THIRD ANNUAL REPORT
ON
GASOLINE COMPOSITION AND
VEHICLE EXHAUST GAS
POLYNUCLEAR AROMATIC CONTENT
Period Ending July 30, 1972
CRC-APRAC Project No. CAPE-6-68
EPA Contract 68-04-0025
Submitted to
The Coordinating Research Council, Inc.
and
Environmental Protection Agency
Esso Research and Engineering Company
Products Research Division
Linden, New Jersey 07036
-------
THIRD ANNUAL REPORT
ON
GASOLINE COMPOSITION AND VEHICLE
EXHAUST GAS POLYNUCLEAR AROMATIC CONTENT
CRC-APRAC Project No. CAPE-6-68
EPA Contract 68-04-0025
Period Ending July 30, 1972
Submitted to
The Coordinating Research Council, Inc.
c/o
Mr. A. E. Zengel, 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
THIRD ANNUAL REPORT
TABLE OF CONTENTS
LIS
LIS
PEE
1.
2.
3.
T OF TABLES
T OF FIGURES
SONNEL
SUMMARY
INTRODUCTION AND OBJECTIVES
EXPERIMENTAL
3.1 Fuels
3.1.1 Emission- Test Fuels
3.1.2 Deposit Fuels
3.1.3 Additions of PNA and High-Boiling
3.2 Test Vehicles
3.2.1 Standard Vehicles
3.2.2 Low- Emission Vehicles
3.4 Analysis of Samnles
iii
iv
V
. . 1
. . 3
. . 7
. . 7
. . 7
. . 7
. . 10
. . 14
. . 14
. . 18
. . 19
. . 20
4. EMISSION RESULTS 22
4.1 Emission of PNA 22
4.1.1 Effects of Current Emission-Control
Systems 29
4.1.2 Effects of Experimental Low-Emission
Systems 29
4.1.3 Effects of PNA in Fuels 33
4.1.3.1 PNA Present in Deposit Fuels . . 33
4.1.3.2 PNA Present in Emission-Test
Fuels 36
4.1.3.3 Effects of Fuel PNA on PNA
in Used Oils 38
4.1.4 Effects of Lead and Phosphorus in
Deposits 38
4.1.5 Between-Test Carryover of Fuel PNA .... 41
4.1.6 Effects of Test-Fuel Aromatics 42
4.1.7 Effects of Two-Ring and Three-Ring
Aromatics 42
4.1.8 Effects of a Heavy Catalytic Naphtha ... 45
4.1.9 Effects of Engine Mechanical Defects ... 47
4.2 Phenol Emission 49
-------
TABLE OF CONTENTS (CONTINUED)
Page
5. DISCUSSION OF EMISSION RESULTS 52
5.1 Control-System Effects on PNA Emission 52
5.2 Effects of PNA Present in Fuel 54
5.3 Effects of Heavy Reformate and Catalytic
Naphtha Fractions 56
5.4 Phenol Emission 58
5.5 Hydrocarbon (NDIR) Emissions of Vehicles 58
6. PNA-SAMPLE COMPOSITION STUDIES 59
6.1 Compositions of Samples of Different Origins .... 59
6.2 CAPE-6-68 PNA Sample Compositions 60
6.2.1 Statistical Analysis 61
6.2.2 Reactivity Differences Between PNA
Species 63
6.3 Ratios Between PNA Species vs. Sampling
Losses 69
6.4 Repeatability of Analyses (Quality Control) 70
7. SELECTIVE LOSSES OF PNA SPECIES IN EXHAUST SAMPLES 73
7.1 Summary of Loss Studies 73
7.2 Evidence for CAPE-6-68 Losses 73
7.3 Confirmatory Tests of Losses 75
7.4 Loss-Reduction Experiments 76
7.5 Future Practice in CAPE-6-68 Sampling 79
7.6 Tracer Injection for Improved Accuracy 81
APPENDIX A
References 83
APPENDIX B
Exhaust-Gas Temperature and Composition Data ....... 85
ii
-------
LIST OF TABLES
Table No. Title Page
1 Fuel Identification 8
2 Deposit-Fuel Inspections 9
3 GC/UV Analyses of Fuel PNA 11
4 Comparison of Heavy Naphthas 12
5 Additions of High-Boiling Aromatics and Polynuclear
Aromatics 13
6 LVMS Analyses of Fractions from a High-Severity
Catalytic Reformate 15
7 Deposit-Fuel and Oil-Consumption Records 17
8 Summary Data Sheet 23
9 Corrected PNA Data for 1966 Plymouth and 1968
Chevrolet 26
10 Changes in PNA Emission with Successive Tests 28
11 PNA-Emission Reduction by Emission-Control
Systems 30
12 PNA Emissions from Experimental Low-Emission
Vehicles 31
13 PNA Emission Results Arranged by Condition and
Origin of Engine Deposits 34
14 Change in Average PNA Emission with Deposit
Condition and PNA Content of Deposit Fuels 35
15 Effects of Fuel PNA on PNA Emission 37
16 Analyses of PNA in Used Oils 39
17 Comparison of PNA Emissions from Deposits with
Varying Lead and Phosphorus 40
18 Fuel Aromatics Effects with Varying Deposit
Conditions 43
19 Effect of Two-Ring and Three-Ring Aromatics on
PNA Emission 44
20 Effect of a Heavy Catalytic Naphtha on PNA
Emission 46
21 Phenol-Emission Correlations with Fuel
Aromatics 51
22 Multiple-Species Analyses of Exhaust Samples 62
23 Ratios Between PNA Species in Exhaust Samples
and in Gasolines 64
24 Quality-Control Sample for GC/UV PNA Analysis 71
25 Recoveries of PNA Species After Various Exposures ... 77
26 Esso Research Tests on PNA Collection Techniques ... 80
B-l Exhaust Gas Temperatures 86
B-2 thro.ugh
B-41 Exhaust Analysis Data (CO, HC, NO) 87-106
iii
-------
LIST OF FIGURES
Figure No. Title Page
1. Effect of Emission Control Systems on
Emission of FNA. 32
2. Effects of Fuel Aroma tics and Emission
Control Systems on Phenol Emission 50
3. Ratios of Benzo(e)pyrene to Other PNA in
Exhaust Samples and Gasolines 65
4. Ratios of Benzo(a)pyrene to Other PNA in
Exhaust Samples and Gasolines 66
5. Ratios of Benz(a)anthracene to Other PNA
in Exhaust Samples and Gasolines 67
iv
-------
PERSONNEL
(Third 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.
Mr. J. S. Ninomiya
Ford Motor Company
Dr. A. J. Pahnke
Petroleum Laboratory
E. I. duPont de Nemours & Co., Inc.
Dr. J. H. Somers (Project Officer for 68-04-0025)
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
* During the project year, Mr. Gerard was succeeded by Dr. S. S. Hetrick,
Mobil Research and Development Corp.
-------
- 1 -
1. SUMMARY
Vehicle emission control systems, fuel composition, and fuel-
derived engine deposits have been studied as variables affecting the
exhaust emissions of eleven species of polynuclear aromatics (PNA) and
phenols. This program was the third year continuation of research sup-
ported jointly by the Coordinating Research Council, Inc., and the U.S.
Environmental Protection Agency.
Effects of Vehicle Emission Control Systems
Current vehicle emission-control systems substantially reduced
the emission of PNA and partially reduced phenol emissions. Two experi-
mental (thermal and catalytic) advanced emission control systems almost
completely eliminated PNA and phenol emissions.
Relative Emissions
Emission-Control System PNA Phenol
None 100 100
Engine-modification (1968) 20-30 100
Engine-modification (1970) 20-30 65
RAM Thermal Reactor 1-2 <0.5
Monel/PTX-5 Dual Catalyst <1 <0.5
Effects of Fuel Composition
The PNA content (four-ring and five-ring aromatics) of gasoline
was the dominant fuel-composition variable affecting PNA emission in both
uncontrolled and emission-controlled vehicles. This effect of existing
fuel PNA was apparent only after engine deposits had formed from use of
the fuel.
Fuel PNA content also dominated the accumulation of PNA in
engine oil, and emission control systems did not control this accumulation.
Engine oil used for 2000 miles under low oil-consumption conditions with
a high-PNA fuel (3 ppm benzo(a)pyrene (BaP) and other PNA) accumulated
12-23 ppm BaP. Small increases (0-2 ppm BaP) in oil PNA also occurred
with low-PNA, high-aromatic fuels.
Two-ring and three-ring aromatics (naphthalene, anthracene,
etc.), present in a low-PNA gasoline as either a high-boiling reformate
fraction or as a heavy catalytic naphtha, caused small immediate increases
in PNA emission. In extended use of the gasoline, there was no further
increase in PNA emission, but small increases in oil PNA occurred.
-------
- 2 -
Phenol emission was proportional to fuel aromatics and
independent of other fuel variables.
Effects of Engine Deposits
Deposits accumulated in the use of high-PNA fuels (3 ppm BaP,
etc.) caused PNA emissions which were 1.9 to 8.6 times the emissions
measured after use of low-PNA (near zero BaP) fuels. Emission levels
were greatest with recently-formed deposits and were reduced by subsequent
use of low-PNA fuels. A mechanism of PNA storage in deposits and later
release is suggested, and further studies are planned. The magnitude of
this fuel PNA effect on PNA emission is undefined for operating conditions
differing from these tests. Intermediate levels of fuel PNA have not
been evaluated. PNA emission was not affected by deposit-fuel lead, with
or without phosphorus.
Effects of Engine Malfunction
Mechanical malfunctions resulting in overchoking at start-up
caused extremely high emissions of PNA and soot and obscured the effects
of fuel properties. Emissions also increased at least temporarily with
higher engine temperatures during a failure of the vacuum spark advance
system.
Composition and Sampling of Emitted PNA
Certain more-easily oxidized PNA species, including BaP and
benz(a)anthracene (BaA), were significantly less abundant in exhaust PNA
samples than in fuel samples, when compared to less-reactive species.
These compositional differences could imply either differences in the
synthesis of different PNA (engine vs. refinery) or differences in the
survival of different PNA that were introduced in a high-PNA fuel or
synthesized in the engine. The highest PNA emissions occurred with de-
posits from high-PNA fuels, implicating survival of fuel PNA as the major
factor causing the differences.
These composition differences were shown by special tests to
be due, in part, to losses of reactive PNA species that occurred after
the sample was collected but before the addition of radiotracers at the
start of analysis. These losses resulted in emission measurements for
BaP and BaA which were low by a factor of about two. The losses do not
invalidate the conclusions reached on the effects of controlled test
variables, since all PNA species responded similarly to these variables.
Loss-reduction techniques for sampling were developed but were not used
because they would introduce an inconsistency with earlier data.
-------
- 3 -
2. INTRODUCTION AND OBJECTIVES
This report covers the work performed on the exhaust emission
of polynuclear aromatic hydrocarbons (PNA) and phenols under CRC-APRAC
Project CAPE-6-68 in the third year of the Project. Government (Environ-
mental Protection Agency) support (33% of total cost) has been previously
provided under Contracts CPA-22-69-56 and CPA-70-104 and is currently
provided under Contract 68-04-0025.
While the present work extends past work, a time-gap occurred
between the end of the second year in February, 1971, and resumption of
contract work on May 6, 1971. During this delay period, several PNA/phenol
tests were carried out at the expense of Esso Research and Engineering
Company. These Esso tests are described here to the extent that they may
affect subsequent contract tests or are essential for data interpretation.
To some extent, these tests were in line with contract plans; in any case,
they were valuable for maintaining test personnel and test capability
during the renewal-negotiation period.
The Second Annual Report (1, la) (References are given in Appen-
dix A) reported on the directly observable effects of fuel aromatics, fuel
PNA, fuel lead, and the presence of a high-boiling naphtha in a low-aromatic
base fuel, and on the effects of vehicle emission-control systems. Also,
an approximately two-fold difference in PNA emission levels was apparently
associated with the composition of engine combustion-chamber deposits.
However, the property of the deposits that controlled the PNA emission
level was undefined. Either base (hydrocarbon) composition, high lead as
against either low lead or zero lead, or a phosphorus additive in one of the
deposit fuels could have been the responsible factor. A review of some of the
published studies on fuel composition and deposits, although related in
some cases to knock, etc., rather than to pollutant emission, suggested the
possibility that some aspect of fuel hydrocarbon composition was a definite
possibility as the responsible factor for indirect, or deposit-related,
effects. In the area of fuel composition, overall aromaticity, the amount
of high-boiling aromatics including PNA, and the presence of other high-
boiling naphthas were considered promising for study.
The broad objectives of third-year work in Project CAPE-6-68
were listed in EPA Contract 68-04-0025 and are further condensed and
summarized in the following.
Deposit studies, with special reference to lead and/or
phosphorus in deposits, but also including non-leaded
deposits.
Studies of experimental low-emission control systems (as to
PNA, phenol, CO, NO, and HC) that may be made available to
(but not developed by) CAPE-6, and specifically including
Esso's thermal reactor system (RAM) and dual-catalyst (Monel +
PTX-5) system. A third system, if made available, through the
Project Group from an external source, was also to be tested.
-------
- 4 -
Validation, in a 1970 test vehicle, of the conclusions from
1966 and 1968 test vehicles. Fuel properties, engine deposits,
and control-system effects were included.
More-detailed studies of fuel-composition effects, beyond
overall aromaticity and with special reference to higher-
boiling hydrocarbons.
Additional tasks, as specified by the Project Group and
Project Officer, and within the limits of available time
and funds.
The objectives summarized above include studies of deposits
with lead and/or phosphorus, and also include studies of non-leaded
deposits. In the area of direct fuel-composition effects, i.e., effects
which appear immediately when a fuel is first tested, special reference
is directed to the PNA-emission effects of higher-boiling hydrocarbons.
As mentioned above, a variety of published studies have indicated that
high-boiling fuel components can influence deposits, and past CAPE-6
work had shown a deposit effect on PNA emission. It was thus appropriate
for CAPE-6 to consider fuel hydrocarbon composition, including high-
boilers, in the context of deposit fuels, as well as emission-test fuels,
i.e., as a factor in defining the PNA-emission level that is associated
with any given set of engine deposits after extended use of the fuel.
Fuel hydrocarbon composition, in a broad sense, can imply
many different fuel characteristics. One practical approach to the
deposit-effect aspects of fuel composition centers logically on high-
boiling materials and, specifically, on fuel PNA, i.e., on four-ring and
five-ring polynuclear aromatics in the fuel. A small-scale survey by Esso
Research, which was made available to CAPE-6 and mentioned in the First
Annual Report (2) of CAPE-6, found a substantial variation in the PNA
content of gasolines, expressed as the content of benzo(a)pyrene (BaP,
five-ring) in ppm or in yg./gallon, where 1 ppm is about 2800 Ug./ gallon.
The maximum observed BaP content was about 8000 yg./gal. (3 ppm), while
the analysis of nationwide composite premium and regular-grade gasoline
samples showed 0.43 and 0.19 ppm, respectively. The same report also
showed that the residue from distilling a catalytic reformate was a rich
source of PNA, including BaP.
In previous CAPE-6 work on deposits, effects were sought in
terms of "leaded" vs. "low-lead" and "no-lead" deposits . Deposit-
fuel PNA was not a controlled variable in the sense that its presence
in a deposit-forming fuel might have affected engine deposits and, in
turn, the emission of PNA. In fact, the deposit fuels used in the first
two years were not inspected for PNA or BaP content before use, although
reasonable estimates can be made. Two commercial premium fuels were used
-------
- 5 -
for deposit formation: one contained lead and phosphorus and is known to
have been formulated with a catalytic reformate which had not been re-run
(distilled) to remove high-boilers. From this knowledge and from tests of
later presumably similar fuels, its BaP content is estimated to have been
2-3 ppm. The other deposit fuel base was lead-free and was used both
as-purchased and with lead added at 0.5 g./gallon. Inspections of the
second of two shipments that were actually used indicates that this fuel
contained only about 0.01 ppm of BaP. This second fuel, whether lead-free
or with a low-lead level, gave deposits in two test vehicles with which
relatively low emissions of PNA were obtained. These differing emission
results between deposits from using two differently-based fuels could thus
be attributed to differences in either phosphorus use or lead use (with
high lead not equal to low lead) or to base-fuel composition. If fuel
composition were involved, fuel PNA content appeared to be a promising
aspect of fuel composition for further study, since the two fuels could
be reasonably assumed to have differed in PNA content (BaP) by a factor
of at least 100. No other fuel-composition differences of this magnitude
(except additives) were recognized for these fuels.
The third-year deposit-study program evaluated PNA emissions
from deposits formed with fuels differing in lead, phosphorus, and PNA
content, with each variable at two levels. The effects of two-ring and
three-ring aromatics and of a heavy catalytic naphtha were also examined.
The same three test vehicles (1966 Plymouth V8, without emission controls
(NC), and 1968 and 1970 EM-controlled Chevrolet V8's) that were used
previously continued in use. Two experimental low-emission vehicles
(RAM thermal reactor and dual-catalyst (Monel + PTX-5)) were loaned by
Esso Research and Engineering Co. for emissions tests under conditions
that were comparable with existing data from the 1968 Chevrolet (basic
vehicle for RAM) and the 1970 Chevrolet (basic vehicle for the catalyst-
equipped vehicle). These experimental vehicles are described in References
(3) and (A) in substantially the forms used for the CAPE-6 PNA emissions
tests.
The third-year program has also provided information in several
areas that were outside of the main program but were nevertheless essen-
tial to a correct interpretation of the experimental data:
Engine-deposit condition, or the amount of previous emission
testing since deposit formation, influenced PNA emission
results.
PNA present in fuel accumulated in engine oil during fuel use.
Engine malfunctions caused significant PNA increases.
Exhaust PNA samples differed from fuel PNA samples, with
exhaust samples relatively deficient in certain more-reactive
PNA species.
-------
- 6 -
More-reactive PNA were partially lost in collecting exhaust
samples. These losses (about 50% for BaP, for example) did
not explain the exhaust/fuel differences, nor did they
invalidate the conclusions on the effects of variables on
PNA emission.
The first three items above are introduced where appropriate, while the
fourth and fifth items are covered separately in Sections 6 and 7 after
the discussion, in Section 5, of the results from the main program.
-------
- 7 -
3. EXPERIMENTAL
3.1 Fuels
The fuels used in the CAPE-6-68 program can be grouped as
emission-test fuels, prepared in relatively small quantities for emission
tests, and deposit fuels, which were used for deposit accumulation and
also for most of the third-year emission testing. Both sets are identi-
fied by the code designations given in Table 1.
3.1.1 Emission-Test Fuels
Two sets of emission-test fuels (1969 and 1970) have been
used, differing primarily in that the 1969 set included a high-boiling
naphtha (mixed heavy catalytic naphtha, polymer, and catalytic C,-'s)
while the 1970 set did not. (Second-year results had found an emissions
difference apparently related to the presence of 16% of this naphtha
in the low-aromatic test fuel.) Both the 1969 and 1970 sets included
low- (11-12%) and high - (46%) aromatic fuels, and the 1970 set was
crossblended for 28% aromatics. At these three aromatics levels, fuel
TEL content (0,0.5, and 3 grams Pb/gal.) and PNA content (as BaP, at
0.01 and 2.9 ppm) were varied. Table 1 shows the combinations used,
and indicates the use of a PNA additive from the distillation of a
catalytic reformate. As prepared, the emission-test fuels were very
low in PNA content. Compositions and inspections for these fuels appear
in Tables B-l to B-3 on pages 52-54 of the Second Annual Report (1) ,
while inspections of individual blending components are in Table B-4
on page 56 of the First Annual Report (2). Fuels from the emission-
test-fuel sets were used for only one-third of the third-year (this
report) emissions tests, primarily to relate an engine-deposit condition
under study to some earlier deposit condition where the same test-fuel
had been used.
3.1.2 Deposit Fuels
Section 2 of this report introduces the composition of a
deposit-formation fuel, and particularly its high-boiling components (PNA,
etc.), as a potential major variable in defining the level of PNA emissions
in exhaust. The code designations for deposit fuels appear in the lower
part of Table 1. These fuels can be grouped in two broad groups with re-
spect to PNA content: (a) near-zero PNA, coded A, A', and the indicated
blends in these base fuels (except A' + PNA, used only in an emission test
and not for deposits), and (b) high-PNA, i.e., 2-3 ppm BaP, coded B, C3P,
D and D3. The designations A, A', B, C, and D indicate different base
fuels (aromatics, olefins, volatility, etc., as given in Table 2) but we
do not, at this time, recognize any substantial effect of these differences,
other than PNA content, in terms of deposit-related PNA emissions. Particu-
lar attention is invited to the fact that the substantial differences in
deposit-fuel PNA contents are not correlated with the boiling character-
istics indicated by the ASTM D86 distillation data in Table 2. The one
exception here is in the data on 20% blends of two different heavy catalytic
-------
- 8 -
TABLE 1
CAPE-6-68
FUEL IDENTIFICATION
(1) Coded Emissions-Test Fuels; the 1969 set contains a high-boiling naphtha
at 16% in the 12%-aromatic and 2% in the 46%-aromatic fuels.
Fuel Code Designation
(a) TEL(b)
11-12%
1969
100
108C
108
Arom.
1970
1LOO
1L03
1L83
Cross-
blend
1970
3000
3000.5
3003
46%
1969
5HO
5H8
Arom.
1970
SHOO
5H03
5H80
5H83
PNA
Low 0
0.5
3
High 0
3
aAdded, as 384+°F VT still bottoms from catalytic reformate (First Annual
Report, page 52), at 22.2 gins/gallon to add SOOO^ig BaP/gallon, or 2.9 ppm.
In ccs. of TEL/gallon, as Motor Mix. (1 cc TEL contains 1.06 gins Pb).
(c)
A similar fuel, 5LOC, of higher mid-fill volatility, was also used in the
1969 set.
(2) Coded Deposit-Formation Fuels, some of which were also used in
PNA emissions tests.
£± is a commercial lead-free high-aromatic premium gasoline which is
low in PNA content; also used with added lead and phosphorus, as
A.5 and A3P. A second batch, A", has been used for blending with
certain high-boiling components. The blends are coded A' + PNA,
A' + V45, A' + HsCN.
B is a lead-free mid-aromatic premium-grade gasoline with substantial
~ PNA content; it is available semi-commercially for engine and vehicle
research only.
£ is a commercial-type premium-grade (when leaded) gasoline, either
purchased or blended by Esso Research from purchased components;
components include non-rerun catalytic reformate, and inspected
samples of the gasoline contain substantial PNA. Used with lead
and phosphorus, as code C3P, in various batches.
D is a fuel similar to £, with higher reformate content and lower
~ volatility. Limited use, without (D_) and with (D3) lead.
Suffix ^ means lead at 2-3 grams Pb/gallon.
_._5_ means lead at 0.5 grams Pb/gallon.
P_ means phosphorus at 0.2 theory (for Pb3(PO^)2) based on
lead, or 0.06 grams/gallon when lead is absent.
-------
- 9 -
»<
o
CM
+
^
4
Z
S£
as
cn
o
rH
z
to
SB
Q
PH
en
u
4J
g
s
01
B!
0
o
rH
vO
O
rH
cn
CO
i i
rH
«
U
H
P
in
rH
1
i-H
rH
6Q
CO
r-:
~<
rH
rH
O 4J
M to
0 -H
g °
^x \o
OO
Pu 1
> Q
3 g
a w
Cd a-cMooco "^
oocriOcnu^cfii immr-rHfop- i t
r-
ON
CO
s-* CO
03 O
CX
cfl T3
> «H
tU 03
^ (S
fetnooooooooomtn >^
«. a*
CU rH (*4 >
H PH U
£
SO
so
CO
SO
0
SO*
m
cr\
in
rH
SO
1
ON
SO
CM
so
CO
en
SO
SO
CM
so
O
O
so
«
rH
a
o
to
3
*t
>>
4J
H
%
0
-31 \O
vD *
"*
p- rH
ON in
**
a> m
r>r o^
CM
vt f>
o oo
CM
CM rH
COrH
CM rH
-» rH
CM en
CO
CM m
r~ o
*j
J- CM
r~ CM
ro
U
H
4J C
§iH
MH
O 01
< 0
B< fr«
n
M
0
1 O
1
O
1 0
-*
§CM
»H
O CM
O
rH
m
O CM
0 1
rH CM
O
8°
rH
v1
0
o o
in
o
0 O
so
6 rH
O. <0
P. 00
. oo
3
»
o fi
0 0
r* m
en oo
o r*^
m -*
GO O
rH ^O
P.
^
z
o
o a. u
n
S ai rH
P..C 3
p. H o
-------
- 10 -
naphthas that are included in the table. The naphtha blend having the
higher final boiling point does differ in PNA from the lower-boiling
blend; this is related to the PNA contents of the two naphthas (Table 3).
The blend with the "winter" (lower-boiling) naphtha was prepared, but
has not been used experimentally in view of results with the higher-boiling
blend.
The PNA contents of the major deposit fuels, of the two heavy
catalytic naphthas, and of selected other fuels are given in greater
detail in Table 3.
The two heavy catalytic naphthas, mentioned above and examined
in the third-year work, are compared in Table 4 with a different high-
boiling naphtha (mixed heavy catalytic, polymer, and C-'s) that was used
in the first-year (1969) emission-test blends. It is clear that the first-
year naphtha was similar to the lower-boiling ("winter") naphtha, rather
than to the recently-tested "summer" naphtha. Also, Table B-3 of the Second
Annual Report shows by mass spectrometry analyses that the earlier naphtha
contributed relatively little additional C^o~^12 aromat;i-cs' This distinction
will be mentioned in the Discussion section. The terms "summer" and
"winter" applied to the third-year naphthas relate to the practice of
some refiners, in some circumstances, to lower the end-point of heavy
catalytic naphtha in the winter to produce more heating oil. The implica-
tion of this small sample of data, that the end-point of a heavy catalytic
naphtha is an indicator of PNA content, should not be generalized without
additional data.
3.1.3 Additions of PNA and High-Boiling Aromatics
The test results presented in Section 4 compare the PNA-emission
effects of three different additions of high-boiling aromatics to a
low-PNA fuel:
(a) Two high-boiling reformate fractions, V4 and V5, from
vacuum distillation after prior atmospheric distillation,
used in their yield proportions.
(b) The same amounts of V4 and V5, with the yield-proportioned
amount of the distillation bottoms, comprising as a whole
the "PNA additive" used in this program.
(c) The "summer" heavy catalytic naphtha referred to in the
preceding Section. This material contained relatively
little PNA.
The use of these three materials increased the high-boiling
aromatics present in fuel blends by amounts which are compared in Table 5.
Vapor cuts V4 and V5, and the heavy naphtha, increased the amounts of
Cin-C,, aromatics, but did not cause substantial increases in PNA. In
the C10-C;L3 range, the use of 20% heavy naphtha caused substantially
greater increases than the use of cuts V4 and V5 at the levels at which
-------
- 11 -
o>
pa
0
CM
in
O
rH
vO
O
OO
ed
\^f
<
3
in
w
,
rH
to
t>0
^~
00
3
A
4J
C
01
o
<
PL,
PL,
> m
d< vO
O
>3
ej*
VD
O
tx.
ON
00
O
Ov
O>
en
O
vo
CM
CM
rH
O
-3
en
o
rH
O
r~~
vo
en
CM
O
CM
O
o
rH
O
en
en
00
o
en
p^
rH
O
vD
00
vO
O
rH
00
CM
rH
O
r~
rH
rH
<±
O
vO
CM
vO
O
vO
O
00
O
CM
O
~3
O
CM
en
O
rH
o
rH
00
r^
O
o\
oo
o
CM
in
m
CM
m
vO
CM
-3
rH
rH
in
a-
en
^
oo
CM
VD
~3
CM
m
m
en
in
m
o\
en
o
-a-
o
en
o
CM
0>
O
CM
-a-
CM
en
-3-
en
VD
o
3-
en
CM
vo
CM
o
CM
r-.
O
CO
O
rH
rH
Ov
O
m
oo
VD
o
co
o
vO
O
CM
O
1
O
r^
en
vo
O
CM
en
O
O
en
in
o
vO
o
m
o
r-
CM
O
O
CM
O
vO
O
oo
o
CM
CM
CM
m
r-
o
vO
o
en
o
en
o
CM
in
m
o
vO
CM
O
O
r* x-x
CM O
m o
O v~'
o
o
CM
CM
O
r^
r-»
vo
O
vo
CM
vO
en
en
01
in
o)
t-,
o.
01
01
o
01
UH
01
m
to
o
01
ji
^
to
en
cu
M
tO
4-1
tO
O
en
Q
B
o
to
00
S TJ
01
3
01
rn
01
)-,
t-l
o
a
-H
01
01
h
o
B
o
s
01
B ^
Ol to
in 11
>* o.
68
TJ
73
tO
-------
- 12 -
TABLE 4
CAPE-6-68
COMPARISON OF HEAVY NAPHTHAS
First-Year, Third-Year
Mixed HCN/Polymer/ Heavy Catalytic Naphthas
Cat. C5's(a) "Winter"^b)"Summer"^
Gravity, Ibs./gal., 60°F. 6.38 6.71 6.73
FIA, % Aromatics 37.2 49.8 44.9
% Olefins 29.8 14.5 21.4
RVP, psi 8.6 1.7 2.6
ASTM Dist, IBP, °F 95 139 138
°F at 10% D + L 155 206 208
30% 222 253 260
50% 280 295 305
80% 345 353 375
90% 371 374 402
95% 390 387 420
FBP 419 423 442
% Recovery 97.0 98.0 98.0
% Residue 1.3 1.0 1.5
PNA,/ig/gal, BaP "Zero"(c) 24 795
BaA "Zero"(c) 17 920
Used at 16 vol % in 12%-aromatic, and 2% in 46%-aromatic
emission-test fuels in the first-year test-fuel set. This
component was excluded in the second-year set.
Both third-year heavy catalytic naphthas were blended at
20 vol % in base fuel A', but only the "summer" naphtha blend
was evaluated in emission tests.
^ This naphtha was not analyzed directly, but a blend (Code 108C)
containing it at 16%, along with a known amount of PNA additive
contained only the PNA due to the additive.
-------
- 13 -
TABLE 5
CAPE 6-68
ADDITIONS OF HIGH-BOILING AROMATICS
AND POLYNUCLEAR AROMATICS
Grains of Aromatics, with Indicated Carbon Numbers,
Added/Gallon of Blend(
In 13.3 g of V4 in 22 g PNA Add.
(384-445°F VT) (Reft. Fracs. V4,
Aromatic and 6.7 g of V5 V5 , and 5754 °FVT
Carbon No. (445-575°F VT) Still-Bottoms}
10 00
11
12
13
14
15
16
PNA Added,
>ug/gal
BaP (C20)
BeP (C2Q)
BaA (C1o)
Pyrene (C,,)
9.2
5.1
2.5
1.5
0.8
0.4
0.2
19.7
None<^
9.2
5.1
2.5
1.5
0.9
0.9
0.9
21.0
8010
6850
6370
32070
In 20% (v.) of
"Summer" Hvy.
Cat. Naphtha
35
32
19
4.7
0.8
0.4
0.2
92.1
159
160
184
468
(a)
(b)
The CiQ-C^g grams/gallon values are based on LVMS analyses
of rerormate vapor fractions and bottoms and LVMS and MS
analyses of heart-cuts and bottoms from two different dis-
tillations of the naphtha. PNA additions (fig/gal) are from
Table 3. The computed LVMS results on PNA may be low by 1
carbon number, because, for example, pyrene (Cj_gHj_Q MW 202)
is reported as C-.cH22, MW 202.
Includes alkyl benzenes, naphthalene, methylindan, tetralin, etc.
analyzed by GC/UV, but the diluted blend does not show PNA
absorbance in the UV region for these materials.
-------
- 14 -
they were used. Maintaining these latter levels, but including the
distillation bottoms (i.e., using the PNA additive) caused no further
change in the C1_-C1_ range but very substantially increased the
^15+ materials» including four-ring and five-ring PNA. The three cases
present different possible real-world situations:
(a) Distillation of catalytic reformate to a high vapor
temperature, with bottoms losses minimal.
(b) Use of reformate without distillation.
(c) Use of a high end-point catalytic naphtha
(distillation is normal practice, but end-point
may vary).
The low voltage mass spectrometer analyses of vacuum distillate
cuts from a catalytic reformate, on which part of the foregoing discussion
is based, are summarized in Table 6.
3.2 Test Vehicles
The third-year program of CAPE-6-68 continued to use the
three standard production vehicles used in the previous work (on loan
from Esso Research and Engineering Company), and also made single-test
use of two advanced experimental low-emission vehicles that were de-
veloped and loaned by Esso Research. In a negotiation period between
the second and third contract years, Esso Research used one of the
standard vehicles for PNA emission tests, including within the test
series a deposit-renewal run on the same fuel used earlier by CAPE-6
and a deposit change through the use of a second deposit fuel. Some
of the data obtained in these tests has been made available to CAPE-6.
3.2.1 Standard Vehicles
The standard vehicles used are listed in the following
table:
STANDARD TEST VEHICLES
Emission-
Make Year Engine CID Control
Plymouth 1966 V8 318 None
Chevrolet 1968 V8 307 Engine Mod.
Chevrolet 1970 V8 350 Engine Mod. ^
(a) With transmission-controlled spark retard at low speeds.
-------
- 15 -
TABLE 6
CAPE-6-68
LOW-VOLTAGE MS ANALYSES OF FRACTIONS FROM
A HIGH-SEVERITY CATALYTIC REFORMATS^
At. Pr. V. Temp.,
Fraction No.:
Total Wt. %, of Aromatics Present in Each Fraction,
Having Indicated Carbon Number
358-362
V-l
Carbon No.
(b)
6
7
8
9
10 (Naphthalene and
C. Benzenes)
4
11
12 (Acenaphthene)
13
14 (Anthracene, Phenanthrene)
15
16 (Pyrene)
17
18 (BaA, Chrysene)
19
20 (BaP, BeP)
21
22
Totals
Vol. % of Feed
362-372
V-2
372-384
V-3
384-445
V-4
445-575
V-5
575 +
Bottoms
0.21
2.93 0.94 0.78
1.88 1.15 0.74
42.58 17.56 5.63
51.99 78.99 88.43
0.32 1.32 4.37
0.03
.rene)
99.94 99.96 99.95
1.00 0.98 0.99
0.03
0.19
0.50
1.88
66.97
27.21
2.88
0.26
99.92
v 0.96
0.10
0.15
0.45
3.66
21.53
32.16
21.36
12.03
5.59
2.34
0.36
0.03
99.76
0.50
0.17
0,06
2.07
25.16
32.22
18.29
8.77
5.91
2.76
2.62
1.67
99.70
O.]2(c)
Mixed for "PNA Additive".
(a)
(b)
Fractions by vacuum distillation of 5% bottoms from previous atmospheric
distillation. Indicated vapor temperatures are atmospheric equivalents of^
temperatures observed at reduced pressures. Fractions are essentially 100%
aromatic.
Reported carbon no. may be low by 1, or 2, if implied hydrogen is high by
12, or 24. Thus, C-.H-.-will be reported as C,2H _, and will inflate C°/,.
Each carbon no. is n^in CH_, wher.e x may = 6, 8, 10, 12, 14, 16, an3 18.
(c)
May be low due to recovery losses.
-------
- 16 -
Deposits were created, or changed, in each vehicle prior to a group of
one or more PNA emission tests by light-duty cyclic operation on a
Mileage-Accumulation Dynamometer (MAD), using a mixed city/suburb/highway
regime at an average speed of 30 MPH, and ending in approximately 20 miles
of fixed-speed highway operation. Weighed oil changes (Commercial 10W-40,
Type SE) and oil filters, and the weights of oil drains and used filters,
were used to obtain oil-consumption data in each MAD-operation period. On
the MAD, oil-drain intervals of 2000 miles were normally used. Several of
the drain-oil samples have been analyzed for PNA content and are reported
in Section 4. Begeman (5) has reported that high oil-consumption leads to
high PNA emission, and that PNA accumulates in used oils. Table 7 summarizes
the deposit-fuel (codes in Table 1), MAD-miles, and oil-consumption histories
for the three vehicles since project inception. With the exception of some
data involving oil leaks, the oil consumption has remained below 0.5 qts./
1000 miles in all vehicles and shows no tendency to increase.
After MAD operation and removal of the used oil and filter, a
fresh oil charge and oil filter were installed before each emission test,
even if several successive tests were carried out before returning to the
MAD. The vehicles were periodically checked with respect to basic spark
timing and idle speed, but carburetor and idle mixture adjustments were
specifically avoided unless a clear malfunction made them necessary.
The first emission test after MAD operation involved either
"stabilized" or "unstabilized" deposits, depending on whether at least seven
7-mode cycles (as required, for example, to obtain emission data on CO,
unburned hydrocarbons, and NO) had been conducted with the exhaust to waste,
before collecting the total exhaust in the PNA sample collector.
PNA emission tests consisted of three cold-start 12-cycle blocks
using 137 -second, 7-mode cycles (20). The test procedure is described in
greater detail in Section 3.3. A continuous sidestream sample, cooled,
filtered, and passed to non-dispersive infrared analyzers, provided emis-
sion data on CO^, CO, HC, and NO; analyses for 02, and HC by a flame ioniza-
tion detector, were also recorded. The infrared readings were, in general,
processed by an on-line computer, so that up to three emissions values for
gaseous pollutants were obtained in each PNA-emission test. The printed
values of corrected (to CO + CO = 15%) concentrations of CO, HC, and NO
for each mode and cycle used in calculating the emissions results were
averaged over the available (1 to 3) blocks that had been computer-analyzed.
These results are tabulated in Appendix B. Although emission standards for
the 1970 vehicle normally require calculations to a gram/mile basis, the
CAPE-6 data on this vehicle were processed only to the concentration basis
in order to permit comparison with the two earlier vehicles. The NO emis-
sions, on which no standards exist, were processed as though they were HC
emissions, i.e., reading concentrations in the same seconds and using the
same weighting factors.
-------
- 17 -
TABLE 7
CAPE-6-68
DEPOSIT-FUEL AND OIL-CONSUMPTION RECORDS
Test
Vehicle
1966 Plym.
1968 Chev.
1970 Chev.
(Deposit Fuels in Order of Use)
Deposit-Fuel
Code
C3p(a)
A
C3P
A. 5
A
D3(b)
S
-------
- 18 -
The computer-printed emission results provide data on the
ppm HC, ppm NO, and % CO emitted (after weighting) and also on the
average correction factors required to force CO + C02 to = 15%. The
average values of these correction factors for each vehicle, treated as
a measure of air-dilution, can provide estimates of the average % 02
in the exhausts of the vehicles. Also, average % CO values can be
obtained for each vehicle, as well as average peak exhaust gas tempera-
tures at the muffler exit (see Appendix B). The averages of % CO,
estimated % 0£, and peak temperatures are summarized below for each
vehicle, using averages of data from all three years.
VEHICLE EXHAUST CONDITIONS
(Three-Year Averages)
Avg. % CO Est. Avg. Avg. Max. °F.,
(7-Mode Test) % 02 Muffler Exit
1966 Plymouth 1.52 0.98(a) 570
1968 Chevrolet 0.75 1.80 594
1970 Chevrolet 0.64 2.56 703
(a) The two-year Plymouth average % 02 was 1.20%; all other
three-year averages are almost identical to the two-year
average values reported in the Second Annual Report.
These measures of exhaust condition are believed to influence the
emission of PNA, but complete data on this point are not available, be-
cause the variations were not systematically explored. However, the PNA
from the 1966 and 1968 vehicles is apparently correlated with % CO, and
PNA emissions were corrected (see Section 4), by the use of a first-year
equation, to standard CO levels in these two cars. (In this regard, CO
is not proposed as a causal factor on PNA, but as a symptom of oxygen
deficiency, which, along with temperature, can be expected to affect PNA
survival and emission.) As in the past, we note that the CO emitted by
the 1966 and 1968 vehicles is abnormally low, so that PNA emissions may
also be depressed from normal levels.
3.2.2 Low-Emission Vehicles
The two prototype experimental low-emission vehicles (a RAM
thermal reactor and a dual-catalyst system) were developed by Esso Research
and Engineering Company.
-------
- 19 -
The RAM system has been described by Lang (3). Briefly, it
provides reduced NOX through the use of rich carburetion and spark-
timing control and rapidly, after startup, achieves almost total
oxidation of CO and HC by reaction with added air in the high-turbulence
toroidal reactors located in the exhaust-manifold positions. The
system used by CAPE-6 was installed in a 1968 Chevrolet with a 307 CID,
V8 engine matched to the standard 1968 Chevrolet used in the regular
CAPE-6-68 program. The deposits present in the RAM vehicle when it
was used by CAPE-6 were derived from using a high-PNA, leaded fuel.
The dual-catalyst system has been described, in the form used
by CAPE-6-68, by Lunt, Bernstein, Hansel, and Holt (4). Briefly, the
system used closely-controlled, slightly-rich carburetion and modified
spark-timing to produce exhaust gas from which NO was removed by
reaction with CO and/or H£ over Monel saddles in a first reactor, fol-
lowed by air addition and a second reactor containing Engelhard PTX-5
platinum-on-ceramic oxidation catalyst for control of CO and HC emissions.
The system employed by CAPE-6-68 was installed in a 1970 Chevrolet with
a 350 CID, V8 engine matched to the 1970 Chevrolet used regularly by
CAPE-6-68. The deposits in the catalyst vehicle, when tested by CAPE-6,
had been derived from use of a high-PNA unleaded gasoline.
The following table summarizes the emissions (CO, HC, NO) ob-
tained from both experimental vehicles. The CVS emissions relate to tests
by Esso Research with these vehicles, using the 1-bag, 23-minute, CVS
test (8). The cyclic-test emissions were obtained on an IR-analyzed
sidestream during collection of the PNA-emission samples by CAPE-6-68
in 7-mode cycles, and were computed with the usual (1968 standards)
weighting procedures. For these results, the instruments were operating
at their lower useful limits throughout most of the test.
EMISSIONS OF EXPERIMENTAL VEHICLES
CVS, gms./mile Cyclic Test
Vehicle CO HC NOX % CO ppm HC ppm NO
'68-RAM 4.2 0.07 1.9 0.07 3 406
'70-CAT 8 0.4 0.7 0.03 12 130
3.3 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 (2).
-------
- 20 -
Operation of the test vehicle on a Clayton chassis dynamometer in three
cold-start blocks of twelve 137-second, 7-mode cycles each covered about
30 miles and used about 2-2.4 gallons (weighed) of fuel. The total
exhaust, except a sidestream for infrared gas analysis, was cooled
and filtered and the condensed water was recovered. The samples analyzed
included the water condensate, 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 sidestream scrubber tower on the filtered exit gas
was not used after the early part of the first year, because results at
that time showed it to be of doubtful value for either PNA or phenols.
As in the past, muffler washings were included in each sample. The mufflers
used were modified by installing drain plugs and flanged inlet and outlet
fittings to facilitate their use and cleaning.
At the conclusion of the production run for a vehicle-
emission sample, standard practice was to seal the collector unit
overnight before removing the filter media and beginning the recovery
of sample material deposited on metal surfaces. Third-year experiments
using known specimens of BaP, BaA, and benzo(e)pyrene (BeP) exposed
to gases at the filter outlet, and also using analyses of separate
samples from different parts of the collector, showed that this procedure
(overnight residence) permitted reactions to occur which destroyed part
of the reactive PNA (BaP and BaA, for example) selectively, leaving the
unreactive species (BeP, etc.) unaffected. Section 7 describes these
experiments and discusses their implications. After removal of the
sample from the collector, radiotracer BaP and BaA were added before con-
centration of the sample by distillation to a small volume for analysis;
once added, the BaP and BaA tracers took any further losses into account,
at least for similarly-reactive species.
3.4 Analysis of Samples
Analytical techniques for phenols were unchanged in the
third year from the descriptions in the first-year report (2). The
aqueous condensate usually contained over 90% of all the phenols, while
the Soxhlet extract of the fiber glass filter media contained no phenols.
Although many different phenols are present, the data are reported in
terms of phenol, i.e., the parent compound. CRC-APRAC CAPE-12-68 at
Esso Research has studied techniques for estimating individual phenols
in the sample. The technique developed is reported in the forthcoming
final report of Project CAPE-12-68, but no more-complex phenol analyses
are presented in this report.
All of the third-year CAPE-6-68 PNA-emission samples, as well
as selected fuels and oils, have been analyzed by the GC/UV technique for
-------
- 21 -
PNA developed by CRC-APRAC Project CAPE-12-68. An early, but still
valid, description of the procedure was given in Appendix E of the
Second Annual Report (1) of Project CAPE-6-68. The forthcoming final
report of CAPE-12-68 (6) also describes the method. With the exception
of occasional missing values, each analysis provided data on the emis-
sion (yg.) or content (yg./gal. or ppm) for eleven PNA species:
Benzo(a)pyrene (BaP) Methyl BaP (MBaP)
Benzo(e)pyrene (BeP) Methyl BeP (MBeP)
Benzo(ghi)perylene (BghiP) Triphenylene
Benz(a)anthracene (BaA) Chrysene
Methyl BaA (MBaA) Pyrene
Dimethyl/Ethyl BaA(DM/EBaA)
It will be recognized that certain of these "species" are in fact
mixtures of several isomers with varying alkyl-group positions.
Future plans call for expanding the analyses to additional species, such
as perylene, coronene, and the benzofluoranthenes.
Multiple-species analyses on PNA in fuels were already given
in Table 3 of this Section, and Table 16 in Section 4 presents data on
used-oil PNA contents. PNA species emitted in exhaust samples are
reported in Table 22 in Section 6, which also discusses the
between-species ratios in samples from different sources. Early
CAPE-6-68 emission samples were analyzed for only BaP and BaA, and
these two species continue to provide the basis for evaluating the
effects of fuel and vehicle variables on PNA emission.
-------
- 22 -
4. EMISSION RESULTS
A summary of data on all tests for the third contract year
is presented in Table 8. This table includes data on observed emis-
sion rates of benzo(a)pyrene (BaP), benz(a)anthracene (BaA), and phenols
(as phenol) 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 in-
cluded(a). Tables B-2 through B-41 in Appendix B 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 B tables present the Federal cycle results
that appear in Table 8.
The GC/UV analytical method for PNA analysis developed by
CRC-APRAC Project CAPE-12-68 has been used on all samples produced in
the contract year. The analysis, in its present form (Second Annual
Report, Appendix E) , gives data on up to 11 PNA species. However, for
both simplicity and consistency with past work, the tests are presented
and analyzed in terms of BaP and BaA analyses. Section 6 presents
the complete PNA analytical results and discusses the interrelation
between various samples (fuels, exhaust samples) from different sources.
Multiple-species PNA analyses for fuels were given in Table 3 and appear
in Table 16 for used oils.
As noted in Section 2, the 1968 Chevrolet test vehicle had been
used for seven PNA-emission tests by Esso Research during the delay period
between the second and third contract years. Three of the tests were made
with A-fuel deposits (low-PNA, no lead) and four with B-fuel deposits
(high-PNA, no lead), which were created by Esso Research in 3800 miles
of light-duty operation on the Mileage Accumulation Dynamometer. All
emission tests used lead-free test fuels. The vehicle was then used in
CAPE-6 testing. Emissions of BaP and BaA from two of the Esso tests
(E-4 and E-6) are made available and used in this report where they con-
tribute to the presentation. Eleven-species PNA analyses for all seven
Esso tests were included in the between-species ratios that were computer-
processed for the presentation in Section 6 on average sample compositions,
4.1 Emission of PNA
Two major points are established by the PNA emission data pre-
sented in this section: (1) The emissions of PNA from both the thermal-
reactor and dual-catalyst experimental vehicles are extremely low.
(2) Engine combustion-chamber deposits and, specifically, the PNA contents
of the fuels from which they are formed, are dominant factors in governing
PNA emission from a given vehicle. (Only deposit fuels with near-zero
and near-maximum PNA contents have been studied, however.) It follows that
(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-lb. class), the factors of 23.8 and 0.0127, respectively,
will convert the tabulated % CO and ppm HC to grams/mile.
-------
- 23 -
for>,i-ivooooococoinoo-3i
cNcocsjcocncncocococococococNC'OCN
OOOO\«-HOOOi-i,-iCM
^-i
OOOOOOOOOQO
-»i--
OOOOOOOOOOO
flctol
PL, O Oq
i ItNOOOO
-
1 3
-?g
-o <
* >
sr
^[
O
COOCOOOOOOOOOOOOO
OOOOOOOOOOO
I I I I I I t I I i I
cOOOOOOOcgOOCvJQOOOxOOOQOr-r^COr^ OOvOOOcMCMOOCOCOr^r-r
-- -
O < U
oj ^ m
PC CM PC
OOOOO'-'OCOCOOO
...
llllicsllCMCMII
OOvOCOOOCMOOOCOOOOO
-
.
Jfncn
< <
, +
<
coc^cnen
-------
- 2A -
c
0
co
0
)-i
3
O
-H
r4
CO
r-l
O
en
en
0
P
0
H
cO
0
4-f
O
cn
«
0
a
0
4_t
c
0
0
r-l
a
0
FQ
r-l
0
4-J
01
rC
4J
3
CO
QJ
E
3
P
en
0
0
CJ
H
0
x:
o
cO
0
0
00
C
4-1
OJ
4-1
4-1
O
0
TJ
J-I
O
c
a
-H
f-J
r"l 0
4-1 3
3 4-4
TJ 4-1 CO
1 CO -H <
4-J CO S
rC QJ 01 PH
OOrH ^4 1
H 0 x:
1-1 AJ x: M
ctj F-i H
4-1 -^ x:
TJ -H O H
0 4-J TH 4-1
N CO 4-1
r-l [_, CO W
r-l 0) J-I 4-1
H D. 0 r-l
xi o a. co
rH )-l O
0 O O
1-4 >i 4-H CO
CO U 1 r-l
4-J CJ
(0 M-l r4 3
AJ r4 CO 4-1
-H O
en r a <
o TJ oj E:
CL CD TJ P-l
0 N 1
O -H 4-1 3
rH en o
H ttj rH
rQ rH
' (TJ 4-(
t-t 4J Q) O
to w x:
Mr AJ 0
^ CO
rQ 0 r4 3
P-l r4 OJ
CO 4-J TJ
W CO CO 4-1
E 4J CO
60 -H 4-J 0
co 3 a-
C O rQ 0
H CL h
0 4-J
rJ P en x:
W QJ 4-)
P AJ -^
3
- 4J TJ
P W 0 p^
O OJ 4-J rH
rH AJ en s-i
rH -H CO
tO C r-l rH
00 O 3
"^ -H 0 CJ
P-4 en x: -H
co en 4-i 4J
CQ -H )-,
E TJ co
4-f 0 0 Q.
O TJ
00,.
. XI O en
00 -U 0 AJ
-t r-l -H
P QJ O
H -o /-x a.
0 CO QJ
cn u AJ TJ
H 0 CO
< D- 4-> 0)
PH ^ C -r4
rH O rH
r-l 0 r4 r4
0 4-J CO X)
3 co cn co
TJ E W
rH E 0) O
O r4 r- 1 4-1
E .. >> P
r*l 0 U 0
en TJ i >
0 E cn
TJ
0 >, 0 00
CJ CO rJ C
' .C i 4J
0 bO cn
d -H J-, 0
4-i XI O AJ
O -H
O 00 3
OJ C I-i
H T-l O *
U 0) QJ cO
QJ 4J0
Q- " cn j_(
co ^-v co O
t-i 3 , 00
3 CJ
co g to
rH 3 0
00-04-1
3 U O O
4-1 CO E
I 0
j-j 0 r*- u
H 00 C
en co C 0
O 0 0) TJ
0 -r-t 0 >
O E cn 0
4J
QJ
C
O
H
en
r-l
E
0
c
co
00
3
4-J
P
0
a
cn
o
o
r I
0
CJ
OJ
m
4-4
o
cn
o
4-4
o
0
U
r--
-1-1
)H
U-i
O
4-1
CO
4-J
CO
0)
0
AJ
3
CX
E
O
U
E
O
Li
4-1
CfJ
0)
(0
U
<
3
0
c
0
AJ
a
E
3
C
o
o
H
o
J£
CJ
0
X
u
o
_
cn
Q
QJ
3
14-1
C
O
CO
0
4-J
CO
3
o
AJ
0
CJ
>>
CJ
CM
>>
x>
TJ
0
N
H
to
cn
TJ
C
CO
TJ
P
0
TJ
cO
00
H
r4
0
1
CO
TJ
P
en
0
O
O
TJ
QJ
AJ
CO
r-l
0
Q-
O
CO
AJ
OJ
fH
O
4-J
0
CM
LTl
cn
en
C
o
O
0
4-1
g
3
nj
00
bO
^
P-i
r;
>
cn
cn
en
o
w
0
r-l
^
0
u
6^
0
00
J-I
0
>
to
4-1
CJ
0
4-1
CO
>:
AJ
P
CO
U
H
S
H
CO
O
P
id
T3
4J
3
r-
r-.
AJ
01
H
C
H
P
O
en
0
0
0
nj
QJ
O
C
0
5
CO
E
4-J
cO
4-i
0
CJ
0
QJ
rJ
4-J
O
0
O
d
00
e
H
O
"o
1-1
0
o
r-
0
4-1
4-1
m
en
a
o
1-1
AJ
r-l
TJ
c
o
o
o
co
en
O
QJ
-a
QJ
0
AJ
Q)
l-i
0
0
4-J
en
to
3
0
4-J
4-J
CO
3
X
0
XT
AJ
3
CQ
0
3
4-1
C
o
,_J
u
H
x:
cu
4-1
O
c
0
4-J
tt)
a-
o
0
U
u
QJ
TJ
01
CJ
OJ
r-l
OJ
i-4
0
CQ
TJ
CO
rH
CO
4-J
0
H
CO
a
o
j-i
0
0
o
o
CO
c
o
(0
}-l
Cn
<
S-i
s
o
rH
3
X1
a>
g
C
TJ
CO
x;
3
CO
X!
j-t
0
>
0
r4
0
0
^
3
rQ
r4
(0
a
cn
o>
4J
0
H
4-1
O
AJ
V4
CO
X!
U
3
0
3
TJ
CO
Xi
CQ
CO
AJ
H
V1
^
0
*. o
0 cn
_M en
0 -r-l
CJ 0
r4
0
J-J
x-s QJ
cn 3
0
4H rH
O
£
4-1 CO
r-l
co O
"O
0 0
O (0
0
TJ l-i
CO 0
x: e
H
CXI TJ
0
4-j cn
OJ co
EH u
4-1
0
4-1
00
c
TJ
OJ
O
0)
rJ
a
m
o
0
4-1
4-1
QJ
0
Xl
H
CO
CO
O
a
0
AJ
cO
H
H
OJ
O
4J
__,
0
3
C
OJ
AJ
W
CO
O
QJ
O
CM
TJ
0
TJ
0
0
rJ
a
cn
OJ
CQ
00
cn
cn
H
O
H
E
0
<
2
P^
00
C
I-I
CO
0)
S-i
0
"O
u
OJ
CO
E
0
0
o
cn
rQ
AJ
0
0
3
3
O
OJ
r-l
0
O
u
3
0
0
4-1
O
CO
J-J
4-1
C
0
o
CO
0
«
r4
Xi
TJ
P
CO
"O
C
o
u
QJ
c
OJ
QJ
3
QJ
Xi
5-J
CO
0
0
o
CO
CO
TJ
0
CO
0
1-1
U
AJ
r4
O
a
0
TJ
0
3
CQ
u
o
rQ
0
O
1
OJ
O
3
4-J
O
4-1
TJ
0
4-1
H
E
cn
0
EH
00
r--.
AJ
0
H
TJ
OJ
TJ
0
U
QJ
S-l
D-
c
O
AJ
CO
N
H
H
rQ
CO
CO,
TJ
C
CO
H
O
a
0
TJ
0
3
4-1
CO
3
0
C
C
o
H
CO
3
E
3
O
O
CO
^
TJ
AJ
x;
00
^
r-l
H
0
JZ
0
cn
en
0
AJ
O
"O
CO
Xi
OJ
0
H
O
cn
r4
O
a
0)
TJ
rH
QJ
3
4-4
FQ
E
O
r-t
4-1
T3
QJ
o
d
rH
CJ
X
01
c
0
QJ
rQ
0
to
x;
4-J
CO
TJ
0
x;
4-1
4-1
3
x>
X?
AJ
3
O
1,
c
r4
r-.
AJ
0
H
O
rH
cO
0
01
0
AJ
0)
00
bO
3
cn
H
00
C
00
c
Q)
CO
H
O
0
TJ
4-1
O
0
CJ
O
tfl
H
CO
>
4-1
M
Xi
60
-H
x;
rH
rH
CO
E
r-l
O
tO
0
co
O
f-J
cn
en
H
E
0
P-
CO
C
3
O
00
ca
o
H
4-1
CO
H
4-1
CO
4-"
P
O
cu
00
cO
>-4
0
>
(0
P
0
-T-l
-H
OJ
O
a
a>
TJ
CO
CO
QJ
CO
a>
rC
O
TJ
0
AJ
3
C
o
u
0
0
X!
TJ
4-1
a
QJ
to
OJ
'
P
o
o
0
in
4-J
0
4J
J-I
0
H
J-I
CO
0
4-1
o
rJ
Q)
e
3
C
0)
00
tO
<£
cn
CO
CO
0
H
TJ
0
TJ
o
0
J-J
a
0
3
4-1
P-.
U
£
3
P
H
cO
N
cO
rQ
T3
Q>
O
O
AJ
0
a
0
TJ
x;
CO
OJ
fc
p
3
O
c
.5
c
D
H
P
O
-H
cn
H
E
0
<:
&
CM
o
4-J
CJ
0
4-1
4-1
0
0
0
4-1
O
"--
8
01
rH
°"
QJ
rH
4-J
4-J
CO
Xi
rJ
>,
0
0
c
o
TJ
0
3
CO
CJ
4-J
3
U
H
4-4
TJ
00
4-J
AJ
QJ
QJ
O
O
4J
3
r-l
O
iH
E
"D
0
ca
0
u
c:
H
ca
cO
E
TJ
S
0
4-1
AJ
CO
3
CO
X!
X
0)
X!
4-1
H
z
pi
TJ
0
CJ
T3
O
4-J
r-l
0
tfl
rC
ca
E
j-i
to
UJ
QJ
01
J-i
x:
AJ
o
01
o
H
o
H
4-1
H
P
00
0
J-I
O
4H
0)
X)
60
C
-H
CO
CJ
Q)
P
r-l
00
c
w
-------
- 25 -
in future very-low-emission vehicles, fuel properties (though not explored
in this program) should not be significant. In current vehicles,
fuel PNA content easily overshadows all other fuel properties that were
examined.
In common with previous practice in this project, the PNA
emissions (BaP and BaA) from the 1966 Plymouth and 1968 Chevrolet were
corrected to standard levels of CO emission (1.5% CO for the 1966 vehicle
and 0.8% for the 1968 vehicle, both by the 7-mode cyclic-test computed
results). No similar correction is available, nor apparently valuable,
for the 1970 Chevrolet test vehicle. The development of the correction
equation and its use are detailed in the First and Second Annual Reports.
Table 9 lists the % CO and observed (from Table 8) and corrected BaP
and BaA emission rates for the two vehicles in the third contract year.
Data for these two vehicles are presented and analyzed in the following
sections on the basis of the CO-corrected emission data.
At this point, it is appropriate to note that certain PNA species,
including BaP and BaA, appear to have been partially lost by chemical
reactions from the emission samples before the addition of radiotracer
PNA which would account for such losses. The demonstration and magnitude
(on the order of 50% for BaP and BaA) of these losses are covered in
Section 7. It is shown that the losses were apparently so reproducible
that the PNA species that were partially lost were no more variable than
the less-reactive loss-free species. Thus changes in the measured amounts
of BaP and BaA emitted when controlled variables were changed appear to
be valid measures of the relative effects of the variables and continue
to be so used. It must be understood, however, that the losses
preclude the assumption that quantitatively correct measures of emissions
of BaP and BaA are available for even the small -3-vehicle sample used
here.
To explain the importance of engine deposits and, in particu-
lar, the PNA content of the fuel used to form deposits, it is necessary
to introduce the concept of "deposit stabilization." Deposits were
accumulated by vehicle operation on a Mileage Accumulation Dynamometer
(MAD) under light-duty (city/suburb/highway cycle, 30 MPH average) con-
ditions, with final operation in a highway-driving, fixed-speed, mode.
Thereafter, each emission test included repeated accelerations in 36
7-mode, 137-second, driving cycles with three cold starts. It has been
found that if an emission test directly followed the light-duty operation,
PNA emissions were very much higher from the "unstabilized" deposits than
if even a limited amount of cyclic emission-test operation had occurred.
In some cases, as little as seven 7-mode cycles (i.e., a CO/HC emissions
test for vehicle-acceptance purposes) appeared to "stabilize" the
deposits in the sense of causing the next-following test to have PNA
emissions not greatly different from those in subsequent tests. More
generally, one 12-cycle block, with the exhaust to waste, was used where
stabilization was desired. Alternatively, any test other than the test
immediately after deposit formation would be a test with stabilized deposits,
-------
- 26 -
TABLE 9
CAPE-6-68
CORRECTED^ PNA DATA FOR 1966 PLYMOUTH
AND 1968 CHEVROLET, THIRD YEAR
PNA Emission, yg./gal.
Test
No.
1966 Plym. 61
62
61B
68
69
73
76
77 (b)
80
81
77B
85
89 (b)
89B(b)
93
96
1968 Chev. 64
65
71
72
78
79 (c)
82
86
90
92
94
Benzo (a)pyrene
% CO
1.59
1.47
1.58
1.49
1.66
1.46
1.41
1.32
1.39
1.41
1.32
1.24
1.34
1.47
1.78
1.84
0.69
0.79
0.68
0.95
0.80
0.68
0.83
0.67
0.83
0.81
0.69
Observed
362
136
141
125
127
321
169
305
98
155
115
98
176
453
105
100
19
16
7.8
36
40
163
79
15
27
18
26
Corrected
335
140
133
126
112
334
184
368
107
169
138
127
206
466
84
76
21
16
8.7
32
40
183
77
17
26
18
29
Benz (a) anthracene
Observed
700
155
167
129
185
360
223
306
99
193
232
66
121
217
139
104
25
30
20
105
100
144
82
29
57
54
64
Corrected
649
160
157
131
163
375
243
368
108
211
278
85
142
224
111
79
28
30
22
93
100
161
80
33
56
54
71
(a) Corrected by first-year equations to 1.5% CO for 1966 Plymouth and to 0.8% CO
for 1968 Chevrolet.
(b) Although included here, certain results may be invalid for reasons such as
choke manfunctions at startup, which are not evident in overall CO emission.
(c) Wide deviation (unexplained) of this test from normal vehicle level forces
exclusion on statistical grounds from averages related to controlled variables.
-------
- 27 -
In the case of deposits from low-PNA fuels, the distinction between
PNA emissions from unstabilized and stabilized deposits was much less,
since the emission levels were much lower. Data illustrating
deposit-stabilization effects will be presented in later sections.
The existence of stabilization as a factor on PNA emission
was an accidental discovery growing out of isolated test experiences
and analysis of the records of testing details. Only in the latter
part of the third year was its true nature realized and the work conducted
accordingly. One early, incorrect, impression was that the highway-mode
ending of deposit accumulation would stabilize the deposits. It does
not, but this was not obvious when using low-PNA deposit fuels, because the
stabilization effect is largely evident with deposits from high-PNA fuels.
Recognition of the deposit stabilization concept leads then to
two questions: (1) Does a gradual decrease in PNA emissions occur with
repeated tests, particularly when repeatedly using low-PNA test fuels
over stabilized deposits from high-PNA deposit fuels? (2) Does fuel PNA
used in a test of a high-PNA fuel carry over to affect the next-following
test? Four selected series of successive tests over deposits derived from
high-PNA fuels, for two of the test vehicles and utilizing data from three
project years, are listed in Table 10. The emission-test fuels used are
specified as to % aromatics and PNA level, which is adequate for the pur-
poses at hand. Three of the four series start with unstabilized deposits.
The importance of this condition in causing high emissions is immediately
apparent, regardless of the test-fuel used. The selected series also
display some downward trend in emissions with time, although the pattern
is broken by changes in test-fuel % aromatics and PNA, so that these
variables do influence the results and were so used in earlier data
analyses. The possibility of overall decline is particularly evident
in the last parts of series I and III in the table, where 28%-aromatics
fuels gave high emissions after a high-PNA test fuel, and then unexpectedly
low emissions. While not absolutely clear, the data in the table also
suggest that some carry-over may occur of PNA from a high-PNA fuel in one
test to a following test. Several examples that may reflect this situation
are cited in Section 4.1.5.
All of these progressive changes (stabilization, gradual
decline, and carry-over of PNA) are consistent with a process of deposi-
tion and subsequent emission as the mechanism by which fuel PNA leads to
emitted PNA. However, they have some obvious implications as to the
caution required in comparing tests that are appreciably separated in
time (not sequential) over the "same" deposits.
-------
- 28 -
TABLE 10
CAPE-6-68
CHANGES IN PNA EMISSION WITH
SUCCESSIVE TESTS OF LOW-PNA FUELS
(1966 and 1968 test vehicles, three years, with deposits formed from high-PNA fuels)
Test
Vehicle
1966
(NC)
Series
No.
II
1968
(EM)
III
IV
Test No.,
In Order
15 (a)
16
17
18
19
20
22
23
26
27
28
61(a)
62
61B(b)
68
69
73(c)
5(d)
6
7
13
14
21
24
25
29
30
31
E-4(a, e)
E-6(e)
64
65
71
72
Aromatics In
Test Fuel
46
12
12
46
46
46
11
11
28
28
28
28
28
28
38
32
28
46
46
12
12
46
46
11
11
28
28
28
32
38
28
46
38
32
CO-Corrected PNA Emission, yg./gal.
Fuel Used In Using Low- Using High-
Deposit
Formation
C3P
PNA Test Fuels PNA Test Fuels
D3
C3P
BaP
360
78
127
125
50
59
78
63
335
140
133
126
32
32
25
30
10.5
36
11
8.9
48
21
16
8.7
BaA
674
182
258
196
109
157
130
86
649
160
157
131
75
63
52
55
40
75
28
20
63
28
30
22
BaP
91
165
37
BaA
212
276
134
112
334(c)
27
36
13
83
32
163
375(c)
85
85
75
248
93
NOTE: The 1970 vehicle does not display a clear trend toward lower emissions with deposit
age, but is sensitive to stabilization and to deposit-fuel PNA content.
(a) Fresh, unstabilized, deposits after light-duty use of indicated fuel; all other
tests were with stabilized deposits.
(b) Before Test 61B, additional use of fuel D3 to check oil consumption was followed
by deposit stabilization.
(c) High emission in Test 73 may reflect the use of two successive high-PNA test
fuels over leaded deposits. See Sections 4.1.3.2 and 4.1.5.
(d) Deposits were stabilized, after accumulation, by CO/HC/NO emission tests before
the PNA Test.
(e) Between second-year and third-year contracts, Esso Research conducted 3 tests
in this vehicle with A-fuel deposits and 4 with B-fuel deposits. Data on
emissions on the first and third B-deposit tests are made available above.
Tests E-5, -6, and -7 gave progressively less PNA, using low-PNA test fuels.
Test 64 then followed under CAPE-6 resumption.
-------
- 29 -
4.1.1 Effects of Current Emission-Control Systems
The 1968 and 1970 Chevrolet test vehicles, with engine-
modification (EM) systems for control of CO and HC emissions, were
found in the first two years of the project to emit about 70% less
BaP and BaA per gallon of fuel than the uncontrolled
(NC) 1966 Plymouth. The conclusions were based on geometric means of
emission ratios between paired tests in the vehicles, with test fuels
and deposits constant between the members of each test pair. Additional
data from the three vehicles are now available, but the new tests
available are not well-suited to precisely paired comparisons. A
different approach is now used, in which the tests in each vehicle,
from all three years were grouped first between the presence of
stabilized or unstabilized deposits and then whether these deposits
were derived from low-PNA or high-PNA deposit fuels. The emission
averages from each group for the EM-controlled vehicles are compared
in Table 11 in terms of the % reduction in PNA relative to the group-
average emissions in the 1966 (NC) vehicle. The numbers in parentheses
indicate the number of EM-vehicle tests averaged for each % reduction
shown; most of the data derive from high-PNA stabilized deposits.
Tables 13 and 14 in Section 4.1.3.1 list the actual grouped data and
group averages used for these %-reduction results. The identities of
the emission-test fuels used in the averaged tests, and any other
variables, are ignored in averaging and are tacitly assumed to average
out as non-essential variables for the purpose at hand, if not under
other circumstances.
The prior figure of a 70% reduction by the control systems of
both cars now appears to be, at least, valid and possibly conservative.
Reductions of 75%-80% for BaP and 70-75% for BaA are indicated in Table 11
for both vehicles, particularly for the stabilized-deposit situations
from which most of the data are. derived.
4.1.2 Effects of Experimental Low-Emission Systems
An order-of-magnitude further reduction in PNA emission was
achieved by both of the experimental future low-emission control systems
(vehicles 1968-RAM and 1970-CAT) loaned by Esso Research. Evaluations
with fuels having high aromatic and high PNA contents, and the presence
of deposits from high-PNA fuels, offered the maximum opportunity for high
PNA emissions, yet the data given in Table 12 show that the emissions
from the experimental vehicles were no more than 1 or 2% of the maximum
values seen in the 1966 Plymouth. In fact, these emissions were lower
than any observed with any other fuel-vehicle combination in the entire
program. Figure 1 shows the data in Table 12 in graphic form. According
to data (Section 6) on other PNA species than BaP and BaA in the RAM-
vehicle exhaust, this system may be less effective in controlling
-------
- 30 -
TABLE 11
CAPE-6-68
PNA-EMISSION REDUCTIONS BY
EMISSION-CONTROL SYSTEMS
Engine
Deposit
Condition
Unstabilized
Stabilized
PNA Level
In Deposit
Fuel
Low
High
Low
High
% Reduction in average PNA Emissions/
gal. of fuel for EM controlled
vehicles, relative to 1966 Nc(a)
Benzo (a)pyrene Benz (a)anthracene
1968 1970 1968 1970
Chev. Chev. Chev. Chev.
77(4)
76(1)
73(10)
78(17)
84(2)
79(1)
84(6)
82(19)
53
63
67
69
54
75
76
72
(a) Expressed as the % reduction based on average emissions (for
specified deposit descriptions) for the number of emission
tests shown in parentheses.
(b) Low-PNA deposit fuels include A, A.5, A3P, A'+V45, and A'+HsCN;
high-PNA fuels include C3P, D3, and B.
-------
- 31 -
TABLE 12
CAPE-6-68
PNA EMISSIONS FROM EXPERIMENTAL LOW-EMISSION VEHICLES
, . BaP and BaA Rates, yg./
Test FuelC ; gallon, in Each Test Vehicle
% Ar.-PNA-TEL
46-3-0
46-3-3
BaP:
BaA:
BaP:
BaA:
1966
NC
-
165
276
1968
EM
_
36
85
1970
EM
14
49
28
63
1968
RAM
-
1.6
6.0
1970
CAT
1.1
1.5
-
(a) When tested, all vehicles had deposits from C3P fuel, except the
1970 catalyst vehicle with B-fuel (no-lead) deposits.
(b) Fuel PNA as BaP in ppm.; TEL in gms. Pb/gal.
-------
- 32 -
JUU
0 200
CO
a.
z'
O
CO
CO
2
JJJ 100
0.
o
%
///
K^] BaP EMISSION
BaA EMISSION
TEST FUELS - 46% AROMATICS
AND 3 ppm BaP
% W, RAM CAT
^ yfr ^x j t
1966 1968 1970 EXPERIMENTAL
NC EM EM
Figure 1
EFFECT OF EMISSION CONTROL SYSTEMS
ON EMISSION OF PNA
-------
- 33 -
less-reactive PNA, such as BeP, than is suggested by the BaP and
BaA results. Additional testing would be desirable, however, because
existing data may reflect engine cranking that was required for one of
the three cold starts.
4.1.3 Effects of PNA in Fuels
The effects of PNA present in fuels can be conveniently
separated into effects related to PNA in deposit-formation fuels,
effects related to PNA in emission-test fuels, and effects of deposit-
fuel PNA on the accumulation of PNA in engine lubricating oil. The
studies of fuel PNA have been limited to two levels: near zero PNA
and PNA near the field-maximum level determined by Esso Research in a
limited New Jersey sampling (2) at the start of Project CAPE-6-68, i.e.,
that level of PNA characterized by a BaP content of about 3 ppm
and which could be introduced by adding 22 grams/gallon of a 384+ °F
VT catalytic-reformate still-bottoms containing naphthalene and higher
aromatics. The fuels studied for "high PNA" (Section 3.1) were either
blended to this level with the still-bottoms additive or already con-
tained approximately this level of PNA.
4.1.3.1 PNA Present in Deposit Fuels
Data showing the importance of some undefined deposit property
in establishing a PNA-emission level for a vehicle were presented in
the Second Year report. This property caused a roughly two-fold difference
in emission levels. The discussion of fuels in the early part of this
present report indicates the probability that this property is the
PNA-level existing in the fuel used to form the deposits, and this hypothe-
sis appears to be confirmed by recent data. Also, in Section 4.1 the
distinction was made between "stabilized" and "unstabilized" deposits,
noting that it was particularly significant for deposits formed from
high-PNA fuels.
Table 13 lists all the valid BaP and BaA emission data for
the three vehicles for the three project years, grouped according to
deposit condition (stabilized or unstabilized) and by the identity (code)
of the deposit-forming fuel. No distinction is made as to the composition
of the emission-test fuel used for each result, even though this does
have an effect. Examination of the table readily shows higher PNA emission
for (1) unstabilized high-PNA-fuel (C3P, D3, B) deposits compared with
stabilized deposits, and (2) for deposits from high-PNA fuels compared
with deposits from low-PNA fuels; and reduced PNA emission for the control-
system vehicles.
Grouping deposit fuels as low PNA (A, and blends in A or A')
or high PNA (C3P, D3, and B) in Table 13 gives the summary in Table 14,
-------
TABLE 13
CAFE-6-68
PNA EMISSION RESULTS(a)
ARRANGED BY
CONDITION AND ORIGIN OF ENGINE DEPOSITS
Deposit-
. . Formation
Deposit Condition At Test Fuel
Unstabillzed A
A.S
ASP
A'+V45
A'+HsCN
C3P
D3
B
Stabilized A
A.S
ASP
C3P
D3
B
(a) A variety of emission-test fuels, not
tests listed.
(b) Emissions for 1966 and 1968 vehicles
Benzo(a)pyrene,
1966
Ply».
54
127
18
-
76
-
360
335
-
56
39
54
56
39
57
62
84
36
29
-
91
78
165
127
125
50
37
59
78
63
-
-
-
-
-
-
-
_
-
140
133
126
112
334
184
107
169
138
-
1968
Chev.
15
17
3.5
-
-
29
-
-
83
12
12
26
18
-
-
-
-
14
3.4
-
32
32
27
25
36
30
10.5
13
36
11
8.9
-
-
-
_
-
-
-
-
_
-
-
-
-
21
16
8.7
32
40
77
differentiated
are corrected to
(c) Fresh deposits, after light-duty mileage-accumulation
1970
Chev.
10
-
-
12
-
-
-
- '
72
_
-
-
-
-
-
-
-
_
-
7.8
20
10.5
12
29
17
19
34
26
26
36
48
20
28
12
14
19
18
17
19
_
_
-
-
-
_
-
-
-
-
in this
uniform
Benz (a)anthracene,
1966
94
85
30
-
79
-
674
649
-
126
82
124
155
51
82
104
111
49
46
-
212
182
276
258
196
109
134
157
130
86
-
-
-
-
-
-
-
-
-
160
157
131
163
375
243
108
211
278
-
1968
Chev.
17
33
13
-
-
71
-
-
248(d)
25
23
56
54
-
-
-
-
15
15
-
75
63
85
52
85
55
40
75
75
28
20
-
-
-
_
-
-
-
-
_
_
-
_
-
28
30
22
93
100
80
table, were used
CO levels.
1970
Chev.
34
-
-
32
-
-
-
-
167
_
-
-
_
-
_
_
-
_
-
22
70
38
28
52
42
57
72
60
51
61
70
31
63
27
49
39
38
60
101
.
_
_
_
-
_
_
_
_
_
for t
, are "unstabllized". Deposit:
(d)
are "stabilized" if operation in at least seven 7-mode 137-second cycles has pre-
ceded the listed PNA emission test.
Test result made available by Esso Research from between-contract use of this vehicle,
including deposit formation (B fuel) and four emission tests.
-------
- 35 -
*
rH
W
hJ
pQ
<:
2
M
U
O
p3
CJ
.
rH
Cfl
60
^
60
3.
»
01
C
at
0
Cfl
^i
42
4J
C
cfl
CO
£
p
fa
!3
0
M
H
«
2
o
fa
1
H
t-H
CO
o
PL,
W
O
fe
o
g
u
g
o
u
^^
55
eu
/-N
^^
N
G
0)
n
»
6£
>
>
<~{
o
0
1^
CO f^ rH Cvl CO ^~
en \o cvi in
rH in Cvl
>
01
rC
U
oo
^ OO en rH ON O^
en *^ en in *
Cvl r*"» rH
e
iH
PL.
vD
v£>
Cvl CN VO en OO O
r~- vo o\ oo
VO 00 rH Cvl
/**\
43
v^
rH
Cfl
60
60
A
0)
c2
0)
1-1
a
cfl
o
N
c
01
PQ
60
<^
^
d)
rd
°
O
«.
s-^
Oi
^-N ^-v ^ rH
Cvl rH rH ^
^ v^ V^1 Cvl 00
rH Cvl in 00 Cvl
rH I-- Cvl
vO
>
01
O
OO
vO
s~\
*^. x*\ ^^ r**
-* -H vo H
vo en CN «^ r^. o*
rH OO rH Cvl
U-l rH
C2
p^
rH
VO
x~\ /~*
/ v /~^ O O»
* Cvl rH rH
Ol OO O rH Cvl «3-
VO -* m Cvl
en m rH cvi
rH
OI
3--v
Jf 1 f^
1 >_^
4J CX
H 3
CO O
0 rl
CX O
01
Q
<: s < s
"^ £3 O ^1 13 O
£5 p ( j "] ^^ f\J ) 1
fXj "*». pU "^-^
43 43 43 43
& 60 60 U GO 00
O *H *H O 'H *H
rJ PC 33 r3 5S 33
G
4-1 O
H -iH
CO 4-1
O -H
CUTS
01 G
a o
U
O
OI
tvl T3
H O)
H -H
43 rH
tfl T-l
4J 43
CO CO
C 4J
p co
4-J
c -3
H CO
01
43 ^1
60
43 0
01 )-i
M Ol
cfl 4D
« 3
*v3
C CI)
cfl 43
4J
en <4H
Q 0
« eu
PH 60
en co
CJ rl
01
CO >
rH Cfl
CU
3 01
M-i 43
4-1
*«S
4-1 CO CO
(3 -H rH
a) a)
4-1 3
G Ol <4H
O 3
CJ rH 4-1
Cfl CO
<; > ai
!Z 4J
PM >
CO rl 4-1
CO O CU
Cfl O iH
rH M
CJ CU Cfl
CU 4-1
M Cfl
Cfl 43
4J 43
Z -H 4-1
0 & -H
CO >
33
o cu
< -rl >
co M
a co 01
(3 -H CO
cfl B 43
ai o
m P-I co
M
- M CO
P
CO ?
rH < 43 O
0) !S 0 43
£14 [14
/ \ /^N
eo ,n
*~s N-X
-------
- 36 -
where the effects of deposits, of deposit-fuel PNA, and of vehicle
controls are evident. These emission averages have already been used
to calculate the reductions in PNA for control-system vehicles (Table 11).
Table 14 includes the ratios of emissions for deposits from high-PNA
fuels to those from low-PNA fuels. The ratios demonstrate the highly
significant conclusion that deposit accumulation from high-PNA fuels
(near the field maximum, or 3 ppm BaP content) causes PNA emissions
that are 2 to 9 times greater than for deposits from low-PNA (near
zero) fuels. The factors range from 5.0 to 8.6 for unstabilized de-
posits (emissions in accelerations after light-duty operation) down to
1.9 to 2.8 for stabilized deposits (i.e., subjected to some prior
accelerations before testing). The factors are reasonably consistent
between vehicles and present a uniform pattern of higher PNA emission
for high-PNA deposits. Where the data are adequate to justify analysis,
t-tests show the within-vehicle average PNA emissions for the two
levels of deposit-fuel PNA to be non-equal at the 90% to 99.9%
confidence levels.
It should be noted that this clear distinction between de-
posits from low-PNA and high-PNA fuels derives from data in which both
the deposit fuels and the emission-test fuels vary over wide limits of
composition in terms of aromatics, lead and phosphorus
contents, and deposit variations within the general class of "stabilized"
are also present. The variability in results introduced by these
sources of variance does not overcome the major effect of deposit-fuel
PNA content. However, it should also be noted that deposit fuels have
not yet been tested with PNA intermediate between the extremes used here.
4.1.3.2 PNA Present in Emission-Test Fuels
In addition to the data presented in the preceding section on
the long-term effects of fuel PNA, data are available on short-term
effects, i.e., with the PNA in the fuel used (about 2 gallons) for the
30-mile emission test. The evidence that fuel PNA affects exhaust PNA
through a deposition > scavenging mechanism raises a question whether
so short a test is an adequate measure of fuel PNA effects. The Second
Year Report used data from three test-fuel pairs in each of three
vehicles to find only a non-significant 8% increase in BaP and a 32%
increase in BaA for fuel PNA (as BaP content) increasing from 0 to 3 ppm.
The tests in each pair (low- and high-PNA) were generally adjacent
in the test sequence so that any effects of gradual deposit changes
would have been minimized. All of those tests were conducted with deposits
from a high-PNA leaded fuel (C3P) present.
In third-year tests, one additional test pair, each involving
the same pair of fuels, is available for each vehicle (Table 15). In
these tests, deposit fuels A and B were used in that order as emission-
test fuels; they are similar except in PNA content. Data for the 1966
Plymouth also extend to a third fuel, D, which is similar to fuel B and
-------
- 37 -
m
H
w
^J
5
H
cd
^^. M-J
^^ r-i
cd cd
4J 00
cd -.
TJ
CO
£5
O
M
CO
CO
M
52
EM
^J
S5
P-I
JJ5
o
00
O 4J -H
I-- CO CO
ON CU O
rH H P-
CU
P
^3
cd
PQ
00 0 1
CO vO
00 1
rH
co
rH
CU
PM
cd
PQ
^
oo r- I
rH rH
CU
m I I-H
a\
O
U
a
o
It 1
4J
CU CM 0
rH f^ O
0 . p
> r-T
CU 1^- CO
J3 4J
O CO iH
4-> CO
00 M O
vo cu &
ON H CU
rH P
S
3 CM O
p^ *
rH 00 CO
PL< VO 4J
, 1
vo co co
vO 4J O
ON CO O.
rH CU CU
H P
^J
R)
PQ
cu
> CO
rH CO in
co vo r~-
H rH CO
«* vo cd cd
CM CO J3
rH CN J
CO W
a) H
i_j
O T>
H C
PU
cd
pq
4J
^M W
CU
r*. CM ^~
CM rH CO
rH rH CO
00 fO cO
CO VD CO PQ
CN vO
ON S
c I-H a
CU O.
3 -0
cr a w
cu cd cd
co
a rH
H CU
N_> 3
4J
CO
cu
H
a
o
H
CO
CO
H
S
^^
js
^N CO
u o
H iH
1 CO
rH ON CM iH
fr« O 4J
1 CM CO (fl
CU 1 1 1 rH
T3 00 CN 00 CU
O CO CO CM Pd
U 1 1 1
^ <: PQ p
vO TJ
vO CU
ON -M
rH Cd
M CO
co co O
Cd Cd MH CO
|H|
» > CO
i-H rH C
-------
- 38 -
provides the unique case of two successive tests with high-PNA fuels.
Another unique feature of the data is the presence of unleaded deposits
in the 1968 Chevrolet.
The data show that in the 1966 and 1970 vehicles with leaded
deposits, fuel B (high PNA) caused no increase in BaP and a 24% or 58%
increase in BaA. These results are consistent with the earlier results
from tests with leaded deposits. The data show much larger increases
(164% to 322%) in two situations, however: lead-free high-PNA (B-fuel)
deposits, and in the second of two tests with high-PNA fuels over
leaded deposits. Apparently, the real, potentially large, effect of
fuel PNA became evident in 30 or 60 miles, depending on whether the
deposits were leaded or not. It follows that the 2000+ miles we have
used for deposits appears more than adequate, but the 30-mile emission
test may not be adequate. Of course, this conclusion is relevant
only to the planning of suitable tests. The large long-term effect of
fuel PNA is clearly the most realistic in terms of vehicle emissions and
air quality.
4.1.3.3 Effect of Fuel PNA on PNA in Used Oils
GC/UV analyses for the PNA content of fresh oil and of used
oils from 2000-mile deposit accumulation runs on a variety of low-PNA
and high-PNA fuels are given in Table 16. It is clear that the high-PNA
fuels (B, D3), in any of the vehicles, caused much more PNA to appear
in the engine oil than was found with the various low-PNA fuels. Even
the latter did not entirely avoid the accumulation of PNA compared to the
near-zero levels of fresh oil, however. The data on three vehicles using
B fuel show that the control (70-80% reduction) in PNA exhaust emissions
seen for the 1968 and 1970 (EM) vehicles is not paralleled by any
substantial reduction in the buildup of oil PNA.
By way of contrast, it should be noted that in 2000 miles of
use of fuels with 3 ppm of BaP, the oil acquires 12 to 23 ppm of BaP.
No data are now available on the results with longer oil drains, nor has
the effect of this used-oil PNA on PNA emission in the exhaust been
determined. All PNA-emission tests have been conducted with a fresh oil
charge and new oil filter.
4.1.4 Effects of Lead and Phosphorus in Deposits
Table 17 presents available (3 years) data from sets of
emission tests with as nearly as possible the same test-fuel used over
different engine deposits at different times in the program. The first
set of tests (1966 Plymouth) shows that the results are equivalent for
high-PNA leaded deposits, with or without phosphorus. The second set
for the Plymouth shows equivalent emissions for 0 and 0.5 grains of
lead/gallon in a low-PNA fuel.
-------
- 39 -
oo
vO VO
iH I
VO
W 1
J H
pQ Q.
<5 «
H U
52
u
CO
33
J"
>
O -H H
s
W co o
co cu <±
E3 rH 1
- J ^^
^n fcS-
13 60
M rH
O
< o H PQ
2; o cd
P-i CM -H CU
CJ *O
fe >-i (J O
O O CU CJ
4-1 g
CO B
W 13 O
CO CU O
>H CO
hJ 3 CO
<| -H
S CO
-'
00
rl
cd
CJ
rH
rl
0
43
CJ
cd
W rH
^ -H
0
i"!
4J
H
!3
TJ
CU
CO
3
/*^
CU
*O
o
"*~S
rH
CU
3
00 >
- 43
CJ
vo e
vO ^»
CL.
o >
r*» cu
- 43
0
9
&
vO >-,
VO rH
- P-I
^
O CU
^6
^
00 0)
VO 43
- CJ
C3
VO S
VO rH
- f*4
43
CO rH
^ *r^
rl O
fe
0)
rH
O
H
43
cu
vo co ^f *^ in f^* *^ m
CMCO cnOiHcvJiHO-^
\Q \^ Q% «^- vo r^1 **^ oo
«MCM OOrHCMrHOCO
^^
f*» vO 1 sf 00 O CO rH
. . . o
OCM OrHOOOO
O O 1
coin oorHvoi^
CMCM CMrH»*CMCM<-
CO in 00
cMin moor«-
rHrH rHrH-*rHOOr«-00
OO
vOvO HmOO 'COvO
rHrH CMrHs*CMCOHCM
co r^ co 1
vor^ rH r^ o
rHrH CM r fO CM CM VO
r-^-d- oocooooo
OO OOCOOOOO
x-s CU
cd cu e3
^s C3 0)
B 0) cj
ex cj cd cu cu
ex cd M c3 c3
rl 43 CU CU
rH CU fi 4-1 e3 >^ >^
H C CU C cd CX CX
O 0) rH Cd x-x x-sx-s
cuu cut>*. x-^cdcdcu
M-I C Cd CM Cd^-"-'>-'
OCUri CUCU ^^NOO
M43 MCX NC3NN
4J >, 4J >>x-\ Cl 0) C C
C Cud CX -H CU43CUCU
CU x^ cd x\ 43 4^ 42 43
4-> Cd X-N CU 60 4-»
C v_x nj ^x<«^ ,^~. >, >>
CJ NN NNCU43CU4343
GC Ce3rl4Jg4J4J
C_5
c
43 CU
CX CX
H CU
rl Cd
4-1 (-1
4J
M
CU M-I
C3 0
cu
co cd
>. rl
rl 4-1
43 O
o cu
CX
S-l CO
o
14H >
J 1
CO
cu cu
CO 43
^t 4-1
rH
cd C
^
x-v
cd
-------
- 40 -
w
r3
PQ
IS
< \
cd
t^s
C
O
iH
co
CO
H
8
w
.Jj
23
P-l
rH
cd
00
^
oc
3.
T3
PJ
^
CO
4J
iH
CO
o
cu
cu
n
M
o
fa
13
CU
CO
[""1
i-H
0)
^
fa
4-1
CO
cu
H
4-1
CO
CU
H
4J
CO
0)
H
CO
PQ
PL,
cd
PQ
c
o
H
4-J
H
T3
e
o
o
4-J
H
O
a
cu
Q
i-H
cu
3
fa
O
iz
cu
CJ
H
J2
CU
Cd
CO 4->
C CO
3 C
3
PL? -
cn en
U O
cn
O O
W 0
m cn
in rH
rH VO
O
53
vO
vO
rH
CT> rH
**^ lO
vo a>
PO co
T3
cu
N tj
rl CU
H N
,£) rH
Cd -rl
4-1 ft
co cd
4J
CO
in
«\
-------
- 41 -
The estimates and averages for the 1968 Chevrolet emissions
show similar equivalence for deposits for C3P (lead and phosphorus)
and B (no additive) high-PNA fuel deposits. Finally, the 1970 Chevrolet
data show that lead and phosphorus have no effect on emissions from low-
PNA fuel deposits. In summary, lead and phosphorus in deposits do not
appear to influence PNA emission.
4.1.5 Between-Test Carry-Over of Fuel PNA
Data in Section 4.1.3.2 and Table 15 suggest a short-term
carry-over effect related to deposits. Leaded deposits appeared to
"hold back" fuel PNA in a 30-mile test so that it appeared strongly in
the subsequent test if use of a high-PNA fuel continued. The long-
term effect would still be an emission increase attributable to fuel
PNA. However, the short-term effect is significant because it implies
the carry-over of high test-fuel PNA from one test to another test in
which the test-fuel PNA might be low. Several examples from Project
data suggest that this does occur:
1) In the 1968 Chevrolet, low-PNA-fuel Tests 13 and 21 show
smaller BaP decreases than BaA decreases, relative to
their respective preceding high-PNA-fuel Tests 7 and 14.
2) In the 1968 Chevrolet, Test 29, low-PNA-fuel, showed
increased BaP, but constant BaA, relative to the
prior Test 25 (on high-PNA-fuel), while the subsequent
Tests 30 and 31 on fuels similar to the Test 29 fuel had
much lower PNA emissions.
3) In the 1970 Chevrolet, Test 53 (low-PNA fuel), preceded
by a high-PNA fuel, Test 52, had higher emissions than
occurred with the similar low-PNA fuel used in Test 49.
4) Similarly, in the 1970 Chevrolet, Test 70, low-PNA fuel,
preceded by high-PNA-fuel Test 67, had greater emissions
than Test 66 using a fuel similar to the Test 70 fuel.
These examples with leaded deposits suggest caution in viewing test re-
sults that may be influenced by PNA carry-over.
The carry-over may apply with unleaded deposits, as well, but
there are less data available. The effect may explain the unexpectedly
high emissions from A fuel (low PNA) after B fuel (high PNA), with B-fuel
deposits, in Tests 78 and 79 in the 1968 Chevrolet. A parallel sequence
and set of test results appears for Tests 76 and 77 in the 1966 Plymouth
(high A-fuel emission after lower B-fuel emission), but the A-fuel test
was characterized by a severe choke malfunction in one cold-start and was
not accepted as a fully valid test.
-------
- 42 -
4.1.6 Effects of Test-Fuel Aromatics
Fuel aromatics effects were evaluated in the first two years
and there is no basis at this time for new conclusions in this area.
Some of the potentially useful new data appear to be influenced by PNA
carry-over (see above), or by changes in the potency of deposits as
PNA sources, i.e., from depletion of the PNA present in the deposits
as a result of its in-test emission.
As an example of PNA-depletion in deposits affecting the
apparent effects of fuel aromatics, Table 18 presents data from three
consecutive tests in two vehicles with the same lead-free, low-PNA
fuels used in both vehicles to provide 28%, 38%, and 46% aromatics.
Both vehicles had stabilized B-fuel deposits (no lead, high PNA) .
In order to "maintain" the deposits, the 1966 Plymouth was operated,
between tests, on fuel B, in emission cycles, with the exhaust to
waste. The 1968 Chevrolet was not operated between tests and had not,
in fact, used a high-PNA fuel since the fourth-preceding test (by Esso
Research, between contracts); i.e., its deposit PNA could well have
been depleted. The Plymouth data in Table 18 are reasonably in line with
the fuel aromatics contents, except for the BaA pattern, and are at a
level comparable to first-year emissions with other high-PNA deposits
in this vehicle. The 1968 Chevrolet data, however, show no aromatics
effects. Instead, a downward trend (PNA depletion) appears (Table 10,
Series IV), reaching the level seen earlier in this vehicle for A-fuel
(low-PNA) deposits. Thus, in six 30-mile tests using low-PNA fuels, the
high-PNA character of the B-fuel deposits in the 1968 Chevrolet had been
lost.
4.1.7 Effects of Two-Ring and Three-Ring Aromatics
Two-ring and three-ring aromatics (atmospheric-pressure boiling
points of 384-575°F) were blended in unleaded low-PNA fuels A and A1
(section 3.1) at the concentrations at which they might occur in a fuel
with 3 ppm BaP (five-ring) derived from non-rerun catalytic reformate.
The PNA (four-ring and five-ring aromatics) were not present in significant
amounts. Fuel-supply limitations forced the use of two different but
similar low-PNA base fuels, A and A' with 38% and 47% aromatics, respectively,
Two vacuum-distillate reformate cuts, V4 and V5, were used together as
the aromatics additive with the designation V45.
Data from three tests related to this variable are summarized
in Table 19. Between the base-case test, No. 85, and the direct-effect
test, No. 93, two tests occurred with severe choke malfunction at startup;
the results of these tests were accordingly rejected. Carburetor and
choke overhaul then inadvertently caused slightly richer (higher % CO)
operation in the subsequent tests. (Prior to these tests, cyclic operation
on A fuel, with exhaust to waste, was used to purge any effects of
-------
- 43 -
oo
i-H
W
hJ
a
en
§
M
H
H
O
CJ
CO
O
Pw
I
EC
H
CO
H
CJ
CO
CJ
M
H
i
S
cd
IMS
CO
cu
rH
CJ
H
42
CU
f*i
4-1
O
pq
C
H
CO
4J
H
CO
o
p
cu
0
rH
CU
3
(X4
I
pq
*o
cu
N
H
rH
H
ft
CO
4-1
co
co
4-1
CO
43 CU
4J H
& «3
CU
« cu
4J JJ
CU 4-1
tH CU
o pq
M
> c
Q) O
r* - '
r^ "M
CJ 4-1
cd
oo l-i
ID CU
rH O
O
~\
CJ
C
O
iH
cd
00
"'l^...
00
3-
<
cd
pq
00 O CM
CM CO CM
P-i
cd
PQ
r^
tH VO
CM rH 00
O
z
4j
CO
CU
H
vO vo r^«
CO
4-1
CO.
CU
H
43 C3
4-1 CU
iH CU
^ J5
4-1
> cu
43 PQ
*3 /"N
^
rH T)
PM CU
CO
ON rH
rH CU
3
p?
N
O
c
o
tH
cd
00
00
^J
Hj
pq
00 rH 00
O rH r-
rH CM CM
PH
cd
pq
r^ cyi oo
O vO CO
rH rH iH
o
4-1
CO
cu
H
pq
o iH r>-
oo oo r~
rH
CU CU
CU 3
J^ pLj
pt4
1 4-1
-a w
cd cu
, c
< o
S5 -H
P-l CO
1 CO
O E3
*-5 W
CO
CJ
H
4J
crt
g
0
**
84
CO VO
CM *
CU
o
CJ
0 0
0 0
o na
CO "">
oo
CO
^
co
4-1
CO
CU
4-1
00
J3
H
CU
^
M
CU
4-1
C
H
4->
3
O
43
4J
H
Jj
A
}J
CU
1-1
0
-a
cu
4-1
CO
J5
cu
43
a
H
*O
CU
Q
}_l
O
J_)
cu
a.
cu
cu
>
CO
4-1
CO
0)
H
cd
st conditions, with
cu
4-1
1
(3
O
H
CO
co
1
o
H
rH
CJ
0
CU
*rt
a
3
CO
cd
>
^
cd
cu
1-1
o
(3
xjT
juj
PH
43
00
H
43
N '
iH
CU
3 CU
M-l 4J
CO
pq cd
[5
C
O O
4-1
c
0 4J
H CO
u 3
cd cd
M ,**!
CU X
a cu
o
X-N
V '
CO
rH
CU
J^
CU
tH
O
CJ
6
o
»4H
(3
3
O
4-1
*o
cu
4-1
0
cu
1-1
M
o
CJ
CO
H
CO
CO
g
cu
^J
53
PM
CJ
-------
- 44 -
CU
tfl
pq
m
oo
cu
cu
cu
CM
oo
vo
T)
0)
T3
cfl
cu
rH
C
O
t-t
c/>
CO
PH
»
O
c/i
O
ON
rH
W
00
VO
VO
H
CJ
3
O
PM
vO
ON
o
O
H
CJ
w
w
pq
O
CJ
vO
vO
ON
CO
oo
in
o
oo
m
oo
pq
ON
oo
ON
oo
en
O
O
C3
O
H
4J
rl
T3
rj
O
CJ
4J
H
CO
O
ft
CU
Q
T)
a
cfl
r 1
cu
PM
4J
CO
CU
H
1
rj
O
H
CO
CO
H
w
4J
CO
CU
H
C
O
i-{
CO
CO
H
13
pj
^
O
PH
e
0
Vi
I4H
CO
4J
H
CO
O
CU
""CJ
13
CU
N
H
rH
H
'cfl
4-1
CO
C
rc}
4->
H
£5
s 1 I
rH CU
0) 3
3 MH
I}.)
<3 ^
CO
4-1
CO
CU
4-1
cu
^
H
4J
CJ
CU
*4-t
CU
13
A
*o
CU
K^
O
f~1
o
rl
cu
(§
4-1
CO
CU
4J
("]
4J
s
A
4J
H
CO
O
cu
13
rH
CU
3
HH
1
CU
e
«
CQ
V
m
»^-
^>
+
«
H
CU
3
UH
1
CJ
^
fN
T3
CU
T3
CU
O
CU
^|
cx
t^">
fl
(ti
0)
4J
3
H
e
H
rH
0)
O
4-1
A
0)
4-1
CO
Cfl
O
4J
4->
CO
3
cd
^C]
^
cu
«
ON
00
*t
ON
00
co
4-1
co
cu
H
MH
o
co
4-1
a
cu
m
M-l
cu
a
o
rl
MH
CO
4-1
H
CO
o
cu
T)
rrt
CU
N
H
rH
H
f\
Cfl
4J
CO
C
rl
4J
H
|j
A
m
>
+
m
x^-
>
.
-------
- 45 -
the defective tests and return to the conditions after Test 85.) The
PNA data in Table 19 are presented in both "observed" and "CO-corrected11
(to 1.5%) bases to show that the same conclusions are reached on either
basis and are not caused by the relatively large corrections for CO
differences. The data are characterized by BaA values that are both
larger (usual)>about equal, and less (rare) than the BaP values, as
well as by the differences in % CO.
In spite of these differences, the data for Tests 93 vs.
85 show only a small change in BaP (a real decrease is unlikely from
the additive, but might reflect a small low-PNA deposit stabiliza-
tion difference); and a modest increase in BaA emission. This result is
similar to the earlier direct-effect result for added fuel PNA, where
the addition involved V45 as well as the four-ring and five-ring
still-bottoms aromatics.
Comparing Tests 96 and 93, with similar CO levels, it is
clear that the deposits from using fuel A1 + V45 for 2000 miles, even
though not stabilized and hence presenting a severe case, did not cause
increased emission of either PNA species. (Used-oil PNA content (Table
16) increased only slightly from use of this fuel.) Thus, while the
presence in fuel of aromatics with two through five rings usually caused
a small direct effect and a large deposit-related (long-term) effect on
PNA emission, the presence of only two-ring and three-ring aromatics
caused a similar short-term effect but no further effect on extended use
to form deposits.
4.1.8 Effects of a Heavy Catalytic Naphtha
The test conducted in the third year included tests in the
1968 Chevrolet of the effect on PNA emission of a 20% blend of a "summer"
heavy catalytic naphtha (442°F end-point by ASTM D86 distillation
(Code HsCN) in a low-PNA unleaded base fuel (Code A'). As discussed in
Section 3.1, the naphtha blend contained significantly more C..,-C,,
aromatics (including two-ring and three-ring aromatics) than would be
present in a "high PNA" fuel but contained relatively little PNA (four-
ring and five-ring aromatics). Thus the heavy catalytic naphtha studies
examined a higher level of two-ring and three-ring aromatics than were
used in the 1966 Plymouth tests described above, and also introduced
other high-boiling materials such as olefins that may be present in the
naphtha.
The data obtained are presented in Table 20. The base-case
test (No. 86) was carried out with low-PNA fuel A for both deposits and
testing, but the lack of additional fuel A forced use of a new very similar
fuel, A', having 47% aromatics (vs. 38% in fuel A). The test sequence
included a test (No. 92) in which this fuel base was blended with a
384°F + reformate bottoms sufficient to add 2.9 ppm of BaP. This test
permits direct comparison of the effects of fuel PNA and of the summer
-------
- 46 -
CO
to
55
O
ct)
o
CM
W
3
oo
1"
vO
W
§
O
M
oo
vO
CTi
rH
cd
^3
PQ
00
ZL
n
CO
T3
01
4-1
O
ru
H PD
r4 Cd
O PQ
o
r^
rH
c
o
H
4-1
H
T3
(3
0
CJ
4-1
H
CO
O
p.
0)
a
4-1
O
(3
A
rH
01
3
14H
<
e T)
O 0)
^ N
i*H -H
H
"H
43 43
CO Cd
01 4J
M CO
fa
VO
m
SO
CM
*
13
O
H
4J
cd
M
CU
a.
o
rl
CU
43
4-1
O
O
(3
A
4-1
CO
0)
4-1
4-1
X
0)
55
«*
in
00
rH
ca o
0) 4-1
rH
(J rj .
>-. M vO
cj 3 oo
4-1
CM 0) 4-1
rH ^ CO
CU
>. 0 H
43 4-1
S-i
T3 rH CU
0) 01 4J
T3 3 HH
01 MH cd
0
cu <3 c
rl 0
a. c -H
O 4J
~ -H
4J 0) T3
CO 4-1 C
0) CO O
4-1 Cfl O
s
4.J 01
X O 43
0) 4J 4-1
SB
H
r^
a>
CM
4-1
0
fi
M
55
CJ
ca
S3
T
4H CO
<
oo
3-
CM
m
00
43
00
rl
43
Nw/
pq
rH
0)
3
<4-4 T3
01
(3 N
O -rl
rH
CO -H
0) 43
H Cfl
H 4-1
0 CO
O 4->
O O
O 13
CM
*
H s-i
0)
-------
- 47 -
heavy catalytic naphtha (HsCN). The naphtha (at 20%) was tested for
both its direct effect (Test 90) over existing deposits and its deposit-
related effect (Test 94).
Comparison of Tests 92 and 86 shows that the PNA additive caused
very little difference in BaP emission and a small increase in BaA emission.
This is in line with earlier experience for short-term effects of added
PNA. The unstabilized deposits in Test 86 were low-PNA deposits . From
earlier data, stabilization of such deposits would have little effect on
these values. Also, the change in aromatics from 38 to 47% (fuel A to A')
has had little effect, since the difference between Tests 86 and 92 is
adequately explained by the PNA additive.
Comparison of Tests 90 and 86 does show an increase in emission
attributable to use of the naphtha blend in a short-term test. Some,
but not all, of the data on blends of a different heavy naphtha in the
Second-Year Report also showed increased emission attributed to that
naphtha. Test 94, after 2000 miles use of the A* + HsCN blend and
without deposit stabilization (i.e., a severe case) shows emissions
elevated only slightly from Test 90, where formation of deposits was
not involved and the deposits present were stabilized. It is apparent
that the heavy catalytic naphtha caused some immediate increase in PNA
emission, but no further substantial increase occurred with further use
of the fuel for deposit accumulation. By contrast, in Table 20, when
this vehicle was operated on high-PNA fuel B for 2000 miles and tested
without stabilization, the BaP and BaA emissions were 83 and 248 yg./gallon,
or about three times as great as from the 20% heavy catalytic naphtha blend
in Test 94.
4.1.9 Effects of Engine Mechanical Defects
In the course of conducting CAPE-6 emission tests, four tests
occurred in which known mechanical defects apparently caused significant
increases in PNA emission, to the extent that the tests were not useful
for their original purposes. The tests do have value, however, as indicat-
ors of the possibility that incorrect engine operation can overcome expected
reductions in PNA emissions from control systems or fuel properties.
In Test 50 (1970 Chevrolet), loss of the vacuum spark advance
hose connection in the third block of the test suppressed NO formation,
strongly oxidized deceleration hydrocarbons, increased fuel consumption,
and elevated the peak exhaust temperature at the muffler exit by about
80°F. (The TCS system of this vehicle normally prevents vacuum advance
below 23 MPH, so the failure described was only effective above that
speed.) The BaP and BaA emission rates of 122 and 110 yg./gal. of fuel,
respectively, were 2 to 3 times as large as would be reasonably expected
for the test fuel (high-PNA) and vehicle. It is abnormal for BaP to
-------
- 48 -
exceed BaA, and strongly suggests the "cooking out" of higher-boiling
BaP from engine deposits because of the high cylinder-wall and gas
temperatures caused by spark retard. The observed increase in PNA
was measured against a background of the preceding two "normal" test
blocks, so that the effect of the malfunction was probably much larger,
when it occurred, than is apparent from the data on the basis available.
The effect might not necessarily continue, however, since the supply
of PNA in deposits would be depleted.
Tests 77, 89, and 89B in the 1966 Plymouth were characterized
by apparent failure of the vacuum-actuated unloader (or "pull-off")
on the automatic choke to open the choke blade correctly when the
engine started. This occurred in one of three cold-starts in Tests 77
and 89, and two of three in 89B. Using recorded CC^ and 02 analyses,
the off-scale CO content of the exhaust in these starts was estimated at
12-14%, accompanied by misfire and high unburned hydrocarbon levels.
The CO and HC remained somewhat elevated into the first acceleration,
by which time the choke had opened to its thermostatically-controlled
setting, and operation became normal. The start-ups had little or no
effect on the overall computed CO emissions. On examination, the PNA
collector system was unusually sooty after these tests. The PNA emis-
sions, even when corrected to uniform CO levels, were quite high for
the test fuels and engine deposits:
PNA EMISSIONS WITH OVERCHOKED STARTUP
(1966 Plymouth)
yg./gal. Choke
Test Test Fuel, and Deposit BaP BaA Operation
77 A, stabilized B-fuel 368 368 1/3 stuck
deposit
89 A' + V45, stabilized A-fuel 206 142 1/3 stuck
deposit
89B Same as 89, but preceded by 466 224 2/3 stuck
cyclic use of A fuel to
eliminate any effect of
89
It is not feasible to define precisely the "normal" values for these
tests, i.e., to evaluate precisely the effect of the malfunctions;
furthermore, the malfunctions were probably not all equal. The implica-
tions of comparing 89B (two starts defective) with 89 (one defective) is
260 yg. of BaP and 82 yg. of BaA/gal. as the effect of one incidence of
-------
- 49 -
overchoking. Values this high are not tenable in the other tests, how-
ever. In fact, the observed results for Test 77 are directionally con-
sistent with a similar test (No. 79) in a parallel test series in the
1968 Chevrolet, where no malfunction occurred, so that the impact of
the Test 77 malfunction actually may have been relatively small. In
any case, the malfunctions were relatively brief, so the PNA emissions
during the defective startups must have been extremely high to affect
the final results of 30-mile tests.
4.2 Phenol Emission
Phenol emissions for the third-year program (Table 8) have been
combined with earlier data in new linear regressions of emission (mg./
gallon) vs. % aromatics in the test fuel, with separate regressions for
each of the three test vehicles. The new regressions are plotted in
Figure 2 and are compared in Table 21 with those presented in the Second-
Year Report. The new regressions do not differ significantly from the
earlier ones, but the errors of estimate are now larger and the % of the
variance in the data explained by each regression has decreased slightly.
Study of the raw data suggests a slight decrease in phenol yields for all
three vehicles at higher aromatics levels had occurred in the third year,
but the effect does not correlate with deposit type or any other known
variable. No new data were obtained at the 11-12% aromatic level. The
dominance of fuel aromatics, and the relatively lower phenol emissions
from the 1970 Chevrolet (63 to 66% as much phenol as from the 1966 vehicle,
for 20%-50% fuel aromatics), continue to be the significant conclusions in
this area from the three standard vehicles.
Both of the experimental low-emission vehicles (thermal reactor
and dual catalyst) gave extremely low phenol emissions of 2 mg./gallon
when tested on 46%-aromatic fuels. These emissions are on the order of
0.3% to 0.5% of the emissions from the base case vehicles (1968 and 1970
Chevrolets) using the same fuels, so that phenol reductions greater
than 99% were achieved by both systems. These results are also shown in
Figure 2.
-------
- 50 -
10 20 30 40 50
% AROMATICS IN FUEL
Figure 2
EFFECTS OF FUEL AROMATICS AND EMISSION
CONTROL SYSTEMS ON PHENOL EMISSION
-------
- 51 -
i-H
CM
W
j_J
a
co
O
H
O
P^
<
Ed
H
H
^
co
00 O
VO H
1 H
T J
W W
p - p^
5 2
u o
o
0
CO
CO
H
1
O
&
Ed
j"r!
PJ
CO
O
4-1
O
M
CO
~
+
CO
II
,_)
CO
00
*
00
g
*
6
o
II 1
c
o
H
4-J
cfl
01
^_l
o
u
01
CJ
cfl
<
1
CN
cfl
4-1
Cfl
Q
cfl
4-1
Cfl
0
0
^f
i-H
1
Csl
to
rH
1
rH
CT> vO
oo o
Oi
rH en
o o
i-H
m
CTi
o
Q^
Q^
o
o
CM
**
00
o
CM
m
a\
oo
-a-
00
ON
0)
0)
o
oo
vO
cfl cfl
0)
en
CO cfl
01 4.)
tH CO
I Q
CM
vO
CM
en
in
00
in r»~
-* CD
oo
m
m o
VD
o
CTi
o
rH
CM
oo
CM
en
O
oo
t-t
CO
OJ
1
en
r<
co
01
1
CM
cO
CO
a
cfl
4-1
CO
Q
en
O
"f
o
CM
00
1
m
^^
rH
CM
r*^
rH
CM
CJX
O
en
O
rH
CO
O"N
o
00
o*\
o
00
vD
vO
en
en
m
CM
r^
oo
CM
m
CJ\
oo
en
/ s
CO
*~s
\o
iH
%
4-1
CO
C
cS
rH
O
M
01
(X
o
rH
CO
01
o
1 1
CO
a
H
M
O
l-l
01
*&
4-1
co
o
MH
14H
01
O
O
a
0
H
4-1
CO
rH
a>
l-l
- M
o
u
en
m
rH
Cfl
00
00
6
M
4->
co
a)
I4H
o
^
o
^J
M
a)
*
4J
to
m
t->
00
01
04
o
4-1
0)
3
TJ
(-1
cfl
^
B-S
vO
rH
co
i-i
H
cfl
CM
cfl
4-1
Cfl
-a
14_|
o
*
o
z
CO
'O
-------
- 52 -
5. DISCUSSION OF EMISSION RESULTS
The results reported in the foregoing section provide new
information in three general areas:
1) the control of PNA and phenol emissions by prototype
very-low-emission vehicles;
2) the importance of four-ring and five-ring aromatics
(PNA), already present in the fuel as a source of
engine deposits which cause relatively high PNA
emissions and also as a source of PNA in used engine
oil; and
3) the apparently much lower effect on PNA emissions
for the presence in fuel of two-ring and three-ring
aromatics from high boiling reformate or catalytic
naphtha fractions.
5.1 Control-System Effects on PNA Emission
Our previous reports (1,2) and publication (la) have reviewed the
data from CAPE-6-68 and from others showing reductions of about 70-80% in
the emission of PNA from vehicles (1968, 1970) having engine-modification
emission-control systems, relative to earlier uncontrolled vehicles. Our
new data continue to support this level of PNA reduction by these systems.
The reductions by EM systems are now shown (Table 11) to occur about
equally on a relative basis for PNA emissions from both unstabilized
deposits (freshly-formed) and stabilized deposits (subjected to prior
cyclic-test accelerations), and for deposits from both high-PNA and low-
PNA fuels. This uniformity of control is a highly significant and es-
sential extension of the earlier CAPE-6-68 conclusions, which were based
only on stabilized-deposit data, because PNA emissions from unstabilized
deposits in 30-mile tests were typically three to four times as great as
from stabilized deposits. Since normal driving cycles must include at
least some emissions from unstabilized deposits, it is essential to know
that current control systems are also effective for this type of emission.
Similarly, it is valuable to recognize that the controls are effective with
different levels of fuel-derived PNA present in deposits, since marketed
gasolines have a wide range of PNA contents.
The data obtained in the dual-catalyst and thermal-reactor test
vehicles indicate the probability that the introduction of either type
of system for meeting future very-low-emission standards can also achieve
very low PNA emission.
-------
- 53 -
For the thermal-reactor (RAM) vehicle, our PNA emissions
(1.6 yg. BaP and 6 yg. BaA/gallon of fuel) are somewhat lower than
those reported by Hoffman, et. al. (9) for a different reactor system.
Hoffman's results on two vehicles (which differed as to BaA emission),
when converted to the same units as we have used, are estimated to be
around 8 yg. of BaP and either 40 or 130 yg. of BaA/gallon. The differ-
ence in PNA emissions of the RAM vehicle and of the system used by Hoffman
may be due to warmup differences. Both start-up and later operation are
necessarily very rich (high CO) with such systems, so that there is likely
to be a very large amount of PNA emitted from the engine cylinders. This
PNA must be destroyed in the reactor, where rapid development of high
temperatures is essential unless the system is to be, on balance, a high
PNA-emitter. The same consideration and differences apply to CO emission.
Using the 7-mode cyclic procedure, the RAM vehicle gave a 0.07% CO result,
which can be calculated (1970 procedure) as equivalent to about 1.7 grams/
mile. (The newer CVS procedures give materially different results.)
Hoffman's two 1970 reactor-equipped vehicles were reported to emit 7 and
8 grams of CO/mile, which also suggests that slower warmup may have oc-
curred with these vehicles than with the RAM vehicle.
The dual-catalyst system used by CAPE-6-68 was even more effective
than the RAM in eliminating PNA, giving emissions of 1.1 yg. BaP and 1.5 yg.
BaA/gallon of fuel. Padrta, et. al. (10) have reported tests with a pro-
totype catalytic-reactor system on a laboratory engine in which over 90%
elimination of all PNA species was achieved, with the PNA removal often
being greater than the removal of other hydrocarbons. They estimated that
an automobile installation would give virtually complete removal of PNA.
Griffing (11), however, reported on PNA removal in two catalyst-equipped
vehicles where PNA reductions of only 69% and 34% were achieved by the
catalysts in hot-cycle tests, and drew the conclusion that PNA reduction
by catalysts may be less than HC reduction. We note, however, that
Griffing's hot-cycle HC emissions, with the catalysts, were 20 and 79 ppm,
while our complete (cold and hot) emission results were only 12 ppm HC.
This raises a question whether the Griffing catalysts were sufficiently
active to give adequate evaluations of the potentialities for PNA removal
in catalytic systems. In any case, one of the Griffing catalyst systems
was reported as emitting 50% more PNA in hot cycles than in cold cycles.
This result is most unexpected and raises some question of the validity
of the Griffing measurements.
-------
- 54 -
5.2 Effects of PNA Present in Fuel
Several reports have been made, by a variety of investigators,
on the effects of PNA present in fuels on the emission of PNA. Hoffman (9)
found an 87% decrease in BaP emission from a fuel which had been distilled
to remove the initial content of 0.7 ppm BaP (and other PNA). Begeman (5)
reported that Indolene, with 4 ppm BaP, gave four or five times as much PNA
emission as a regular-grade gasoline with 1.1 ppm BaP. Begeman (12) also
reported the partial (0.1% to 0.2%) survival and emission of radiotracer
BaP that had been added to fuel. Other radioactive PNA species were also
emitted, while about 5% of the BaP in the fuel that was used was found to
have accumulated in the engine oil, along with other radioactive PNA.
Griffing (13) found evidence of a carryover effect of the PNA present in
Indolene fuel to a subsequent low-PNA fuel test, causing erratic results.
Thereafter, care was taken to avoid this problem by conditioning the test
vehicles on low-PNA fuels, and the evaluations of other fuel variables,
such as aromatic content or lead, were done on low-PNA fuels exclusively.
An earlier single-cylinder study by Griffing (14) had shown increased BaP
emission from a high-PNA fuel at several mixture ratios.
In our CAPE-6-68 studies of the directly-observed effects of
fuel PNA (3 ppm BaP vs. "0"), we have used isolated, relatively short
tests and reported (1, la) only nonsignificant (8%) increases in BaP and
small (32%) increases in BaA. We continue to find (Table 15) this type
of effect for similar tests, but two sequential high-PNA fuel tests, or
tests with unleaded deposits present, showed much larger increases in PNA
emission from the use of a high-PNA fuel. We have also presented com-
parisons, in Section 4, of certain sets of our data which show evidence
of a PNA-carryover effect and also of the gradual depletion of high-PNA
deposits toward a condition of low potential for PNA emission. The
extreme examples of this depletion are the observed large decreases in
emissions between unstabilized and stabilized deposits from high-PNA
fuels.
These experiences in CAPE-6-68 point up the value of the
precautions taken by Griffing to avoid PNA carryover and PNA deposits derived
from fuel PNA. The Griffing precautions, if taken by CAPE-6-68, would
probably have improved the design of our experiments and simplified the
interpretation of our data. However, they might also have resulted in the
failure to detect the dominant, deposit-related part played by fuel PNA
content in determining the exhaust emission of PNA. In view of the short
tests used, in which fuel PNA generally did not become fully effective,
-------
- 55 -
the effect of fuel PNA would continue to be estimated at the very modest
levels which were initially reported. Instead, we now find that, depending
on deposit conditions at the time of testing, the presence of PNA (at 3
ppm BaP) in the fuel used for the deposits can cause from 2 to 9 times as
much PNA emission as would occur from deposits formed from a "zero"-PNA
fuel. This effect can easily obscure the effects of other variables, such
as fuel aromaticity, which have received the major attention in most pub-
lished studies of PNA emission from vehicles or PNA formation in sta-
tionary flames and pyrolysis tubes.
It appears, in fact, that fuel aromaticity is practically
irrelevant unless fuel PNA is minimized. If fuel PNA were minimized,
together with the gradual shift to emission-controlled vehicles, the entire
level of automotive PNA emission would be so drastically reduced that
aromaticity effects would have little or no practical significance for the
atmospheric PNA burden.
The accumulation of PNA in used engine oil is an additional
effect of the presence of PNA in fuel. Our results (Table 16) show that
the use of fuels with 3 ppm BaP for 2000 miles introduced 12-23 ppm (16
ppm average) BaP in the used oil. Approximating the oil charge at 9 Ibs.
and the fuel consumed at 900 Ibs., we find that 5.3% of the BaP in the
fuel has apparently remained in the oil. This is in excellent agreement
with the report by Begeman (12) on the use of a fuel with added radio-
tracer BaP, that "slightly more than 5% of the total amount in the gasoline
accumulated in the crankcase oil."
The significance of PNA accumulated in engine oil is not clear
at this time. All of the PNA-emission tests by CAPE-6-68 up to the present
time have used a fresh oil charge. Future work will evaluate this practice
in comparison with leaving the used oil in the engine for the test, and
will include cases where the used oil is both high and low in PNA content.
The gradual accumulation of PNA in extended oil use (up to 8000 miles) will
also be monitored.
The occupational health-hazards related to used engine oils and
the disposal practices for used oils appear worthy of examination. The
disposal question has been raised recently by Moran (15) in the context
of oil incineration.
-------
- 56 -
5.3 Effects of Heavy Reformats and Catalytic Naphtha Fractions
Two-ring and three-ring aromatics (naphthalenes, tetralins, indans,
anthracene, phenanthrene, etc.) were evaluated by additions to a low-PNA
base fuel in two different ways:
(1) The catalytic reformate vacuum distillate fractions (1.5% of
the feed) having atmospheric-pressure vapor temperatures from
about 384°F to 575°F were added at 0.7 wt.%- The resulting
blend simulated a gasoline made with about 50% of a distilled
high-octane catalytic reformate boiling up to 575°F, but
containing little or no four-ring and five-ring aromatics.
(2) A heavy catalytic naphtha with an end-point of 442°F (ASTM
D-86 distillation) was added at 20% by volume. The choice of
naphtha end-point matched the practice of some refiners during
periods of lowered heating-oil demand; with higher demand, the
naphtha end-point (and C^g-C^ aromatics content) may be
decreased. The use of 20% of the naphtha for blending reflected
both the need to maintain a reasonably normal volatility for the
blend, 80% of which was an existing low-PNA gasoline, and the
fact that marketed gasolines (both premium and regular) may
contain from 0 to at least as high as 40% heavy catalytic
naphtha. Practical limits are imposed by both octane quality
and the legally-specified (in some states) maximum end-point
(by D-86) of 437°F.
Adding the reformate-fraction introduced essentially pure
aromatics, while the naphtha addition introduced a significantly larger
amount of C1f.-C , aromatics, along with other hydrocarbons such as olefins,
indans, and tetralins and also a small amount of four-ring and five-ring
(C1,+) aromatics (PNA). The presence of PNA in heavy catalytic naphthas
can be expected to vary with distillation conditions, such as entrainment
of heavy material at high rates of throughput. The ASTM D-86 end-point is
not in general a useful guide to PNA content. As pointed out in Section 3,
finished gasolines of nearly equal end-points may vary widely in PNA
content. Hoffman (9) reported analyses of two heavy catalytic naphthas
which contained no BaP or BaA; end-points were not given. In CAPE-6-68
we found (Section 3) that two heavy naphthas differing by about 20° in
end-point had about 30-fold to 50-fold differences in BaP and BaA
contents, but with the PNA in the higher-boiling naphtha still only about
10% of the level found in various high-PNA gasolines blended from cata-
lytic reformate.
-------
- 57 -
Both methods of addition of two-ring and three-ring aromatics
(as reformate fractions and as a heavy catalytic naphtha) were found to
have very similar effects on PNA emission. When tested over existing low-
PNA engine deposits, both blends gave small increases in BaP and/or BaA
emissions. These results suggest that two-ring and three-ring aromatics
provide relatively potent starting materials for the synthesis of PNA in an
engine. By comparison, much larger increases in single-ring aromatics are
required to produce comparable increases in PNA emission. The two blends,
when tested over their own, unstabilized (and, hence, of maximum PNA-emission
potential) deposits from 2000 miles of operation gave about the same
emissions as were seen in the tests from low-PNA deposits, i.e., the ac-
cumulation of deposits from these fuel blends did not cause further in-
creases in PNA emission. This is in sharp contrast to the cases in which
the deposit-formation fuels were high in PNA content (four-ring and five-
ring C..,+ aromatics) and much higher PNA emissions were obtained from the
unstabilized deposits which they formed.
It is appropriate to discuss briefly the tests reported earlier
(1) on the effects of a different high-boiling naphtha on PNA emission.
This naphtha (heavy catalytic naphtha + polymer gasoline + C^'s), when
used at 16% in a low-aromatic fuel, appeared to cause 40% to 109% more
PNA emission than occurred with a similar fuel without the naphtha.
This effect was observed only in the 1966 and 1968 vehicles and was
actually reversed in the 1970 vehicle. In all cases, the deposits pre-
sent had been previously formed from a leaded high-PNA fuel. In Section
3.1 of this report, we have indicated that this naphtha contributed no
PNA and only a little C^o~ci2 aromatics; its end-point (D-86) was only
419°F. In view of our later results, it is questionable whether it would
be likely to have had, at 16%, an increased-emission effect similar to
that observed for the 442°F end-point naphtha at 20% which contributed
more high-boiling aromatics.
Examination of the sequential test histories for the 1966 and
1968 vehicles shows that the with-naphtha (higher-PNA) tests preceded the
without-naphtha (lower-PNA) tests, with several intervening tests all
over the same deposits. We now conclude that the apparent effects were
due in part to gradual declines in the deposits with successive tests.
The reversed effect in the 1970 vehicle also appears to reflect the
sequence of testing. We thus conclude that the previously-reported
high-boiling naphtha effect may have been, to at least some extent, a
spurious effect, even though it was consistent in two of three vehicles
with the effect seen later for the higher-boiling heavy catalytic naphtha
that contained significantly more heavy aromatics.
-------
- 58 -
5.4 Phenol Emission
The data obtained from the experimental low-emission vehicles
indicate that phenol emissions should be essentially eliminated in
vehicles which achieve the very low emissions of CO, HC, and NO required
in the 1975-76 period. However, existing emission-controlled vehicles
(1968 and 1970) continue to show relatively small changes relative to
the uncontrolled vehicle (1966) used as a base case (Figure 2, Section 4.2),
The similarity of the phenol emissions from the 1966 and 1968
vehicles, with lower emissions from the 1970 vehicle, may be related to
the relatively higher exhaust temperature and/or oxygen content of the
exhaust gas from the 1970 vehicle. The temperatures of the 1966- and
1968-vehicle exhausts were more nearly equal. If this distinction is
relevant, it implies a fairly uniform phenol production between the
three vehicles, but greater destruction in the exhaust system before
emission from the 1970 vehicle. The two experimental low-emission vehicles
(thermal reactor and catalytic) present the extreme cases of this suggested
post-engine oxidation.
5.5 Hydrocarbon (NDIR) Emissions of Vehicles
A review was made of the histories of the three test vehicles
with respect to HC(IR) emission over the three project years. First-year
and second-year distinctions due to test-fuel aromatic content (affecting
instrument response) and to deposit type were less clear in the third-
year data, partly because of operation in a narrower range of fuel
aromaticity. The previously-reported higher emissions of HC from leaded
deposits than from unleaded deposits were no longer evident. However, all
emissions of HC in the third year appear to have decreased gradually,
and to be incorrect to some extent, because of an undetected progressive
change in the shape of the HC(IR) instrument's calibration curve. (Periodic
standard samples and span gases weie, generally, at the wrong HC levels to
detect the change.) Replacement of the IR detector cell and recalibration
provided somewhat higher HC emission results in the last three tests of
the report period than in recently-preceding tests. However, even after
the instrument change, the level remains lower than was common in the
first two years.
No correlation of NDIR HC emissions with PNA emissions was found.
In fact, an NDIR instrument (hexane-sensitized) responds less to aromatic
than paraffinic hydrocarbons, while PNA emissions increase with aromaticity.
Thus, no simple correlation could be expected.
-------
- 59 -
6. PNA SAMPLE COMPOSITION STUDIES
6.1 Compositions of Samples of Different Origins
The ratios between various PNA species in samples of different
origin have been widely used to characterize the samples, i.e., to develop
a "profile" of the sample that is presumably related to its origin.
Commins' ' has reported ratio differences for pyrolyses at different tem-
peratures and fuel/air ratios. Oro, et al.^ ', have reported ratio differ-
ences with temperature in pyrolyzing isoprene, while Cleary^ ' has made a
similar report on the off-gases from a brick kiln. Badger, et al. ', re-
ported compositional changes with temperature in the pyrolysis of n-butyl-
benzene, along with optimum temperatures for the formation of different PNA
species.
These papers relate, implicitly, to differences in formation, or
synthesis, of PNA with temperature, etc. An inseparable aspect of such studies,
however, is the question of relative reactivities of different PNA species for
reactions such as oxidation, i.e., the relative extents of survival down to
the final analysis, once the individual species are synthesized. (Or, as may
be the case in some CAPE-6-68 samples, once the species are introduced into a
system (engine), from which they may be emitted at some later time.) Tipson^ '
has reviewed the oxidation of PNA, including molecular-orbital calculations of
the relative reactivities of several PNA species that are regularly determined
in CAPE-6-68 analyses. (Laboratory oxidation studies cited by Tipson support
the molecular-orbital predictions.) Tipson's relative reactivities are com-
pared with Badger's optimum-formation (from butylbenzene) temperatures for
several CAPE-6-68 species in the following table.
FORMATION TEMPERATURES AND REACTIVITIES
FOR SELECTED PNA SPECIES
Badger's Optimum Tipson's Relative
. . Formation Reactivity, in
Compound^ Temp., °C Arbitrary Units
Benzo(a)pyrene 710 12
Benz(a)anthracene 660 5
Pyrene 740 3
Phenanthrene 660 1
Benzo(e)pyrene 710
-------
- 60 -
It is clear that formation temperatures and reactivities do not correlate.
For example, the pairs BaA and chrysene, and BaP and BeP, include widely
different reactivities but equal optimum-formation temperatures. Either
property might reasonably have some influence on the amount of a given PNA
species that is found in an exhaust sample, since both synthesis of PNA from
smaller molecules and survival of fuel PNA are involved. Of the two pro-
perties, survival rate, or relative reactivity, appears more likely to
govern for exhaust samples since the PNA must exit through the hot vehicle
exhaust system, with at least some (1% or more) oxygen usually present even
before discharge to the air or collection in a cooler-filter unit. If this
presumption is correct, we might expect to find progressive changes, between
vehicles having different exhaust conditions, in the ratios of, say, the re-
active BaP and BaA to their respective, less-reactive isomers, BeP and chry-
sene. Even larger differences might be expected between the species ratios
for fuel samples and those for exhaust samples.
It is necessary to keep in mind, however, a rather severe limita-
tion on all studies of PNA sample composition, or "profiles". The findings
of the studies will inevitably reflect any losses of different species in
sampling, or differences in analytical accuracy, that may occur after the emis-
sion or formation of the PNA sample.
6.2 CAPE-6-68 PNA Sample Compositions
The CAPE-6-68 exhaust samples were collected, with rapid cooling,
in a large cooler-filter unit. Washings of unit surfaces and extracts of
aqueous condensate and of filter media were combined and known amounts of
radiotracer BaP and BaA were added before boildown for analysis. The GC
peaks ultimately trapped and assayed by UV absorbance were scaled-up to ori-
ginal values by tracer counting, using BaP* for BaP and BeP (one GC peak) and
BaA* for BaA, chrysene, and triphenylene (in another GC peak). The assays of
all other peaks were scaled by whichever tracer had the better recovery, i.e.,
they took the lesser of two alternate values. Almost invariably, BaA*, rather
than BaP*, was the tracer used for all species except BaP and BeP.
Two assumptions are implicit: there must be no loss before tracer
addition, and all losses after addition must parallel the tracer losses (which
includes all factors in the GC analysis that might cause differences in the
trapping efficiency or UV-assaying of different GC peaks). In fact, it was
generally found from tracer counting that the fraction of BaA* added that was
recovered in its GC peak averaged about 28% greater than the fraction of BaP*
in its peak. The "correct" choice for scaling other peaks is thus not clear.
Furthermore, experiments reported in Section 7 indicate that there probably
were losses of some of the more reactive PNA species from, at least, the filter-
media portions of the CAPE-6-68 emission samples. However, it will be shown
that the ratios of reactive PNA species to non-reactive species are no more
variable than the ratios between two non-reactive species. From this it ap-
pears that, while losses of reactive species may cause inaccuracies in the ab-
solute PNA emission rates, they should not invalidate conclusions as to the
relative effects of changes in controlled variables on these rates.
-------
- 61 -
The question of PNA losses during sampling with equipment such
as has been used in CAPE-6-68 has been raised by Griffing^11' in a discus-
sion of Reference (la). The suggested injection of BaP* (tracer) directly
into the exhaust during sampling would require the solution of many tech-
nical problems and could, at best, quantify only the emission of BaP. A
rigorous measure of the emission of other PNA would still not be available
because of the many possible variations in the ultimate recovery efficiency
for each species.
6.2.1 Statistical Analysis
In the following presentation, we have analyzed the composition
data from CAPE-6-68 emission tests and fuels in terms of between-species
ratios to detect significant differences that may be related to the source
of the sample. In doing so, we recognize the uncertainty, implied above,
whether any given difference is in fact due to the origin of the sample or,
rather, to selective losses, before final analysis, that may themselves vary
with the origin or later history of the sample.
The data input used for the analysis included:
(1) Eleven-species analyses of emission samples from
all accepted (no mechanical defect) CAPE-6-68 tests from No. 32 through
No. 96 in the three standard vehicles, and similar data for seven Esso-
funded tests in the 1968 Chevrolet. The three vehicles (1966, 1968, 1970)
had, respectively, 21, 24, and 27 tests with usable data. Analysis programs
were used which could accept the existence of numerous missing values, es-
pecially in the lower-numbered tests. The input data appear in Table D-2 of
Reference (1) and in Table 22 of this, report.
(2) Nine-species (invariably lacking chrysene and triphenyl-
ene) analyses on eight fuels from Table 3 of this report and two fuels anal-
yzed ex-contract by Esso Research. Of the fuels listed in Table 3, the first,
tenth, and eleventh were excluded as being directly related to other fuels in
the table.
A separate analysis was done for the data from each vehicle or
from ten fuels, using the following steps:
(1) Ratios of BaP, BeP, and BaA, each to ten other
species, were computed and converted to common
logarithms (base 10) .
(2) A multiple-regression analysis yielded as gen-
eral statistics: the mean log-^Q, the standard
deviation (s) in the logiQ, and the actual number
(n) of input values (reflecting missing values)
for each ratio.
-------
- 62 -
TABLE 22
CAPE-6-68 PNA EMISSION DATA THIRD YEAR
MULTIPLE-SPECIES ANALYSES OF EXHAUST SAMPLES BY THE GC/UV METHOD OF CAPE-12-68
CAPE-6
Test Test
Vehicle
'66 Plym.
(NC)
'68 Chev.
(EM)
'70 Chev.
(EM)
'68 Chev.
(RAM)
'70 Chev.
(Monel/PTX-5)
No.
61
61B
62
68
69
73
76
77
80
81
77B
85
89
89 B
93
96
64
65 (^
71
72
78
79
82
86
90
92
94
66
67
70
74
75
83
84
87
88
91
95
60
63
BaP
712
277
266
246
248
639
323
585
200
308
226
197
349
911
210
197
38
32
16
74
81
343
165
30
54
38
53
28
32
46
40
39
44
96
27
18
24
177
4.5
2.8
BeP
3672
920
1048
1000
822
2405
1020
2292
574
1110
897
624
1172
2659
451
484
164
127
72
289
836
1233
769
98
164
156
193
118
177
219
192
196
474
628
139
78
121
814
82
5.
BghiP
2217
916
1011
754
1545
1254
850
2007
606
1414
942
1075
1623
1886
786
758
240
229
112
349
384
488
1102
239
334
78
189
163
236
277
324
184
245
611
296
155
225
369
47
.3 2.8
Total Test
BaA
1377
328
304
254
360
718
427
583
203
384
455
132
240
437
278
204
50
59
40
216
203
302
171
59
114
113
132
64
110
92
85
135
239
254
74
50
80
413
17
3.9
Chrysene
2998
633
489
536
693
1400
700
1148
350
856
886
369
745
1611
423
395
134
126
87
303
622
701
297
153
282
218
222
131
241
198
146
225
377
473
204
108
181
662
(b)
5.7
Yield. UR (for 36 137-Second Cycles)
Pyrene
4382
3362
2588
2334
7043
4184
2956
6265
1670
4410
3471
2531
8163
18099
5557
5853
615
815
510
1597
1515
953
847
896
1256
1454
1237
466
1330
814
719
1043
1299
1525
688
620
776
2000
117
64
Triphenylene
382
124
101
110
146
215
152
306
93
179
146
90
179
321
100
110
39
37
35
72
79
158
59
42
59
52
49
37
62
53
45
52
75
94
52
38
56
134
(b)
2.2
Me-BaA
983
210
246
203
289
680
324
362
145
329
344
76
59
81
64
76
31
25
13
119
112
371
137
29
31
50
59
31
61
43
36
69
83
103
21
16
24
355
(b)
1.6
DM/Et-BaA
468
55
71
52
44
294
107
113
24.4
72
101
27
23
102
20
26
8.6
8.0
3.3
28.7
2.5
192
51
5.1
8.7
6.9
14
6.2
14
9.5
12
18
20
Lost
3.7
4.7
6.9
96
5.0
0.0
MeBaP
353
63
66
63
72
186
94
155
105
102
61
79
46
96
23
79
12
7.4
4.7
14
0
174
52
6.2
10.5
5.6
15
7.7
8.3
12
7.7
9.9
11
16
5.3
2.9
5.0
62
2.6
0.1
Me-BeP
2004
390
526
442
367
1260
506
1017
397
648
532
351
230
308
80
159
69
51
21
142
493
708
526
24
44
46
69
41
64
72
61
71
86
118
27
14
20
477
10.7
1.5
(a)
(b)
Only 24 cycles were run to conserve fuel, but yields are adjusted to 36-cycle basis.
Methyl- and dimethylpyrenes interfered with determinations of several species.
NOTE: Tables 3 (Fuels) and 16 (Oils) present additional GC/UV PNA analyses.
-------
- 63 -
(3) Antilogs of the mean log^Q values and of
the mean log^o i 2 ^/>/"£ were printed as
geometric mean ratios and 95% confidence
limits on the means, i.e., the upper and
lower limits of the mean ratios.
Table 23 presents the geometric mean ratios and 95% confidence
limits, as defined in (3) above, and Figures 3 through 5 display the ratios
and limits on logarithmic scales.
6.2.2 Reactivity Differences Betweeen PNA Species
In the absence of chrysene and triphenylene data for the fuels,
benzo(e)pyrene (BeP) is the least reactive species on which fuels analyses are
available. As a result, the most convenient way to examine the PNA ratios to
detect reactivity effects is the use of Figure 3, where BeP is common to all
ratios. Inspection of the fuels ratios (BeP to other species) in Figure 3, as
against the three sets of vehicle ratios, shows that, in fuels, the mean ratios
of BeP to BaP, BaA, MeBaA, DM/EtBaA, and MeBaP are all significantly different
from their values in the three vehicles. Conversely, ratios of BeP to BghiP,
pyrene, and MeBeP are the same in fuels as in the vehicle exhausts, except that,
in the 1970 vehicle exhaust, BeP /MeBeP is different from its value in fuels.
(This vehicle's exhaust is the hottest and has the highest oxygen content of
the three.) The implication here is that the severe conditions of passing
through an engine destroys the reactive BaP, BaA and their alkyl derivatives,
relative to the less-reactive BeP, MeBeP, pyrene, and BghiP, so that the ratios
for reactive species in the exhaust differ from those in the fuels. The magni-
tudes of changes in geometric mean BeP ratios (Figure 3) from "fuels" to
"1970 vehicle" were calculated as the differences between the mean logs of the
ratios. These differences were then ranked as measures of reactivity, with the
result that the nine available species were ranked in decreasing order of reac-
tivity as follows :
Rel. Reactivity
Decreasing Reactivity (Tipson)
MeBaP
DM/EtBaA
MeBaA
BaP 12
BaA 5
MeBeP
Pyrene 3
BghiP
BeP
The order derived in this way from the CAPE-6 data exactly matches the order
cited by Tipson for four of the species, as calculated from molecular-orbital
theory. The alkyl derivatives, as expected from the chemistry of simple aro-
matic hydrocarbons, are found to be more reactive than the parent compounds.
-------
- 64 -
S
a
H
CO
H
g
1
0
Q
CO
H
iJ
CO
H
CO
5
s
U
B
M
CO
U
H
M
PH
CO
PH
W
pa
t/3
p,
3
/^
i
s
01
5
:e) limits
-confidenc
^8
in
ON
s-^
n
cu
o
g
l-l
01
p.
§
a
g
to
5
2
(Stated s
i
(I)
H
U
H
18
-a-
H
4-1
to
rH
4J
CO
g
S,
lerived fi
lalyses.
" 5
CO
H H
tu
d 3
CO
«H
O P.
oi B
O to
are inclx
emission s
ct
CO
CO
4-J
CO
fi
S
o
h
<«
CO
01
to
£
CO
^
PH
g
Xt-v
a°.
P<0
S"'
£; <~«
CO
a cu
8|
CO CO
01 CO
5o
H H
O v--
to
o
/-\
4J
O to
i*-.
/-N
VO 4J
o\ to
H 01
g:
^1 CM
IW '
Samples
Plymouth
n
0)
S
£1
§
01
£
iJ
corner* vo mcoo** VDO\CM vo vomoovo P^HO P^- p^^a-p^m
CTtOOOO OO OOCOHCO OO>in VOCT*P^ CMCOH in inHp^vo cMP^eyi vo vo^-a-vo
cocom -^ OCTVP^O vomcM co p^mcoco P^HCO co voooinco
OOOOO CM CftvOCTtCTt ONCOOO CM OOvDCTtOO HCMO CM OHsTCM
CO CM ^ -a1 P^ P^ VD CO p^ CO 00 o^ p^ vo H CO P-i O vO VO P^ vo H VO
OvDVO in CT»CT>CMCJ\ COvOO m OOvOOH HinO (T> OvP^vDCM
ON VD vO H VO ON vO in 00 vO P^ H VO ON ^ *^ ON O 00 H 00 "^ VD VO
OOO O OHrHH OOO O OHHH OHO O OtNHH
P^ON^HCOCOOP^OOP^ OP^ONHinOOOCOOCO HcslvOP^COOOOOOOOP^
OOOOOHH-a--3-O vOHCMHOmvOvOP~-* CMOOOOCMCMOCMH
CM CM H H
cocMcofMvoO iH *^ rH 00 lO O> CN \O O*l r^* t*1* lO O^> O "1 CN O
CNlH-tfrHOvOr^-OO** -^-r^vOP^HvOCOOOC^O vOCOCO-^-O-3'iniOCMOO
OOOOOOOfOfOO POOrHOOCslCOt-tcNCNj rHOOOOrHi-tmiTto
f-H i-H
mcniofOcocNjvrcMc^rH cniriooootNcMinoovo ^p^-^-cOi-Hinvo-^i^o
ooooocM«-iu-i O\ rH rH vO vO r** «tf CO &\ ^O rH r^- CO O IO vO rH O ^ O^ CO i/^ rH VO t^1* OO 00 CN
OOOOOrHrH^TCOO cOOCMrHO^O^tini-HCM rHOOOOCsJ'Hm^O
rH rH
eSCN\OCOOu">^COr--*.tf' OOr^-OOrHCO\OCOtOOO rHCOfO^OO^fNCNlrHin
OOOOOrHOCOcMO CMO, a) o. tj w
01
g
H
0) >>
tu tu oi a
PH CO PX!<4P3P4PH
js CiH-ripa-ipapa
-------
- 65 -
<
z.
a-co
LJ *!;
I- O
O CO
O (Tj
CO LU
, Qi CO
SJ >LU
5, °- -1
- T °-
LL 3 ^
o <
Nl CO
CD <"
LL <
O 1
CO LU
o
-
-I
f.
-
-
If!
O
0
on
NO NO rv
ON O ON *
"" " LL
Q-
re
CD
CL
CU
CD
JT
in
NOgOJ.
0- ON 0- «
"" LL
Q.
^p
CD*
Q.
>
o
Q-
0)
CD
ll'l
NO COO V,
ON ON ON *
u_
CL"
Q^
CL)
CD
y
o oo o
o *o t*««
O* O 0
1
cu
r"
a.
f±
CL"
O)
CD
Hi
NO CO O V)
0-0- 0-
"~ LL
re
CD
CD
CL
(L)
CD
1
I
o oo o w
o* o- o *
u_
1
UJ
'
Q
a.
CD
I
I1
L
I
NO CO O «
ON 0. ON *
" '"'" LL
CL
re
CD
V
Q^
01
CD
"t
-
-
NO CO O vi
NO NO t^N ~~
ON ON O *
" LL
Q.
CU
CD
S
o_
cu
CD
i-l
O
O
O
o
CM
0
! |
0
xf
O
o
r-t
O
*
O
o
o
o
CM
O
0
i l
O
o
V)
o
re
o;
c
re
cu
o
cu
cu
o
o
O
o
-------
- 66 -
^co
LJ 2
O CO
Si >{S
j*1 O EL
u. Ss
N c/)
^
UJ 1-
GO CO
^
CO LU
O1^
p~ 9
\ *^-
<~
C£
^
T
1
[[
*
M
O
CJ
cc
CN CN CN
sO OO O »
^ so JV.
u_
DL
CD
"X^
Q.
re
CO
T
1
'n
CN CN CN .
sO CO O «
'O sQ ^. "T
0.0-0.
u.
a.
CJ)
CO
c^
re
GO
II
CN CN CN
sO 00 O V>
sO >O 1^-
O Os O. *
*~ li-
re
GO
Q^
re
GO
I
II
CN CN CN
sO OO O
so so rv
>-,
^=
S,,^
0^
re
GO
T
liT
M-
|
O OO O W
o .0 r>v -T
o. o. o.
u.
41
C
O)
^
Q.
Q^
nj
m
j
^ CO O
o. o. o.
CD
O)
"^,
C
c~
Q.
K
Q^
co
CO
f
HlT
«O CO O W
Os O CK »
U.
^f
(0
00
0)
^
Q^
(0
OQ
Hi,
i
^ 09 O w
O" O. O« *
u_
^
re
GO
UJ
^^
^
O
^~
re
GO
H1
I
O 00 O M
o ^ r*.
O. O. O.
U_
Q.
00
*
u.
CL
(U
GO
0)
X^
oT
re
GO
v_/
1-1
o
o
o
0
CM
O
O
i-i
CD
0 |
CM' |
_ re
O (U
i i ^
u
'»_
.W
-------
<
Q.
o: oo
LJ LJ
X Z
0^
°<
O
LJ
IT) ^<
Si £LU
= F_i
.2> ^0.
LL ^^
^x <;
.5 CO
Ml-
^ CO
T-
° LJ
CO -^
P
Ct:
-
T|
1 1
1
M
i
-
_
m
O
O
o:
^- ^ t^o
\O 00 O w
$££:li
*~ u_
Q.
\
re
GQ
I
,1.
o co rx o
-o oo o *
£$£"
" U-
0.
o- o- o- 5
*~ u_
o- o
CD
r*
O 00 0 W
cSo^S:"*
LL
CL
re
GQ
0)
^
re
GQ
Tn
1
_. 1
1 1
Tl
F:
"» 00 IX 0
SO 00 O M
0^0-0-
u.
CL
03
(U
^
re
CO
O
o
1 1
q
o
o
o
CM
o
o
I 1
o
*
CO O
cJ "re
C£
0 g
«^
0.4 1
f Geometric l\
o
u>
CVJ -J
o ;§
O)
_, o
s
0 -S
VI
0
o
q §
0 0s
CM
O
l-H
O
o
-------
- 68 -
It is clear from Figures 3 through 5 that there is much less
between-vehicle difference in the value of any given mean species ratio than
exists between the value of a ratio for fuels and its value for the three
vehicles as a group. The chemical "severities" of the three exhausts are
roughly equal, but differ substantially from the conditions for the synthesis
of PNA found in fuels. The figures do indicate however, that some ratios
changed significantly between vehicles, while many others changed by less-than
significant amounts. It is reasonable to assume that the maximum difference
in exhaust severity, or maximum ratio changes, will occur between the 1966 and
1970 vehicles. Ranking of the magnitudes and directions of these "1966 vs.
1970" ratio changes in Figures 3 through 5 permits additional estimates of
the relative reactivities of the different PNA species, as was done above for
the "fuels vs. vehicles" comparisons.
These new estimates are listed, with the previous estimate, in the
following table.
RELATIVE REACTIVITIES OF PNA
(Decreasing Reactivity)
Basis for Estimates
Most Reactive
Fuels vs. 1970
Vehicle, Fig. 3
. (Be?)
MeBaP
DM/EtBaA
MeBaA
BaP
BaA
t MeBeP ,
Pyrene
BghiP
(BeP)
1966 vs. 1970 Vehicles
Least Reactive
Fig. 3
(BeP) .
MeBaP
BaP ,
Pyrene
MeBeP
BghiP
DM/EtBaA
MeBaA
Chrysene
(BeP)
BaA
Triphen.
Fig. 4
(BaP)
MeBaP
(BaP)
MeBeP
Pyrene
DM/EtBaA
BghiP
MeBaA
BeP '
Chrysene
BaA
Triphen.
Fig. 5
(BaA)
MeBaP
MeBeP
BaP
DM/EtBaA
Pyrene
> MeBaA ,
BghiP
BeP
Chrysene
(BaA)
Triphen.
The three rankings based on between-vehicle changes are very similar and would
be identical if no species results had been missing in the original input data.
Only the species beyond the horizontal lines reacted to a significantly greater
(or less) extent than the reference compound (in parentheses) in moving from
-------
- 69 -
the "severity" of the 1966 exhaust to that of the 1970 exhaust. The overall
trend is as expected: BaP and MeBaP are most reactive and triphenylene has the
least reactivity. Compared with the theoretical predictions and with the
fuel-vehicle ranking, there are also some unexpected changes, however. BaA
now seems to be much less reactive than BaP and is similar to chrysene and
triphenylene; the alkyl derivatives of BaA also show decreased reactivities.
Conversely, pyrene, BghiP, and MeBeP show increased reactivities.
The meaning of these changed rankings is not clear. One implication
might be that an emission-controlled vehicle (1970, e.g.,) would control BaA
emission somewhat less well than BaP emission relative to an uncontrolled
vehicle. The data in Table 11 suggest a slight trend in this direction, but
the difference is not large. Relative to the 1966 vehicle, the 1970 vehicle
reduced BaP by 79-84% and reduced BaA by 54-76%, but with most of the BaA
reductions at 72-76%. Along the same line, comparisons of species ratios for
the experimental thermal-reactor vehicle (RAM) with the mean ratios for the
1968 Chevrolet show that the RAM apparently controlled the emissions of BeP
and BaA a little less effectively than those of BaP, etc. In the overall
view, however, the emission control systems (both current EM and experimental)
reduced all PNA species so much, relative to the uncontrolled vehicle, that
any fine-scale variations in reduction are not really significant.
This study of PNA ratios has a broader implication than relative re-
activities , in any case. The values of each ratio for exhaust samples, from
all three vehicles and many tests, were relatively constant, while differing
significantly from their values in fuels. This suggests that different sources
(coke ovens, coal furnaces, etc.) may give different characteristic PNA ratios.
This suggests, in turn, that it might be possible to define the source of air-
borne PNA in a given location by comparison of its ratios with those from
likely sources in the area. This possibility has already been suggested by
Commins^"'. Valid results would, of course, require maximum precautions
against losses of certain species during sample collection. This has already
been noted as an apparent problem in the CAPE-6-68 work, for example.
6.3 Ratios Between PNA Species vs. Sampling Losses
Throughout Project CAPE-6-68, the effects of controlled variables
have been measured by differences in measured emissions of BaP and BaA. We
now see that BaP and BaA are among the most reactive of all the PNA species
for which analyses are available. (Only BaP and BaA analyses were available
on early tests.) Also, in Section 7, experiments are reported that indicate
that losses of BaP and BaA, in particular, were very likely to have occurred
during collection of CAPE-6 samples, apparently because of their reactivities.
On the other hand, tracer BaP and BaA were used in sample analysis, so that
any losses of BaP and BaA after tracer addition would be taken into account
more accurately than the losses of other species.
-------
- 70 -
It is thus appropriate to inquire whether, because of reactivity
and losses, BaP and BaP analyses are more variable than the analyses of
less-reactive species. For this purpose, a variety of PNA species ratios
were selected which included species with reactivities (in the ranking given
by Tipson^7') ranging from "both high" to "both very low". For each vehicle,
computer analysis gave the number of values and the standard deviation in the
determination of a single value of the login of each selected ratio. These
standard deviations were then pooled over the three vehicles to provide an
overall measure of variability in the determination of each ratio. The
following table summarizes the results obtained.
VARIABILITY IN PNA RATIOS
No. of
Standard Dev. Ratios
Ratio in Log..n of Ratio Reactivities in Pool
BaP/BaA 0.1548 Both High 72
BaP/Chrysene 0.1647 High/Very Low 72
BeP/MeBaP 0.1476 Low/Very High 48
BaA/BeP 0.1773 High/Low 70
Chrysene/BeP 0.1801 Very Low/Low 70
BaP/MeBaP 0.1346 Both High 48
BeP/MeBeP 0.1602 Both Low 49
ChrysYTriphenylene 0.1490 Both Very Low 61
There is no evident tendency for higher variability in ratios with mixed
reactivities. This implies that losses, or other sources of inaccuracy in
the exhaust emission results, are no more of a problem for BaP and BaA than
for the other, less-reactive species, in the analyses. It follows that the
Project's conclusions as to the effects of controlled variables are no less
valid because they were based on the measured emissions of two of the most
reactive species.
6.4 Repeatability of Analyses (Quality Control)
The repeatability of analyses of PNA in a given sample has been
routinely monitored at Esso Research. The results obtained are clearly
relevant to the questions of variability and accuracy discussed above, but
also depend on the stability of stored samples. Table 24 summarizes re-
sults obtained over an 18-month period on samples of the supernatant liquid
above solids removed from a heat-exchanger that had been in extended use to
-------
- 71 -
CO
CM
cu
rH
3
H
CA
O
O
00
CO
CO
cu
CO
CO CO
s L
o -u
r-l -H
H-l 10
o
13 D.
a> 8
> O
Q O
a> o
a) co
§3
M.C
S-fi
O CO
5
M
CO
O
CA
g
O
co
O
m
00
a
CO
4-1
r^
vO
m
**
co
CM
rH
O
^
cu
rH
cx
CO
CN
CM
rH
*^^
O
rH
CM
**^
VO
rH
£
CM
rH
^^
CM
rH
CM
^^.
O
rH
rH
O
rH
^^
VO
rH
ON
tH
^^
m
rH
r-.
in
*^
^
rH
a
$
cu
CO
Q
r-~ co
ON sr
T** rH
VO rH
ON CO
O CM
f** rH
VO rH
ON ON
-* CM
CO rH
vO rH
CM rH
^3 1^
CO O
VO rH
rH ON
00 H
00 CM
vO rH
ON oo
sr si-
ON rH
VO rH
00 00
O 0
ON CM
VO rH
0)
a
CU
M
oo
CM
CM
rH
oo
VO
rH
rH
CU
g
to
r?
6
CM
r^
CM
CM
m
CM
CO
vO
CM
m
CM
00
m
CM
^
CO
CM
VO
VO
CM
0)
Pi
cu
rH
^
g
*C
a
H
H
o
o
vO
r^
si-
m
CO
CO
m
vO
rH
m
vO
O
vO
rH
CM
VO
^
a\
m
^i
CO
PQ
rH
^
4J
^
vO
CO
rH
f»-
O
rH
vO
rH
rH
vO
CM
rH
-*
rH
m
CM
rH
O
CM
CM
^J
CO
PQ
4-1
w
^
H
O
vO
O\
rH
rH
rH
in
o
rH
rH
CM
rH
^
^j.
rH
vO
ST
rH
oo
m
rH
(X,
CO
PQ
sr
o
m
rH
00
^f
vO
rH
ON
0
vO
rH
^
|^
CO
rH
00
CTi
m
rH
00
m
m
rH
^j.
m
rH
(Xl
0)
PQ
OO
rH
ON
I'-
ll"!
00
rH
(^
m
rH
VO
00
m
rH
VO
O
CO
rH
ON
in
m
rH
ao
m
in
rH
0)
c
cu
PM
H
A
f,
00
o
N
a
CU
PQ
ON
CM
O
CM
ON
O
^
CM
00
cn
ON
CM
OO
CM
PM
tO
PQ
rH
^
g
^
CO
CO
vO
rH
ON
m
ON
oo
m
CO
in
m
o
m
vO
rH
VO
vO
\O
vO
vO
fn
0)
PQ
rH
^
4-1
^
-------
- 72 -
process raw exhaust to composite-sample bags at an emissions dynamometer
installation. The liquid samples were all removed at the same time and
stored in the dark, with air present, in nearly-full sealed amber bottles.
Each analysis uses new additions of tracer BaP and tracer BaA. The data
show generally excellent repeatability for many of the species determined.
However, BaP and BghiP (but not MeBaP) show evidence of losses with time.
At present, we can only suggest that these are true losses due to oxidation.
The amounts of BaP and MeBaP in the sample have always been low, relative
to both BeP and BaA. This is in line with their reactivities and the long
exposure of the original solids to a wide variety of exhausts transitting
the exchanger. The further losses of BaP (but not MeBaP) and, recently, of
BghiP on storage in solution have no obvious explanation.
-------
- 73 -
7. SELECTIVE LOSSES OF PNA SPECIES IN EXHAUST SAMPLES
7.1 Summary of Loss Studies
Within the framework of the CAPE-6-68 Program, several emission-
test samples have been analyzed for PNA content in separate parts derived
from (1) extracts of filter media at the collector outlet, (2) extracts of
aqueous condensate at the heat-exchanger outlet, and (3) the washings from
metal surfaces extending from the vehicle muffler to the filter frame, i.e.,
the balance of a normal sample, which is made up of a composite of the three
parts. The relative amounts of different PNA species found in these differ-
ent areas, together with exposures of pure PNA specimens (BaP, BaA, and benzo
(e)Pyrene (BeP)) to the environment of the cool, filtered, outlet gases, prove
that very substantial losses of certain reactive PNA species deposited on the
outlet filter do occur in normal CAPE-6 sample-production procedures. Speci-
fically, the apparent total yields of BaP and BaA and their alkyl derivatives
are estimated to be only about 50% of the amounts actually emitted from the
vehicle. The more stable BeP does not undergo this type of loss. In further
work, growing out of this discovery and supported substantially by Esso
Research and Engineering, an apparently effective technique for avoiding these
losses has been demonstrated. This Section reports the experimental work
related to these losses and the avoidance of the losses.
7.2 Evidence for CAPE-6-68 Losses
In Section 6 of this report, we have already mentioned the review
by Tipson''' of published theoretical and experimental studies on the relative
reactivities of a variety of PNA species, including six determined in the CAPE-6
exhaust samples:
RELATIVE REACTIVITIES OF CAPE-6 PNA
(From Tipson, for Ortho Reactions)
PNA Species Relative Reactivity
Benzo(a)pyrene 12
Benz(a)anthracene 5
Pyrene 3
Benzo (e)pyrene <1
Chrysene 0.09
Triphenylene 10~6
A similar ranking can be developed for single-point attack on the molecules.
Apparently, BaP and BaA, which are the primary species determined in CAPE-6,
are especially likely to react and, hence, to be lost from the samples. The
alkyl derivatives of BaP and BaA can be expected to react similarly to their
parent compound s.
-------
- 74 -
Accidents affecting parts of the samples from CAPE-6 Tests 68, 78,
and 79 made it necessary to separately analyze portions of the samples for
Tests 69 (water extract) and 80 (filter extract), and three-part analyses
of the samples for Tests 83 and 85 were also made to substantiate the indi-
cations of the Test 69 and 80 analyses. For each test, a total yield for
each species is available (Table 22, Section 6) as the sum of the parts.
In the following table, the amount of each species in each location in the
apparatus is stated as a % of the total.
PNA SPECIES LOCATION FOR DIFFERING REACTIVITIES
7» Of Total In Each Location
(a)
Water Extract
Filter Extract
PNA Species
BaP
MeBaP
BaA
MeBaA
DM/EBaA
BghiP
Pyrene
BeP
MeBeP
Chrysene
Triphenylene
T69
13.9
10.7
14.1
13.3
8.9
7.1
7.5
6.7
7.2
6.4
5.5
T85
12.5
13.9
13.5
14.9
19.5
6.6
6.5
6.2
7.4
4.5
4.3
T83
38.5
23.2
24.3
27.1
26.1
15.7
20.1
11.7
9.6
16.1
15.4
T80
14.0 (b)
6.7 (b)
18.0
13.4
37.7
34.7
43.0
47.4
32.9
45.6
57.4
T85
T83
2.0 (b)
3.1 (b)
8.0
6.0
5.2CB)
34.4
62.1
53.7
50.2
66.3
66.9
17.0 (b)
9.0 (b)
32.8
34.7
23.1
27.4
44.5
67.7
37.6
53.3
54.6
(a)
'Tests 69, 80, and 85 in 1966 Plymouth; 83 in 1970 Chevrolet. The balance
of each sample is in the washings of metal surfaces, etc.
Low UV absorbance of solution of trapped GC peaks reduces accuracy.
These results suggest superficially that BaP and BaA and their derivatives
are more effectively collected in the aqueous condensate than are BeP, chry-
sene, and triphenylene. This conclusion is difficult to accept, however,
because these isomers (BaP and BeP; BaA, chrysene, and triphenylene) are not
even separated, in general, by gas chromatography. The correct meaning of
the results, in view of the reactivity differences, is rather that the reac-
tive BaP, BaA, and their derivatives that are collected on the filter media
are largely destroyed by reactions, either during or after collection of
the sample. This depresses the totals and inflates the apparent % found in
the aqueous condensate (and suspended soot particles) .
-------
- 75 -
If it is assumed that the condensate for a given test should con-
tain about the same fraction of all species, and that no losses occur in
the condensate, then it follows that the true total yields of BaP, BaA, and
their derivatives were about twice as great as the apparent total yields.
Furthermore, since more than one-half of the total of each stable PNA
species is found on the filter, but little of the reactive species, it
follows that losses of as much as 907» of the reactive PNA that are on the
filter may be occurring. It is significant that BeP/BaP ratios, etc., in
the filter are invariably high.
7.3 Confirmatory Tests of Losses
In order to describe experiments to confirm this hypothesis, and
to develop a technique for loss-prevention, it is necessary to review briefly
our standard CAPE-6 procedure. The vehicle was connected, without known
leaks or openings, to the PNA collector inlet (diverter valve) before start-
up. After starting, the diverter was turned from waste to the collector and
was returned to waste (collector sealed) at the end of each of the three cold-
start 12-cycle blocks used for sample production. The collector outlet was
connected continuously to a positive-displacement pump with automatic air-
bleed, which ensured against any back-pressure on the vehicle from the
pressure-drop of the collector system under varying gas flows. After the
third block, which occurred late in a working day, the pump was detached and
the collector was sealed by a blanking plate, leaving the sample exposed to
gases from the final deceleration. The system was opened, and the filter media
removed to a glass jar, no sooner than the next day. In some cases longer
residence occurred before filter removal, but no records are available
on this point. The aqueous condensate was collected in a sidearm receiver,
where it was isolated from the main gas flow, and was removed, after each block,
to a larger vessel, along with any entrained soot. An after-test acetone
rinse of the receiver vessel was combined with subsequent washings from metal sur-
faces because acetone interferes with analysis if present in the condensate. A
diagram and photographs of the equipment appear in the First Annual Report.
The standard experiment used for loss-investigation and loss-pre-
vention experiments has consisted of hanging two (replicate) Dexiglas (abso-
lute filter media) strips, 12" x 1.5", on the outlet side of the sample collec-
tor filter. Each strip contained separate spots, about 1" in diameter, of pure
PNA (BaP, BaA, BeP); typically, 10/ig of each was applied in cyclohexane solu-
tion. BeP was included in only a few experiments. Reference strips, held
in the dark, or exposed to the same light but not to exhaust gas, and/or reference
solutions at standard dilutions of the same amount of PNA, were also prepared,
with some variations in practice in different experiments. In some experiments,
a small amount of n-hexadecane, as an inert medium for the PNA, was incorporated
in the spots, but this had no apparent effect.
After a test, the strips were removed (usually the next day, but
within an hour in some later tests), cut in squares, and each square extracted,
filtered, and diluted to 30 ml. with cyclohexane for UV absorbance measure-
ment. Blank squares were also extracted (no spectra, but a background enve-
lope) and the extracts used in the reference cell of the UV instrument. PNA
recovery was judged by the absorbances of defined peaks (383, 332, and 289 nm
-------
- 76 -
for BaP, BeP, and BaA) above standard baselines. Even with severe losses,
no new spectra were observed, nor did the spectral curves change beyond
lowering and slight broadening of peaks.
The data obtained in these tests are summarized in Table 25.
The table includes results from exposing a concentrate (for GC/UV analysis)
left from CAPE-6 Test 73. This concentrate, in solution, had definite UV
peaks and measurable tracer activity. This was also true of the extracted
spots after exposure. As noted in footnote (d) of the table, the activity
and prior analysis permitted some additional interpretations of the results
to be made.
Recoveries of PNA by extraction of reference (unexposed) spots did
not exceed 907, for BaP and BaA, but were very good for BeP and for the Test
73 composite sample. Some low recoveries (to 25% for BaA) in light-exposed
reference strips remain unexplained, since carefully-controlled light-exposure
experiments involving more-than-usual light exposure showed no effects for
different kinds of light.
The most significant conclusions to be derived from Table 25 are
that losses of up to 95% of the BaP and 80% of the BaA, but little or no
loss of BeP, do occur in the filter-box environment. These conclusions are
in line with the reactivity predictions and with the interpretation of the
GC/UV sample analyses and locational variations presented earlier in this
Appendix. The losses evident in the exposure data on the Test 73 concentrate
are somewhat at variance with the analyses of extracts of the main filters,
however. It appears that BaP, MeBaP, and benzo(ghi)perylene (BghiP) were
all lost in the exposure, yet substantial BghiP was found in the filters of
Tests 80, 83, and 85. (Tipson does not present estimates of the reactivity
of BghiP.) Also, the exposed Test 73 concentrate did not seem to lose BaA
and its derivatives, whereas they were apparently reduced in the filter ex-
tracts of the tests, and were lost to some extent, but less than BaP, in the
exposed strips of Dexiglas.
7.4 Loss-Reduction Experiments
Table 25 includes results on PNA loss from exposures in Test 89B
that are particularly meaningful as guides to a successful solution to the
loss of PNA from filter samples. Independently of the other test objectives,
including the exposures, a large-volume (to 170 CFM) flow of air through the
entire collector system was created for several minutes at the end of the
test. This was done to obtain pressure-drop data at various fixed flow rates.
The PNA-exposure specimens were removed from the filter on the following day.
In contrast to later tests without the air-flushing, these specimens showed
no loss of BaA and significantly reduced losses of BaP. This suggested that
purging shut-down gases from the system was desirable. (In this case, shut-
down gases had been in contact with the specimens during vehicle cool-down
after test-blocks 1 and 2.)
With this suggestion, Esso Research has used the PNA collector sys-
tem in slightly different tests which have demonstrated a substantial reduction
in the loss problem. A different test vehicle, with uniform deposits and with
-------
- 77 -
* _.
a CM
0 CO g ON
SZ CU /> | 1 |
CJ CO CO T> vO
4-1 C 0) CO ' ON
co co B c-g I
CU ,£1 -H
H rl rH
0 CU
S co co
O J3 CO
u a j3 8
"" > T> C
^ & 01 00
> B rH
P 60 -H CO
-a-
p^
1 1 1 iH
r-.
-» B MH
CJ -H OJ
U CO 13
3
rl 01
0 " >
MH -a- o s
CO
cu 4-1 in
4J CO CO . CO
cd oj j«! ^s co
in
00
CO
00
rl H CO
>% 4-1 OJ B
CO B B P.IH
CO 0) *rl
CO O 01 >,
CO B T) 4J rl
O CU -H 01 B
> o to c > B
P O -H O /-»
< P. MH CJ CO 13
IH !S X 0) OJ 00 ^
O PH 0) Tl pd CO
o
1 1 1 -
o
MH
.^
Tj
01 rH
4-1 H
O 01 01
CO CJ B
rl 01
4J O CJ
X CJ B-S Id
CO CU B tl
WO) - J3
PS Tl tl CO 4J
to B 0) P B
CO cd MH -H cd
O CU IH ^N
PH " ti 4-1 cd
X 13 CO ^
W CJ OJ N
co j2 cd B
CO O 4J -HO)
1=1 P. Tl CO
O X B ,01
H 0) i-l S
-a- co -a- ON
< r*- co cs
ON vo in CM
< « tl tl
> tl 01 01
01 4J 4-1 0)
P fe MH £*
1 < X 01 P.
VO CO S 4-1 s-,
I co cd cd cd cd
W W rH H CJ ^
PH H OOJ3 -H O
< O iH rH N
O W X TJ P B
PL* ol OJ 3 CU
CO O co Q pd
fi in H
in CM in rH
-a- cy> m rH
-a- lH rH
o
PH 01 ti
Tl MH
PH 01 C cu
O 4J CO CO B
*rl 01 OJ
CO CO MH -H tl
U O O tl >,
HP, 01 P
P3 01 4-1 > ^
W TJ O O CU
t> CO CJ *x
O B ti cu o
W B X B
pti *H 01 <1J
CJ pa
co CM -a-
H VO O
O O ON
H rH
01 Ol
P*,rJ
CO 4J
CJ 4J vD
CO -H 1
C co in ,rj
O OJ 1 ^^
H iH CM
4J p. (0
CO S Tl 4-1
IH CO CU ^3
CO CO CO 00
P. O -H
0) MH PrlH
tl 0 X
P, 01 CO
tl O CO O
(0 13 01 iH
4J -^ rH H
MH p, oj
id B >
§
iH s~> T3 O
"CM Q)
r*- ^ vi f-H
O Q
«'~\ J3 X « W)
00 CM 4J Q) 4J -U M-l P3 »
H <-* PQO CX!M<>N
»CN ONB OJQOOJcO^l
-a- ^ oo B 1-1 co to
O - O iH JD Tt
H C >> P. O O 0)
H rH @ O 4J ^
CO CO O cd rH OJ tl
4J B 1 4J 3 4-1
CO IH t* ^N B *O CO
CU O & H O 01
4J K 3 4J CO TJ
O CJ CO -H
OJ B tl -rl CO
tl . OJ CO OJ CO
[^ 34J OOSfiOS
r~ co co B OB
^^ OOJ HrHCJCO^-s
O CM p, 4-1 > 1 & ^A
4JrH X COCOOJtlcd
^^ oioi ,r3cotiooi
in jz cj cu co &.
CM <3 4-1 " & rQ
gco o cd co
tl 4-1 O CO 00
cu j3 -a- 4-1 8 co
aj 4-1 oojc ,c B ^
J3 MH -H tb
H CO rH *H 00 IH
W rH 00 CU
to 4-1 CM 4-1
0) G O tl 4-1
CO 4J CU O OJ CO
4J 3 CJ TJ MH J3 iH
tl B CO B 4J
CO iH OJ CO - 0)
rH 4-18 tl CO MH £
ON CO O 4-1 O 4->
^ rH 3 CO CJ
OCM COCd H4JCUUHMH
4JrH tl MHBMHrHO
* MH OJ OJ MH CO
CO O> ^CJOJ.COI
m OJ Q co cj
CM CO O 0) tl 4J B
ti O 3 CU
B tl O O O TJ
HO -a- 3 -.a -H
MH Cu iH CO CO >
00 MH OJ 01
B ti O 4J
H i-l 13 S B CO iH
Jd CO rH O 01 J2 iH
O O IH 4J CO
f MH OOO CU
r- o o 3 B G oo
O rH 3 CO O CO M
O £? 0) T
MH > TJ O CO OH
O CU O >-, CO
OJ 4-1 rH M
cu oo cj o B cd a)
CO |H CO -a* rH B S
3 CO ti Cu *H CO -f-
CO lH - 4J MH PH
O 4-1 X " > CO
0)COXOJ4JOt> CO
r, 60 J2 -H "*^
,C -rl TJ 00 4J CJ 0
>->4-lCG-HCOO4J
4-1 -H IH CO rH CO
O & 0) G H IH OJ
O > TJ 3 P. O 3
CO TJ O 01 CO -H T)
OJ tl * tl
CO.G4-1COTJCU PiCO
CO CO CO P. 0) 3 iH
J5 3 cd 0) 'UJ O* 'O
iH 01 IH 3 CO B 0)
OJ MH rH p.iH p. id CJ
iH O O B
P. CO 4J 01 B 1 CO CO
BCOcdtl-H>4-lJ3
CO S 01 b C IH
/-N co cu S co 3 O
CO 4J CJ 01 O CO
OJ B-rlcOCOtltlOjO
H HGiHBSOJ co
p. c03p
-------
- 78 -
the same fuel used for both deposits and emission tests, was operated in
three single 23-tninute cold-start tests with the 1972 emission-test driv-
ing cycle. Duplicate speciments of BaP and BaA were present in the filter-
box for each test. The start-up for each test had the diverter valve already
open to the collector, but different procedures for sample collection were
used thereafter:
(1) Like CAPE-6, the collector was sealed at shutdown and opened
the following day.
(2) At shutdown, room air was drawn through the system (tailpipe
disconnected) for 4 minutes at 175 CFM (measured after water-
vapor removal in the cold exchanger), and the specimens were
removed within one hour.
(3) A fixed outlet gas flow of 120 CFM was maintained throughout
the test, combining (at a large T) air filtered through
charcoal and paper (CVS filter element) with whatever exhaust
was produced by the vehicle; i.e., a "CVS system" at 120 CFM
total flow, followed by cooling and filtering, was created.
After the test and 1 minute of air flow, the exposure speci-
mens were removed within one hour.
The data obtained in these experiments are presented below. It
ESSO RESEARCH TESTS ON PNA LOSSES
(Single 23-Minute 1972 FTP Tests)
Recovered
Specimen
Ref. Solution
Held in Dark
Specimen #1
#2
(1)
Like
CAPE-
6
100
91
32
35
BaP
(2)
Flush
and
Remove
100
-
72
68
(3)
Dilution
and
Remove
100
89
89
91
(1)
Like
CAPE-
6
100
91
67
62
BaA
(2)
Flush
and
Remove
100
-
88
86
(3)
Dilution
and
Remove
100
90
88
87
appears from the data that both of the changes in procedure were effective in
avoiding the loss of BaA from exposed specimens, while the CVS-type procedure
(constant dilution) was slightly superior with respect to avoiding BaP loss.
It was felt, in planning these experiments, that the air-diluent
cooling (item (3) above) and further cooling in the heat exchanger would mini-
mize any losses of PNA from oxidation reactions. It was recognized, however,
that air dilution to a total flow of only 120 CFM (to avoid excessive pressure
drop in the system) would not give substantial diluent cooling in acceleration
driving modes.
-------
- 79 -
In general, this assumption appears likely to have been justified,
but the relevant data require some qualification to arrive at this view.
Esso Research has made available in Table 26 the emissions results, in
^jg/gal of fuel, for each of these three vehicle tests. The results for the
air-dilution experiments (Column 3) are unexpectedly low for all PNA species,
whether reactive or not. As noted in footnote (b), this is tentatively
ascribed to an unexplained failure to reproduce vehicle conditions. (An
across-the-board loss of all species by reactions with the diluent air is not
consistent with known differences in PNA reactivities.) Arbitrarily doubling
all of the PNA-emission values in the third column gives the "corrected" values
in the fourth column of Table 26. Comparison of these values with those in the
second column of figures indicates that only BghiP, MeBaA, and MeBaP appear to
have been depressed by reactions with diluent air. Of these, both the MeBaA
and MeBaP results are of relatively poor accuracy (footnote (c)). Their GC
peaks, as well as their UV spectra, did show that their emission rates were
indeed low, however, so the apparent losses cannot be fully discounted.
With the above qualification and the presumption that the three tests
can be validly compared, then the most significant result from Table E-2 Is the
evidence of losses in the procedure that most nearly equalled CAPE-6 operation.
All species except triphenylene, chrysene, and BeP underwent apparent losses
relative to the air-flush test of 34% (pyrene) to 74% (DM/EBaA), with the
losses in BaP and BaA amounting to 66% and 53%, respectively. Losses of this
magnitude are consistent with the previously-described analyses of emission
samples from different parts of the collector unit, and with the losses of pure
PNA exposed in the filter-box outlet gas.
7.5 Future Practice in CAPE-6-68 Sampling
These experiments suggest that air-flushing after each vehicle opera-
tion period, and early filter removal, would be a relatively simple but improved
procedure for future use in CAPE-6 or any similar project. However, such a
change would increase the expected results of any test and would destroy its
comparability to earlier data. For this reason, the limited further research
planned by CAPE-6 will continue to use the existing procedure, without air-
flushing, and to base conclusions primarily on changes in BaP and BaA emissions.
This decision is fully consistent with the objective of CAPE-6-68
to evaluate the relative effects on vehicle PNA emission for various changes
in controlled fuel and vehicle variables. Section 6 has already shown that
the ratios between reactive species (BaP, BaA, etc.) and non-reactive species
(BeP, etc.) are no more variable than the ratios between two non-reactive
species. This implies either reproducible losses of reactive species, or com-
pensating sources of variability in determining non-reactive species. In
either case, the use of either type of PNA should give equally valid measures
of the effects of variables.
-------
- 80 -
VO
CM
W
rJ
P?.
<
H
01
G
H
60
C
Ol
1 J
H
3
CO
w
D
o-
M
K
U
W
EH
55
O
H
LLECT
o
c_>
13
01
4-1
co
01
4J
*l
01
T-l
CJ
/^
CO
4->
CO
01
^J
PH
EH
fe
CM
r^
rH
01
4J
3
C
H
e
en
CM
0)
i 1
60
c
3
CO
c
H
13
01
CJ
3
13
O
(H
a.
01
M
o
4)
CO
£3
rH
41
PL,
<4-l
O
rH
CO
60
6t
G
H
CO
G
O
iH
CO
CO
H
g
(3
*»
G
O
H
3
i~H
H
Q
i-l
H
<
4J
ed
43
CO
3
rH
M-l
!H
H
-,
rH
v^
cfl
W
43
T3
01
4-1
rrec
o
u
/ * /"" \
0 CJ
^ 00 v-^ 0
* CO * CM CM
'CM O CM rH 00 O
CnrHcnvOCMrHvOCM
13
01
S-,
01
CO
43
O
^~\ / >
0 0
V ' ^^- Q
r^
cfl O
W E
41
> OS
*O
C
w
,*-S
CJ
^ r^ o o
gcMoiinomvo
OOrH^rm-3-rHvOrH
4J
43
13 60
41 -i-l
rH C
cd ^
01 4)
CO >
0
"o1
* VO CNI u-,
-inovor^-oo
encMrHcMrHvd-^-,H
CM O
en CO
O
41 CX
G 41
O 13
60
G.S
rH CO
CO '
" 41
CO Lj "
0) o
CO o
£5 41 10
PH -rj OJ1
M e. 0)
CM CQ
S
01
CO
>,
IH
43
01
01
C
01
43
CX
CO
3 co
ed 01
43 4)
X M
41 60
cfl
13
G TO
CO 41
H CO
rH O 13
01 01 C
3 CX 3
>4-l CO O
CX
VO M E
1 O O
Pd i-i CJ
Pt, efl
< E -< IH
01 0 C
43 Cfl
E 6043
3 C 4J
C -H
^ 41
4) G >
60 efl -H
H VJ 4-1
co a
rH CU 50
4: 41
efl H M
C 01
H Vi
4-1 O
co C E
41 41
H 4-1
O C 13
41 O 01
CX U 4-1
CO O
G 41
MH 01 CX
0 60 X
>-. 01
W X
O) O CO
60 Cfl
G TJ
cfl C >
43 ed 01
U M
01 Cfl
41 >-,
> 3 CO
H 4-1 01
4J Cfl >
cd t-l -H
rH 41 -M
41 CX efl
H E >
41 -iH
<4-l 4-1 ^
O 41
C TJ
CO -H
G rH
O 60 >.
H G M
4-1 iH rH
efl >, cfl
> V4
M CO t3
41 > G
CO CO
43 CO
O 4J «
CO S
G 3 01
O CO -H
43 >
T3 X 4)
01 Ol »-l
CO
CO 43 CO
43 4-1 -
H C
co 3 O
H Cfl
. d,
60 CO -H
C 01 EH
H rH
^! CX43
G E "->
cfl efl -H
OS W S
1?
v? ,
C cj
4) 4)
H M
4J O M
CO 1-1 O
O <4H O
S 0)
rH "O 01
CO iH
W 43
CO C 4J
H 4)
G 135
g CXi-l
3 Q. 1
rH CO
O
cj c .
ed co
a c
M 41 O
H 4-1 i-t
43 CO 4J
«-> 03 iH
c -a
OJ 01 G
43 CX O
4J 0 0
C O 4)
H rH CO
CO 0 4J
x-\ 01 i-l CO
CO 3 43 01
41 rH 0) 4-1
H efl >
CJ > M
01 41 41
CXTJ CJ ,fi
CO 41 3 4J
W -O O
rH O O
rH 4) M 01
efl M CX,r3
' IH 41 4J
O M
rH CJ O
4) O 4J
> 01 -W
rHlH 4, g
, MO
- CX efl 1-1
CO CO 4-1
41 t>% CJ
CO 41 rH efl
8 > 01 41
H > H
rH 4-1 -H 1
O 4-1 C
C
PL, M C
1 41 4)
00 C 4J 43
GO ti
H C «
T> to IH
H M 41 O
O O i-l MH
> >4-l 0
efl 41 co
4-1 CX 4J
MH O CO rH
O 41 3
(J rH CO
41 M rH 41
)-i O CO M
3 0
co q C 01
efl -H O 43
41 4J
S r^fr*
rH O «
eo cu m c
Vi 0
co 3 M-i i-l
s
tl
fll
S
9*
(4_|
Q
fll
W
CO
S
*V
QJ
0)
o
»^«
Ss.
0
n)
M
0
CJ
fj
-------
- 81 -
7.6 Tracer Injection for Improved Accuracy
The foregoing work on sampling losses was largely carried out in
the period following the presentation of SAE Paper 720210 (Reference la) in
January, 1972, and the earlier work had not been fully evaluated by that
time. In prepared comments on that presentation, Griffing^ ' questioned
the accuracy of the CAPE-6 emission results and suggested that losses were
occurring after sample collection. Griffing had injected tracer BaP, vapor-
ized in hot nitrogen, into vehicle exhaust enroute to a large partially air-
filled plastic bag. After a necessarily brief vehicle run, a portion of the
bag's contents, pumped through a filter, yielded a sample for BaP analysis and
counting. These results, with the known tracer BaP addition, gave estimates
of the BaP emitted by the vehicle. Apparently no other PNA species were
determined.
The Griffing procedure requires, of course, that the fates of the
thoroughly-mixed engine-emitted BaP and tracer BaP be the same after injec-
tion, including residence in the sample bag. Hoffman'^' had reported results,
using a collector like the CAPE-6 collector, which raised some question as to
the validity of this assumption in the case of exhaust from the tailpipe of a
test vehicle on a dynamometer. Specifically, Hoffman indicated variable
specific activities (DPM/ug) for BaP from different parts of his collector sys-
tem. By analogy, whatever "part" was sampled from the collector bag in the
Griffing procedure could control the result obtained. Griffing^^) had not
ignored this question, however: Different amounts of injected tracer and dif-
ferent volumes of gas filtered from the bag were said to give equivalent BaP
emission results in replicate vehicle tests. Also, Griffing noted that BaP
losses did actually occur during residence in the collector bag.
Griffing suggested^^) that total-collection systems, such as were
used by Begeman^', Hof fman' , and CAPE-6-68, should be validated by the addi-
tion of a tracer-BaP-in-nitrogen system for tailpipe injection. Assuming the
solution of the very substantial technical problems implied, including radia-
tion hazards, and assuming that the implications of Hoffman's experience (see
above) can be avoided by using a total sample, the suggestion has some merit.
However, the CAPE-6 loss-experiment results, showing losses which vary for dif-
ferent PNA species, make it clear that the injection technique cannot be ex-
tended rigorously beyond the actual species that are injected. Furthermore,
some assurance is required that, if more than one radioactive species is in-
jected, there will be no thermally-induced interconversion or degradation of
one active species to another.
In any case, and without taking into account the losses in the CAPE-6
procedure, the Griffing data and the CAPE-6 data do provide very similar esti-
mates of the relative effects of different controlled variables on BaP emis-
sion. The discrepancies exist, as might be expected, in the estimates of
absolute (ng/gallon or yug/mile) effects. Griffing does question^ ' the ability
-------
- 82 -
of very-low-emission control systems to control PNA to the extent that
CAPE-6 has described in this report. However, we have already indicated in
Section 5 of this report the belief that the control systems used by Griffing
for data that imply this question were not, in fact, as effective as those
used by CAPE-6 so far as the control of CO and HC are concerned. Thus, they
should not properly be compared as to PNA control.
The Griffing suggestion that PNA are being lost in the CAPE-6 pro-
cedure was, on balance, a valuable contribution in spite of the foregoing dis-
cussion. It served as a stimulus to carry out a variety of significant ex-
periments on PNA losses and these, in turn, provided a much-improved insight
into the limitations of CAPE-6 results in an absolute sense.
-------
- 83 -
APPENDIX A
REFERENCES
(1) Gross, G. P., "Second Annual Report on Gasoline Composition and
Vehicle Exhaust Gas Polynuclear Aromatic Content," (CRC-APRAC
Project CAPE-6-68), Period Ending April 15, 1971. (NTIS Accession
No. PB-209-955.)
(la) Gross, G. P., "The Effect of Fuel and Vehicle Variables on Poly-
nuclear Aromatic Hydrocarbon and Phenol Emissions," Paper No. 720210
at the SAE Automotive Engineering Congress and Exposition, Detroit,
Mich., January 10-14, 1972. (Accepted for publication in 1972 SAE
Transactions.)
(2) 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. (NTIS Accession No.
PB-200-266.)
(3) Lang., R. J., "A Well-Mixed Thermal Reactor System for Automotive
Emission Control," SAE Transactions, 80_, 2148-2155 (1971).
(4) Lunt, R. S., Ill, Bernstein, L. S., Hansel, J. G., and Holt, E. L.,
"Application of a Monel-Platinum Dual Catalyst System to Automotive
Emission Control," Paper No. 720209 at the SAE Automotive Engineering
Congress and Exposition, Detroit, Mich., January 10-14, 1972.
(Accepted for publication in 1972 SAE Transactions.)
(5) Begeman, C. R., and Colucci, J. M., "Polynuclear Aromatic Hydro-
carbon Emissions from Automotive Engines," SAE Transactions, 79,
1682-1698 (1970).
(6) Brown, R. A., et al., "Final Report on Rapid Methods of Analysis
for Trace Quantities of Polynuclear Aromatic Hydrocarbons in
Automobile Exhaust, Gasoline, and Crankcase Oil, and Phenols in
Automobile Exhaust," (CRC-APRAC Project CAPE-12-68), Period of
February, 1969, to December, 1971. Report in preparation.
(7) Tipson, R. S., "Oxidation of Polycyclic, Aromatic Hydrocarbons - A
Review of the Literature," U. S. National Bureau of Standards
Monograph 87, issued September 17, 1965.
(8) Federal Register, 35_ (219) Part II, p. 17288 et seq., November 10, 1970.
(9) Hoffman, C. S., Jr., 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).
-------
- 84 -
(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., Prepared Discussion of Reference la (SAE Paper 720210),
presented at the SAE Automotive Engineering Congress and Exposition,
Detroit, Mich., January 10-14, 1972. (To be published in 1972 SAE
Transactions.)
(12) Begeman, C. R., and Colucci, J. M., "Benzo(a)pyrene in Gasoline
Partially Persists in Automobile Exhaust," (letter), Science, 161, 271
(1968).
(13) Griffing, M. E., Maler, A. R., Borland, J. E., and Decker, R. R.,
"Applying 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).
(14) 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, Vol. 14, No. 3, page B-162).
(15) Moran, J. B., Prepared Discussion of Reference la (SAE Paper 720210),
presented at the SAE Automotive Engineering Congress and Exposition,
Detroit, Mich., January 10-14, 1972. (To be published in 1972 SAE
Transactions.)
(16) Commins, B. T., "Formation of Polycyclic Aromatic Hydrocarbons During
Pyrolysis and Combustion of Hydrocarbons," Atmospheric Environment, 3_,
565-572 (1969).
(17) Oro, J., Han, J., and Zkatkis, A., "Application of High Resolution
Gas Chromatography - Mass Spectrophotometry to the Analysis of the
Pyrolysis Products of Isoprene," Analytical Chemistry, 39, 27-32
(1967).
(18) Cleary, G. J., "Polycyclic Hydrocarbon Ratios: Their Use in Studying
the Sequence of Combustion in a Hand-fired Intermittent Brick Kiln,"
Proceedings of International Clean Air Congress, October, 1966, pp.
111-114, National Society for Clean Air, London.
(19) Badger, G. M., Kimber, R. W. L., and Novotny, J., "The Formation of
Aromatic Hydrocarbons at High Temperatures. XXI. The Pyrolysis of
n-Butylbenzene over a Range of Temperatures from 300° to 900° (C) at
50° Intervals." Australian J. of Chem., 17_, 778-786 (1964).
(20) Federal Register, 211 (61) Part II, p. 5170 et seq., March 30, 1966.
-------
- 85 -
APPENDIX B
EXHAUST-GAS TEMPERATURE AND COMPOSITION DATA
Table B-l summarizes the hot-cycle muffler-exit exhaust-gas
temperatures measured in the latter part of each emission-test block.
(For details of test equipment and procedures, see the previous Annual
Reports (1, 2) and Sections 3.2.1 and 3.3 of this report.) Averages
of these gas temperatures were used in Section 3.2.1 as measures of
vehicle exhaust conditions.
Average (over one to three of the three test blocks) values
of corrected gas analyses (see Section 3.2.1) are given in Tables B-2
through B-A1 for the CO, HC, and NO emissions in each mode of selected
cold and hot cycles in each PNA-emission test. The composite results
for CO, HC, and NO for each test also appear in Table 8.
-------
- 86 -
TABLE B-l
EXHAUST-GAS TEMPERATURES
(Third-year data. Measured in hot cycles at muffler exit.)
Muffler Exit, °F., min./max.
Vehicle Test No.
1966 Plymouth
(NC)
1968 Chevrolet
(EM)
1970 Chevrolet
(EM)
61
61B
62
68
69
73
76
77
77B
80
81
85
89
89B
93
96
64
65(a)
71
72
78
79
82
86
90(b)
92
94
66
67
70
74
75
83
84
87
88
91
95
(a) Test included only two blocks.
(b) Thermocouple instrument defective in part of test.
Block 1
452/565
/570
440/555
470/580
467/597
458/571
447/560
451/575
442/555
435/545
457/569
441/560
446/563
460/581
466/592
456/578
476/584
A on /c;Qn
HOU/ jyu
495/610
478/596
483/595
473/585
464/575
472/574
483/595
485/590
577/711
583/715
578/715
585/719
576/707
555/675
546/680
562/691
562/690
556/667
585/695
Block 2
456/569
450/575
446/560
465/580
463/594
458/575
445/562
455/575
446/560
433/546
461/569
441/561
451/578
462/585
467/594
456/580
482/587
A fin /^QO
*+OU / D s\J
496/610
480/603
482/600
470/580
476/576
470/574
487/600
487/592
583/717
577/715
580/720
582/720
574/699
561/680
546/675
560/695
573/697
560/672
585/695
Block 3
443/562
450/575
445/558
467/583
462/590
458/573
446/563
455/575
446/560
433/545
456/567
445/565
448/570
463/595
460/588
463/582
481/585
I
1
494/608
479/595
472/587
472/585
462/571
475/577
Afi9 l^ift"}
HOi / _)O /
475/580
482/592
588/711
573/710
576/715
580/725
575/704
560/680
545/675
556/695
569/690
572/676
593/703
-------
)-> -O
ss
X
I
ol.
E p
(U M
CD 01
M W
01 TO
3 )-<
-------
t£> Q
W V)
8
X
-------
^
I
fX, M
< tn
3 SM
-------
g
H
3
o I
3 3
-------
- 91 -
a
I
I
-------
I
-------
i
H
8
+
8
o I.
g P.
-------
SCO
^
-------
- 95 -
« O
fL, W M-l
W O
-------
P- M *H
-------
- 97 -
I
O I.
2 CX
-------
sJ S 12
g 3 S
2 8
: i
B ^o
ao
5
I
-------
- 99 -
1
^
t! 8
oi
a o.
-------
- 100 -
si
-------
& D.C.
-------
- 102 -
a
o I.
05 «<
f S
VO Q
W en
Oj M
S 2
-------
- 101 -
a co
ft. M
< W
3 >.
o §.
2 a.
£
S o
4J
» g]
M -O (it
22 *
0) *J
m - o
O nH
O U CO
60 V
,H -H 60
X n)
W Li
83S
O ct rt
H >
PQ i-l 4J
IH nj QJ
U nH CQ
t-l 4-1
d) o) -a- c
-a -Of
3SS-
O W ,
a; GO O
iSS-S
-------
-------
- 105 -
« g
en o
I
^
i
-------
- 106 -
S
I
<^
I
ol.
------- |