0.1
2
0.9
BghiP
10
7
15
7
0.7
23
2
9
0.9
1
6
2
4
3
content as
RPA Chry
17
6
5
4
8
4
4
3
0.6
0.7
0.5
1
2
<1
0.7
1
1
0.5
% of
sene
14
3
4
2
6
2
7
2
0.4
0.5
1
0.9
0.5
0.9
0.8
0.3
Total PNA
Pyrene Triphenylcru
10 10
4
3
0.2
0.1
6
4
1 2
0.4
0.9
0.3
0.5
0.4
0.2
0.1
1.1
0.8
0.3
(a)
Abnormalities in this test (see text) may influence the results reported
here.
-------
- 85 -
APPENDIX D
CORRELATION AND APPLICATION OF THE
GC/UV PNA METHOD
During the second year of CAPE-6-68, CRC-APRAC Project CAPE-
12-68, also at Esso Research, developed a GC/UV method for analysis
of exhaust-tar solutions (CAPE-6 samples), gasolines, or oils for
PNA content. The method is described in Appendix E. Major differ-
ences from the first-year procedures given in Reference (1) include
resumption of the addition of radiotracer spikes before, rather than
after, boildown, with new precautions to assay the exact amount of
activity added, caustic extraction to remove possible interfering
materials, a modified alumina-column separation and elution, use of the
programmed GC separation to yield trapped peaks (in place of TLC
separation and developing, sectioning, and extracting the TLC bands),
and the availability of assays on eleven different PNA species
(formerly, only BaP and BaA). The assays are more accurate for five
of the species than for the remaining six, as indicated in a footnote
on Table D-2; the analysis for benzo (g,h,i) perylene has been
particularly likely to vary, or to be in error, according to a recent
examination of standard-sample results.
Parallel analyses, by CAPE-12-68 using GC/UV and by CAPE-
6-68 using the TLC method of the first year, were carried out on
samples from 12 CAPE-6 tests. In certain tests, equal halves of the
original sample have been independently spiked (after boildown, rather
than before) with radiotracers, and two independent TLC analyses have
been carried out on the two samples, with GC analysis of one. Table
D-l compares the test yields, in micrograms of BaP and BaA,calculated
from results of the two methods. The regression equations relating
these results, and the average ratios of results, were presented in
Section 3.3 of the main text of this report. Following the results
of this comparison, the GC/UV method was adopted for CAPE-6 use and
gradually expanded to include up to eleven PNA species.
The available multiple-species analyses are given in Table
D-2, covering 29 vehicle tests. The alkylated BaA, BaP, and BeP
species are not precisely defined as to the points of substitution on
the parent structures; it follows that if the UV spectra differ
slightly between isomers, the absorptivity values (Appendix E) may not
be precisely correct. Certain of these isomers elute in adjacent GC
peaks which are separately trapped and the UV absorbances then added
together. Moreover, where several species (BaA, chrysene, triphenylene)
occur in one GC peak, the UV absorbance for one species at its selected
wave-length may require correction for absorbance due to another
species, which is assayed by the absorbance at some other wave-length.
Disregarding, of necessity, these sources of error, the
ratios between some of the prominent PNA species have been calculated
and are given in Table D-3. Each ratio has a fairly uniform value,
with the exception of BeP/BaP (i.e., benzo (e) pyrene/benzo (a) pyrene).
-------
- 86 -
This ratio is apparently higher in data from the 1970 Chevrolet than
from the other two cars, and especially so if the data from "dubious"
tests are excluded. It is not clear whether this represents a differ-
ence between carburetion or other characteristics of the vehicles, or
a difference in the deposits present when the tests were run. For the
data in the 1970 vehicle, the BeP/BaP ratios from low-aromatic fuels
appear to be higher when the high-boiling naphtha was present (1969
fuel set) than when it was absent.
Because of the possibilities for error noted above, and on
Table D-2, conclusions are probably not justified until more multiple-
species analyses are available from future work.
-------
- 87 -
Table D-l
PNA ANALYSES BY TLC AND GC METHODS
Yield from Test, yg
CAPE-6
Test No.
16
19
22
23
24
26
27
28
29
30
31
32
Sample
No.
1
2
1
2
1
2
1
2
TLC
172
279
86
70
20
105
83
130
131
129
68
73
12
10
16
39
BaP
GC
184
281
95
89, 89
24
102
140
133
82
(a) ^
21
40
BaA
TLC
402
567
189
255
77
270
253
270
218
213 .
209 (a)
164
46
55
45
62
GC
361
464
162
217, 239
60
272
234
183
167
48
47
68
(a)
Value rejected as derived from low, or impure,
UV spectrum on extract of TLC band.
-------
- 88 -
Table D-2
CAPE-6-68 PNA EMISSION DATA
MULTIPLE-SPECIES ANALYSES OF EXHAUST SAMPLES
BY THE GC/UV METHOD OF CAPE-12-68(f )
CAPE-6
Teat
No.
32
34
37
41
42
43
54
33
35
36
38
39
40
44
45
46
47
48
49
50
51
51B
52
53
55
56
57
58
59
BaP
40
73
62
71
98
86
260
7.9
30
5.8
30
21
19
48
91
26
31
70
42
305
245
52
90
64
64
92
112
48
63
BeP
157
377
236
145
176
603
54
28
168
51
53
281
186
247
247
392
262
1171
1057
306
437
482
362
507
518
467
312
BghiP
188
127
195
324
216
1318
35
81
134
208
45
67
115
236
1254
1121
334
431
565
483
278
766
54
223
BaA
68
102
97
94
142
143
398
26
32
25
34
46
39
170
144
94
71
124
101
275
294
154
187
146
125
158
163
74
143
Total Test
Chrysene
206
167
180
171
316
414
862
50
72
69
81
87
108
276
201
182
163
232
199
560
597
238
310
290
258
330
310
175
303
Yield, Including Muffler Washings, UK.
Pyrene
1727
1140
1220
1283
1514
3911
362
640
832
1088
533
757
425
568
626
1472
515
665
1496
1273
702
1143
1525
552
1386
Triphenylene Me-BaA DM/Et-BaA Me-BaP Me-BeP
71
61
47
145
30
28
26
28
45
69
55
73 116 25 36 293
76
62 53 17 17 131
46 78 21 24 262
76 81 17 29 266
46 22 6 11 71
72 61 12 9 104
Notes
(a)
(d)
(d)
(e)
(b)
(d)
(d)
(d)
(d)
(d)
(c)
(e)
(e)
(e)
^All tests in this group are in the 1966 Plymouth.
C^All tests in this group are in the 1968 Chevrolet.
^C'A11 tests in this group are in the 1970 Chevrolet.
^Analyst reports a contaminant (dioctylphthalate?) leading to lower precision in GC peaks and 0V spectra and
sometimes requiring additional column separations. Accuracies of BaA, chrysene, and triphenylene are
particularly affected. The contaminant has now been avoided.
^Tests of doubtful validity for reasons discussed in Appendix C.
^BaP and BeP elute together, as do BaA, chrysene, and triphenylene. Both of these GC peaks contain tracers
(BaP and BaA); the accuracy of these five species is thus superior to that of all other species, for which
the tracer showing the greater recovery is the basis for scale-up to total yield, even though it occurs in a
different GC peak than the species being estimated.
-------
- 89 -
Table D-3
CAPE-6-68
RATIOS OF PNA SPECIES
IN EXHAUST SAMPLES
Test
No.
32
34
37
41
42
43
54(3)
33
35
36
38
39
40
44
45 (a)
46
47
48
49
Test
Vehicle
'66 Plym.
51
51B
52
53
55
56
57
58
59
'68 Chev.
'70 Chev.
(GC/UV Analyses)
Total-Yield Ratios
BeP/BaP
3.92
6.08
3.32
1.48
2.05
2.32
1.80
4.83
5.60
2.43
2.79
5.85
2.04
9.50
7.97
5.60
6.24
3.84
4.31
5.88
4.86
7.53
5.66
5.51
4.63
9.73
4.95
BaA/BaP
1.70
1.40
1.56
1.32
1.45
1.66
1.53
3.29
1.07
4.31
1.13
2.19
2.05
3.54
1.58
3.62
2.29
1.77
2.40
0.90
1.20
2.96
2.08
2.28
1.95
1.72
1.46
1.54
2.27
Chrysene/BaA
3.03
1.64
1.86
1.82
2.23
2.90
2.17
1.92
2.25
2.76
2.38
1.89
2.77
1.62
1.40
1.94
2.30
1.87
1.97
2.04
2.03
1.55
1.66
1.99
2.06
2.09
1.90
2.36
2.12
Triphenylene/BaA
1.04
0.60
0.48
0.36
1.15
0.88
1.04
0.82
0.98
1.77
0.54
0.39
0.52
0.50
0.29
0.47
0.62
0.50
(a)
Test of dubious validity. See Appendix C.
-------
- 90 -
Appendix E
Analytical Procedures
(Adapted from CAPE-12-68 Description)
Introduction
This method was developed to determine polynuclear aromatic
hydrocarbons in gasoline, crankcase oil, and tar from auto exhaust;
the tar-sample preparation is described in Reference (1).
The method covers polynuclear aromatics (PNA) ranging from
pyrene to benzo(g, h, i) perylene. As described herein, eleven PNA's are
measured, including: Pyrene, benz(a)anthracene, chrysene, triphenylene,
methylbenz(a)anthracene, dimethyl and/or ethylbenz(a)anthracene, benzo(a)-
pyrene, (benzo(e)pyrene, methylbenzo(a)pyrenes, methylbenzo(e)pyrenes and
benzo(ghi)perylene. Additional PNA's can be included. The procedure was
successfully demonstrated for gasoline, exhaust tars and crankcase oils.
Summary
A sample to be analyzed is spiked with known quantities of
carbon 14 labeled benzo(a)anthracene (BaA) and benzo(a)pyrene (BaP). A
caustic treat is done and then PNA hydrocarbons are removed from the total
sample by separation on a column of partially deactivated alumina. The
solvents used are cyclohexane, cyclohexane-benzene, benzene, and benzene-
methanol. The fraction containing the PNA's is reduced to a minimum
volume by evaporation on a steam bath. An aliquot of this sample is
injected into a gas chromatograph and fractions are collected at predeter-
mined retention times. Quantitative measurement of each Pna is based on UV
spectra of collected fractions.
Apparatus & Isotope Dilution
1. The Gas Chromatograph is equipped with a flame ionization
detector, a linear temperature programmer and a flow controller. The
injection port consisted of an aluminum block with a stainless steel tube
insert (quartz sleeve). The chromatograph is modified by the addition
of a by-pass and trap so that 15% of the effluent of the column will go
to the detector and 85% to the trap. A Perkin-Elmer 900 chromatograph
was used in the development of this method, but equivalent instruments
should perform satisfactorily. A diagram of the trap is presented in
Figure E-l.
-------
- 91 -
Figure E-I
GAS CHROMATOGRAPH TRAPPING ASSEMBLY
1/8" COLUMN
STANDARD PE900
BULKHEAD & SPLITTER
1/16" LINE
TO FLAME
DETECTOR
SWAGE LOK
200-9-316
SWAGELOK
<"100-R-2-316
INSULATION
1/4" O.D. TUBE 6" LONG
EXPANDED TO MATCH
SWAGELOK UNION
S.S. TRAP
1/8" TUBING 7" LONG
-------
- 92 -
2 Recorder. 0-1 millivolt, 2 min/inch chart speed.
}, Trapping Tubes. 20 or more stainless steel tubes 7" long
by 1/8" dianeter.
A. Column, Ten feet of 1/8" O.D. stainless steel tubing packed
with 2 percent SE-30 (GC Grade) on Chromosorb G (acid washed and D.MCS
treated), 80/100 ir.esh. The prepared packing is available from Supelco
Inc., Bellefonte, Pennsylvania, and other suppliers.
5. Syringe. 10 microliter. The syringes, Cat. #701N, manufactured
by the Hamilton Co., P. 0. Box 307, Whittier California 90608, and Model C
of Precision Sampling Corp., P.O. Box 15119, Baton Rouge, La., are satisfactory.
6. Spectrophotometer, spectral range 225 nanometers-400
nanometers with spectral slit width of 2 nanometers or less. Under
instrument operating conditions for these absorbance measurements, the
spectrophotometer shall also meet the follovring performance requirements:
absorbance repeatability, +0.01 at 0.4 absorbance; absorbance accuracy,
+ 0.05 at 0.4 absorbance; wavelength repeatability, +_ 0.2 nanometer;
wavelength accuracy, +_ 1.0 nanometer.
7. Spectrophotometric Cells, fused quartz cells having optical
path lengths of 1.000+ 0.005 cm. and 5 cm. cells. The 5-Cm cells are micro-
cells and contain about 3 ml of solution.
£. Graduate Glass, 5 ml.
9. yials, one- and three-dram with cap.
10. Steam Bath equipped with nitrogen outlet for purging vials.
11. Medicine Dropper, with tip drawn out.
12. Special Low UV Emission Room Lights, Westinghouse W-40-gold
fluorescent lamps.
13. Chromatographic Column, two feet long 0.5 inch OD and 0.43
inch ID with 200 ml. receiving bulb on top and stopcock with Teflon barrel
on the bottom.
14. Radiation Counters used are Intertechnique SL-20 and Packard
Tri-Carb.
15. Counting Vials, 20 ml. with cap.
-------
- 93 -
REAGENT AND MATERIALS
Regeant grade chemicals must be used tnroughout. The polynuclear
compounds were used as purchased. Compounds of comparable purity from other
suppliers may be employed.
1. Cyclohexane As a large amount of cyclohexane is used, it
must be of highest purity. To check purity, evaporate 180 ml. down to
5 ml. Run a UV scan on this residue in a one-cm, cell from 280-400
nm. The absorbance should not exceed 0.01 units. To purify, percolate
through activated silica gel, Grade 12, in a glass column, 90 cm. long
and 5-8 cm. diameter. Satisfactory solvents may be purchased from
several sources including Burdick and Jackson laboratories, Inc.,
Muskegon, Michigan 49442. Their products are called "Distilled in
Glass Solvents."
2. Toluene,
3. Acetone.
4. Sodium hydroxide.
5. Hydrochloric Acid.
6. Benzene, Burdick and Jackson Laboratories, Inc.
7. Methanol, Burdick and Jackson Laboratories, Inc.
8. Benz(a)anthracene C10H10, MW 228, Eastman 4672.
lo i/
9. Benzo(a)pyrene C20H12' M
-------
- 94 -
Weigh exactly 25.0 ings, of each of the foor compounds listed as
Items 4 through 7. Place in a 25-ml. volumetric flask, add 20 ml. of
toluene and swirl until compounds are in solution. Then make up to vol-
ume. This will contain 1 microgram of each compound per microliter of
solution. Pour into a small, narrow neck brown bottle and keep in a
cool, dark place, preferably a refrigerator.
13. Organic Counting Solution 8 gm of BBOT in a liter of toluene.
METHOD
Precautionary Note; Because of the sensitivity of the test,
the possibility of errors arising from contamination is great. It is of
the greatest importance that all glassware be scrupulously cleaned to re-
move all organic matter such as oil, grease, detergent residues, etc.
Examine all glassware, including stoppers and stopcocks, under ultra-
violet light to detect any residual fluorescent contamination. As a pre-
cautionary measure it is recommended practice to rinse all glassware with
purified isooctane immediately before use. No grease is to be used on
stopcocks or joints. Great care to avoid contamination of samples in
handling and to assure absence of any extraneous material arising from
inadequate packaging or storing is essential. Because some of the poly-
nuclear hydrocarbons sought in this test are very susceptible to photo-
oxidation, the entire procedure is to be carried out under subdued light.
Avoid use of fluorescenc lamps, special yellow lights are available.
Tygon tubing must absolutely be avoided , to eliminate Di(iso)octyl phthalate
(DOP) contamination which interferes in both the GC and UV.
Radioactive Materials must be handled and disposed by accepted
methods. Glassware should be rinsed with chloroform-acetone solvent and
washed in dichromate-sulfuric acid cleaning solution.
A. Distillation of Exhaust - Tar Solution
Before commencing the distillation prepare the radioactive
BaA and BaP spikes. Add 50 A of BaA to a 25 ml. volumetric flask
and 50 X of BaP to a second. Make up to volume with cyclohexane and remove
duplicate 100 Xaliquots from each for counting. Pour the contents of t
flasks into the still and rinse out well with cyclohexane and acetone
Conduct the distillation as described in Reference (l)v.
B. Caustic Extraction
1. The volume of still bottoms will be about one liter. Take exactly
one half of this and place in a one-liter separatory funnel.
-------
- 95 -
2. Add 50 ml of 0.5 N aqueous sodium hydroxide solution, shake one minute,
and remove lower aqueous phase.
3. Repeat the extractions with 30 and 20 ml portions of sodium hydroxide.
4. Wash with 50 ml of water, then 50 ml of 1/10 N HC1.
5. Finally wash with 25 ml of water.
6. Evaporate cyclohexane to about 150 ml and filter, if necessary.
7. Evaporate filtrate to below 50 ml, place in a 50 ml volumetric flask
and make up to volume with cyclohexane.
8. Take 25 ml for placing on the alumina chromatographic column.
C. Column Chromatographic Separation.
1. A column of 2 ft. length and 0.43" I.D.. is prepared with de-
activated alumina. Seventy-five grams of Woelm Neutral alumina are dried
at 150°C. in an oven 1 hour. The alumina is transferred to a bottle and
allowed to reach room temperature. 1.50 ml of water are added dropwise
with shaking. The bottle is capped and placed on a paint shaker for 15
minutes to obtain a uniform mixture.
2. A glass wool plug is placed in the end of the column and the
alumina is poured in along with gentle tapping to pack the alumina. The
column is filled to within one inch of the top and another glass wool plug
is placed on top.
3. Ten ml. of cyclohexane are placed on top of the column and
allowed to run into the column under nitrogen pressure of 2 pounds. The
sample in cyclohexane is then poured into the 200 ml, bulb on top and
allowed to run into the column. 100 ml cyclohexane is added to the bulb.
4. When the cyclyhexane has run into the column below the top
glass wool plug, 100 ml of cyclohexane benzene (4:1) are added and run into
the column. The first one hundred (cut #1) ml of cyclohexane is collected
and set aside. The following solutions are collected in 3 dram vials and
capped. Each should contain 10 ml of solution.
-------
- 96 -
5. After the cyclohexane benzene (4/1) is below the top of the
glass wool plug, 100 ml of benzene are added to the column. Collection of
10 ml fractions is continued.
6. After all the benzene has passed the top glass wool plug,
100 ml of benzene methanol (1:1) is added to the column. The benzene
methanol front on the column is followed by the movement of an orange
colored band usually in the sample. In addition, the front can be detected
by a warm front moving down the column which is apparent to the touch
or followed with UV lamp. When the front reaches the end of the column
a two phase system becomes visible in the collecting vial. At this point,
the remaining solution is collected in one large bottle and allowed to run
off the column (cut 3). Column should not be stopped or allowed to run
dry between solvent additions.
7. The solutions in the individual 3 dram vials are scanned in
order of elution on a uv spectrophotometer. The peak at 340 nm is used
as a guide in determining the start of elution of tetracyclic PNA compounds.
The solution whose spectrum shows a significant 340-peak is the start and
all remaining solutions are combined as the PNA fraction (cut #2).
D.. Concentration of PNA's
1. Transfer the contents of the first 4 vials of cut #2 to a
150 ml beaker and rinse each vial twice with 1 to 2 ml of cyclohexane and
combine with the contents of the beaker.
2. Place the 150 ml beaker on the steam bath under a small jet
of nitrogen.
3. As evaporation progresses, add the contents of the remaining
vials with similar rinsing to the beaker. Do not evaporate to less than
two mis and in no case let the contents go to dryness.
4. Transfer the concentrated solution to a one-dram vial.
Wash beaker with cyclohexane and add to vial. If it ±s necessary to obtain
the weight of the extracted material, the tare weight of the vial should
be determined to the nearest 0.1 milligram.
5. Place the vial in a 30 ml beaker for support. The beaker
is placed on a steam bath under a gentle jet of nitrogen.
-------
- 97 -
6. Rinse the 150 ml beaker with 3 or 4 small portions of cyclo-
hexane (~1 ml) and combine with the contents of the one-dram vial as
evaporation proceeds.
7. If, as is preferable, it is unnecessary to obtain the weight
of the PNA residue, evaporate the solution down to about 50 .ul (0.05 ml).
Add about 0.5 ml of toluene and again evaporate to 50 jul.
8. If the weight of the residue must be obtained, exercise
great care as the solution approaches dryness. Keep the vial under constant
observation and the instant all the solvent has been removed, cool and weigh
the vial.
9. Return the vial to the beaker under nitrogen for an additional
minute, cool and reweigh. Repeat this procedure until a constant weight is
reached. For the GC analysis add 10 to 20 microliters of toluene to lower
the viscosity of the sample. Cap and save.
E. Gas Chroraatographic Analysis
Parameters
The following parameters are employed in developing the gas
chromatographic separation with the Perkin-Elmer 900. Other instruments
might use different conditions.
Carrier Gas - Helium
Flow Rate - 30 ml./min.
Hydrogen and Air - at manufacturers
recommendation, or optimum rates.
Injection Port - 300°C.
Detector - 340°C.
Program - 175°C to 300°C at 4° per
minute. The temperature is held at
300° until all peaks are eluted, but
in any case, for 20 minutes.
Dual columns are not used.
-------
- 98 -
Column Conditioning
1. The Supelco (Jo. states that only minimum conditioning is
needed on the column specified in (4) of apparatus. Connect the column
to the inlet of the GC but not with the detector.
2. Pass helium through at 30 ml. per minute program at one
degree per minute from 150°C to 275°C. Keep at 275 for 30 minutes. Cool
and connect to the detector.
3. If SE 30 as supplied by the manufacturer is used in place of GC
grade, connect as in (1). Purge with helium for 10 minutes and then put a cap
on the exit end, keeping a helium pressure on the column.
4. Raise the oven to 350°C and bake for 3 hours.
5. Cool to room temperature, remove cap, pass helium through
at 30 ml. per minute and continue conditioning column as shown below
without attaching to the detector.
Temp., °C. 150 200 250 275
Time, Min. 30 30 30 60
6. Cool to room temperature and attach to detector. The column
is now ready for use.
Performance Test
This test is used to measure retention times and demonstrate
suitable recovery of individual PNA's. Recoveries of > 83% are required.
With the column conditioned and the parameters set as described in 1, pre-
pare to inject a sample.
-------
- 99 -
7. With the 10 ul syringe, take the Test Blend and flush
out twice.
8- The third tine fill the syringe and hold vertically point
up. Stick the tip in a rubber septum and depress the plunger to compress
air bubbles which rise to the top.
9- Now advance the plunder to the desired arr.ount, for instance
2 }i\i the rubber is removed, and the plunger retracted until the air-liquid
meniscus enters the glass bore.
10. Note the volume from meniscus to plunger.
11. Insert the syringe up to the hilt in the GC injection port,
depress the plunger. Start the recorder.
12. Remove the syringe from the port and retract the plunger.
13. Note the volume remaining (generally about 0.2 ul) and sub-
tract from the previously noted volume (4) to determine the net volume
injected.
14. Have stainless steel trapping tubes and one-dram vials
available. The vials should be supported in small holes drilled in a
board.
14A. An ice bath surrounds the stainless steel insert as a
precaution.
15. As a peak approaches, insert a tube. Remove as the trace
returns to the baseline. Mark the peak No. I and place the tube in a
one-dram vial labelled I.
16. Continue taking cuts, putting each tube in a separate
vial labelled 2, 3 etc.
17. Typical retention times obtained in setting up this method
are shown in Table E-l.
-------
- 100 -
Table E-l
OBSERVED RETENTION TIMES
Compound Retention Time, Min.
Pyrene 13.6
Triphenylene 20.0
Benz(a)anthracene 20.1
Chryscne 20.4
Methylbenz(a)anthracenes 23.3
Dimethyl/ethylbenz(a)anthracenes 25.0
Benzo(e)pyrcne 27.2
Benzo(a)pyrene 27
McthylbenZo(e)Pyrcnes(0 ) ^
Methylbenzo(a)pyrenesv3/ )
Benzo(g,h,i)pcrylene 33.3
(a) Measured in three different peaks.
18. When the run is completed take the first tube and place in
a 5-al graduace. With an elongated medicine dropper add enough cyclohexane
through the top of the tube to fill the spectrophotometrie cell to be
used. In developing this method, 3.8 ml. was a convenient volume to em-
ploy, but all cuts must be made up to the same volume.
19. Pour the solution into the original vial, cap and protect
from light. Rinse out the graduate with cyclohexane and use for the next
tube. Save the fractions for UV, and calculation of recoveries.
20- When the run is completed, measure and record the retention
tine in minutes from injection point for each compound including the in-
ternal standard.
Analysis of Samples
2l. From the vial containing the PNA concentrate from the alumina
column, take 5-10^pl of sample and inject into the GC unit. Trap the cuts
at the retention times obtained on the Test Blend runs and also trap any
other peak that might be of interest. The cuts of. the samples must be made
up to the same volume employed for the test blend.
-------
- 101 -
F. Ultraviolet Spectrophotometric Analysis
Calibration
1. Prepare solutions of known concentrations of each PNA in
cyclohexane, e.g., 2.4 ;ug/ml.
2. With cyclohexane in the reference cell, obtain the spectrum
of each solution from 400 to 225 nm.
3. Rinse the cell thoroughly with cyclohexane between spectra
and remove the last traces of solvent by vacuum or a gentle jet of air.
4. From the spectra obtain the absorbances of the various com-
pounds at the wavelengths shown in Table E-2. Calculate absorptivity from
the known concentrations. Typical calibration data are also shown in the
table.
Analysis of Samples
5. Obtain the UV spectrum of each selected fraction as directed
in step C21. Measure the absorbance at the specified wavelength. Several
of the trapped fractions will contain more than one PNA species. Tri-
phenylene, benz(a)anthracene, and chrysene occur together and are analyzed
in their common spectrum by determining BaA, then correcting the 269 nm
(chrysene) absorbance for BaA and the 259 nm absorbance for both BaA and
chrysene; the coefficients are given in Table E-2. The fractions for the
methyl- and the dimethyl/ethyl-BaA isomers include the alkyl derivatives
of chrysene and triphenylene; calibration data are lacking for the latter,
but their presence does not interfere with determining the BaA derivatives
with the tabulated coefficients. BaP and BeP are found together but do
not interfere when absorbances at 383 and 333 nm, respectively, are used.
Methyl-BaP and - BeP occur in three GC fractions yielding three UV spectra
with slightly varying peak locations due to isomer effects; the three
absorbances near 384 nm are added for MeBaP, and the three near 335 nm for
MeBeP. Pyrene and benzo(ghi)perylene occur in peaks separated from
interfering species.
6. One ml of the fraction containing BaA is placed in a counting
vial along with 9 ml of toluene and 10 ml of counting solution. Counts
per minute are measured using a scintillation counter. The same procedure
is followed with one ml of the BaP fraction.
-------
- 102 -
TABLE E-2
UV CALIBRATION DATA FOR SOME PNA HYDROCARBONS
Wavelength nm Di- Benzo-
Tri- methyl (g,h,i
Chry- phenyl Methyl /ethyl Methyl Methyl peryl-
Peak Base Line Pyrene BaA sene ene BaA BaA BaP BeP EaP BeP ene
336
289
269
259
327-343
283-295
263-277
252-263
0.240
0.332
0.039 0.464
0.015 0.107
0.480
2933 285-300 0.27
2933 285-300
383 373-390
333 325-338
384b 373-390
335 325-340
382 370-390 0.074
Occurs in range, 290-293 nm, depending on isomers.
Occurs in range, 383-385 nm, depending on isomers.
-------
- 103 -
To calculate the amount of the various compounds in the entire
sample, the aliquot factor (activity ratio) must be determined based on
the radioactivity in 1 ml of the fractions containing the internal standards
QN
" (2)
Q = activity added to sample, DPMV '
M = activity ratio
K = average CPH per ml
L = Background, , CPM
N = counting efficiency CPM per DPM
CPM is counts per minute
(2)
DPM is disintegrations per minute (a measure of radioactivity).
The concentration of each component is determined from the UV
absorbance.
D =
EC
D - concentrations of each PNA (yg/ml)
A = absorbance of each PNA
B = absorptivity of each PNA (ml/yg-cm)
C = cell length (cm)
The weight of added radiotracer is determined,
X = weight of radiotracer added, yg
S = specific activity (DPM/yg)
-------
- 104 -
Calculate weight of PNA in the sample
T = MD - X
T = weight of PNA in the sample.
Where more than one internal standard is used, all components
eluting in the same peak with a standard are calculated based on that
standard's activity ratio. All other components are calculated using
the smallest activity ratio since this represents the maximum recovery
of standard.
Calculations are now handled on an IBM 1800 for 11 of the PNA's,
------- |