-------
SOF COMPOSITION, (% MASS)
g
m
n
o
s
2 ^
g O
m
z
1
0)
*
z
1
•M
O
m
z
1
tn
ss
m
z
a
c
CD
1
to
ib
*
m
1
ro
*
4h
"
)
(
C
O
o
1
*
> TJ
0 >
a n
1 1
_A cn
jk "
5 *
> T
0 >
0 31
1 1
u> J5
" ^
^
•> >
3 31
1
» $
f X
2
3D
g
^
>
Q
ro
<0
0 =
- \
* •
> _j
n x
J z
1
ft *»
B in
3)
Z
1
O)
s
-1
D
£
•i
s
X
O
1
-t
3
Z
1
W
;£
O
X
•<
1
u
O
X
1
(O
0
o
X
1
ro
§
U
o
CD
O
I—
I
O
3D
SOF COMPOSITION, (% MASS)
s
m
O
O
O
>
m
z
1
to
23
1
tn
bo
0
>
cn
•SJ
*~
O
D
CD
O>
1
1
bo
S-
o
o
1
CO
f [
j.
cn i
tn -
JS j
rn
1
™
n
>
w
*3
^
>
O
I
W
T>
>
1
CO
*~
> -
31 3
O *
\
b> -
S? S
i s
1
Si IO
- ro
5-
0 > 2 >
! 8 5 §
i ' i
•3 09 at *•
* bn b cn
^
2)
I
W
* In
H °
? 5
1
! 2
*"
i i
i
sS S
5 S
i i
S ?
* iS
o
of ?
^33 W
O C»
X
z
1
ro
I
Z
1
01
I
Z
1
CO
i
i
c»
ro
> O I
§2f
iis
o ^< >
2 m m
- > S
o r-
03
m
?!!
si?
0 > Z
o "" w
z 5
>
1 5
1 f
-1 DO
ss
33 o
I
CO
O
o
m
CD
§
3)
TJ
CO
bo
CD
O
~
>
m
-------
67
transitional and oxygenated fractions actually increase with injection retard.
One possible explanation of this discrepancy could be that when fuel molecules
thermally decompose, free radicals can be formed. These radicals readily
combine with monatomic oxygen to form the acidic, oxygenated, and transitional
subfractions. At high temperatures, N2 will dissociate to N free radicals which
have a higher affinity for 0 radicals than the organics. Thus two competing
reactions can take place:
1) 0 + N ->NO
2) 0 + 'Rf ->ACD, TRN, OXY
where 'R' denotes an organic radical and ACD, TRN, and OXY are, respectively,
the acidic, oxygenated, and transitional subfractions. Since monatomic nitrogen
has a higher heat of formation than monatomic oxygen, 473 MJ/kmol as opposed to
250 MJ/kmol, N will dissociate at higher temperatures than 0^, Reduced
temperatures, by retarded injection, would decrease the relative amounts of N
and make more 0 available for the second reaction. This is evidenced by the
reduced NO emissions at retarded injection timings. Therefore, even though the
reactions are much more involved than indicated here, the increase in oxygenated
and transitional fractions at mode 9 can be explained simplistically by the idea
of the competition for monatomic oxygen between N and organic radicals.
The effect of injection timing on the biological activity of the SOF is
shown in Figure 3.1.17. The SA of the mode 3 fraction is increased as timing is
retarded, while at mode 9 the SA decreases. As mentioned earlier, these changes
in SA magnitude are considered to be slight and may not represent significant
differences. Both modes show reduced BSSA at retarded timings due to decreased
SOF emissions.
The relative magnitudes of the BSNO , BSFC, BSTPM, BSSA, and BSNO for
various injection timings for modes 3 and 9 are shown in Figures 3.1.18a and
-------
SPECIFIC ACTIVITY
(REV/MG)
BSSA
(KREV/KW—HR)
ro
ov
C rr O
C H- i-n
rr <
o
m o
o
O _L
Z CD
s
BRAKE SPECIFIC EMISSIONS
o s n
3 CL rt
cn n
- c/i o
|12»BTC
! 17" BTC $
I 22" BTC I
] 12' BTC
O oo
| 12* BTC
MT" BTC
22° BTC
12«BTC
17- BTC
22" BTC
< m
O
CD O
O ni
5 •=•
X 3>
BRAKE SPECIFIC EMISSIONS
12- BTC 5 o>
17- BTC
22- BTC
22'BTC
I o
12' BTC
17« BTC
1 22- BTC
00
-------
69
.4
_ .35
oc
I .3
a
I '25
m
.2
16
14
a"
I "
O 10
X
o
z 8
00
6
4
3.8 IN. BOWL HIGH RATE
RETARDED INJECTION TIMING
A 12- BTC
0.5 1.0 1.5 2.0
BSTPM (G/KW-HR)
Fig. 3.1.19 - Effect of injection timing on BSTPM
and BSNQx and BSFC reJatIcnshlps, 3.8
in. bowl, high rate. Each curve contains
points from all 6 modes
3.0
2.5
> 1.5
1.0
0.5
_
/
- o
I
\
0\ D
V
V
o\
— i 1 1 1 1 r—
INJECTION TIMING ('BTC)
17
\ 22 3.8 IN. BOWL.
\ HIGH RATE
\
\
\
\
V
\
\
\
\
\
I \
v. ^
0 n A A
I 1 1 1 1 1
0.2 0.4 0.6 0.8 1.0
BSN02 (G/KW-HR)
1.2
Fig. 3.1.20 - Effect of injection timinR on BSTPM-
BSFC-BSN02 tradeoff, hi?h rate, 3.8 in.
bowl. Each curve contains points from
all 6 modes
8 10 1?.
BSNOx (G/KW-HR)
Fig. 3.1.21 - Effect of injection timing on BSTPM-'
BSFC-BSNO tradeoff, 3.8 in. bowl, high
rate. Each curve contains points from
all 6 modes
-------
70
3.1.18b. The mode 3 results show that as BSFC and BSTPM are reduced, the BSNO
x
and BSNO are increased. These trends were shown for most of the modes tested.
However, under certain conditions, such as at mode 10, BSNO , BSNO-, and BSTPM
increase together while BSFC decreases.
Figure 3.1.19 contains plots of BSNO and BSFC vs. BSTPM. Each curve in
this and the following figures contains six points at a given timing which
correspond to the six modes tested. The major effects of retarding the
injection timing were to lower the BSNO levels while the BSTPM and BSFC changed
X
only slightly. To further analyze these relationships, all three variables
should be plotted on the same graph. The normalized values of BSFC and BSTPM
were combined and are plotted vs. BSNO and BSNO in Figures 3.1.20 and 3.1.21.
^ X
The curve closest to the origin represents the timing which results in the
lowest BSTPM-BSFC-BSNO or BSTPM-BSFC-BSNO relationship for these variables.
^- X
Another method of determining the injection timing which has the lowest
BSFC-BSTPM-BSNO (and BSNO ) combination would be to find the average normalized
X £~
distance of the data points from the origin. From the examination of these data
(Table 3.1.6), it is apparent that the system approaches the optimum conditions
as the static timing is retarded. This is due to the pronounced effect of
timing on the BSNO and BSNO emissions.
X ^-
-------
71
Table 3.1.6 - Effect of Injection Timing on Combined
Emissions and Performance, 3.8 in. Bowl,
High Rate
Combined Combined
Static Injection
Timing (°BTC)
12
17
22
BSFC, BSTPM, 9SN02
Parameter
1.55
2.18
2.19
BSFC, BSTPM,
Parameter
1.66
1.84
1.98
,BSNO
(a;
(a) See Table 3.1.5 for explanation of parameter
Injection Rate vs. Injection Timing
The relative effects of injection rate vs. timing can be examined by
comparing the percent change of emissions and BSFC resulting from increased rate
of injection to those resulting from changes in injection timing. This is best
accomplished by selecting an interval of injection timing which produces the
same relative change in one of the measured variables as does a change in the
rate of injection. It was found that a five degree advance in injection timing
had about the same effect as the increased rate of injection. That is, a five
degree advance in timing increased BSNO emissions by an average of 28%, while a
X
21% increase in the mean injection rate increased BSNO by an average of 25%.
X
Note that the average percent change represents the value obtained when
averaging over all six modes tested. The effect of injection timing on the
average percent change of BSNO was nearly linear, but the effect was non-linear
X
for BSSolids, BSSOF, BSN02 and BSSA. Figures 3.1.22 and 3.1.23 illustrate the
percent change in the above variables for retardation of timing from 22°BTC.
The individual points plotted at a given timing each represent a different
engine mode. Table 3.1.7 is a listing of the values of the average percentages
as well as those due to the change in injection rate. From these data it was
-------
72
Ul t- 9
O «B *
"
-50
-75
O CD
"
25
BSNOy
J.« M. BOWL, HIGH RATE
SOLD POINTS: 12M RPN
OPEN POINTS: 2100 RPM
•ug
<9 CO
-25
-50
« .25
-50
Ul
O
50
25
tug
O CO
22 20 18 16 14 12
INJECTION TIMING CBTC)
U
a
O
o
*
•25
-50
-75
3.8 IN. BOWL, HIGH RATE
SOLID POINTS: 1260 RPM
OPEN POINTS: 2100 RPM
50
O
E
Ul
O
-2S
22 20 18 16 14
INJECTION TIMING (°BTC)
12
Fig. 3.1.22 - Effect of injection timing on changes
in BSFC and emissions when retarded from
22'BTC, 3.8 in. bowl, high rate
Fig. 3.1.23 - Effect of injection timing on changes
in emissions and biological activity when
retarded from 22"BTC, 3.8 in. bowl, high
rate
-------
73
determined that for a given change in BSNO , the effect of injection timing was
X
greater than the increase in rate of injection for BSFC, BSSOF, and BSSA.
Timing has a lesser effect than injection rate for BSSOLIDS and BSN09. Although
the biological activity was affected more by injection rate, this was offset by
changes in the amount of SOF. Thus the BSSA (which is the product of SA and
BSSOF), showed a greater effect due to injection timing.
The effect of increased rate of injection was to reduce BSFC and TPM while
increasing NO and NO . The effect of retarded injection timing was to increase
BSFC and TPM while reducing NO and NO . As previously discussed, an increase
A. ^-
Table 3.1.7 - Effect of Fuel Injection Rate and
Timing on Percent Change in Engine
Performance and Emissions; Values are
Averages for all 6 EPA Modes Tested
Rate Increased
21 ± 9.1%
Timing Retarded
from 22° to 17°BTC
Timing Retarded
from 17° to 12°BTC
BSFC
BSSOLIDS
BSSOF
BSNO
X
BSNO
BSSA
-0.4 ±
-36.4 ±
-27 ±
25 ±
143 ±
88 ±
1.4
19
21
12
75
34
3.4 ±
25 ±
-59 ±
-21 ±
-47 ±
-47.8 ±
2.5
17
24
5
24
17
9.3
23.5
-38
-37
-68
-34
± 3.:
± 29
± 31
± 8
± 24
± 34
in the rate of injection combined with a retard in injection timing could be
used to improve the particulate characteristics of an engine without adversely
affecting the BSNO , BSNO , or BSFC. Data from the engine for the 3.8 in. bowl
X £•
with the low rate of injection are plotted in Figures 3.1.24a and b. The rate
of injection was then increased and the timing retarded. The figure shows that
the BSFC and BSNO data from the low rate system lie between the values
-------
74
200
_ 150
CO
i 5 100
5- 50
0
u. S
o
50
100
50
0
1000
500
0
O I O-35
UL >
S I 0.30
O
ENGINE SPEED: 2100 RPM 3.8 IN. BOWL-
— LOW RATE HIGH RATE
19-BTC
°~-~-~
M2«BTC
ir* 12*
19*
12*
17-BTC
._-^^r_-0 19-BTC
' 12- BTC
12«BTC
A 19. BTC
17' BTiC
400 600 800
BMEP (KPA)
1000
150
CO
£§100
— 50
0
CO
0- 50
in o
5. 0
150
0*2
si a 50
O »* ""
"" 0
1500
z a
" 500
0
a:
X
o I
g| -25
2 20
ENGINE SPEED: 1260 RPM
3.8 IN. BOWL
— HIGH RATE ^^^^ 1"* "
~^**^ BTC
D BTC-
-
_
""
$- — -rfv 19 BTC
17'»12«BTC
19« BTC"
^^Z^-'-^** BTC"
A ^A^*^**^ "
• Q~_ :— '=B- 17- BTC.
t7*BTC
a
n^^^^,.— & 12* BTC
*™
12- BTC
A 19-BTC_
B^^sr^^-Z^I^^-- .IT* BTC
lA-1 1 1 1 l__
400 600 800 1000
BMEP (KPA)
1200
a) 1260 rpm
b) 2100 rpm
Fig. 3.1.24 - Effect of increased rate of injection
and retarded injection timing, 3.8 in. bowl
-------
75
corresponding to the high rate of injection data. Also, the smoke and TPM are
lower for the latter data. Thus, it appears that the rate of injection, or more
appropriately the rate of air-fuel mixing, has a greater effect on particulate
than static injection timing. Similarly, NO is affected more by injection
X
timing than by mixing rates.
Up to this point the emphasis has been on the effects of various factors on
the trends and relationships experienced by the engine performance and emissions
for individual modes. However, one must also be concerned about the magnitude
of the composite emissions from a number of modes. Table 3.1.8 contains the
weighted brake-specific emissions for all six EPA modes for several combinations
of variables. Each mode was weighted equally (16.7%). The table shows that the
overall emissions follow the same trends as were discussed earlier, i.e., the
increased bowl diameter increases particulate and BSFC, but lowered BSNO and
X
BSSA; the increased injection rate lowered BSTPM and BSFC, but increased BSNO
X
and BSSA; and retarded timing increased BSTPM and BSFC while reducing BSNO and
X
BSSA. However, due to the difference in the magnitudes of the BSTPM and BSNO
emissions, one could consider the effect of timing on BSTPM to be minimal. For
fixed injection timing, the BSNO is highest at low engine speeds and decreases
X
as speed is increased, with load having little effect (Table Al of Appendix A).
A variable timing advance unit that would advance timing at higher speeds could
be used to yield fairly flat BSNO levels. This would allow the rates of
X
injection to be increased for particulate control with a minimum of timing
retard at low speeds to provide acceptable BSNO levels and good BSFC.
X
-------
76
Table 3.1.8 - Weighted Brake Specific Emissions for all 6 EPA Modes
Tested, (Unless Otherwise Noted) Each Weighted 16.7%
Engine/Inj .
System
Static Timing
(°BTC)
BSFC
(kg/kw-hr)
Mode 5 Smoke
(% Opacity)
BSSolids
(g/kw-hr)
BSSOF
(g/kw-hr)
BSSO,
(g/kw-hr)
BSTPM
(g/kw-hr)
BSNO
(g?kw-hr)
BSNO
(g/kw-hr)
BSN02
(gZkw-hr)
BSSA1 '
Baseline -
Low rate. .
3.3 in.bowlU;
19
0.252
9.5
0.34
0.09
0.04
0.49
8.00
4.88
0.54
123.0
Low rate, .
3.8 in. bowl (
19
0.258
7.4
0.84
0.09
0.02
1.03
7.34
4.65
0.22
25.0
\
High
22
0.248
3.5
0.32
0.14
0.04
0.51
11.18
6.81
0.72
60.0
rate, 3.
19
0.257
6.8
0.51
0.08
0.04
0.63
9.15
5.64
0.51
47.0
t -\
8 in. bowl (c)
17
0.251
3.5
0.43
0.08
0.05
0.55
8.58
5.34
0.40
29.5
12
0.276
6.1
0.49
0.03
0.04
0.57
5.39
3.45
0.09
19.5
(a) APE-BB 'low rate' pump, 3.3 in bowl, 5x.32 mm nozzle
(b) APE-BB 'low rate' pump, 3.8 in bowl, 6x.29 mm nozzle.
(c) APE-6G 'high rate' pump, 3.8 in bowl, 6x.29 mm nozzle
(d) Average includes modes 3 and 9 only, each weighted 50%
-------
77
Effect of Higher Sac-Volume Nozzles
As mentioned in the Background section on fuel injection modifications, the
nozzle orifice volume is a design parameter which affects particulate emissions.
To observe the effect of this change, the Mack engine with the high rate (6G)
pump was outfitted with injector nozzles of 0.77 mm sac volume for comparison
3
with the baseline volume nozzles of 0.32 mm . The orifice number (6) and their
diameter (0.29 mm) remained unchanged with the new nozzles.
Table 3.1.9 gives the brake-specific emissions for both sets of injector
nozzles. These tests were run in close succession to guarantee that ambient
test conditions were a negligible factor in creating observed differences. The
increase in HC caused by the change in sac volume was over 100%, while the
particulate solids increase was 78%. The SOF increased much less than either of
these (only 10%). This suggests that the molecular weight of the excess
carbonaceous emissions is either in the volatile hydrocarbon region or in the
3 3
Table 3.1.9 - Performance Comparison between 0.32 mm and 0.77 mm
Sac Volume Nozzles. Mack ENDT-676 No. 2 Fuel, APE-6G Pump,
22° ETC Timing, EPA Mode 3 (all units in grams/kw-hr ± 1
standard deviation)
3 3
Parameter 0.32 mm nozzles 0.77 nan nozzles
BSFC 330 ± 2 338 ± 1
BSNO 9.80 ± 0.17 9.46 ± 0.18
BSNO .06 ± 0.00 .06 ± 0.00
BSNO 6.35 ± 0.11 6.13 ± 0.11
BSHC 1.20 ± 0.02 2.58 ± 0.15
BSTPM 1.138 ± 0.056 1.815 ± 0.053
BSSOLIDS 0.873 ± 0.043 1.551 ± 0.043
BSSOF 0.1768 ± 0.0173 0.1949 ± 0.0161
0.064 ± 0.002 0.069 ± 0.003
-------
78
highly polymerized solid carbon region. A comparison of the chemical
subfractions present in the SOF (Table 3.1.10) shows that the paraffinic
fraction decreased both in absolute amount and in percentage of SOF, while
acidics, transitionals, and oxygenates all increased. These results show that
unburned fuel from larger sac volume nozzles leaves the engine after
considerable oxidation, cracking and dehydrogenation.
Table 3.1.10 - Fraction results for standard (0.32 mm ) and high
sac-volume (0.77 mm ) injector nozzles; same test
conditions as in Table 3.1.9
Subfraction
Ether insoluble
Basic
Acidic
Parraf in
Aromatic
Transitional
Oxygenated
0.32 mm
% of SOF
7.5
1.6
14.7
53.8
4.0
1.2
12.4
Hexane Insoluble 5.0
The Effect of
Injection Timing
0. 77 mm
% of SOF
6.6
1.6
17.3
44.5
4.9
2.9
16.2
5.9
and Ultra-High
0.32 _mm
mg/m
0.80
0.17
1.56
5.71
0.42
0.13
1.31
0.53
Injection
0.77 mm
mg/m
0.77
0.18
2.01
5.16
0.60
0.33
1.88
0.68
Rates
Figure 3.1.25 shows the effects of the ultra-high-rate shuttle pump on
BSFC, NO , SOLIDS, SOF, and TPM emissions. The increase of 200% in injection
X
rate and 32% in peak injection pressure at mode 3 resulted in a 5% increase in
BSFC over the high rate of injection. Advanced injection timing reduced BSFC
somewhat, but not to the level experienced with the lower injection rate.
Increased BSFC is likely due to the excessive work required to generate the high
injection rates, as well as possible spray overpenetration and impingement on
-------
79
CO £•
o S
•3 a
Ss
o
MODE 3, 3.8 IN. BOWL
0.23 _
10 14 18 22
INJECTION TIMING (°BTC)
Fig. 3.1.25 - Effect of ultr.i-hiph injection
rate and timing on BSFC and emissions,
3.8 in. bowl
1.0
a 3.5
§ J.O
3
w 2.5
3 2.0
0
1 W
a l.o
O.S
0
ENGINE SPEED:1260 RPM
MODE 3. 3.8 IN. BOWL
~
-
SOF
S04
SOLIDS
SOF
S04
COUDS
SOF
S04
SOLIDS
SOF
S04
SOLIDS
SOF
S04
SOLIDS
-
IrBTC • irBTC 14'BTC ' 12*810 fBfO
HIGH BATE L ULTRA-HIGH HATE-
Fig. 3.1.26 - Effect of ultra-high injection rate
and timing on brake specific partic-
ulate emissions, 3.8 in. bowl
-------
80
the combustion chamber wall; such fuel should contribute to increased SOF and HC
emissions, but this was not observed.
A more likely reason for increased BSFC may be an increase in the amount of
fuel injected during the ignition delay, since this should result in more
premixed combustion and less diffusion combustion, yielding lower SOF and HC.
Although this rate increase will also raise the amount of fuel premixed past the
lean limit of combustion, the decrease in SOF and HC with the ultra-high rate
suggest that increased premixed combustion predominates over mixture
over-leaning.
Figure 3.1.26 shows that the decrease in BSTPM with the ultra-high
injection rate is due largely to much decreased BSSolids, with BSSOF actually
increasing over that experienced with the high rate. Retarded injection timing
reduced both BSSOF and BSSOLIDS, except for the 8°BTC point, which resulted in
very high SOF levels, probably due to retarding the injection past the minimum
ignition delay. Excessive combustion-generated noise experienced with this
timing supports the premise that high SOF levels originate from excessive fuel
detonating due to accumulation after the end of ignition delay.
The decrease in solids emissions with timing retard, rather than advance,
as for lower injection rates, may be due to both the higher degree of premixed
combustion and lower cycle temperatures at retarded timings. Khan et al. (21)
found similar results when investigating the effects of injection timing on
smoke. With the Mack engine, however, the minimum ignition delay appears
between 8° and 12° BTC, and a more plausible explanation for lower solids near
maximum retard could be the injection of fuel into an increasingly dense air
charge. Increased air density would counteract any overpenetration effects
caused by the extremely high injection pressures experienced with the shuttle
pump.
-------
81
The effect of ultra-high rates of injection on sulfate levels with timing
retard is minimal. This indicates that a high-sulfur fuel might give sulfate
levels approximating the total mass of the TPM at an injection timing of minimum
TPM emissions.
Figure 3.1.27 shows that the ultra-high injection rates had little effect
on the paraffinic subfraction; there is little change in the total percentage of
subfractions containing oxygen, while amounts of individual oxygen-containing
subfractions were either increased or decreased with the ultra-high rate. In
terms of total SOF, however, the 17° timing gave a greater amount of all
subfractions except the aromatics and transitionals with the ultra-high rate
pump than with the high rate pump (Figure 3.1.28). A notable feature of the SOF
for the ultra-high irate at the 8° timing is the extremely high ether insoluble
content. While detailed characterization of this subfraction was not performed,
the character of this subfraction in previous work (12) suggests that it may be
high molecular weight material containing oxygen and aliphatic as well as
aromatic character.
Figures C2 through C12 of the Appendices show the trend in chemical
subfractions for the high rate (6G) pump at modes 3, 5, 9, and 11, and for the
ultra-high rate (shuttle) pump at mode 3, as a function of injection timing.
These figures give percent concentration of total SOF for eight subfractions as
a function of timing (Figures C2-C6) and as a function of load (Figures C7-C12).
It should be mentioned that the fractionation procedure is described by
Funkenbusch et al. (12) is a manual procedure and is prone to subjective
variability which decreases measurement accuracy. Additionally, the measurement
of SOF by particulate extraction adds a variability. Still, there are obvious
similarities between the high rate and ultra-high rate pumps as timing is
retarded at mode 3. Paraffins were generally decreased, both as a proportion of
-------
82
V)
w
z
g
H
00
o
0.
5
O
O
LL.
O
100
80 —
60 —
40 -
20 —
—
—
HIN-HEXANE INSOLUBLE OXY-OXYGENATED ARO-AROMATIC
BAS-BASIC EIN-ETHHR INSOLUBLE TRN TRANSITIONAL
PAR-PARAFFIN ACD-ACDIC
HIN-4.5
OXY-11.1
— TRN-2.R
PAR-11.5
ACD-16.9
EIN-46.6
OXY-16.5
TRN-4.3
ARO-27.3
PAR-25.1
ACD-4.8
EIN-16.6
OXY-18.1
TRN-3.0
ARO-14.1
PAR-S2.4
ACD-5.1
EIN-4.5
OXY-16.9
TRN-9.4
ARO-5.7
PAR-46.3
ACD-6.2
EIN-11.6
8°BTC 12°BTC 14"BTC 17" BTC
4 III TRA-HlfiH RATP— k
OXY-27.5
TRN-3.2
" ARO-2.2
PAR-46.1
ACD-3.9
BAS-6.8
EIN-5.8
17" BTC
• HIGH RATE
Fig. 3.1.27 - Effect of ultra-high injection rate
and timing on the chemical compos-
ition of the SOF, 3.8 in. bowl, mode 3
14
12
ct
5
i 10
z
o
5 8
cc
111 6
O
z
o
« •
5
2
0
HIN-0.7 —
OXY-1.7-
TRN-0.4 —
ARO-0.7 —
r
PAR-1.7 J
ACD-2.6
BAS-0.3 —
-
EIN-7.0-1
••
•
HIN-0.3 —
TRN-0.5—
ARO-3.0 —
PAB-2.8 —
ACD-0.5 —
BAS-0.3 —
l-IN-1.8 —
OXY-2 2 —
TRN-0.4 —
ARO-1.7 —
PAR-6.3 —
BAS-0.2-L
E1N-0.5 —
OXY-2.5 -
TRN-1 4 —
ARO-0.8-
PAR-6.7-
ACO-0.9 —
BAS-0.4-
EIN-1.7-
12
14
17
INJECTION TIMING («BTC)
Fig. 3.1.28 - Effect of injection timing on the
mass emissions of the chemical sub-
fractions, mode 3, 3.8 in. bowl,
ultra-high rate
-------
83
total SOF, and in terms of total mass and power-normalized emissions, as timing
was retarded. This is consistent with less fuel mixed past the lean combustion
limit, to leave the cylinder unchanged; injection rate does not seem to affect
this phenomenon greatly. Oxygenated 'Compounds are also decreased as timing is
retarded for both pumps, which is contrary to the mode 9 observation with the
high-rate (6G) pump. This suggests that mode 3 conditions are characterized by
less fuel mixed past the lean combustion limit than mode 9 (due largely to less
air swirl), and the phenomenon of fewer lean regions with timing retard
predominates over the competition of free radicals for monatomic oxygen
mentioned previously for mode 9 conditions.
Another similarity between the two injection rates at mode 3 include a
general increase in ether insolubles with retard past 14°. This subfraction
could represent partially pyrolyzed fuel that originates from incomplete
atomization due to relatively poorer mixing at retarded timings. It is not
pyrolyzed completely to carbon because peak temperatures do not remain in the
solids formation region (2000°-2400°F) long enough for this to occur.
If one observes the trends of subfraction percentages as a function of mode
for the 6G (high rate) pump, one sees that the oxygenated fraction is relatively
independent of injection timing. Percent oxygenates in the SOF was generally
highest at mode 9, with modes 11, 3 and 5 showing decreasing amounts of
oxygenates. This independence of percent composition with timing was not
observed with any other subfraction and may be due in part to more reproducible
mass determination for this subfraction.
Although only two points are available for the load-composition trends, the
consistency of decreased transititonals and oxygenateds as well as increased
hexane insolubles and aromatics with increased load at intermediate speed
(Figures C7-C9) is surprising. This suggests that timing is relatively
-------
84
unimportant in the direct relationship between increased load and increased
dehydrogenation of SOF precursors. The dehydrogenation of fuel should increase
insoluble and aromatic subtractions, while addition of oxygen (favored at
lighter loads) will increase transitional and oxygenated subfractions. Rated
speed modes (figures C10-C12) did not show these trends.
The biological activity of the SOF with ultra-high injection rates was
increased over that with high rates, and was increased as timing was retarded
(Figure 3.1.29). This is contrary to the timing-specific activity trend of the
high injection rates. For both pumps, however, the SA follows the same trends
as NO. concentrations, which supports the hypothesis that nitration of organics
may increase the activity of the SOF. As before, the acidic and transitional
were the most active subfractions, while paraffins and aromatics were
essentially unresponsive.
Figures 3.1.30 and 3.1.31 show the relationship between NO and N09,
X f-
respectively, with the BSFC and BSTPM as well as the BSFC-BSTPM combination
parameter for both pumps, as timing is varied. The BSNO and BSNO- trade-off
X £~
with BSFC and BSTPM shows the ultra-high rate pump to be better from an overall
standpoint; but if fuel consumption alone is considered, the ultra-high rate
pump is inferior to the high rate pump.
Recommended Fuel Injection Characteristics
Theoretically, increased mean injection rates and pressures, by allowing
retarded timings, can improve BSFC by permitting the thermodynamic ideal of
constant volume combustion to be approached. Additionally, high injection
pressures result in: short ignition delays, leading to reduced "lean limit"
HC's; reduced peak cylinder temperatures and NO emissions; and reduced
X
smoke/particulate levels.
-------
85
MODE 3. 1260 RPM, 3.8 IN. BOWL
sr
to x
in :r
10 c
i
o
sl
400
300
200
100
0
4
3
2
1
o"
J I 1 1 1 1 1 1 1
.00
""•"••s^ ULTRA-HIGH RATE
'. ""-v
D ^J3
HIGH RATE
T Q
D
~"""~-~.^ ULTRA-HIGH RATE
•~ HIGH HATE 0 . ..P
. I I i i r i i i i
10 12 14 16 18 20 22 - 24
INJECTION TIMING (BTC)
Fig. 3.1.29 - Effect of ultra-high injection rate
and timing on biological activity,
3.8 in. bowl, mode 3, tester strain
TA1GO, WiLnuuL 3—9
MODE 3, 1260 RPM, 3.8 IN. BOWL
MODE 3, 1260 RPM, 3.8 IN. BOWL
o
00
m
1.6
1.5
1.4
1.3
.28
.55
.35
.25 -
ULTRA-HIGH HATE
ULTRA-HIGH RATE
- D-
HIGH RATE
ULTRA-HIGH RATE
I I "
OC
- .26 O
a
10
CQ
10
12
14
BSNOx (G/KW-HR)
Fig. 3.1.30 - Effect of ultra-high injection
rate and timing on BSTPM,BSNOX,
and BSFC relationships, 3.8 in.
bowl, timing variable
•6 .8 1.0
(g/kw-hr)
Fig. 3.1.31 - Effect of injection rate on
BSTPM, BSNO,, and BSFC relation-
ships, 3.8 in. bowl, timing vari-
able
-------
86
If mean injection rates are important, the instantaneous rates are perhaps
even more important. The instantaneous rates may be analyzed using a rate vs.
time curve, as in Figure 3.1.32. This shows the two injection rate shapes
generally used in fuel injection systems. Advantages of the triangular shape
(Figure 3.1.32a) are low HC and noise due to low initial rates, while
disadvantages include smoke and solids contributed by initial and final rates.
The square rate shape (Figure 3.1.32b) provides lower peak pressures (and NO ),
X
and can reduce EC's by providing a rapid final cut-off; however, it has the
disadvantage of a high initial rate of injection, leading to noise, HC
emissions, and structural loading.
Figure 3.1.33 shows the local histories of $, temperature, solids, HC, and
NO masses expected to be contained in an infinitesimal packet of injected fuel.
This approach for combustion analysis was developed as a basis for computer
modeling (22, 24), and these histories have been based on information gained
throughout the study. Histories of three different fuel packets are shown, each
formed during a different phase of the injection process. Fuel injected during
the initial stages of injection will follow either path 1 or path 1', depending
upon initial injection pressures. If the initial pressure is too high, path 1
will be followed. The packet mixes past the lean limit of combustion, does not
ignite, and contributes to HC emissions. Packet 1' is followed if initial
pressures are too low. This packet fails to mix well and $ remains high. The
packet contributes to solids formation since its temperature never rises into
the solids oxidation region.
Packet 2 is formed during the main period of injection. Since it is
injected during the peak cycle temperature, it is heated well above the solids
formation region and particulate oxidation occurs. At this point, NO formation
-------
87
<
o
5
cc
o
I
o
o z
Z8
TDC
CRANK ANGLE
a) Triangular Pulse
<
O
cc
o
o
Ul
-20
TDC
CRANK ANGLE
b) Square Pulse
8d
3S
w o
£
^ 2400
| 2000
IS
INJECTION RATE
3
^ '
TDC
80
CRANK ANGLE (DEGREE)
a 3 1 32 - Schematic o£ the two general
' rate shapes of the injection rate
vs. crank angle curve
Fig. 3.1.33 - Histories of infinitesimal fuel-
air packets formed at various stages
of injection
-------
88
begins until the temperature drops below 2000°F. Due to the high temperatures
and <(>£l, this packet does not contribute to HC or SOF emissions.
Packet number 3 is formed near the end of injection under conditions of low
injection rate and pressure. Solids form rapidly due to low and high
temperatures; these solids will be exhausted from the cylinder. Little or no HC
or SOF emissions result from this packet due to high temperatures.
This discussion demonstrates the importance of temperature on HC, solids,
and NO emissions. Temperature, however, is dependent upon rates of fuel
X
delivery and mixing. Figure 3.1.34 shows the ideal rate shape based upon
fundamental considerations, while Figure 3.1.35 is a schematic of ideal
temperature histories resulting from this ideal rate shape; packet numbers refer
to the same injection period as they did in the previous discussion.
It has been found, for the Mack engine, that injection of fuel at pressures
less than about 40 mPa will result in excessive smoke levels (54). Therefore,
the initial injection period should just exceed this level until after ignition,
to ensure atomization. A packet of fuel injected during this period will follow
path 1 and autoignite at the end of the ignition delay. The temperature rises
rapidly and then falls due to relatively cool surroundings, contributing little
to solids and NO formation. Since peak temperatures are very high, all HC
emissions should be consumed. The duration of initial injection corresponds
approximately to the ignition delay period, and may be reduced relative to the
total injection duration as load is increased for a turbocharged engine.
Parker (55) has suggested a maximum injection pressure during the main
injection period of 100 mPa. Peak pressures should be as high as possible to
minimize the total injection period, but should not cause loss in efficiency due
to excessive energy requirements. At light loads, Parker has suggested mean
injection pressure of 50 mPa to avoid overpenetration, with pressures increasing
-------
INJECTION RATE
MM'/STROKE
1
1
01
m
3 ra M
TO )-• 3
i_n c
O Ui 0
-
o c- ^
7T OJ
O ft rf
O D T3
• W rl
r-j •• o
»o <
^J•
CL
I
O
Z
o
r-
m
3
m
O
33
m
m
o
o
o
m
en
Ol
CO
o
m
3,
O
o
m
_JO
01
o
o
O
>
a
oo
-------
90
IU
oc
H-
UJ 2400
Q.
UJ
*"" 2000
O
o
- 200
s
5
o
20
it
O
^ 10
5
I I I I I
PACKET 1
REGION OF NO
PAPKFT 7 FORMATION AND
A SOLIDS OXIDATION
' I SOLIDS
/ \ FORMATION
I / \ REGION A
\ / Nx/ \ PACKETS
V s ^^^ ^
I 1^
I / ^
/ /
I*'
^:
^*~ ~~^
./^ ACCUMULATED FUEL
^^^
- PACKET NO.
2
~ J 1 3 INJECTION RATE
) i T i i i i
TDC 20 40 60 80
CRANK ANGLE (DEGREE)
Fig.
3.1.35 - Proposed temperature ranges for
optimum BSFC and emissions
-------
91
rapidly with speed and/or load exceeding 40% of the rated value. Fuel injected
during the peak of such an injection cycle will follow path 2 and pass through
the solids formation region. If air-fuel mixing is sufficient to give a rapid
decrease in $, little net solids formation should result. By retarding the
timing as late as possible without reducing the packet's expansion ratio, peak
temperature and NO emissions are reduced.
The cut-off during the final period of injection would ideally be
instantaneous, and should not exceed one degree crank angle (55). Fuel injected
during this period will follow path 3 and will always form solids due to high
cylinder temperature. Perhaps the only way to reduce solids formation from such
packets is to decrease their number by as rapid an injection cut-off as
possible.
Injector nozzle orifice number and their size determine spray geometry, and
higher swirl speeds at higher speeds and loads may require more, nozzles of
smaller diameter to maximize air utilization and reduce spray overpenetration.
A variable orifice nozzle should enable one to use constant injection pressures
over a wider speed range and maintain pptimum mixing and penetration
characteristics. If the same injector nozzle orifices are used with the ideal
injection rate curve as were used for the low rate and shuttle pumps during this
study, the recommended injection rates (as calculated from coefficients of
3
discharge of the nozzles) would be 9.4 and 13.3 mm /°CA. These rates correspond
to 40 and 100 mPa injection pressure, respectively. Although the nozzles used
were not variable-orifice in nature, these pressures should give spray
geometries appropriate for swirl conditions experienced during a moderate
portion of the load cycle.
-------
92
Optimum Injection Characteristics
With the recommended injection rate shape as shown in Figure 3.1.34, BSFC
should be kept low by minimizing energy required to generate injection
pressures. Particulate emissions may be controlled by: 1) using a zero sac
volume nozzle in combination with a minimum of fuel injected before ignition to
reduce SOF; and 2) injecting as quickly as possible with rapid mixing to reduce
solids. Since initial rates of injection largely determine the SOF-solids trade
off, the initial rate should be just high enough to avoid poor mixing yet not so
high as to contribute to solids. To reduce NO emissions, lower maximum cycle
X
temperatures should be reached. This is achieved by retarded timing and by
injection rates high enough to avoid particulate formation but low enough to
maintain reasonable fuel consumption.
While a variable orifice nozzle may not be necessary, its elimination would
require variable injection pressure for varying speeds, and would reduce control
over the solids. Variable timing control may not be critical, but it would
allow better control of NO . Achieving so high a degree of control over the
X
injection characteristics may not be possible without a set of electronically
triggered, hydraulic actuated unit injectors. This type of injector uses a
high-pressure (< 20 mPa) fuel supply to operate a differential area piston, and
the injection pressure is a function of the differential areas of the piston and
the supply pressure. Although the shuttle pump used in this study approached
the mean pressures required, it used a triangular pulse. Although a square-rate
cam might improve BSTPM and HC by giving a fast injection cutoff, the hydraulic
actuated injectors might still be the best system.
Assuming that the ignition delay, in crank angle degrees, is constant (56),
one may use the measured fuel consumption of the various modes to predict the
injection duration and hence the mean injection rates the ideal system would
supply. Also, assuming that equal changes in injection rate result in similar
-------
93
emissions characteristics, one may predict the increase in BSNO that is
X
expected. This would allow predicting the amount of injection timing retard
necessary to reduce the BSNO to the desired level. Next, by assuming the
X
emissions trends to remain similar to those tested previously, one could
calculate the BSSolids and BSFC that would result from this system.
This has been done, and the results are shown in Table 3.1.11. One should
be very cautious when using these predicted values, as many assumptions have
been made. These assumptions greatly simplify the complex process of diesel
combustion, and are based on a limited amount of experimental data.
Table 3.1.11 - Fuel Injection Rates, Injection Timing, and Emissions
Expected to Result from Proposed Ideal Rate Characteristics
Expected Timing for Expected Expected
Fuel F/0mean F/0(a' BSNOx @19 ETC BSNOx -10 solids BSFC
Mode
3
4
5
9
10
11
(mm )
69
124
180
155
110
77
(°CA)
7.3
10
14
12.2
8.9
8.2
(mm /°CA)
9.4
12.5
12.8
12.7
12.4
9.4
(% increase) I
22
29
20
63
79
59
(g/kw-hr)
12.65
15.1
13.8
12.5
12.3
9.9
(g/kw-hr)
14
13
14
15
16
19
(g/kw-hr)
.27
.20
.42
.02
0
0
(kg/kw-1
.228
.220
.221
.264
.278
.337
(a) relative to the high-rate, 3.8 inch bowl system at 19°BTC
The injection timings were selected so as to yield a flat BSNO level over
X
the EPA modes listed. The BSSOLIDS and BSFC values are the predicted values
resulting from the ideal mean rates of injection and the injection timing.
These values have been averaged over the six modes and compared to the averages
resulting from the 3.8 in. bowl, high rate system at 19°BTC timing. The overall
result is that for a BSNO level of 10 g/kW-hr, the recommended injection
X
characteristics resulted in a 70% reduction in BSSolids and should also reduce
the SOF significantly. This would then yield an overall reduction in the TPM.
-------
94
The BSFC should remain almost constant, as compared to the high rate system. It
should be noted that these BSSOLIDS levels were predicted based on past trends,
but if the correlation for Dent's mixing rate is used (84), similar BSSOLIDS are
predicted.
It is expected, however, that the elimination of the high peak injection
rates and pressures will yield lower NO levels than those indicated. This
would allow advancing the injection timing until NO reaches the maximum desired
level and result in improving the predicted BSSOLIDS and BSNO levels.
X
-------
95
CATERPILLAR 3208 - AFTERTREATMENT DEVICE STUDY
Engine Specifications
The engine used for all studies of exhaust aftertreatment was a Caterpillar
3208 medium-duty diesel truck engine. The engine was received new and was
run-in for a total of 50 hours prior to emissions measurements. The engine
specifications are given in Table 3.2.1.
Table 3.2.1. Caterpillar Engine Specifications
Engine Manufacturer and Model: Caterpillar 3208
Engine Type: V-8, Direct Injection, Naturally Aspirated
Bore X Stroke: 114mm x 127mm (4.5 in. x 5.0 in.)
Displacement: 10.4 liter (636 cu.in.)
Rated Power: 157 KW (210 hp) at 2800 rpm
Rated Torque: 654 N M (485 FT LBF) at 1400 rpm
A total of four different aftertreatment devices were utilized with this
engine. These will be discussed separately, and the relative merits of each
device will be discussed in the second section of the conclusions.
Englehard PTX Oxidation Catalyst
Engine Test Set-Up - The first aftertreatment device (an oxidation
catalyst) chosen for evaluation was an Engelhard Corporation model
PTX-D-616-300NKG platinum exhaust gas purifier. The specifications of the
catalyst are given in Table 3.2.2. This catalytic converter is considered
typical of oxidation catalysts that could possibly be used on diesel powered
vehicles in the near future. Since the Caterpillar 3208 engine is a V-8, a
separate catalytic converter was installed for each bank of cylinders.
-------
96
Table 3.2.2. Oxidation Catalyst Specifications (57)
Manufacturer Engelhard Industries Div.
Engelhard Carp.
Model PTX-D 616-300NKG
Catalytic Agent Platinum
Catalyst Loading <3500 gm/m
Wash Coat Alumina
Substrate Material Cordierite
Substrate Type Laminar Flow Monolith
Substrate Cell Size 46.5 cells/cm
Substrate Size 14.6 cm dia x 15.24 cm long
The Engelhard PTX catalysts were installed as close as possible to the
exhaust manifolds of the engine. This was done to allow the exhaust to enter
the catalyst at the highest possible temperature. Blank tubes were fabricated
to replace the catalytic converters for baseline testing. Uncatalyzed monoliths
had been furnished by Engelhard to be used for baseline testing, but it was felt
that their use would not result in an accurate evaluation of the catalyst and
substrate as a system. The addition of an uncatalyzed monolith to the exhaust
system would have resulted in a large increase in surface area that the exhaust
components would have been subjected to, and may have resulted in the catalysis
or quenching of reactions occurring in the exhaust system. Approximately 80
hours of use were accumulated on the catalysts before actual testing was begun,
as recommended by the catalyst manufacturer (57).
It was discovered during preliminary testing that the exhaust system can
act as a large storage place for diesel particulates. Of interest was the
amount of sulfates stored when operating with the catalyst installed. Large
variations in total mass and sulfate concentrations were observed during early
testing when tests without the oxidation catalysts were conducted soon after
tests with the catalyst had been completed. It is believed that these
variations were caused by the re-entrainment of particulate matter (probably
sulfates (25,26))that had been deposited on the exhaust system walls from
-------
97
prior tests with the catalysts installed. Exhaust system conditioning times
were then increased when tests were to be conducted at different conditions than
those run immediately prior to allow for system equilibrium to be obtained at
the new test conditions. The conditioning time was increased to at least 2
hours when a mode change was made and to at least 8 hours when the catalysts
were either removed or installed.
The three fuels used for this evaluation were AMOCO Premier No. 2 (58), a
specially blended No. 1 fuel supplied by Chevron Research Company (59), and
shale-derived fuel oil supplied by the Department of Energy (DOE). The
properties of these fuels were given in Table 2.2. All of the shale oil
properties were determined by Chevron (59) , except for the percent sulfur and
nitrogen (by DOE). AMOCO Premier No. 2 fuel was chosen because it represents a
typical No. 2 grade fuel with approximately the national average sulfur content
of 0.23% by weight (60). The Chevron No. 1 fuel was chosen for this evaluation
because its lower sulfur content (0.04% by weight) may be desirable for use in a
catalytic converter equipped diesel-powered vehicle. These two fuels were
similar to those used by Frisch et al. in a previous study of fuel effects done
at MTU (12,14). The shale oil represents a highly refined fuel which could
become a future alternative to the extensive use of petroleum products in
diesel-powered vehicles. This particular fuel was refined from Paraho crude
shale oil by Standard Oil of Ohio as part of a joint program with DOE and the
U.S. Navy (61). In general, it meets military specifications and has good
storage capabilities.
A test matrix was devised to allow for comparison of the three test fuels.
Because of the limited quantity of shale fuel available for evaluation, testing
was limited to Mode 4 of the EPA 13-mode steady-state cycle. Results reported
earlier (31) indicated that the oxidation catalyst used for the evaluation had
-------
Table 3.2.3. Test Matrix for PTX Oxidation Catalyst Evaluation
EPA MODE 3459 10 11 3454
SPEED (rpm) 1680 1680 1680 2800 2800 2800 1680 1680 1680 1680
LOAD (N-m) 160 320 480 399 266 133 160 320 480 320
BMEP (kPa) 192 383 575 485 322 161 192 383 575 383
FUEL No. 2 No. 2 No. 2 No. 2 No. 2 No. 2 No. 1 No. 1 No. 1 Shale
CATALYST w/o with w/o with w/o with w/o with w/o with w/o with w/o with w/o with w/o with w/o with
Total
particulate,
SOF, and XXXXXXXXXXXXXXXXXXXX
SO mass
emissions
HC, NO, and
NO mass X XX XX XX XX XX XX XX XX XX X
emissions
Chemical
charac- x XX XXXX XX XX XX XX XX XX X
terization
of SOF
Ames
mutagenicity x XX XX XX XX XX XX XX XX XX X
bioassay
on SOF
Ames
mutagenicity XX X XX XX
bioassay on
subfractions QO
-------
99
the greatest effect on the physical, chemical, and biological character of the
emissions at Mode 4. This condition (50% rated load and 1680 rpm) was chosen
for the fuel evaluation. The quantity of No. 1 fuel available was also limited,
allowing for tests with and without the catalyst for this fuel only at Modes 3,
4, and 5. Sufficient quantities of No. 2 fuel were available for testing at
modes 9, 10, and 11 in addition to modes 3, 4, and 5.
The experimental test matrix is given in Table 3.2.3. Due to the failure
of an injection pump during testing of shale fuel, the baseline emissions with
which shale fuel emissions should be compared are those run prior to testing of
the close-coupled catalyst, discussed in the next subsection. The dilution and
filter sampling parameters observed with No. 2 fuel are listed in Tables 3.2.5
to 3.2.7. The divergence of the mode 9, without catalyst test from the EPA
specified maximum sampling temperature of 52 C was due to unusually high
dilution tunnel inlet temperature. Since the sample temperature was not too far
above 52 C, the results obtained are acceptable for research purposes.
Table 3.2.4. Ambient Test Conditions
Ambient Temp. Barometric Pressure Specific Humidity
(°C) (kPa) (g H20/kg air)
MODE unc. cat. unc. cat. unc. cat.
3
4
5
9
10
11
29
31
31
37
25
32
30
32
31
22
27
27
98.9
99.2
99.0
99.4
101.5
99.4
99.9
99.6
100.3
98.6
98.9
98.8
4.5
2.0
3.1
6.2
2.1
9.8
3.0
2.7
2.4
7.4
4.3
2.3
-------
100
Table 3.2.5. Dilution Tunnel Conditions
MODE
Volume Dilution
Ratio
unc. cat.
3
4
5
9
10
11
15.0
14.7
14.9
15.3
14.9
15.0
15.1
14.7
14.4
15.0
15.0
14.2
lilutio
n Temp.
NO ^Filter
( C) X(ppm)
unc.
39
48
51
60
50
49
cat.
40
47
50
52
51
48
unc.
32.7
64.2
84.3
63.5
48.1
29.8
cat.
33.4
60.6
80.9
73.7
48.4
28.7
NO ©Filter
(ppm)
unc.
2.0
1.3
1.3
0.5
1.0
0.9
cat.
1.4
19.5
8.4
1.1
1.7
3.9
Table 3.2.6. 47mm Filter Operating Conditions
Sample Time Sample Rate Face Velocity
(min.) (% Isokinetic) (m/s)
MODE unc .
cat.
3
4
5
9
10
11
32
42
40
11
39
21
30
53
32
15
10
16
unc.
73
86
83
81
79
81
cat.
84
84
83
61
98
59
unc.
0.28
0.33
0.32
0.31
0.30
0.31
cat.
0.33
0.32
0.31
0.24
0.35
0.23
'i
0
0
0
0
0
0
Iter
m|
cm
unc.
.175
.201
.216
.218
.424
.318
Loading
TPM
stain
cat.
0.122
1.299
1.324
0.493
0.268
0.341
Total Mass
unc.
2
2
3
3
6
4
•
,
•
*
•
a
54
92
14
17
16
62
(mg)
cat .
1
18
19
7
3
4
•
.
,
.
•
B
77
86
22
16
89
95
Table 3.2.7. Engine & Catalyst Operating Conditions
MODE
3
4
5
9
10
11
A/F
unc.
66
41
29
24
36
51
cat.
67
41
31
25
35
51
Exh.
('
unc.
252
365
499
642
474
377
Tempva'
3c)
cat.
258
372
481
666
490
374
Back Pressure
(kPa)
unc. cat.
101 104
101 104
102 105
104 108
105 106
104 106
Exh.O v
(%J
16.1
13.1
10.8
8.4
11.9
14.7
Cat. Temp
(°C)
269
383
498
695
507
391
Space Vel.
x 10
(hr X)
167
200
242
450
349
295
(a)
(b)
Measured at exhaust manifold (inlet to catalyst)
, . Calculated assuming ideal combustion
M«««..^«J a(- center of catalyst monolith
-------
101
Effect of Catalyst on Mass Emissions - As shown in Table 3.2.7 and in
Figures 3.2.1 and 3.2.2, the use of the catalyst had little effect on the
operating parameters of the engine. The greatest effects the catalyst had were
a slight increase in exhaust backpressure and an increase in exhaust opacity at
modes 4 and 10 (50% rated load at intermediate and rated speeds) because of
increased particulate emissions. The use of the catalyst had no effect on the
brake specific fuel consumption or on the air/fuel (A/F) ratio of the engine.
Figures 3.2.3a and b show the effect of fuel on NO , NO and equivalent NO
X £.
emissions measured with and without the catalysts, respectively using No. 1 and
No. 2 fuels. Figure 3.2.3c shows these same emissions for rated speed modes,
with No. 2 fuel only. In general, the NO , NO, and equivalent NO emissions
X £-
using the No. 1 fuel follow the same trends as those observed using the No. 2
fuel for both catalyzed and uncatalyzed tests. The NO , NO, and NO- emissions
X L.
measured both with and without the catalysts using No. 1 fuel were all lower
than those measured using the No. 2 fuel. This difference is probably due to
changes in the characteristics of the fuels. For example, although both fuels
had similar cetane indexes (indicating similar ignition delay times assuming the
cetane index approximates the true cetane number), the No. 1 fuel contains a
lower percentage of aromatic than the No. 2 fuel. Fuel aromatic content has
been shown to affect NO emissions of diesel engines (62,63,64).
X
The brake specific NO and NO concentrations by volume (ppm) were
X X
unchanged by the catalyst. The levels of N0_ emissions were, with the exception
of mode 3, increased by the catalyst, depending on the engine operating
condition. This increase in NO was at the expense of NO emission rates and
concentrations, which were reduced.
The changes in conversion of NO to NO- by the catalyst (neglecting other
changes in exhaust gas composition) as related to catalyst temperature are shown
in Figure 3.2.4. The conversion of NO to NO- by the catalyst (calculated as the
-------
•n
£
EXAUST OPACITY <%)
I
•o
BRAKE SPECIFIC FUEL CONSUMPTION (kg/kW x hr)
to
in
W
o
£0
en
3,
X >
to
m
•u
o
hO
-------
103
~ rVr
£
^
o>
CO
g
to
CO
5
ui
CM
o
z
Q
0
X
o
z
o
LU
O
LU
D.
CO
LU
EC
CO
15
10
1.0
0.5
•Vr
NOx NO. 2 FUEL
[r —
r
NOX NO. 1 FUEL
NO NO. 2 FUEL
CATERPILLAR 3208
CATALYZED
1660 RPM
Fig
200 300 400 500 600
BMEP (KPa)
3 2 3a - Effect of fuel on brake-specific
' ' NO, and N02 with the PTX oxidation
catalysts
15
10
2 5
I
2
to
o
55
to
j. NOX NO. 1 FUEL"
: o-
- NO NO. 1 FUEL*
1.0
CM
O
z
o
o
z
X
o
O 0.5
Q.
to
Ul
cc
m
NOX NO. 2 FUEL
—
-A
NO NO. 2 FUEL
-O
CATERPILLAR 3208
UNCATALYZED
1680 RPM
NO2 NO. 2 FUEL
200
300 400
'BMEP
500
600
Fig.3 .2.3b - Effect of fuel on brake-specific NO ,
NO, and N02 without the FIX oxidation*
catalysts
UNCATALY2EO-
3NO CATALYZED
Fig. X2.3c - Brake specific NO, NO2, and NOX emis-
sions with and without oxidation catalysts,
rated speed (2800 rpra)
-------
104
CATERPILLAR 3208
CATALYZED
NO. 1 FUEL
O NO.2FUEL
U )EPAMODE
\I STANDARD ERROR
CO,
M
O
6
z
U.
O
O
to
cc
Ul
O
U
•10
CAT
200 300 400 500 600 700
CATALYST BED TEMPERATURE (°C)
Fig. 3.2.4 - Conversion of NO -> NO^ by catalyst as
a function of temperature at the center
of the the monolith, At based on space
velocity using gross catalyst volume,
NO- no finalized by the factor
NO /NO to account for effect of
, xunc ,xcat
changes in NO concentration
-------
105
fractions of NO converted per unit time)* peaks at approximately 400°C. The
catalyst actually reduces some NO back to NO at temperatures below 275°C.
Fuel change effect on HC emissions (with and without the catalysts) is
shown in Figure 3.2.5a, while Figure 3.2.5b shows the effect of the oxidation
catalyst at the rated-speed modes using No. 2 fuel only. The catalyst was
effective in reducing the HC emissions of the engine for all modes and both
fuels. There was no change in fuel-metering components or calibration, as the
same nozzles were used with each fuel. The uncatalyzed brake-specific HC
emissions were lower when No. 1 fuel was used. This is possibly due to the
higher volatility of the fuel resulting in more complete combustion because of
a higher rate of mixing of the more volatile No. 1 fuel. The brake-specific
HC emissions when the catalysts were used were higher for the No. 1 fuel than
the No. 2 fuel, indicating the percent reduction in HC emissions was not as
great with the No. 1 fuel as compared to No. 2 fuel. This indicates that some
physical or chemical characteristic of the No. 1 fuel affects the conversion of
gaseous HC in the catalysts.
The brake-specific particulate emissions measured at 1680 and 2800 rpm are
shown in Figures 3.2.6a and b, respectively. For both uncatalyzed and catalyzed
tests, the use of No. 1 fuel resulted in lower TPM, sulfate, and solid emissions
than No. 2 when the catalysts were used. This follows the similar trend in HC
emissions discussed above.
*A more elegant method of expressing this quantity would be the form
gm moles NO reacted
sec-gm Ft
However, this does not account for the changes in inlet NO concentration and
mass flux. The gram of Pt term is not included because it is constant throughout
these tests and is a proprietary value. We believe the method of presentation
used adequately illustrates the general effects of temperature on
the conversion of species on or above the catalytic surfaces.
-------
BRAKE SPECIFIC HC EMISSIONS (g/kW x hr)
ro
b
ro
in
u
b
CO
en
» re
7? **
n ^
CO
2
m
•o
tJ
09
BRAKE SPECIFIC HC EMISSION (g/KWxhr)
ro
b
Ol
b
>
0
WdU
BO
-------
107
2.0
a>
UJ
3
o
cc
a.
O
C
o
u
a.
co
UJ
cc
co
1.0
NO.1
. O
UJ
N
>•
^.
- O
Z 0
3 Ul
f— i N
u.
O
CO
J
K-
0
U.
0
CO
N0.2
o
UJ
N
^
^
< Q
O UJ
§S
Q.
O
CO
<
< "•
o °
£ o"
O (/)
CO / '
/ Q
— ' Ij
— /co
3
(269)
NO.1
o
UJ
j Q
2 N
< >
O "*
VsP <
VS. °
\ ^
N
4
NO.2 NO.1 x
u.
O
"u7
o
CO
l}
U ^J
O
CO
Q
O
CO
i CO
\
U. UJ
O N
CO >
^
h-
O
Z
a
u_ OO O
O^ CO CO CO
OT^f=4\ 1
\ i'» \\r4-
"<*• \ 9 \ —
o— ^p>- y.. \
— ' ' i I I I I
5
N02,
1^-
O
f/1
a
O
OT
o
N
<
O
(383) (498)
EPA MODE
(CATALYST BED TEMPERATURE.-C)
Fig. 3.2.6a - Effect of fuel and ?TX oxidation catalysts
on brake specific particulate emissions,
1680 rpm
-------
BRAKE SPECIFIC PARTICULATE EMISSIONS (g/kW x hr)
11 10 9
(391) (507) (695)
EPA MODE
(CATALYST BED TEMPERATURE (°C))
so;
SOLID SOF UNCATALYZED SOF
SOLID I SO4 I CAT.
^ W O3
f: co ro
O o o
O CO
CO -n ~C ^
SOLID SOF UNCATALYZED . SOF ^ 5
SOLID I S04 I CATALYZED F
o
m
m
S04 SOF
SOLID I UNCATALYZED JSOF
SOLID S04 CATALYZED
o
00
-------
109
The effects of the oxidation catalyst on the mass emission rates of
particulate matter are also given in Figures 3.2.6a and b. As discussed
earlier, the total particulate emissions are composed of the sum of the SOP,
sulfate, and solids emitted by the engine.
The use of the catalyst reduces the amount of SOF emissions of the engine
at all of the test modes. Since the amount of SOF found in the particulate at a
given engine speed and load is a function of the dilution parameters and raw
exhaust HC concentration, the reduction in particulate SOF is related to the
reduction in HFID HC emissions by the catalyst.
As shown in Figures 3.2.6a and b, the amounts of sulfate emissions are
increased with the catalyst installed. This phenomenon has been observed with
both catalyst equipped spark ignition and diesel engines. The percent
conversion of fuel sulfur to exhaust sulfate on a mole basis is shown in Figure
3.2.7.
The use of the catalyst also increases the amounts of total particulate
emissions at 5 of the 6 modes tested. The only mode that showed a decrease in
total particulate emissions is mode 3., This mode is characterized by a low
exhaust temperature, which reduces the catalyst's activation of the oxidation of
SO. to SO., and also by a relatively high baseline SOF emission rate, which
causes even a moderate reduction in SOF (61%) to result in a significant
decrease in total particulate emissions.
Close examination of Figures 3.2.6a and b reveals that the increase in
total particulate emissions is not due entirely to the observed increase in
sulfate emissions alone. An increase in the unextractable residue (largely
solid) of the particulate at most of the modes of the test cycle used was
observed.
-------
110
UJ
_J
<
O
100
O
O
d, 50
ff
V)
X
UJ
o
UJ
u.
u.
o
z
o
DC
UJ
Z
o
o
cc
10
1.0
2800 RPM
CATALYZED
RPM
CATALYZED
3208 CATERPILLAR DIESEL
NO. 2 FUEL
2800 RPM
UNCATALYZED
1680 RPM
UNCATALYZED
200
300 400
BMEP (kPa)
500
600
Fig. 3.2.7 - Percent conversion of fuel sulfur to
exhaust sulfate (on a mole basis) with and
without the PTX oxidation catalysts
-------
Ill
The solid portion may contain species such as graphitic carbon, high molecular
weight hydrocarbons, and other compounds that are insoluble in DCM, and that may
or may not be soluble in the water/isopropyl alcohol solution used for the
sulfate extraction. Some of these compounds might be liquids, but the bulk of
them are expected to be solid. The increase in opacity observed at modes 4 and
10 with the catalyst installed would suggest that the measured increase in
"solid" emissions could be due to an increase in these types of compounds.
The increase in backpressure with the catalyst installed (Table 3.2.7)
might cause the increase in solids. It is not felt that the_backpressure
increase was large enough to account for the measured change in solids. If the
backpressure increase was acting to increase the solid portion, the effect
should be similar to that of exhaust gas recirculation (EGR). The effects of
EGR on a diesel engine include a decrease in the air/fuel ratio and a decrease
in NO emissions; these phenomena were not observed in these tests. Therefore we
x v
feel that the slight increase in backpressure caused by the catalyst was
insignificant to the results.
The increase in solids caused by the catalyst may be due to dehydrogenation
of the organic compounds present in the exhaust resulting in the formation of
solid carbon and low H/C ratio hydrocarbons. The definition of oxidation of an
organic compound includes both the addition of oxygen to the molecule
(oxygenation) or the removal of hydrogen from the molecule (dehydrogenation).
An order of magnitude study was performed to determine if the change due to the
installation of the catalyst in the amounts of HC measured in the raw exhaust
was sufficient to account for the increased amount of solids measured at modes
A, 5, 9, 10, and 11. The HC measurements done with the HFID both with and
without the catalyst at each of these modes were used for this study. Since the
-------
112
response of a HFID is in ppm carbon present in the gas phase HC, the difference
in HFID response from uncatalyzed to catalyzed exhaust should be related to, and
not more than, the measured increase in solids due to the catalyst. The solid
portion of the particulate is probably composed of very low H/C ratio compounds,
therefore a H/C ratio of 0.014 by weight (65) was assigned to the compounds
measured by the change in HFID response. This available "carbonaceous material"
which would be measured as increased solid particulate matter due to the
catalyst was then calculated from the change in HFID response and compared to
the actual measured increase in solids (as defined by Equation 3 of the
Experimental Section), as shown in Figure 3.2.8. With the exception of mode 9,
there was enough available gaseous phase HC's in the exhaust to account for the
increase in solids due to dehydrogenation in the catalytic converters.
The formation and subsequent condensation of aromatic free radicals above
the catalyst sites would result in the formation of low H/C ratio compounds (66)
that would be detected as solids. The formation of these radicals is an
endothermic process (67) , with the required heat being supplied by exothermic
reactions occurring on the catalyst sites. The oxidations of NO, SO , CO, and
HC's on the catalyst sites are all highly exothermic. The formation of benzene
radicals by hydrogen dissociation and subsequent condensation to diphenyl has
been studied and found to occur both with and without the presence of oxygen
over platinum catalysts (68).
Figure 3.2.9 repeats the mode 4 and 5 data of Figure 14 in addition to
showing the effect of fuels on the measured increase in solids due to the
catalyst, for modes 4 and 5, without regard for other changes in exhaust
composition. The use of No. 2 fuel resulted in a greater increase in solid
emissions. The No. 2 fuel has a higher aromatic content as well as a higher
sulfur content than the No. 1 fuel (Table 2.2). We believe that the increase in
-------
113
3.U
CO
^
o
O
K-
UJ 2.0
0
_J X
Is
cc^»
<^" 1.0
0.
o
_J
o
t7)
Z
UJ
0
^™ rt
3208 CATERPILLAR DIESEL
NO.
S-IV
2
FUEL
EASURED INCREASE IN £,
SOLID PARTICULATE DUE
TO
CATALYST.
C-AVAILABLE CARBON TO
FORM SOLID CALCULATED
FROM HFID MEASUREMENTS
" IN UNCATALYZED
AND
CATALYZED EXHAUST.
m
m,
CO
O
< ° 4
(/) (\
m
5
CO
O
CO
o
CO
r
9 10 11
O EPA MODE
Fig. 3.2.8 - Comparison of measured increase in unex-
tractable residue (solids) with the PTX
catalysts to the calrulated -fnrrpnsp HUP
to HFID HC decrease assuming .014 ratio
compounds
co
0
UJ
0
3 :
O:
1.0
UJ -e 0.75
0.5
o_
Q
_t
O
CO
z 0.25
ui
O
Z
I
o
3208 CATERPILLAR DIESEL
COMPARISON OF FUELS
S - MEASURED INCREASE IN SOLID PARTICULATE
DUE TO CATALYST
C-AVAILABLE CARBON TO FORM SOLID
CALCULATED FROM HFID MEASUREMENTS
IN UNCATALYZED AND CATALYZED EXHAUST.
CM
0
o
CO
i o
CO
M
6
Z
EPA MODE
Fig. 3.2.9 - Comparison of measured increase in un-
extractable residue (solids) with the PTX
catalysts to the calculated increase
due to HFID HC decrease, showing effects
of fuels
-------
114
solid emissions is primarily due to the formation of aromatic free radicals
above the catalyst and the subsequent condensation of these radicals into very
high weight polynuclear aromatics.
The data in Figure 3.2.10 show a slightly higher catalyst temperature from
the inlet to the center of the monolith for all modes of operation for No. 2
fuel than for No. 1 fuel. Since No. 2 fuel has a higher sulfur content than No.
1 fuel, this temperature increase could be partly attributed to SO oxidation.
A three degree temperature increase over the entire catalyst bed would not
likely promote any radical formation. It is likely, however, that at the active
catalyst sites local temperatures are much higher than the average catalyst
temperature that one would measure with a thermocouple.
An increase in solids emissions from the formation of insoluble metal
sulfates has been suggested. While this might contribute a small amount to the
solids increase, it is unlikely that it is significant since approximately 90
percent of the fuel sulfur not emitted as gaseous compounds can be accounted for
by analysis as soluble sulfate in the particulate (15).
The conversion of exhaust HC to solid in the catalytic converters using
each fuel is shown in Figure 3.2.11, neglecting the influence of other changes
in exhaust composition. The conversion of HC to solids was greater for the No.
2 fuel than for the No. 1 fuel. For both fuels, the conversion of HC to solids
increased with temperature; and at catalyst bed temperatures below 275 C, the
catalytic converters apparently reduced the amount of solid emissions. The
effect of the oxidation of SO to SO in the catalysts on the formation of
solids was determined by dividing the fraction of HC converted to solids by the
amount of S0» per unit volume (as ppm) formed in the catalysts. As shown in
Figure 3.2.12, this normalization results in similar conversions of HC to solids
for both fuels, which shows that the amount of heat released in the catalyst by
-------
115
90
<
u
z
K
U
ec
£
15
10
CATERPILLAR 3208
CATALYZED
1680 RPM
NO. 1 FUEL
200 300 400 500
BMEP (kPa)
600
Fig. 3.2.10 - Effect of fuel changes -_n the temp-
erature rise from the exhaust manifold
to center of PTX catalyst: monolith
(O
HI
DC
111
Q.
O
_1
o
V)
o
u.
O
•z.
o
U)
a.
in
>
O
o
CATERPILLAR 3208
CATALYZED
1680 RPM
• NO. 1FUEL
O NO.2FUEL
( )EPA MODE
{STANDARD ERROR
SOLIDout- SOLIDin
CONVERSION
HCin x At CAT
20 -
0 -
200 300 400 500 600 700
CATALYST BED TEMPERATURE (°C)
Fig. 3.2.11 - Solid formation by PTX catalysts.
Calculated assuming an H/C ratio of .014 by
weight for solid and HC, Atcat based on
space velocity using gross catalyst
volume
-------
116
the oxidation of SO to SO possibly affects the formation of solids in the
catalytic converters. The conversion of HC to solids normalized to SO
production for all test conditions using No. 1 and No. 2 fuels is shown in
Figure 3.2.13.
Although, as shown in Figure 3.2.14, the percent conversion of fuel sulfur
to sulfate was similar for both fuels with and without the catalysts, the lower
sulfur content of the No. 1 resulted in lower sulfate emissions both with and
without the catalysts in use.
The effect of catalyst bed temperature on the conversion of exhaust SO to
sulfate (neglecting other changes in exhaust composition) is shown in Figure
1
3.2.15. Because the actual concentration of SO- in the raw exhaust entering the
catalysts was not measured, it was calculated by assuming that except for the
percent of fuel sulfur in the form of sulfate measured without the catalysts
installed, the entire fuel sulfur content of the fuel was in the form of SO .
The conversion of S0_ to S0_ in the catalyst peaked at 500 C, which coincides
with the peak conversion temperature of SO to sulfuric acid used in the contact
process (69).
Figures 3.2.16a and b show the SOF composition for the modes investigated,
using No. 1 and No. 2 fuels, respectively. Use of the oxidation catalyst caused
an increase in the percentage oxygenates for 5 of the 6 modes tested. The only
mode for which oxygenated compounds were a smaller percentage of the total in
catalyzed exhaust was mode 9. This may be due to a long storage time (greater
than 3 months) for a portion of the mode 9 filters. Even though the samples
were stored frozen, they were thawed and opened repeatedly for the addition of
SOF as it became available. This was necessary due to the small amount of SOF
collectible on each filter before plugging, requiring addition of SOF from a
-------
117
15
CO
O
W
Ou
a.
LU
s
I-
H
2
OC
Q
O
LL
O
(0
DC
O
O
-1
-2
-3
-4
CATERPILLAR 3208
CATALYZED
- 1680RPM
• NO. 1FUEL
O NO. 2FUEL
(> EPA MODE,
_I
STANDARD
ERROR
HCxSOJx A t
In CAT-
* SO3 PRODUCED
IN CATALYST (IN PPM)
200 300 400 500 600 700
CATALYST BED TEMPERATURE (°C)
Fig. 3.2.12 - Solid formation in PTX catalyst nor-
malized to amount of SO-j formed. Cal-
culated assuming an H/C ratio of .014
by weight for solid and HC, At t based
on space velocity using gross catalyst
volume
O
V)
0.
CL
in
LU
Z
o
o
10
-5
CATERPILLAR 3208
CATALYZED
• NO. 1 FUEL
- O NO. 2FUEL
() EPA MODE
J STANDARD ERROR
CONVERSION/=
HCin x S03 x AtCAf
S03 PRODUCED IN CATALYST (IN PPM)
200 300 400 500 600 700
CATALYST BED TEMPERATURE (°C)
Fig. 3.2.13 - Solid formation in PTX catalyst nor-
malized to amount of SOj formed for all
tests. Calculated assuming an H/C ratio
of .014 by weight for solid and HC, Atcat
based on space velocity using gross cata-
lyst volume
-------
118
100
UJ
u.
(O
O
ec
5> 10
g
u.
U-
O
O
55
ec
ui
O
O
ec
^ 1.0
X
CATERPILLAR 3208
1680 RPM
NO. 2 FUEL
UNC.
______ A
^~~ NO. 1 FUEL
UNC.
75
200 300 400 500 600
BMEP (KPa)
co
n
O
CM
O
co
u.
O
z
o
V)
DC
UJ
>
Z
o
o
50
25
CATERPILLAR 3208
CATALYZED
CONVERSION =
• NO.1 FUEL
O NO.2FUEL
() EPA MODE
SO
3out
- SO
3in
S02inx
200 300 400 500 600 700
CATALYST BED TEMPERATURE (°C)
fig. 3.2.14 - Effect of fuel on the molar conversion
of fuel sulfur to SO^
Fig. 3.2.15 - Conversion of SO, (calculated) •* SOj
by catalyst as a function of tempera-
ture, At. , based on space velocity using
gross catalyst volume.
-------
119
large number of filters. Otherwise, no reasonable explanation is available for
this discrepancy.
Another effect of the oxidation catalyst on the SOF was the general
decrease in the percentage of paraffins and increase in the percentage of
aromatic compounds with its use. The only modes for which this was not observed
were the 50% load modes (modes 4 and 10), and mode 5 with No. 1 fuel. With the
exception of the 25% load modes (modes 3 and 11), an increase in the percentage
of the acidic compounds present in the SOF was observed with the catalyst in
use. Since acidic compounds are also partially oxygenated, one might expect the
high load modes (higher exhaust and catalyst bed temperature) to produce the
greatest total increase in oxygenates and acidics. This was not observed, but
the differences in residence time in the catalysts from mode to mode (higher
load results in decreased residence time; see Table 3.2.7), and the difficulty
encountered with the mode 9 sample may have obscured this trend.
The effects of fuel changes for fuel No. 1 and No. 2 on the percent
composition of the eight fractions composing the SOF for tests with and without
the catalysts are also given in Figures 3.2.16a and b respectively. Figures
3.2.17 a, b, and c present these data on a mass emissions basis. The
concentrations of SOF in the exhaust are independent of the fuel type for fuels
1 and 2 and depends only on the mode and whether or not the catalyst is used.
This independence of the effect of fuel properties on the eight fractions that
make up the SOF is also apparent.
The percent reductions of the major components of the SOF by the oxidation
catalyst are listed in Table 3.2.8. The aromatic and paraffinic fractions show
that fuel No. 1 and No. 2 have identical trends toward higher percent reductions
going from mode 3 to mode 5. The percent reductions of the transitional
fraction and acidic fractions show no consistent trends, while the percent
-------
MASS CONCENTRATION (mglM3)
MASS CONCENTRATION (ms/m3)
MASS CONCENTRATION (mg/m3)
NJ
O
-------
SOF COMPOSITION (% BY MASS)
O ft
O f-h
cn H
O X
^
o
<: x
o o'
*
» *
:
O C
O
O
m
m
7,
C
r
o
0
» I
z
o
o
\
C.
z
0
§
1"
^ w
o
2
C
z
o
S
m
z
C
z
o
S
r
2
Is i •
^ O 3*
Wo 31
• > O > 3) 31
(rt O 9 O Z
| M
m o> >' -o >'-i j >^
z Crt O 3} o Z 3) Oz
m o> > a > 2
= > 0 > 0 >
— wow o /3> ^x
CO > TJ X>"
£3 S §
58 S £ |
- «o » O 2:
30 31 X =
i r -i
> H O -
33 33 X S
0 z •< :
!
5 1
X
O x
X —
^ "^
'H o
31 X
Z -<
X =
-< '
2 < •?
a ^> n
> x x
2 -< >
m
:
1
> -o >
030
s?-i;
S | 2
2 ^
0
m CD
? S^
i x >
t/>
o
C
w
m
1
;
SOF COMPOSITION (% MASS)
§
c^ o.
co y
O rr
o>
3 «3
o
o
m
Zvi o
|/
,1,
\ 1
m CD ^
5>0
z o
. >
> 3D
3D O
1
1 / /
•O > H
> 3D 3D
3D O Z
1 \\
V ', >4
n > 3D 3D
0 3D OZ
> -t 03
D 30 X =
if //
i §
I /
. !;
^ i
o
x
•<
UNC* g S"3?
** * > ngm
CAT. | o J < > O
> z mo p
UNC. g K so
W
CAT. nmBI>,S
* ^t a > TI O
UNC. £ § i P
CAT. N |
z
-------
122
reduction of the oxygenated fraction decreases going from mode 3 to mode 5. The
results are consistent for these two fuels also with the greater ease of
oxidation of the paraffinic and aromatic fractions over the already partially
oxidized transitional, oxygenated and acidic fractions. Some of the aromatic
and paraffinic material will be converted to transitional, oxygenated or acidic
material leading to apparent lower percent reduction of these fractions.
The similarity in behavior of the aromatic and paraffinic fractions for
these two fuels is somewhat surprising. Evidently the effectiveness of the
catalyst is sufficient to negate differences in the thermodynamic stabilities of
aromatics and alkanes.
Table 3.2.8. Percent Reduction by Mass of SOF Fractions by Oxidation Catalysts
No. 1 Mode
3
4
5
No. 2 3
4
5
Shale Oil 4
HIN
100
(a)
71
67
49
68
80
OXY
60
72
27
55
53
40
88
Fraction
TRN
62
68
80
70
80
59
85
ARO
65
85
89
29
94
(a)
36
PAR
70
69
87
56
84
91
89
ACD
65
55
62
74
65
32
94
fa)
Measured amounts too small for meaningful calculations
Figure 3.2.18 is a gas chromatogram of the No. 2 fuel, mode 11 uncatalyzed
aromatic subtraction. Numbered peaks correspond to compounds identified by mass
spectrometry, as listed in Table 3.2.9. Compounds no. 1 and no. 17 are possibly
not derived from engine particulate samples. Compound no. 1 is a preservative
in the diethyl ether used for fractionation, while compound no. 17 is a
component of many plastic materials. The compounds listed have been previously
-------
123
3208 CATERPILLAR DIESEL
NO. 2 FUEL
MODE 11 AROMATIC SUBFRACTION
ATTENUATION 8
1.8 pi OF 10% SOLUTION IN ISO-OCTANE
17. ATTENUATION 64
30 35 40
RETENTION TIME (WIN )
Fig. 3.2.18 - Gas chronatogram of mode 11 uncatalyzed
aromatic subf roc tier,, l.'tisbcrcd peaks cor-
respond to compounds listed in table 3-2.9
-------
124
Table 3.2.9.
Tentative Identities of Compounds in Mode 11 Aromatic
Subtraction (No. 2 fuel) as Determined by Gas
Chromatography-Mass Spectrometry
Peak
Number
(figure)
1
2
3
4a
5a
6a
7a
8a
9
10
11
12, , ,
(a,b)
14(b>
15
16
17
Retention
Time (min.)
19.2
28.6
31.2
32.0
32.8
33.0
33.6
33.8
34.8
35.6
36.3
37.6
38.3
39.9
41.2
48.0
53.0
Molecular
Compound Weight
2,6-di(_t-butyl)-4-methyl phenol 220
Phenanthrene 178
Monomethyl dibenzothiophene isomer 198
Monomethyl dibenzothiophene isomer 198
Monomethyl phenanthrene isomer 192
Monomethyl phenanthrene isomer 192
Monomethyl phenanthrene isomer 192
Monomethyl phenanthrene isomer 192
Dimethyl naphthothiophene isomer 212
Dimethyl naphthothiophene isomer 212
Dimethyl naphthothiophene isomer 212
Dimethyl phenanthrene isomer 206
Fluoranthene 202
Pyrene 202
Trimethyl phenanthrene isomer 220
Benzo(ghi)fluoranthene 226
Phthalate ester (insignificant mass ion) -
Known mutagens (13)
Despite nearly identical mass spectra, identity of these compounds was
confirmed by retention times
-------
125
identified by Brooks et al. (7) and Schuetzle et al. (11). Compounds 4-8 and 13
are known mutagens (7), but the mutagenic activity of this aromatic subfraction
was not fully determined.
Ames Bioassay Variability and Strain Differences - The statistical analysis
of the effect of different assay dates upon the Ames assay data for the same
extract is presented in Table 3.2.10. Mode 5 catalyzed SOF and mode 11
catalyzed oxygenated subfraction, both without the S-9 activation mixture, were
each assayed on two different dates using identical concentrations within each
sample and the Ames tester strain types TA98, TA100, TA1537 and TA1538.
For each strain/mode/date combination a dose-response curve was established
and the slope of its linear portion (in Cartesian form) calculated. The two
slopes from the different dates within a strain and mode were then compared
using the two sample t-statistics. The null hypothesis (HQ) of no difference
between the two slopes from the two assay dates within each strain/mode was
tested against the alternate hypothesis (H ), which states that a difference
3
dees exist between the slopes from the different assay dates with each
strain/mode. For Table 3.2.10 (and Table 3.2.11), the linear model R = a + b ' C
was used for the testing of differences between slopes. The analyses involved a
t-statistic that was calculated based on the differences of the two slopes
divided by their pooled residual mean square.
In addition to each date's slope and the calculated t-statistic for each
pair of dates within a strain and mode, Table 3.2.10 gives the probability that
the differences which exist between the slopes being compared are due to chance.
Also shown in Table 3.2.10 is the value of the 5% two-tailed t-statistic with 14
degrees of freedom. The Ha was rejected at the 5% level for all comparisons
except for the TA100 mode 11 catalyzed oxygenated subfraction.
-------
Table 3.2.10. Effect of Assay Date on Ames Test Results
Hypotheses Tested: H = No difference between experimental dates within sample and strain.
H° = There is a difference between experiment dates within sample and strain.
3.
Q
Ames Strain
Mode
Extract
TA98
5
SOF
11
OXY
TA100
5 11
SOF OXY
TA1537
5 11
SOF OXY
TA1538
5 11
SOF OXY
Date of Q72980 081480 072380 080780 072080 081480 072380 080780 072980 081480 072380 080780 072980 081480 072380 080780
Experiment
_, d REV , ,,
Slope 4.45
v yg
t-statistic
Probability
.05(2), 14
4.64 1.64
1.39
.1295
.50
-------
Table 3.2.11. Effect of Experimental System Variability on Ames Test Results
Hypotheses Tested: H = No difference between samples collected and extracted on different days within strains.
H° = There is a difference between samples collected and extracted on different days
Q
within strains.
a
Ames Strain
Mode
Sample
c REV
F Ug
t-statistic
Probability
d
.05(2) ,14
^Without S-9
TA98
3 11 3
11 11 I .
.469 .541 .506 .605 .809 1.
1.385 1.251 .1010
.10
-------
128
Table 3.2.11 shows the effect of varying all phases (collection,
extraction, and assaying) of diesel exhaust particulate testing. Samples 3-1 and
3-2 were collected at mode 3 while samples 11-1 and 11-2 were collected at mode
11. The distinction between samples 3-1 and 3-2 (also samples 11-1 and 11-2) is
that they were collected on different days, extracted on different days, and
assayed in separate experiments.
For each strain and mode combination (without the S-9 activation system), a
dose-response curve was obtained and the slope of its linear portion (in
Cartesian form) was calculated. For each strain, sample 3-1 was compared with
3-2 and sample 11-1 with 11-2 for significant differences between their slopes.
The two sample t-statistic is used for testing the null hypothesis of no
difference against the alternate hypothesis that a difference exists.
In addition to the calculated t-statistic, Table 4.2.11 gives the
probability that the differences which exist between the slopes are due to
chance. Also shown is the tabulated t-value of the 5% two-tailed t-statistic
with 14 degrees of freedom. The H was rejected at the 5% level for all cases
3.
except for strain TA1537 with samples 11-1 and 11-2.
The statistical analysis therefore indicates that the Ames bioassay data
were reproducible in 14 of 16 comparisons, whether the same SOF samples were
repeated or whether SOF samples from different engine runs were tested. The
difference found with strain TA1537 tested against two mode 11 uncatalyzed SOF
samples (Table 3.2.11) may be due to the fact that this strain has relatively
low-level response (compared to strains TA98, TA100, and TA1538). The slope
differences were therefore likely due to natural variation in the Ames bioassay.
Strain TA100 provides high-level response with usually good reproducibility.
The slopes presented in Table 3.2.10 for repeats of mode 11 catalyzed SOF
oxygenated subfraction were not significant at the a = 0.1 significance level
-------
129
and therefore the differences probably also represent natural variation in the
test results.
Figure 3.2.19 shows the response of all five Salmonella typhimurium tester
strain genotypes to the catalyzed mode 11 SOF. Dose-response data indicate the
mutagenic components of the diesel SOF are not base-pair substitution mutagens
as no mutagenic response was demonstrated by strain TA1535, which is the only
strain used responding solely to base-pair substitution mutagens. The greatest
response was observed with strains TA100 and TA98, followed by strains TA1538
and TA1537. All of these strains are frameshift mutagen detectors. Similar
results have been reported by other investigators working with a variety of
diesel and gasoline engines (8,9,49,70). For the purpose of further discussion,
only the TA100 tester strain data are presented in graphical form. All test
results are listed in Appendix B.
Effect of Catalyst on Ames Biological Activity - The dose-response curves
presented in Figure 3.2.20 for strain TA100 and mode 11 SOF (without S-9
activation) indicate that the catalyzed mode 11 SOF (on a revertant/yg of SOF
basis) resulted in a greater mutagenic response than the uncatalyzed mode 11 SOF
sample. This effect was generally observed with other engine modes, as
demonstrated by the dose-response results from modes 4 and 10 catalyzed and
uncatalyzed (Figure 3.2.21). The only exception occurred with mode 3, where the
catalyzed SOF produced fewer revertants/yg of SOF than the uncatalyzed SOF.
Additionally, the SOF from the catalyzed engine modes often displayed a toxic
effect which was not found with the uncatalyzed modes (Figure 3.2.21, modes 4
uncatalyzed and 4 catalyzed).
Some of the SOF samples were tested for mutagenic activity in the presence
of the S-9 activation mixture. As demonstrated by strain TA100 in Figure
3.2.20, the results from mode 11 and mode 11 catalyzed, with and without the S-9
-------
130
i
£>
_ 600
o
85
3208 CATERPILLAR DIESEL
NO 2 FUEL
MODE 11 CATALYZED TOTAL SOF
WITHOUT S-9
I STANDARD DEVIATION FOR
I THREE REPLICATE PLATES
£=> ~«0-T---° T.A1535
_
O.
(A
100
UJ
DC
a
Z 400
—I 1—I I I I III
3208 CATERPILLAR DIESEL
NO. 2 FUEL
AMES TESTER STRAIN TA100
MODE 11 TOTAL SOF
I STANDARD DEVIATION FOR
} THREE REPLICATE PLATES
I CATALYZED
l WITHOUT S-9
UNCATALYZEO
WITH S-9
60 100 300
DOSE (ug/PLATE)
600
Fig. 3.2.19 - Response of all five Ames tester strains
to mode 11 catalyzed SOF
10 30 go too • 300 600 IOOQ
DOSE (iig/PLATE)
Fig. 3.2.20 - Effects of PTX catalyst and S-9 mammalian
activation system on mode 11 SOF
o.
5
i-
z
o
%
u
oc
3208 CATERPILLAR DIESEL
NO 2 FUEL
AMES TESTER STRAIN TA100
TOTAL SOF
WITHOUT S-9
[STANDARD DEVIATION FOR
[THREE REPLICATE PLATES
MODE 4
CATALYZED
,^:-'
WlX^.-' MODE 4
••• »QT UNCATALYZED
10 100 300
DOSE (ug /PLATE)
Fig. 3.2.21 - Response of Ames tester strain TA100
to SOF from modes 4 and 11 PTX catalyzed
and uncatalyzed
-------
131
mixture, indicate that the mutagenic activity is direct-acting (i.e., metabolic
activation is not required) and the addition of the S-9 mixture causes a general
decrease in mutagenic response. These responses have also been noted for SOF
from other diesel and fuel systems (7,9,47,70). Tests with other samples and
tester strains showed similar results (Appendix B).
The effects of fuel changes on the TA98-SOF dose-response curves with and
without the oxidation catalyst at modes 3, 4, and 5 using No. 2 and 1 fuels are
shown in Figures 3.2.22a, b, and c, respectively. Figure 3.2.23 shows the
effect of the catalyst at rated-speed (2800 rpm) modes for No. 2 fuel. (It
should be noted that, in order to reduce variability in the Ames bioassay
results, catalyzed and uncatalyzed SOF samples from the same fuel and mode were
tested on the same day. A similar format was used when testing the
subtractions.) The mode 3 dose-response curves in Figure 3.2.22a are unique
with respect to their response to the oxidation catalyst. Only at mode 3 is the
biological activity of the No. 1 and No. 2 fuels catalyzed SOF dose-response
curves lower than their uncatalyzed SOF dose-response curves. The No. 1 fuel
catalyzed SOF has the lowest activity at mode 3, while the uncatalyzed SOF from
the No. 2 fuel has the highest activity. The No. 2 fuel's catalyzed and
uncatalyzed SOF appears to produce a higher response than does the No. 1 fuel at
mode 3.
At modes 4 and 5 (Figures 3.2.22b and c), the catalyzed SOF displayed far
greater mutagenic activity than did the uncatalyzed SOF. At all rated speed
(2800 rpm) modes, the catalyzed SOF also displayed greater mutagenic activity
than did the uncatalyzed SOF (Figure 3.2.23). When comparing within modes, the
No. 2 fuel (catalyzed or uncatalyzed) is generally more active than the No. 1
fuel.
-------
132
103
102
10
CATERPILLAR 3208, MODE 3
NOS. 1 AND 2 FUELS
CATALYZED AND UNCATALYZED SOF
STRAIN TA98, WITHOUT S-9
UNC.*- UNCATALYZED
CAT.**—CATALYZED
NO. 2 FUEL, UNC.*
. 2 FUEL,
CAT.* *
. 1 FUEL, CAT.
10"
5
_
a.
to
oc
W
UJ
cc
CATERPILLAR 3208, MODE4
NOS.1 AND 2FUELS
CATALYZED AND UNCATALYZED SOF
STRAIN TA98, WITHOUT S-9
UNCAT *-UNCATALYZED
CAT * *-CATALYZED
NO 2FUEL.CAT*
/ NO 1 FUEL.UNCAT
300
1000
Fig.
37.5 75 150 300 600
SOF CONCENTRATION (^ig/PLATE)
3.2.22a- Comparison of No. 1 and No. 2 fuel SOF
from Mode 3 uncatalyzcd and Mode 3
fTf. catalyzed using TA98 without S-9
activation
3 10 30 100
SOF CONCENTRATION
Fig. 3.2.22b - Comparison of No. 1 and No. 2 fuel from
Mode 4 uncatalyzed and Mode 4 PTX catalyzed
using TA98 without S-9 activation
10"
2
5. 103
V)
CC
LU
gj 102
CC
10
CATERPILLAR 3208, MODE 5
NOS. 1 AND 2 FUELS
CATALYZED AND UNCATALYZED SOF
STRAIN TA98, WITHOUT S-9
UNC. * —UNCATALYZED
_ CAT.* * —CATALYZED
I
UJ
1
OC
LU
103
10
CATERPILLAR 3208, NO. 2 FUEL
2000 rpm (MODES 9,10, and 11)
CATAl YZED AND UNCATALYZED SOF
STRAIN TA98, WITHOUT S-9
9-MODE 9,UNCATALYZED 9C- MODE 9.CATALYZED
10- MODE 10,UNCATALYZED 10C- MODE 10,CATALYZED
11- MODE 11, UNCATALYZED 11C- MODE 11, CATALYZED
10C
-10
75 105 300
SOF CONCENTRATION
600
1200
37.5 75.0 150 300 600
SOF CONCENTRATION (^g/PLATE)
Fig. 3.2.22c - Comparison of No. 1 and No. 2 fuel SOF
from Mode 5 uncatalyzed and Mode 5 PTX
catalyzed using TA 98 without S-9 activation
Fig. 3.2.23 - Comparison of No. 2 fuel SOF from 2800
rpm modes uncatalyzed and PTX catalyzed
using TA98 without S-9 activation
-------
133
Figure 3.2.24 shows the effect of the catalyst on the mutagenic SA of the
SOF and on the BSSA at 2800 rpm (rated speed). While the SA of the SOF is only
slightly increased with increasing engine load at this speed, the BSSA are
decreased due to much lower SOF emissions at higher loads, with and without the
catalyst..
Figures 3.2.25a and b show the effect of fuel changes on the mutagenic-SA
of the SOF and on the BSSA eraissions with and without the catalyst. (Specific
activity calculations were performed using the log dose-revertant transformation
of the liner regression model, and were not standardized with respect to lower
and upper doses taken in the regression.) The use of the No. 1 fuel resulted in
lower SOF mutagenic-SA and in lower BSSA compared to the No. 2 fuel. Fuel
changes have been shown to affect SOF mutagenic-specific activity (47).
A study was conducted to determine whether various sampling and exhaust
system variables could be empirically correlated to the specific activity as
determined by the Ames bioassay. The quantity SA*, defined as non-paraffinic
SOF specific activity, was calculated using the following equation:
SA* = SA 10°
100 - % Paraffins
This value was used rather than the specific activity (SA) of the SOF because
paraffins comprise a large portion of the SOF and are inactive using the Ames
bioassay. Table 3.2.12 lists the SA* for Strains TA98 and TA100 along with the
respective sampling and exhaust system variables. Using the STATJOB subroutine
REGAN2 (71), the correlation between a number of independent variables and one
dependent variable can be determined, as well as those variables which are
significant to the correlation. For the TA100 data, only the raw N0_
concentration displays significance at the a = .05 level. The TA98 data showed
that exhaust volume flowrate, raw exhaust temperature and HC concentration in
the raw exhaust are significant factors at the a = .05 level, while raw NO-
-------
134
x
UJ
z
i-
z
<
I-
cc
Ul
cc
o
u.
o
Ul
a.
CO
i
UJ
DC
CO
10'
CATALYZED
UNCATALYZED
CATERPILLAR 3208
2800 RPM
NO.2 FUEL
AMES TESTER STRAIN TA98
WITHOUT S-9
CATALYZED
UNCATALYZED
en
,3.
Ul
CC
>
I-
o
o
Ul
0.
V)
0.5
0.1
200
300
BMEP(KPa)
400
Fig. 3.2.24 - Effect of PTX catalysts on SOF muta-
genic-specific activity and brake
specific revertant emissions, 2800 rpm
-------
BRAKE SPECIFIC REVERTANT EMISSIONS
(REV/kWxhr)
° i, i.
u
o
o
CO
m
^1
yr
•o
to
TT
' -'''I'
p
in
at
o
SPECIFIC ACTIVITY (REV/Mg)
»-
X fi> rt
O tu
O H- 00
rt p H- M
BRAKE SPECIFIC REVERTANT EMISSIONS
(REV/kWxhr)
LO
U1
SPECIFIC ACTIVITY (REV/^g)
-------
Table 3.2.12.
Sampling and Exhaust System Variables Associated with Non-Paraffin
Specific Activities (SA*) No. 2 Fuel
MODE PTX
3
3
4
4
5
5
9
9
10
10
11
11
3
3
3
11
11
11
X
X
X
X
TA98
SA*
(Rev/yg)
1.07
1.08
0.92
8.20
1.82
6.01
0.61
1.27
0.74
2.22
0.47
2.14
1.13
.95
NA
1.26
.77
.45
TA100
SA*
(Rev/yg)
2.07
1.06
1.37
9.50
1.53
4.91
0.90
3.82
0.78
2.98
1.17
4.90
1.21
0.89
0.2
1.53
1.17
1.15
N°2dil
(ppm)
2.0
1.4
1.3
19.3
1.3
8.4
0.5
1.1
1.0
1.7
0.9
3.9
8.2
2.5
1.2
2.5
0.5
0.3
NO.
2 raw
(ppm)
30.1
20.6
19.1
286
19.3
12.0
8.0
16.3
14.3
25.8
13.6
55
41
38
33
13
8
8
S°4dil
(mg/Actm )
.1
.7
.2
7.9
.4
16.8
.6
7.0
.3
10.1
.1
9.3
.6
.2
.1
.7
.13
.1
SO.
4 raw
3
(mg/Actm )
.9
5.8
2.1
58.5
23
107.6
3.5
36.6
2.0
64.2
1.0
64.1
1.9
1.8
1.0
2.0
1.0
1.2
~exh
^ min '
14
14
17
17
21
21
34
34
29
29
24
24
14
14
15
23
24
25
) Tdil(K)
312
313
321
320
324
323
333
324
323
324
322
321
335
314
308
358
322
315
T (K)
raw
525
531
638
645
111
754
915
939
747
763
650
647
523
532
528
652
650
674
Sample Time
(min)
60
120
240
45
45
30
30
15
120
10
20
20
45
90
135
10
20
40
HC
raw
(ppmC)
210
55
150
25
100
17
115
25
320
40
310
60
210
210
210
310
310
310
-------
137
concentration is significant at the a = .1 level. This study tended to show
that the raw exhaust variables that would be associated with exhaust system
reactions to be more significant than diluted variables and sampling time which
might be associated with artifact formation.
Effects of Shale Fuel on PTX Catalyst Operation - As mentioned previously,
the emissions data from the PTX catalyst run with shale fuel are not strictly
comparable to the baseline emissions used for No. 1 and No. 2 fuels due to the
replacement of a failed injection pump in early 1981. Table 3.2.13 compares
engine data for the two baselines at modes 3, 4, and 5, although the only mode
for which shale fuel data were obtained using the PTX catalyst was mode 4. An
appreciable decrease in NO levels was seen for the more recent baseline, with
lesser increases or decreases in various particulate components of the exhaust.
Figure 3.2.26 shows the change in brake-specific NO , NO , and NO emissions
X ^
for the shale fuel with and without the PTX catalyst. As with the No. 1 and No.
2 fuels, the catalyst had little effect on NO emissions but increased N00
X Z
emissions, with resultant decrease in NO emissions. Figure 3.2.27 shows BSHC
and BSTPM (including SOF and solids; sulfate was negligible) for mode 4 using
shale fuel. As with the other fuels, the HC and SOF were reduced with the shale
fuel by use of the oxidation catalyst; however, the TPM was also reduced despite
an increase in the solids content with the oxidation catalyst, due primarily to
the extremely low sulfur content of the shale fuel.
Figure 3.2.28 illustrates the SOF composition of samples collected using
shale fuel with and without the oxidation catalysts at mode 4. The uncatalyzed
baseline values obtained with No. 2 fuel after the injection pump change are
shown for comparison. Use of shale fuel resulted in lower percentage paraffins,
oxygenates, and transitionals; and higher aromatics and acidics for both
catalyzed and uncatalyzed shale fuel, compared with uncatalyzed No. 2 fuel. The
percentage of aromatics using shale fuel and the oxidation catalyst was higher
-------
138
Table 3.2.13, Baseline Emissions Comparison
Caterpillar 3208, No. 2 Fuel (all units in grams/kw-hr ± 1 standard deviation)
Mode Emission March-April 1980 July 1981
3 BS NO 13.65 ^0.26 10.24 ^ 0.03
BS NO* 0.84 +_ 0.09 1.16 +_ 0.22
BS NO 8.36 +_ 0.21 5.92 ^0.16
BS HC 2.01 jf 0.00 1.92 ^ 0.21
BSTPM 0.889 +_ 0.034 1.047 +_ 0.031
BS Solids 0.117 0.084
BS SOF= 0.753 +_ 0.03 0.932 _+ 0.030
BS S04 0.019 _+ 0.003 0.031 _+ 0.001
4 BS NO 12.69 _+ 0.18 10.39 +_ 0.13
BS NO* 0.26 +. 0.03 0.85 _+ 0.05
BS NO 8.11 jfO.ll 6.22 jf 0.07
BS HC NA 0.77 ^ 0.02
BS TPM 0.331 +_ 0.009 0.482 +_ 0.030
BS Solids 0.086 0.069
BS SOF= 0.215 +_ 0.003 0.377 ^ 0.024
BS SO^ 0.035 jf 0.008 0.036 _+ 0.002
5 BS NO 12.07 +_ 0.36 9.69 +_ 0.08
BS NO^ 0.18 +_ 0.05 0.36 +_ 0.01
BS NO 7.75 jf 0.27 6.08 +_ 0.05
BS HC NA 0.25 +_ 0
BS TPM 0.266 +• 0.006 0.241 +_ 0.005
BS Solids 0.202 0.172
BS SOF= 0.032 +_ 0.004 0.038 +_ 0.001
BS SO. 0.032 + 0.013 0.031 + 0.002
-------
BRAKE- SPECIFIC EMISSIONS (g/KWx h rt
!-• a
c s
— OJ >
iSS
I I /
m
== > O
II
> -I O 4-
3)33 X £
OZ -< Z
I _ I I
Z>O
tna
9?? ?
i iips
. x . «p lilli 2
^^O 3; _j —t z: ca rn H
^-^^ 5: oo^f^ 2-S
t"1 [71 O "^
> O CJr-
m
61 TJ
O H
O- X
lUNC.
I CAT.
nUNC.
MASS CONCENTRATION (mg/m3)
H H-
X O
W OJ H1
rt O
fu rr O
BRAKE-SPECIFIC EMISSIONS (G/KW-hr)
53
CO
UNC.
CAT.
so
>
3}
-SOLID
SOF
U'NC.
CAT.
-------
140
than for most SOF examined. A different analysis of this data, as shown in
Figure 3.2.29 and the last line in Table 3.2.8, also demonstrates the stability
of shale fuel aromatics at mode 4. With shale fuel, every fraction but the
aromatic fraction is reduced by the catalyst to a greater percentage than for
fuels No. 1 and No. 2. The reason for the lower percent reduction of the
aromatics must be due to more stable types of aromatics in the SOF. The lower
percent conversion of the aromatics to oxygenated, transitional and acidic
fractions for the shale fuel is indicated by the higher percent reduction of
each of these three fractions. The paraffinic fraction of the shale fuel is
reduced by the same percentage as for fuels No. 1 and No. 2.
The mode 4 results in Table B7 of the appendix show that the SA of SOF from
all these fuels is similar, and may not be statistically different when using
the catalyst; however, the uncatalyzed sample shows the shale fuel to have a
noticeably higher SA than the No. 1 or No. 2 fuels. Figure 3.2.30 is a
comparison of the Ames dose-response curves for these fuels. The catalyzed
samples showed higher SA than the uncatalyzed samples for all these fuels, with
No. 1 fuel showing the lowest activity for either catalyzed or uncatalyzed SOF.
The biological activity of the acidic and oxygenated SOF subfractions from
the Nos. 1 and 2 fuels and shale oil fuel at EPA mode 4, with and without the
oxidation catalyst, are shown in Figures 3.2.31a and b, respectively.
Dose-response curves for the other subfractions and all SA calculations are
presented in Appendix B Figures B2 - B7 and Table B4. In all cases, except one,
the use of an oxidation catalyst increased the SA of the mode 4 SOF subfractions
from the three fuels. The exception is the No. 2 fuel's aromatic subfraction;
(Figure B2) however, little confidence can be placed on the aromatic subfraction
data because of its poor r value (Table B4). In most cases, the No. 2 fuel and
the shale fuel had a higher SA than did the No. 1 fuel.
-------
141
2600
2400
2000
1600
800
400
NO 2FUELCAT * *
CATERPILLAR 3208. MODE 4
CATALYZED AND UNCATALYZEO SOF
STRAIN TA98, WfTHOUTS-9
UNC *—UNCATALYZED
CAT * *-CATALYZED
NO 1 FUEL UNC
3 10 30 100 300 1000 3000
SOF CONCENTRATION (^lg/PLATE)
Fig. 3.2.30 - Comparison of fuel effects on dose-
response curves of Ames tester strain
TA98, with and without PTX catalysts
3200
2800
^2400
I
~ 2000
to 1600
1200
800
400
CATERPILLAR 3208, MODE 4
ACIDIC FRACTION
CATALYZED AND UNCATALYZED/
STRAIN TA98, WITHOUT S-9
UNC *—UNCATALYZED
CAT * *—CATALYZED
NO 2 FUEL CAT * *
2800
2400
2000
1600
< 1200
-------
142
Johnson-Matthey Close-Coupled Port Catalysts
The second aftertreatment device investigated was a close-coupled exhaust
9
port catalyst. The specifications of this catalyst are given in Table 3.2.14
Caterpillar Engine Company donated a set of modified cylinder heads with these
catalysts installed for evaluation (one per cylinder, eight in all). Modes 3,
4, 5, 9, 10 and 11 of the EPA 13 mode cycle were investigated using these
catalysts and the same No. 2 fuel as was used in the evaluation of the PTX
oxidation catalyst; shale fuel was also employed at modes 3, 4, and 5. Table
3.2.15 is the test matrix used for evaluation of these catalysts.
Table 3.2.14 Close Coupled Exhaust Port Catalyst Specifications
Manufacturer Johnson-Matthey
Catalytic Agent Platinum
Substrate Fecralloy
Substrate Type Laminar Flow Monolith
Substrate Size 1.8" Diameter x 2.0" Long
Substrate Cell Size 0.8 mm x 1.5 mm
Effective Open Volume 89%
Effects of the Close-Coupled Catalysts on Engine Operating Conditions (Both
Fuels) and Gaseous Emissions(No. 2 Fuel) - After installing the close-coupled
catalyst heads, the engine was run-in using No. 2 fuel until emissions stability
was reached, which required 20 hours. An additional 8 hours of running at mode 4
was performed to confirm the stability of emissions.
Table 3.2.16 gives catalyst operating conditions at all 6 EPA modes
tested, along with BSFC and exhaust temperature measured just outside the
exhaust manifold. The catalyst temperature was measured on the outer face,
rather than the center of the monolith, as was the case for the PTX downstream
-------
143
Table 3.2.15 Test Matrix for Close-Coupled Port Catalyst Evaluation
EPA Mode 345
Speed (RPM) 1680 1680 168'0
Load (N-M) 160 320 48:0
BMEP (kPa) 192 383 575
Fuel No. 2 No. 2 No. 2
Catalyst w/o w w/o w w/o w
Total Particulate
SOF, and SO, x x x x x x
mass emissions
HC, NO, and
No mass x x x x x x
V
emissions
Chemical
Characterization x x x x x x
of SOF
Ames Mutagenicity
Bioassay x x x x x x
on SOF
9 10
2800 2800
399 266
485 372
No. 2 No. 2
w/o w w/o w
XX XX
XX XX
XX XX
XX XX
11 3
2800 1680
133 160
161 192
No. 2 Shale
w/o w w/o w
XX XX
XX XX
X X
X X
4 5
1680 1680
320 480
383 575
Shale Shale
w/o w w/o w
XX XX
XX XX
Ames Mutagenicity
Bioassay
on subfractions
x x
-------
Table 3.2.16 Engine and Catalyst Operating Parameters, Close-Coupled Catalyst
Catalyst Catalyst Baseline
Exhaust Catalyst Space Residence Exhaust Exhaust Baseline
Temperature Temperature Velocity Time Opacity Opacity BSFC BSFC
EPA Mode, Fuel °C °C Volumes/sec msec % % kg/kw-hr kg/kw-hr
3
4
5
9
10
11
3
4
5
(a)
, No. 2
, No. 2
, No. 2
, No. 2
, No. 2
, No. 2
, Shale
, Shale
, Shale
Not Available
243
353
483
587
477
363
269
380
514
288
406
530
690
545
454
286
409
558
404
488
582
1093
901
773
393
484
584
2
2
1
0
1
1
2
2
1
.48
.05
.72
.92
.11
.29
.54
.07
.71
0.3
0.5
1.4
5.5
3.5
4.6
0.3
0.7
2.0
0
0
1
4
3
3
N.A.
0
N
.3
.9
.9
.9
.8
.2
(a)
.4
.A.
0.291
0.229
0.218
0.272
0.290
0.399
0.283
0.226
0.215
0.287
0.231
0.217
0.252
0.282
0.371
N.A, ^
0.229
N.A.
-------
145
Of
I
2
^
\
CD
ID
CO
z:
o
CJ
LU
13
LJL
CO
QQ
o
o
o
10
CO
CD
CD
OO
O
10
CM
T 1 1 1 1 1 1 1 1 r
BS FUEL CONSUMPTION VS. BMEP
CflTERPILLHR 3208 BRSELINE
-------
146
o
I
I
CD
CO
2
O
i — i
CO
CO
UJ
X
o
Q
z
CE
C\J
O
O
2
CO
CO
I
o
"ii i 1 1 r
CRTERPILLflR 3208
EPfl MODES 3.1.5
0 .0
NOX uncatalyzed
\
N02 uncatalyzed
NO catalyzed
NO uncatalyzed
N02 catalyzed
1
I L
200 .0
400 .0
1
600 .0
BRflKE MERN EFFECTIVE PRESSURE (KPR)
FIG.3.2.33aBRRKE SPECIFIC NO, N02. RND NOX NITH RND
WITHOUT CLOSED-COUPLED CRTRLYSTS
NO.2 FUEL
-------
147
o
IE
I
IS
^
\
CD
o
I—I
en
en
i—i
z:
LU
X
O
Q
Z
CE
C\J
O
O
en
m
o
o
i 1 1 1 1 i
CflTERPILLRR 3208
EPR MODES 3.10.11
1 T
0 .0
NOX catalyzed
NOV uncatalyze
NO catalyzed
-e
N02 catalyzed
NC>2 "fincatalyzed
j i [_
I
I
200 .0
400 .0
600 .0
BRRKE MERN EFFECTIVE PRESSURE (KPR)
FIG.3.2.33bBRRKE SPECIFIC NO, N02, RND NOX WITH RND
NITHOUT CLOSED-COUPLED CRTRLYSTS
NO.2 FUEL
-------
CHRNGE IN N02 EMISSIONS (PERCENT)
0 .0
500 .0
1000 .0
T 1 1 1 r
1500 .0
CD
to
c_ m ~n o
o z o o
J_ CD TO IS.
z m -o
CO F" —I 3D
O IE 51 70
Z 3D O i-<
! 70 CO
IZtDDQ
3D 3D Z
-H 1) -H
H —I 3D O
31 X I H
m -<
-< D co n
O -H 31
m Z CO 3D
X Z Z
37 U") O CD
3D —\ Z m
d 70 CO
co m —i
—I 3D 31 --<
is m z
~o
o o n z
70 3D 3D O
-H — | —| IX)
3D
CD i— co
3D -< r^o
-H CO o
3D -H co
n co ••
CO 3D
—\ Z
co a
n
3D
H
3D
co
CD
m
a
rn
2
T)
m
70
3D
70
m
o
m
CD
o
CD
IX)
CD
CD
O
CD
CD
CD
CD
CD
•
CD
CO
CD
CD
CD
00
-------
149
CJ
UJ
CO
•*•
o
LJ
_J
O
C\J
O
CO
LU
_J
O
CM
O
O
I—
O
o
•
CD
CD
CM
O
•
CD
O
•
O
OJ
T
T
N02 FORMflTION RflTE THROUGH CflTRLYST
VS. CRTRLYST BED TEMPERRTURE
ENGLEHHRD PTX DOWNSTRERM CRTHLYSTS RND
JOHNSON-MflTTHEV EXKflUST PORT CflTRLYSTS.
CRTERPILLRR 3208. MODES 3,4,5,9,10,11
Engelhard PTX
1680 rpm
Port Catalyst
1680 rpm
Engelhard PTX
2800 rpm
Port Catalyst
2800 rpm
0 .0
200 .0
400 .0
600 .0
800 .0
CRTRLYST BED TEMPERRTURE (DEG C)
FIG.3.2.34bFORMRTION RRTES OF N02 IN CRTRLYSTS:
ENGLEHRRD PTX CRTRLYSTS RND JOHNSON-
MRTTHEY EXHRUST PORT CRTRLYSTS ON THE
CRT 3208, EPR MODES 3,4,5,9,10,11
-------
150
CD
\-/
CO
•z.
o
I — I
CO
CO
I — I
21
LU
O
CO
LY
CE
O
O
CO
CD
O
o
•
CO
o
o
•
CM
O
o
T I I I -1 1 1 1 1 1 1 ,-
BS HYDROCRRBON EMISSIONS VS. BMEP
CflTERPILLRR 3208 BflSELINE (NO CONTROL)
RND EQUIPPED WITH EXHRUST PORT CflTflLYSTS
EPR MODES 3.1,5.9, 10, 11
2800 rpm
baseline
2800 rpm
port
catalyst
J I l_
J I
1680 rpm
0 .0
200 .0
400 .0
600 .0
BRRKE MEflN EFFECTIVE PRESSURE (KPR)
FIG.3.2.35THE EFFECT OF EXHRUST PORT CRTflLYSTS ON
HYDROCRRBON EMISSIONS FROM THE CRT 3208
CRTRLYSTS: JOHNSON-MRTTHEY FECRRLLOY
SUPPORTED OXIDRTION PORT CRTRLYSTS
-------
151
catalyst. Fuel consumption at intermediate speed modes was 2% or less higher
than that of the baseline, probably due to the relatively small increase in
exhaust in back pressure of the engine with the close-coupled catalysts
installed (Figure 3.2.32). Rated speed modes showed a 5-8% increase in BSFC
relative to the baseline engine, due to likely higher back pressure increases
resulting from increased exhaust flowrate. Exhaust opacity was decreased or
unchanged by use of the catalysts at all modes but 9 and 11.
Figure 3.2.33a and 3.2.33b show the effect of the catalyst on
brake-specific NO , NO. and NO emissions, with No. 2 fuel. The NO and NO
x / x
levels were only slightly higher at all six modes using the port catalysts,
while N02 levels were higher for modes 5, 9 and 10 (the high-temperature modes)
only. This compares favorably with PTX downstream catalysts, where NO- was
higher for catalyzed exhaust for all modes but mode 3. The trend for both
intermediate and rated speeds for the port catalyst was a decrease in NO at
light loads, changing with load so that NO is increased at high load. This
contrasts with the Englehard PTX, which showed large increases in all modes
except mode 3. Figure 3.2.34a shows the percent increase in NO for both catalysts,
and figure 3.2.34b demonstrates the fundamental difference in NO oxidation
behavior between the PTX and port catalysts. Only at modes 5 and 9 do the port
catalysts show a significant increase in NO oxidation rate over the PTX
catalysts. This may be related to the non-linear operating temperature
difference between the two devices, to be discussed later. Figure 3.2.35
shows the BSHC emissions for the baseline engine and engine with port catalysts.
Reduction of HC at the high-speed modes was greater than at the low-speed modes,
with HC levels showing a much greater reduction at mode 11 than at mode 3
despite the two-fold greater residence time at mode 3. This suggests that the
catalyst temperature may be more important in hydrocarbon oxidation than
-------
152
catalyst residence time. Figure 3.2.36 shows that the PTX catalyst was much
more effective at reducing hydrocarbons despite its slightly lower operating
temperature. The increase in port catalyst effectiveness at higher load
(temperatures) is shown by comparing temperature differences between the
catalysts, as in Figure 3.2.37. Since radiative heat transfer is proportional
to absolute temperature raised to the fourth power, the oxidative advantage of
close-coupled catalyst over a downstream catalyst is fully realized for hottest
running conditions.
Effects on Particulate Emissions - %. 2 Fuel - Brake-specific particulate
emissions for the close-coupled catalysts are shown with corresponding baseline
emissions in figures 3.2.38a and 3.2.38b. The changes in total particulate
emissions ranged from a 25 percent decrease at mode 3 to a 178 percent increase
at mode 5.
The solid fraction emission trend was generally opposite the total
particulate trend as the range was from a 36 percent decrease at mode 5
(apparent oxidation of some of the solid) to a 65 percent increase at mode 3.
The supposed dehydrogenation of gas-phase hydrocarbons to solid, as observed
with the PTX catalysts, was not observed to a significant extent with the port
catalysts. Mode 5 showed a measurable decrease in solids, and it can be
surmised that oxidation to CO^ and H_0 predominates over dehydrogenation to
solid at this mode, while other modes show only slight predominance of the
dehydrogenation observed with the PTX catalysts.
The SOF emission reductions were generally less than those of the PTX
catalyst (paralleling the gas phase hydrocarbon trend). The sulfate fraction
emissions were greatly increased as expected due to the oxidation of SO to S0_
through the catalyst and subsequent hydrolysis of SO to sulfuric acid. However,
-------
CHRNGE IN HYDROCRRBON EMISSIONS (X)
c_ m
o z
x o
z m
CO T
O X
Z X)
L ^
ID ~
-H ~0
H H
X X
m
o
rn 21
X Z
X CO
X) —I
O 70
co m
-H ID
"0
o o
70 X
O
ID
l~
3) -<
—\ CO
ID —\
r~ co
jf
CO ID
—I 2:
CO O
~n
i—i
o
to
KJ
OS
XI O
o o
70 12
T)
H X
i: TO
O K-H
CO
o o
X) Z
-H
X) O
i—n
co o
—i x
CO X)
z
O CD
z m
co
—i
—T— | |
m z
o x
x -<
—i o
7(3
<^o o
ro o
CD ID
co TO
• • ro
o
co
n
ID
-i
ID
CO
CD
m
o
m
is.
"0
m
70
ID
70
m
CD
m
CD
o
CD
o
ro
CD
CD
CD
-f=-
CD
O
•
O
CO
CD
O
•
CD
CO
CD
O
CD
-100 .0
-80.0 -60 .0 -40 .0 -20 .0
0 .0
X
CTi
oo
o
O
l-S
rt
n
O
rt
en
rt
3
o
a
m
c-m
oz
icn
^:m
cor~
01
Z3D
XiO
oo
7O3
X)
-HX)
2170
O'-
CO
3i rio
5^ X)Z
com '
-'i X)O
So i n
— (TJ
— i-H
IX
m
33
r>r
33-<
CO O
—I X
CO ID
~7_
OO
zm
CO
rnz
ox
X)-<
-HO
70
co o
ro n
01)
CO TO
.. CO
o
CO
Ln
u>
-------
154
U
CD
LU
Q
X
I—
Q_
O
CD
O
LO
O
•
CD
O
CO
CD
•
O
CM
CD
•
CD
0 .0
OPERRTING TEMPERRTURE DIFFERENCES
FOR TWO CflTflLYSTS ON THE CflT 3208
ENSELHRRD PI A DOWN3TRERM CHTHLYST HMD
JOHNSCN-mTTHEY EXHflUST PORT CflTflLYSTS
EPfl MODES 3,4.5.9. 10. 11
(2800 rpm)
(1680 rpm)
j i
200.0
400 .0
600 .0
BRRKE MEriiM EFFECTIVE PRESSURE (KPfl)
FIG.3.2.370PERRTING TEMPERflTURE DIFFERENCES
FOR TWO CflTflLYSTS ON THE CflT 3208=
ENGELHRRD PTX DOWNSTRERM CRTRLYST flND
JOHNSON-MRTTHEY EXHRUST PORT CRTPLYSTS
-------
155
.084
.139
EPA Mode
Base]ine
(g/kW-hr)
1.047
.790
EPA Mode 3
Port Catalyst
(g/kW-hr)
.069
.074
EPA Mode
Baseline
(g/kW-hr)
.482
EPA Mode 4
Port Catalyst
(g/kW-hr)
.172
112
EPA Mode
Baseline
(g/kW-hr)
.709
EPA Mode 5
Port Catalyst
(g/kW-hr)
Key: Q Total Particulate Matter (TPM)
£ Solid Fraction (SOL) of TPM
@ Soluble Organic Fraction (SOF) of TPM
%fc Sulfate Fraction (SO4) of TPM
Control Device Indication
Fig3.2.38aThe Effect of Johnson-Matthey Exhaust Port Catalysts
on Caterpillar 3208 Particulate Emissions at 1680 rpm
(1 Fecralloy based Oxidation Catalyst per Exhaust Port)
-------
156
.730
.786
EPA Mode 9
Baseline
(g/kw-hr)
EPA Mode 9
Port Catalyst
(g/kW-hr)
.626
.693
.792
EPA Mode 10
Baseline
(g/kW-hr)
EPA Mode 10
Port Catalyst
(g/kW-hr)
.857
1.473
1.864
EPA Mode
Baseline
(g/kW-hr)
11
EPA Mode 11
Port Catalyst
(g/kW-hr)
.038
.325
Key:
Total Particulate Matter (TPM)
Solid Fraction (SOL) of TPM
Soluble Organic Fraction (SOF) of TPM
Sulfate Fraction (SO4) of TPM
Control Device Indication
Fig.3.2.38bThe Effect of Johnson-Matthey Exhaust Port Catalysts
on Caterpillar 3208 Particulate Emissions at 2800 rpm
(1 Fecralloy based Oxidation Catalyst per Exhaust Port)
-------
157
the increases in sulfates were never as large as with the PTX catalysts. Figure
3.2.39 shows the molar percent conversion of fuel sulfur to sulfate for the
two catalysts. Both catalysts show the same character of reaching a maximum at
a similar intermediate temperature but the port catalyst shows significantly
lower conversion. Figure 3.2.40a plots the same variable (molar conversion of
fuel sulfur) against catalyst residence time. One can envision a response
surface (molar conversion of fuel sulfur as a function of catalyst residence
time and catalyst temperature) which would consist of a diagonal ridge (see
Figure 3.2.41b). A summary of the effects of both catalysts on fuel consumption
and emissions is given in Table 3.2.17.
The chemical fractionation results are presented in Figures 3.2.42 and
3.2.43 and Table C^4 of the Appendix as the brake specific emissions (mg/kw-hr)
for each subfraction. This table also gives percent composition of the SOF for
the fractions. From the data it can be seen that the catalyst is very selective
with respect to the subfractions. A comparison of the selectivity with respect
to the subtractions of the port catalysts and PTX catalysts is given in Table
3.2.18. There are many cases where there are differences between the two
catalysts such as one catalyst showing an increase in a subfraction while the
other showed a decrease. The most unusual case is mode 10 with the port
catalysts which as Figure 3.2.43 shows had an unusually small decrease in total
SOF. The data of Table 3.2.18 shows this to be the result of the reaction of
hydrocarbons to other types of hydrocarbons rather than to CO and HO. An
increase was observed in the ether insoluble and acidic subfractions with an
four-and-a-half-fold increase in the aromatic subfractions. Although the solid
and SOF were nearly unchanged at mode 10 with respect to the baseline emissions,
-------
158
o
CO
o
f-
Z)
LL
_J
ID
CO
UJ
ID
O
o
o
Q/
CL
_J
O
n r
~i r
~i i r
~i r
MOLRR CONVERSION OF FUEL SULFUR TO S04
VS. CRTflLYST BED TEMPERflTURE.
ENGLEHRRD PTX DOWNSTRERM CRTRLYSTS RND
JOHNSON-MRTTHEY EXHRUST PORT CRTRLYSTS.
CRTERPILLflR 3208. MODES 3.4.5,9.10.11
Engelhard PTX
2800 rpm
Engelhard PTX
1680 rpm
Port Catalyst
2800 rpm
Port Catalyst
1680 rpm
0 .0
200 .0
400 .0
600 .0
800 .0
CfiTfiLYST BED TEMPERflTURE (DEG C)
FIG.3.2.39 SULFUR TO SULFflTE % VS. CflTRLYST TEMP
ENGLEHRRD PTX CRTRLYSTS RNO JOHNSON-
MRTTHEY EXHRUST PORT CRTRLYSTS ON THE
CRT 3208, EPR MODES 3.4,5,9,10,11
-------
159
o
CO
o
o
CO
LJ
ID
U_
O
o
u
CE
_l
O
o
1—T
~r
TT
"1 I T
MOLRR CONVERSION OF FUEL SULFUR TO S04 -
VS. CflTRLYST BED RESIDENCE TIMES.
ENGLEHflRD PTX DOWNSTREflM CflTHLYSTS flND
JOHNSON-MRTTHEY EXHflUST PORT CflTRLYSTS.
CflTERPILLRR 3208, MODES 3.t.5.3. 10,11
Engelhard PTX
1680 rpm
Port Catalysl
1680 rpm
Engelhard PTX
2800 rpm
Port Catalyst 2800 rpm
I I I I I
10
-i
10
10
10
CflTflLYST RESIDENCE TIME (MILLISECONDS)
FI6.3.2.40SULFUR TO SULFflTE % VS. RESIDENCE TIME
ENGLEHRRD PTX CRTflLYSTS RND JOHNSON-
MflTTHEY EXHRUST PORT CRTRLYSTS ON THE
CRT 3208, EPR MODES 3,4,5,9,10,11
-------
160
r
0)
S
Fig. 3.2.41 Response Surface of Molar Percent Conversion of Fuel
Sulfur to Sulfate Through an Oxidation Catalyst on
a Diesel Engine, Data from a Cat 3208 with two types
of Catalysts: Engelhard PTX Downstream Catalyst
& Johnson-Matthey Close-Coupled Exhaust Port Catalyst,
-------
Table 3.2.17 Comparison of Two Catalysts for Diesel Emission Control:
Englehard PTX (Downstream) and Johnson-Matthey Close-Coupled (exhaust port)
PERCENT CHANGE (FROM BASELINE) IN
BRAKE SPECIFIC FUEL CONSUMPTION AND EMISSIONS, No. 2 FUEL
Particulate
EPA
Mode
3
4
5
9
10
11
Brake
Specific
Fuel
Consumption
PTX
+2.4
+ 1.3
-1.4
+6.3
-3.7
-9.8
Port
+ 1.3
-0.6
+0.6
+8.0
+2.8
+7.7
Total
PTX
-36
+415
+637
+ 137
+289
+79
Port
-25
_i
+ 178
+38
+50
+5
Solid
Fraction
PTX
-27
+533
+ 138
+88
+227
+ 15
Port
+65
+8
-36
-1
+5
+61
Soluble
Organic
Fraction
PTX
-59
-76
-67
-75
-88
-84
Port
-35
-69
-40
-61
-14
-73
Sulfate
Fraction NO
X
PTX Port PTX Port
+800 +36 +1 +17
+3500 +700 -4 +2
+4833 +1630 -7 +27
+1033 +826 -31 +21
+3425 +1180 +2 +16
+8167 +796 -11 +13
Gaseous
N02
PTX Port
-32 -41
+1420 -29
+530 +189
+110 +164
+39 +28
+320 -42
NO
PTX
+4
-33
-15
+30
+ 1
-21
Gas Phase
Hydrocarbons
Port
+25
+5
+20
+ 17
+ 16
+ 17
PTX
-88
-85
-84
-76
-87
-79
Port
-8
-49
-56
-56
-49
-40
PTX - Caterpillar 3208 with Engelhard PTX downstream catalysts
Port - Caterpillar 3208 with Johnson-Matthey Close-Coupled Exhaust Port Catalysts
-------
Table 3.2.18 Comparison of the Selectivity of Catalysts with Respect to Soluble Organic Fraction
Catalysts: Englehard PTX (downstream) and Johnson-Matthey Close-Coupled (exh. Port)
Engine: Caterpillar 3208 Operated at EPA Modes 3, 4, 5, 9, 10, 11 (Dilution Ratio 15:1)
Percent Change in Mass Concentration
EPA MODE
3 4 5 9 10 11
Subfraction
Ether
Insoluble
Basic
Acidic
Paraffin
Aromatic
Transitional
Oxygenated
Hexane
Insoluble
PTX Port PTX Port PTX Port PTX Port PTX Port PTX Port
-59 +55 -65 -46 -79 -28 +125 +35 -99 +36 -72 +10
-87 +7 -63 -92 -48 -85 -68 +17 -76 -72 -91 -90
-78 -19 -60 -53 -31 +31 -74 -83 -85 +42 -90 -75
-63 -47 -83 -74 -93 -46 -54 -80 -93 -37 -86 -79
-42 -42 -93 +12 -8 +86 -38 -60 -92 +462 -80 -46
-76 -31 -79 -78 -69 -67 +60 -48 -97 -56 -87 -80
-62 -22 -52 -76 -54 -57 -91 -55 -90 +4 -80 -80
-73 +11 -45 -25 -75 -86 -95 -89 -89 -46 -57 -43
PTX - Cat 3208 with Engelhard PTX downstream catalysts
Port - Cat 3208 with Johnson-Matthey Close-Coupled Exhaust Port Catalysts
K5
-------
BRRKE SPECIFIC EMISSION (MG/KW+HR)
200.0 400.0 600.0 800.0 1000.0
-n
i — i
•
u>
•
•
-P-
N3
3)0 Z I W
73 r~ m ...
-HC-I -v
*-* oo x m T
o r~ m ~n
cm ~n
r~ n m
3D O X 0
— i 73 m — i
m CD x ni
3D i— i O "^
~n i — i ~n ~r t-^
~\ o r~ m o -.,
-H X 0 ^
fl '1 ( ) _J_ m
/U AJ CJ _D
• 3D 3 C
n ~o co
— 1 O -i
i— i CO
O — ' T>
z. -H o w
>-< 73 vjl
00-1 ^ -o
-TlZ
o
n o 3D
3D "H -H
— 1 ID
-H r~
oo X -<
r<> m co
0 — |
°D CO
t
j
t
(-
^
1 —
1
7 P
L
4
IE
1
£
. . . 1 . , 1 , , , 1 . . , 1 . . . , , - ,
V-1 i i *u isls °^ ,.
--^m _,m
*sss mco
"0^m ~n —
I^zpj mm
S^°a (—!(-) _
|| i ' " ' ' J | UrT"T8 CD "— <
_ "-TI, J i r no-! -J- '
5W>lliJ^'h3Otc "'^o — 1 l~~l
:i i i i °D r n
„ . 1 J 1 oo -n ^ >
i - • 1 i ' ' m? O CO
I U)-H '
— ~4T CO C —
CCD
__.._.. -« •» «» oo r~ .
.1 — - — | iii -urn
— ~\ 1 Hr> o T> n AJ
^ i i « i — IGD •
• ' _\ -> m h—i j)
==J_I O2. .
2>n3*;oxx COO —
zinOTi3Z-< —
i I i i i i i i
mto>-o>HOX -ri
' r -i>o>a>ajxm 'i
1 | X — 33O1>-
-------
BRflKE SPECIFIC EMISSION (MG/KN+HR)
200.0
400 .0
u>
NJ
-P-
TJ (/) O -i
T) 0 Z X
TO r~ m
i /— i
1 1_ i
•— • CD x m
o r~ m TI
cm -n
[— n m
TJ o x n
— i TO rn -H
m CD 2
TJ ^ O
:z ~z. n ~n
T> •— < T)
— i o r~ m
-H X
m ~n n x
70 70 0 T)
. TI X C
O TJ CO
-H 0 -\
*-> CO
O •— TJ
~Z- -H 0
>— " 70
o o — i
*n ~~^
1 1 -i_
n
n o TJ
T) TI — 1
-1 TJ
1 r—
— 1 1
co X -<
rv> m CO
0 —1
oo CO
-o
2
O
O
m
-------
165
an increase in both aromatics and ether insolubles suggests that dehydrogenation
to solid may predominate over oxidation to CO and HO (or other, oxygenated
subtractions) at this mode.
Effects on Emissions - Shale Fuel - Figure 3.2.44 shows brake-specific
NO , NO, and NO- emissions for the close-coupled port catalysts at modes 3, 4,
X £-
and 5 using shale fuel. Baseline (uncatalyzed) values were obtained at mode 4
only. Shale fuel appears to give BSNO and BSNO values slightly lower than No.
X
2 fuel, with BSNO emissions higher than No. 2 fuel, except for mode 5.
Baseline emissions with No. 2 fuel at mode 4 were essentially unchanged from
those experienced with the port catalysts.
Figures 3.2.45 and 3.2.46 show BSHC and BSTPM for shale fuel at modes
3, 4 and 5 compared with the same modes using No. 2 fuel, for the port
catalysts. Shale fuel gave slightly higher HC emissions at all 3 modes, and the
SOF was higher at modes 3 and 4 (mode 5 SOF was insignificantly lower for shale
fuel). Neglecting sulfate, the total particulate was lower at modes 3 and 4 for
shale than for No. 2 fuel. Significantly lower solid emissions at mode 5 for
the No. 2 fuel were not enough to offset the very high sulfate emissions at mode
5 for the No. 2 fuel and BSTPM for shale fuel was approximately half the No. 2
fuel emissions. The significantly higher solids emissions at mode 5 for shale
fuel could be due to a 28°C higher catalyst temperature, but is more likely due
to fuel properties. As discussed previously, the only mode for which solids
showed a significant decrease over baseline emissions using No. 2 fuel was mode
5. Also, it was mentioned in the discussion of the PTX catalyst that aromatic
compounds may be much more stable in the shale fuel than in the No. 2 fuel.
-------
166
(Y.
IE
I
IK
CD
CO
O
i—i
CO
CO
I—I
LU
X
o
z
Q
O
Z
o
z
CO
00
o
0 .0
CRTERPILLRR 3208
EPfl MODES 3.*.5
Baseline (no control) and equipped
with exhaust port catalysts
Catalyzed, No. 2 Fuel
N0x:
NO:
Catalyzed, Shale Fuel
Catalyzed, Shale Fuel
Catalyzed, Shale Fuel
N02: Un
Catalyzed, No. 2 Fuel
Ilncatalyzed,
Shale Fuel
i i
i i
i i
1
200 .0
400 .0
600 .0
BRflKE MEflN EFFECTIVE PRESSURE (KPfl)
FIG.3.2.44BRflKE SPECIFIC NO, N02 . flND NOX NITH RND
WITHOUT CLOSE-COUPLED CRTRLYSTS
SHRLE FUEL; NO.2 FLJ£|_
-------
167
I
I
3
^
\
CD
CO
O
i—i
CO
CO
I—I
UJ
U
CJ
i—i
LL
i—i
U
Q_
CO
LU
CD
i 1 1 1 i i i
CflTERPILLflR 3208
Equipped with Johnson-Matthey
exhaust port catalysts
Shale Fuel, Catalyzed
No. 2 Fuel, Catalyzed
0 .0
200 .0
400 .0
600 .0
BRflKE MERN EFFECTIVE PRESSURE (KPfl)
FIG. 3.2.45 EFFECT OF SHflLE FUEL ON HYDROCRRBON
EMISSIONS USING CLOSED-COUPLED PORT
CRTRLYSTS
-------
EPA Wbde 3
No. 2 Fuel
(g/kw-hr)
.139 '
087
.790
.626
168
EPA Mode 3
Shale Fuel
(g/kW-hr)
EPA Mode 4
No. 2 Fuel
(g/kW-hr)
.221
EPA Mode 4
Shale Fuel
(g/kW-hr)
.224
EPA Uode 5
No. 2 Fuel
(g/kW-hr)
266
Key: (2)Total Particulate Matter (TPM)
@Solid Fraction (SOLID) of TPM
^Soluble Organic Fraction (SOF) of .TPM
^(Sulfate Fraction (SO^) of TPM
Fuel Change Indication
EPA Mode 5
Shale Fuel
(g/kW-hr)
Fig. 3.2.46 The Effect of Shale Fuel, Relative to No. 2 Fuel,
on Caterpillar 3208 Particulate Emissions at 1680 rpm
(1 Fecralloy based Oxidation Catalyst per Exhaust Port)
-------
169
This premise is supported by the observation of predominantly dehydrogenation
(to solid) in the case of shale fuel, as opposed to predominantly oxidation (to
CO and HO) for No. 2 fuel. The very slight reduction of solids using shale
fuel with the port catalysts at modes 3 and 4 suggests that there is a threshold
temperature (near 500°C) above which catalysis of fuel-dependent oxidation and
dehydrogenation occurs.
Ames Bioassay Results - The mutagenic specific activities for both the
baseline and the port catalyst SOF using No. 2 fuel are plotted in Figure
3.2.47t The port catalysts produced SOF of higher specific activity at all
intermediate speed modes but produced a decrease in SOF specific activity at all
rated speed modes. This is contrary to the PTX catalyst which produced
increases at all modes.
The results of Ames tests on the subfractions of the SOF for both baseline
and port catalysts at mode 4 are given in Table 3.2.19. The log-log SA
parameter is compared with the SA calculated from a log-linear dose-response
curve; little difference is seen. There is an increase in SA for most of the
subfractions of the SOF due to the port catalyst which is consistent with the
significant increase in SA of the total port catalyst SOF.
Specific activities for the SOF at all 6 EPA modes tested are shown in
Table 3.2.20. BSSA (shown graphically in Figure 3.2.48) are calculated based
upon log-log SA parameters. Limits of the linear dose-response region were
chosen uniformly at C = 18.75 and C = 600 pg/plate except for the mode 4 basic
subfraction, where sample mass limited C to 300 yg/plate • Mode 5 SA with the
port catalysts was much higher than that for any other mode, with or without the
catalysts. This indicates that temperature and residence time are sufficient at
this mode to produce significantly higher levels of mutagenic compounds.
-------
170
CE
(Y.
CD
O
C£
O
UJ
U
CE
O
O
LU
CL-
IO
CD
LO
•
OJ
CD
CD
•
OJ
CD
in
CD
CD
CD
LO
~I I I ' I ~\ I I I 1 T
SOF SPECIFIC nCTIVITY VS. BMEP
CflTERPILLRR 3208 BflSELINE (NO CONTROL)
flND EQUIPPED WITH EXHflUST PORT CflTHLYSTS
EPfl MODES 3.1,5,9.10,11
1680 rpm
port catalyst
1680 rpm
baseline
2800 rpm
port catalyst
J I L
J L
J I L
1
0 .0
200 .0
400 .0
600 .0
BRflKE MEflN EFFECTIVE PRESSURE (KPfi)
FIG.3.2.47THE EFFECT OF EXHflUST PORT CflTflLYSTS ON
SPECIFIC RCTIVITY OF THE CRT 3208 SOF.
CRTRLYSTS: JOHNSON-MRTTHEY FECRRLLOY
SUPPORTED OXIDRTION PORT CRTRLYSTS
-------
171
Table 3.2.19 Ames Dose-Response Statistics for EPA Mode 4 Subtractions,
No. 2 fuel. C = 18.75 Ug/plate; C_ = 600 yg/plate except Basic-Port
Catalyst (C = 300). Assayed 11/03/81; TA-98 w/o S-9.
SA
of Total SOF
DUU L LclUUiUll
EIN
EIN
BAS
BAS
ACD
ACD
PRF
PRF
ARM
ARM
TRN
TRN
OXY
OXY
HIN
BIN
Baseline
, N Johnson-
SASL '
SL c
.L/CV -LL-C
B">
Po»
B
P
B
P
B
P
B
P
B
P
B
P
B
P
Engine (no
SL h
0.5343
1.0449
0.0849
0.4565
1.4749
2.5373
0.0616
0.0177
0.3843
0.0587
0.1234
0.3055
0.4442
0.6541
0.5647
0.4403
control)
Matthey Close-coupled port
login L2 , Model:
2 " \
ioa(c2b- GI
, Model: Rev
t£ tj^i T T * ' "O '°
0.5459
1.1231
0.0876
0.4780
1.5909
2.8736
0.0548
0.0187
0.3954
0.0563
0.1830
0.2982
0.4678
0.6629
0.4641
0.4210
catalyst
Rev = a + b l°gin (mass)
= a (mass)
8.4
15.1
2.6
0.7
4.1
6.5
56.6
48.5
2.3
8.7
5.5
4.0
16.2
13.1
1.4
3.4
-------
172
Table 3.2.20 Ames Dose-Response Statistics for Total SOF; C =18.75,
C2 = 600 yg/plate throughout. Assayed 1/11/82; TA-98 without S-9. See
Previous Table for Explanation of Parameters.
EPA
Mode
3
3
4
4
5
5
9
9
10
10
11
11
Control
Device
B
P
B
P
B
P
B
P
B
P
B
P
SASL
Rev/yg
0.4393
0.5709
0.3696
1.0957
1.2247
1.9416
0.8372
0.4602
0.7456
0.1845
0.4433
0.3959
SALL
Rev/yg
0.4116
0.5369
0.3296
1.0355
1.3224
2.1738
0.8211
0.5094
0.7367
0.1733
0.4450
0.3873
BSSOF
g/kw-hr
0.932
0.609
0.377
0.115
0.38
0.124
0.026
0.011
0.133
0.120
0.969
0.281
BSSAVd'
kRev/kw-hr
383.6
327.0
124.3
119.1
50.3
52.2
21.4
5.6
9.8
20.8
431.2
108.8
(a) BSSA = (SALL) (BSSOF) (lOy g/g) (kRev/lOOORev)
-------
173
LY
ic
i
is
LJ
LY
U
cr
U
U
UJ
D_
CO
CO
CD
CD
CD
CD
CD
OO
o
•
CD
O
CM
CD
O
T 1 1 1 1 1 I I I I
BS SPECIFIC RCTIVITY VS. BMEP
CfiTERPILLRR 3208 BRSELINE (NO CONTROL)
flND EQUIPPED WITH EXHflUST PORT CRTflLYSTS
EPR MODES 3.4,5.9, 10, 11
2800 rpm
baseline
1680 rpm port catalyst
1680 rpm baseline
2800 rpm
port
catalyst
J I L
0 .0
200 .0
400 .0
600 .0
BRRKE MERN EFFECTIVE PRESSURE (KPR)
FIG.3.2.48 THE EFFECT OF EXHRUST PORT CRTRLYSTS ON
BS SPECIFIC RCTIVITY OF THE CRT 3208.
CRTRLYSTS: JOHNSON-MRTTHEY FECRRLLOY
SUPPORTED OXIDRTION PORT CRTRLYSTS
-------
174
Corning Ceramic Trap
The third aftertreatment device investigated was a ceramic trap, from
Corning Glfss, material designation EX-47. Table 3.2.20 gives the
specifications of the tiaps.
Table 3.2.20 Corning Ceramic Trap Specifications
Manufacturer Corning Glass Company
Material Modified Cordierite (2MgO'2Al 0 "5S10 )
Porosity 0.50
Cell Wall Thickness 0.017"
Mean Pore Diameter 12 - ]3 ym
Cell.Size 15.5 cells/cm2
Nominal Dimensions 14.37 cm dia x 30.48 cm long
Six Corning traps were obtained from Corning Glass Works (three of the six
uncatalyzed, three of the six catalyzed), which were fitted and sealed in
metal containers with 3M Interam which is a mat material (heat expanding)
developed for automotive catalytic converters (72). The metal containers (one
trap per cylinder bank) were easily removable from the exhaust system to permit
regeneration in an oven if necessary. Due to durability problems and
anticipated high sulfate output, the catalyzed traps were not utilized for test
purposes.
Table 3.2.21 is the test matrix utilized for the Corning Traps. No. 2
fuel was used for all engine tests. Due to extremely low particulate and SOF
emissions, no sampling was conducted at modes 5, 9 or 10 for chemical
fractionation, which requires large amounts of SOF.
-------
Table 3.2.21 Test Matrix for Corning Trap Evaluation
EPA Mode
Speed (RPM)
Load (N-M)
. BMEP (mPa)
Trap
3
1680
160
192
w/o w
4
1680
320
383
w/o w
5
1680
480
575
w/o w
9
2800
399
485
w/o w
10
2800
266
322
W/O '
Total Particulate,
SOF, and SO, mass emissions x x
HC, NO, and NO
Mass Emissions
x
11
2800
133
161
XX XX XX XX XX
XX XX XX XX XX XX
Chemical Characterization
of SOF
Ames Mutagenicity
Bioassay on SOF
Ames Mutagenicity
Bioassay on Subfractions
X X
xx xx
x x
X X
x x
-------
176
Effect of Corning Traps on Engine Operation and Mass Emissions - The
results of the Corning trap tests must be viewed with caution because of the
time-dependent behavior of the traps during the engine tests. Depending on the
engine output of particulate matter and the exhaust temperature, the pressure
drop across the traps varied with time. This is illustrated dramatically in
Figure 3.2.49 with similar behavior among the intermediate speed modes (3, 4,
5) and vastly different behavior among the rated speed modes (9, 10, 11). The
intermediate speed modes are characterized by fairly low particulate rates and
exhaust temperatures below the temperature necessary for rapid oxidation of
filtered particulate (i.e. fairly low but steady rates of trap pressure drop
increase). The rated speed modes are characterized by high particulate rates
and a range of exhaust temperatures bracketing the rapid particulate oxidation
rate temperature. Mode 9 has an exhaust temperature of 625°C which apparently
produced an oxidation rate equal to the high particulate filtration rate (zero
increase in pressure drop with time). Mode 11 has an exhaust temperature below
the rapid oxidation temperature and therefore suffered a high pressure drop
increase rate due to its high particulate filtration rate. Mode 10 began its
time dependent behavior similar to mode 11 but had an initial temperature much
higher than mode 11, and the friction of increasing pressure drop gradually
raised the temperature (Figure 3.2.50). When the mode 10 exhaust temperature
reached 925°F (496°C) at time =110 minutes, the pressure drop slowly decreased.
This suggests that 925°F (496°C) is sufficient to produce rapid particulate
oxidation in the Corning traps (at least for the oxygen and hydrocarbon
concentrations of mode 10 and the heat transfer characteristics of the trap
container system).
-------
177
UJ
I—
CE
Q_
O
C£
O
LJ
CO
CO
LD
O
O
o
O
00
O
O
Cvj
O
•
O
O
T
T
TIME DEPENDENT BEHRVIOR OF CORNING TRRP-
PRESSURE DROP RCROSS TRRP VS. TIME
CORNING TRflP <5.66 X 12.) EX-17 MflTERIRL
CflTERPILLflR 3208 (ONE TRflP/CYL SflNK)
EPR MODES 3.4,5.3. 10. 11
.Mode 11
Mode 3
I ' I L
1
l 1 I
0.0
100 .0
200 .0
ENGINE RUN TIME (MINUTES)
FIG.3.2.49TIME DEPENDENT BEHRVIOR OF CORNING TRRP
CORNING TRRP <5 .66 X 12.) EX-47 MRTERIRL
CRTERPILLRR 3208 CONE TRRP/CYL
EPR MODES 3.4,5,9, 10, 11
BflNK)
-------
178
U.
CD
UJ
O
•w
UJ
ID
I—
CT
UJ
Q_
51
UJ
CO
ID
CE
1C
X
UJ
o
o
CM
O
CD
O
o
o
CD
•
CD
CD
O
•
O
o
00
o
o
I I I —I I 1 I 1 1 , 1 , 1—
TIME DEPENDENT BEHRVIOR OF CORNING TRRP
ENGINE EXHRUST TEMPERRTURE VS. TIME
rnDMTur: TDC.D /c rr v ,n ^ «-u .,-, ^.^I-^-I^T^. ^^— • i J. I i i—
CORNING TRRP (5.66 X 12.) EX-47 MRTERIflL
CflTERPILLRR 3208 (ONE TRflP/CYL BflNK)
EPB MODES 3.10.11
Mode 9
Mode 10
Mode 11
j i i
I ' L
1
I I
0.0
100 .0
200 .0
ENGINE RUN TIME (MINUTES)
FIG.3.2.50TIME DEPENDENT BEHRVIOR OF CORNING TRRP
CORNING TRRP (5.66 X 12.) EX-4? MRTERIRL
CRTERPILLRR
EPR MODES
3208 (ON
9.10
TRRP/CYL
11
BRNK)
-------
179
The time dependent pressure drop produced a time dependent fuel consumption
as shown in Figure 3.2.51 for the rated speed modes. The average fuel
consumption increase rates were: 0 for mode 9; +2.6 percent/hour for mode 10
(initial phase); and +3.4 percent/hour for mode 11.
In the remainder of the Corning trap results presented, the time dependent
behavior has been eliminated by taking the average of the results of three
sequential one hour tests.
The effect of the higher pressure drop in the exhaust system resulted in
higher fuel consumption at all modes as shown in Figure 3.2.52 The greatest
increase in fuel consumption was at mode 11 which corresponds to the highest
average pressure drop mode.
The gas phase hydrocarbon emissions are shown with the baseline in Figure
3.2.53 Fairly significant reductions are observed at the rated speed modes
with smaller reductions at intermediate speed. This reduction in hydrocarbons
is consistent with past results reported where bare (uncatalyzed) ceramic
catalytic converter monoliths produced reductions due to the catalytic effect of
the large, hot surfaces. This effect should be even greater in the Corning
traps due to the intimate gas-solid contact as the exhaust passes through the
porous walls. The increase in hydrocarbons observed at mode 3 is due to an
unanticipated upward shift in baseline hydrocarbon emissions at this mode rather
than an increase through the traps. The reduction in hydrocarbons through the
traps appears to be a function of load (temperature) as shown in Figure 3.2.54
which supports the hypothesis that oxidation of hydrocarbons is occurring
through the trap rather than adsorption onto the trapped particulate.
The NO and NO emissions were not significantly changed as a result of
X
installation of the traps. The BSNO emissions showed decreases through the
traps under all conditions except mode 3 (see Figure 3.2.55). This is not
-------
BS FUEL CONSUMPTION (KG/KU-HR)
en
rn n
"O ID
33 -H
m
IK TO
O "0
O I-H
m r~
co r
ID
TO
o
o
TO
z
m
IX)
o
CO
CO
O
m
TO
ID
"0
o
-<
CD
m
~o
—i m
TO z
D CD
TI m
z
/•^, I...-J
en
CD rn
en x
ID
X <
o
TO
O
rn
x n
i o
00 I-H
ID 12 Z
z ID cn
C- m —i
70 70
—• ID
X) "0
m
CD
z
m
70
d
12
m
m
cn
CD
O
.300
.350
CD
CD
-i 1 1 r
~\ r
3
O
a
n>
.450
o
a
s
o
a
(D
S5SSDO
m
i
o -
e <
O —
OO
Z TO
CO
CIO
12 "n
TI
—in
•—o
OTO
-------
BS FUEL CONSUMPTION (KG/KW-HR)
CD
Ln
K>
m —\ o —i
• 73 Z 31
CD 3D • m
en -g (jj
co co m
o
i n
o
X 73
z
a m
m n
z m
O X
31 I
^ -o
O
-H m
73 CO
3D >—
TI CD
O 3D
~< —I
O O
o TI
CO O
d O
CD
o
O CD
z:
~D
O ID
"n TO
-H
-H »-H
n: n
m cz
r~
O 3D
3D —I
—i m
co —I
ro ^j
o 3D
oo T)
• CO
OD
^1
3D
7\
rn
m
3D
m
~n
~n
m
n
m
~D
73
m
co
co
m
Tl
3D
O
•
CD
.250
.300
.350
.400
ro
o
o
•
o
4=-
O
O
en
o
o
•
o
CD
CO
c:
m
r~
n
o
z
CO
c:
iz
T)
<
CO
DO -
^
m
~D
00
-------
BS HYDROCRRBON EMISSIONS (G/KW-HR)
C71 -H
• 70
en ID
CO TJ
CO
i—i ..
z:
o
I O
o
X 70
_
CD
»—i
z m
o x
IT
•n
I—I
CD
O -H
Z T.
m
x
-< m
o ~n
o rn
o o
TJ -H
70
co o
o ~n
z
n
rn o
is TO
, , -7
o
-H m
TO co
TJ •-*
TJ CD
^ Z
O TJ
-< -H
r—— , ,
O
CO Z
TJ
Z
7\
CO
o
z
CO
o
CD
70
n
o
i-o
oo
ID
T)
CO
DO
70
Z)
7\
m
m
m
-n
~n
m
o
m
TJ
70
m
co
co
c:
70
m
TJ
TJ
O
•
O
1 .00
2 .00
3 .00
o
CD
4=-
CD
CD
CD
CD
CD
T
I
N)
00
ft O
n o
cr
DJ oo
t/1 O
0) o
M
H- ^
3 tJ
n> 3
I
"OXZ 3>
05 I O-l
« m
3^=m^j
O- «DT)
D C— i
-
X
CD
O Tl
1 o 01
too
co o
n> z
00
CO
70
O
n
OD
O
co
CO
t— *
o
z
CO
CO
OD
12
m
TJ
oo
-------
183
o
Cd
CO
z
o
m
CK
cr
o
o
&.
a
o
•
o
UJ
CD
z
FIG.3.2.54PERCENT CHRNGE IN GRSEOUS HYDROCRRBONS
DUE TO CORNING TRRPS ON THE CRT 3208
TRRPS: CORNING EX-47 DESIGNRTION
5.66 INCH X 12 INCH <1 TRRP/CYL BRNK)
-------
184
consistent with the observed oxidation of hydrocarbons through the traps. Since
the hot trap surfaces lead to hydrocarbon oxidation, it is expected that they
would also oxidize NO to NO . The increase in back pressure with time may cause
an effect similar to EGR in that partial retention of oxygen-poor exhaust in the
cylinder dilutes the incoming air charge and lowers peak combustion temperatures
and NO levels. A second possibility for reduction in N0_ levels is possible
^- L.
physical or chemical adsorption of NO onto the particulate, as evidenced by
work of Gibson et. al. (73) and investigated by Ahmed (74) in the context of
sampling errors. Ahmad found that the predicted activation energy for NO
X
adsorption agreed with a measured activation energy of a physical adsorption
process; however, these studies were conducted at room temperature. Greater
adsorption of NO at higher temperatures should indicate the extent of NO
X X
chemisorption. Yet another possibility for lowered NO levels using the trap is
reaction of N0_ with water vapor according to
3N02 + H20 -> 2HN03 + NO.
The increase in exhaust back pressure and increase in mixing caused by abrupt
flow changes through the trap walls could increase the residence time
sufficiently to favor this reaction, which proceeds efficiently without a
catalyst (75).
The particulate emission results are summarized in Figures 3.2.56a and
3.2.56b. The TPM emissions were reduced at every engine condition ranging from
an essentially zero reduction of 0.7 percent at mode 3 to a 97.2 percent
reduction at mode 9.
The solid fraction of the particulate was of most interest because it is
the only component of the particulate which the trap is fundamentally capable of
trapping. Some amounts of hydrocarbons could be adsorbed into the dead ended
pores of the trap and the trapped particulate itself but this would be minimal
-------
BRRKE SPECIFIC N02 EMISSIONS (G/KW-HR
01 -H
• 70
en ID
o~> T)
CO
i—-t ..
z
o
X O
o
X 70
ro Z
CD
t—H
z m
n x
O
-H m
70 CO
ID —'
T) CD
\ Z
O 3D
-< -I
O
CD
ID
Z
7\
CD
•
u>
•
to
*
Ui
Ln
O —I
z :r
m
z
o m
m m
iz n
>-< —i
co
CO O
K—4 Y\
o
z n
co o
70
~n z
7D •—'
o z
IS CD
—I T»
IL ID
m TO
-H
O —i
ID O
co ID
ro —i
CD m
oo
. —I
70
D
~O
CO
CD
70
3)
7\
m
m
ID
m
~n
~n
m
n
m
70
m
co
co
c
70
m
T)
JD
O
•
o
.25
.50
,75
1 .00
O
O
CD
CD
•
O
i—r
i—i—r
i—i—
1 .25
"T^
rt oo
i^ o
D) O
1
tr to
fa CD
en o
(D o
M
H- n
3 ^
(D 3
v ro -HCO
u> I
• ^ CD
— O03J
OZOui
• tnxim
^ zr
too
o
i i I L
_i L
1 L
' L
70
ID
7\
m
co
TI
m
n
O
Z
o
m
CO
CO
I — I
o
z
CO
CO
•
00
n •
m
-o
J - 1 - 1 - L.
00
Ul
-------
186
.084
OSS
EPA Mode
Baseline
(g/kW-hr)
1.039
EPA Mode 3
Corning Trap
(g/kW-hr)
.069
EPA Mode 4
Baseline
(g/kW-hr)
.022
.173
EPA Mode 4
Corning Trap
(g/kW-hr)
.172
EPA Mode
Baseline
(g/kW-hr)
.005
.015
EPA Mode 5
Corning Trap
(g/kW-hr)
Key i O Total Particulate Matter (TPM)
£ Solid Fraction (SOL) of TPM
® Soluble Organic Fraction (SOF) of TPM
$fo Sulfate Fraction (SO4) of TPM
—-- — .Control Device Indication
Fig.302.56aThe Effect of Corning EX-47 Ceramic Particulate Traps
on Caterpillar 3208 Particulate Emissions at 1680 rpm
(1-5.66" X 12" Corning EX-47 trap per cylinder bank)
-------
187
.730
009
EPA Mode
Baseline
(g/kW-hr)
.022
EPA Mode 9
Corning Trap
(g/kW-hr)
.626
.008
EPA Mode
Baseline
(g/kW-hr)
10
022
EPA Mode 10
Corning Trap
(g/kW-hr)
.857
EPA Mode
Baseline
(g/kW-hr)
11
.038
.049
.139
EPA Mode 11
Corning Trap
(g/kW-hr)
.024
K«y:
Total Particulate Matter (TPM)
Solid Fraction (SOL) of TPM
Soluble Organic Fraction (SOF) of TPM
Sulfate Fraction (SO4) of TPM
Control Device Indication
Fig,3.2.56bThe Effect of Corning EX-47 Ceramic Particulate Traps
on Caterpillar 3208 Particulate Emissions at 2800 rpm
(1-5.66" X 12" Corning EX-47 trap per cylinder bank)
-------
188
at typical locations of the trap near the engine and at moderate to high loads.
The trap would reach a steady state eventually when the adsorption capacity is
reached. Likewise sulfates could become attached to the trapped solids. The
percent reductions in SOL are plotted in Figure 3.2.57. The rated speed
reductions are all greater than 95% while the reductions at intermediate speed
are strongly a function of load, varying from 0.7% at mode 3 to 97% at mode 5.
An effort at explaining this behavior was made by developing a model of the
Corning trap based on membrane type filter theory. The development and
application of the membrane filter model is covered in detail in a later
section.
The SOF emissions were reduced at every mode except mode 3 (the same one
where an increase in hydrocarbons was observed). The SOF reductions cannot be
explained by the hydrocarbon reductions alone. The SOF reductions are parallel
to the SOL reductions which tends to support the accepted adsorption theory of
SOF formation. This adsorption theory application will be treated in the
Discussion section.
The sulfate fraction (S0~) was reduced at every mode from 42% to 76%
(significant in every case). Some possible explanations for this observation
are: sulfates becoming attached to the trapped solid particulates, lower sulfate
conversion in the exhaust system/dilution tunnel due to the lower particle
concentrations reducing the known catalytic effect of particles on SO. formation
(48), or lower humidity during the trap test resulting in lower SO,(48). It is
likely that the observed sulfate reductions were the result of a combination of
these factors. A summary of the effects of the Corning traps on fuel
consumption and gaseous and particulate emissions is given in Table 3.2.22.
The effect of the Corning traps on the chemical composition of the SOF is
illustrated in Fig. 3.2.58a and 3.2.58b and tabulated in Appendix C. From the
-------
Table 3.2.22 - Summary of the Effects of Corning Ceramic Particulate Traps on Caterpillar 3208;
Fuel Consumption and Emissions
PERCENT CHANGE (FROM BASELINE) IN
BRAKE SPECIFIC FUEL CONSUMPTION AND EMISSIONS
Particulate
Gaseous
Soluble
Gas Phase
EPA
Mode
3
4
5
9
10
11
BSFC
+4.9
+4.8
+2.8
+4.0
+2.1
+ 11.1
Total
-0.7
-64.1
-93.8
-97.2
-97.2
-93.1
Solid
-34.5
-68.1
-97.0
-98.7
-98.8
-94.7
Organic
+4.3
-63.7
-94.7
-98.0
-95.6
-93.6
Sulfate
-61.3
-63.9
-75.5
-69.2
-73.3
-42.7
NO
X
+5.3
-6.5
-5.6
-12.2
-1.0
+6.6
N02
+ 1.7
-57.0
-61.1
-60.0
-52.0
-76.9
NO
+5.7
-2.3
-3.3
-10.7
-0.6
+ 12.5
Hydrocarbons
+ 12.0
-23.4
-28.0
-59.3
-58.9
-29.0
00
-------
PERCENT REDUCTION IN SOLID PflRTICULRTES
CD
•
u>
•
NO
•
t_n
en
o —r
• 70 Z: 31
CD 3D m
CT) "TJ O
co 3) m
i—< > • —i ~TI
z ~n
n co m
3: n ^ n
o o -H
x TO c»
z o
*-• >—* co ~TI
ro z: O
en H n
I—< l-H O
z m o TO
n x z
31 I TI H-.
-f^ 3) Z
^ -N) 70 Q
H- —)
O *-• "0
-H m n 3D
70 CO d 70
TJ CD
\ Z
O 3D PI
-< H CO
CD
3)
Z
7\
O
n
c:
r~
3)
—I
m
70
Z)
"D
CO
CD
70
3D
7\
m
m
m
m
n
m
T)
73
m
co
co
c
70
m
3)
o
•
CD
20.0
40 .0
60 .0
80 .0
100 .0
ro
o
CD
o
o
CD
O
CD
cr
ro
H
ro
cr
tn
en
O
tr1
13 cr
>-( d)
— rt X"
H- ro
o
d cn
jn ro
rt O
ro H-
HI
ro H-
3 o
p-
(0 Cfl
w O
P- M
-O H-
3 Cb
tn
cr
(A
en
o
en
(D
ro
M
n
ro
.
ro
a
c
o
rt
p-
O
3
cr
01
en
O
cr
&
tn
ro
cr
05
en
O
rt
t-l
i
I
^: ou")
fti^om
zr
CD O
3> Z
* P
m
70
o
m
70
m •
a
d
n -
o
z:
co
o
r~
i—i
a
co
CO
v
CO
^
m
-------
191
O
0
<^v 0
f^^^ * '
X
+
2
^ o
\
CD o
z o
•w 00
O
CO o
CO
"-« O
z: o
LU cx>
CJ
i— «
U.
|-H o
LU o
9z °
CO ^t-
LU
<£
£ -
O
f— 1
Cvl
(1(7
• i i
1
' ' ' ' I
- BRRKE SPECIFIC SOLUBLE ORGRNIC
1111
FRRCTION -
. RND THE EIGHT SOF SUBFRRCTIONS
BASELINE ENGINE (NO CONTROL) RNO WITH
CORNJN5 EX-17 CERflMIC PRRTICULflTE
TRflPS ON THE CflT 3208 <1 THHP/CYL BHNK)
EPfl MOOES 3.1.5.9. ID. II
-
-
—
••••
HXI-*
OXY— J
TRN 1
— ARM-
PRF-
ACD
BAS-
FTN-i
—
—
— —
B - BASELINE HXI-HEXANE INSOLUBLE .
-i OXY-OXYGENATED
TRN-TRANSITIONAL
P - PORT ARM-AROMATIC
CATALYST PRF-PARAFFIN
T - CORNINfi EIN-ETHER INSOLUBLE
LHXI—
TRAP
LOXY— ,
L
... TPM
— ARM —
___
1 l\L
Apn
i-BAS— i
— EIN-i
••1H
""^^™
HXI 1
nw. 1
OXY- I
p-TRN— J
ARM J
i f\FM -<\ |
i\\. .PHP— Bj
r^ACD-i i
| ^^ !
rl^l L
IrEIN-j |JT*^*1
^.Jl [H i ,
_^
-
„
—
_
pHXI
-OXY
-TFSN —
— ARM
-PRF
-ACD
-HAS
t?TM
E.XIN
--^-. i '
B T
3
B T
EPA MODE
B T
5
.s 2 58a~THE EFFECT OF CORNING CERflMIC TRflPS
ON THE CHEMICRL COMPOSITION OF THE
SOLUBLE ORGflNIC FRRCTION OF CRT 3208
PRRTICULRTE MRTTER .
-------
192
+
2
^
\
CD
z;
-z.
o
CO
CO
(—<
s:
LU
O
h—1
LU
t—t
u
LU
CU
cn
LU
(T
OQ
o
o
o
o
o
o
0
CO
o
o
o
CO
CD
o
o
^
o
o
0
CsJ
' 1
i 1 1 1 j 1 1 1 1
1 1 t HI
- BRRKE SPECIFIC SOLUBLE ORGRNIC FRRCTION -
. RND THE EIGHT SOF SUBFRRCTIONS
BflSELINE ENGINE (NO CONTROL) ftNO WITH
CORNING EX-t7 CERftMIC PflRTICULRTE
TRfiPS ON THE CRT 3208 U TRRP/CYL BRNIO
EPfi MOOES 3.1.5.3. 10.11
™
B - BASELINE HXI-HEXANE INSOLUBLE P
OXY-OXYGENATEO
P - PORT TRN-TRANSITIONAL
rArLvcr ARM- AROMATIC
CATALYST PRF-PARAFF1N
ACD-ACIDIC
T - CORN 1 Mr B AS-BASIC
TRAP EIN-ETHER INSOLUBLE
•«•••
m. ...
r\v\7
^ ^ VJAX
PTCF
ACD—,
BAS-J
EHSh|L
Jg^L^
r-HXI-"
rHXI- 1
™ — i
TV Dim
'-^ L
-ACD — | ' "'
•HAS— |
-EIN-] i
-OXY— 1
-TRN — 1
-ARM
ACD— |
-BAS-,
EIN
r-
-
•—
—
ITVT
11AJ.
nw
ADM
. PRF
pACD
B T
9
B T
10
EPA MODE
B T
11
FIG.3.2.58bTHE EFFECT OF CORNING CERflMIC TRRPS
ON THE CHEMICRL COMPOSITION OF THE
SOLUBLE ORGRNIC FRRCTION OF CRT 3208
PRRTICULRTE MRTTER.
-------
193
data present, it appears that there was some selectivity (different percent
changes for different subfractions) due to preferential oxidation in the trap
and/or selective adsorption/condensation in the dilution tunnel.
The results of the Ames bioassays on the total SOF are presented in Figure
3.2.59. The traps increased the SA of the SOF over the baseline significantly
at all modes except mode 3. This is in contrast to the port catalysts which
decreased the specific activity at the rated speed modes. The SA of the SOF for
the baseline and the Corning traps are given in Table 3.2.23, and for mode 4
subfractions in Table 3.2.24. The SA of some of the trap subfractions are of
unprecedented magnitude in the experience of the diesel SOF Ames testing at MTU.
The SA of the Corning trap oxygenated subfraction actually approaches the
activity of 2-nitrofluorene, a highly mutagenic pure compound used as a positive
control in the Ames bioassay.
The BSSA was, however, significantly decreased at the rated speed modes
with mixed results at the intermediate speed (see Figure 3.2.60). In both of
the 75 percent load modes (5 and 9) the BSSA was reduced to essentially zero.
Discussion of Trap Behavior - Based upon the membrane filter theory
reviewed in the Background section, a computer model was developed to explain
the Corning trap behavior theoretically. Using the assumed properties of the
trap materal (effective porosity, mean pore diamter) and the assumed flow
conditions of exhaust flow rate, density, viscosity, and capillary conduit flow
(76), this model predicted much higher pressure drops than the pressure drops
measured across the traps during the engine tests. It was felt that the
reported porosity of 50% represented total porosity, and that the effective
porosity of 4.88% as calculated using the pressure-flow characteristics of
Mogaka et al (34) for a clean trap represented the effective trap porosity,
discounting dead-ended and isolated pores.
-------
194
CD
o
UJ
\^
>-
CJ
CT
O
U
UJ
n_
CO
o
ir>
•
OJ
o
o
•
OJ
o
LO
CD
CO
O
LO
T I I 1 1 1 1 j , ,
SOF SPECIFIC nCTIVITY VS. BMEP
CRTERPILLRR 3208 BflSELINE (NO CONTROL)
flND EQUIPPED WITH CORNING TRRPS
EX-H7. 5.66 X 12 CONE TRfiP/CYL BflNK>
EPR MODES 3.1.5.9.10.1J
2800 rpm
trap
1680 rpm
trap
2800 rpm
baseline
1680 rpm
baseline
J I L
J I L
1
I L
1
0.0
200.0
400 .0
600 .0
BRflKE MEflN EFFECTIVE PRESSURE (KPfl)
FIG.3.2.59 THE EFFECT OF CORNING PflRTICULRTE TRRPS
ON SPECIFIC RCTIVITY OF CRT 3208 SOF.
TRRPS: CORNING EX-47 DESIGNRTION
5.66 INCH X 12 INCH (1 TRRP/CYL BRNK)
-------
195
Table 3.2.23 Cat 3208 - Corning Trap (Uncatalyzed, EX-47) Ames Bioassay
Results on Total SOF C = 18.75 yg, C2 = 600 yg Throughout
Assayed on Same Day (TA98 w/o S-9)
EPA Mode
Control
o T} V
3
4
4
5
5
9
9
10
10
11
11
-T(
- B
- T
- B
- T
- B
- T
- B
- T
- B
- T
(a)
(b)
(c)
(d)
SASL(c) SALL(d) BSS°F BSSA(e)
Device (Rev/yg) (Rev/yg) (g/kw-hr) (kRev/kw-hr)
a) 0.426 0.390 0.932 363
b) 0.410
0.257
1.402
1.212
1.411
0.842
2.325
0.725
1.499
0.378
1.199
B = Cat 3208 Basel
T = Cat 3208 with
SA log
(Vci)
SATT = 10a (G2b- C
LL ,
UoCA = fCA. \ fV.^C
0.375 0
0.245 0
1.407 0
1.426 0
1.637 0
0.860 0
2.640 0
0.799 0
1.741 0
0.336 0
1.302 0
ine Engine (No Control)
Corning Uncatalyzed Traps,
°2 , Model: Rev = a + b
Cl
1 >
T 9 Model • Rev — a. vinsss}
m^ nr^6yg.^ ( KRev \
.972
.377
.137
.038
.002
.026
.001
.127
.006
.969
.067
Ex-47
log (mass)
b
365
92
193
54
3.3
22
2.6
101
10.7
325
87
g ' V1000 Rev'
-------
196
Table 3.2.24 Caterpillar 3208, EPA Mode 4, SOF Subfractions, Ames Bioassay
Results for Baseline and Corning Trap (Uncatalyzed). C , C as
shown Assayed on the Same Day (TA 98 w/o S-9)
Subfraction
Device
BIN -
EIN -
BAS -
BAS -
ACD -
ACD -
PRF -
PRF -
ARM -
ARM -
TRN -
TRN -
OXY -
OXY -
HIN -
HIN -
B =
T =
B
T
B
T
B
T
B
T
B
T
B
T
B
T
B
T
Cat
Cat
Cl
( g)
18.75
0.20
18.75
2.34
18.75
18.75
18.75
18.75
18.75
18.75
18.75
18.75
18.75
0.05
18.75
18.75
3208 Baseline
3208 with Cori
C2
( 8)
600.
150.
37.5
75.
300.
600.
600.
600.
600.
300.
600.
600.
600.
18.75
150.
150.
Engine (No
ling Uncatal
SASL
Rev/ g
0.484
22.64
0.373
8.19
0.657
1.78
0.09
0.005
0.503
0.131
0.066
0.504
0.362
153.4
0.838
0.598
Control)
Lyzed Traps,
SALL
Rev/ g
0.489
32.41
0.372
8.18
0.672
2.02
0.07
0.005
0.592
0.130
0.065
0.526
0.329
190.6
0.708
0.595
EX-47
% of
Total SOF
8.36
3.25
2.63
0.54
4.10
9.43
56.61
72.95
2.33
2.84
5.50
3.50
16.21
6.50
1.37
0.99
-------
197
UJ
u
OL
U
U
LJ
Q_
CO
C/)
CD
o
o
•=}-
o
•
o
o
CO
o
o
CM
O
o
T I I I I I I
BS SPECIFIC RCTIVITY
CRTERPILLRR 3208 BfiSELINE (NO CONTROL)
fiND EQUIPPED VJITH CORNING TRRPS
EX-47. 5.66 X 12 (ONE TRflP/CYL BRNK)
EPfi MODES 3.4,5.9.10.11
2800 rpm
• baseline
T~
vs
BMEP
2800 rpm
trap
1680 rpm
trap
1680 rpm
aseline
0 .0
200 .0
400 .0
600 .0
BRRKE MEflN EFFECTIVE PRESSURE (KPR)
FIG.3.2.60 THE EFFECT OF CORNING PflRTICULHTE TRRPS
ON BS SPECIFIC RCTIVITY OF CRT THE 3208.
TRRPS: CORNING EX-47 DESIGNRTION
5.66 INCH X 12 INCH (1 TRRP/CYL BRNK)
-------
198
Using the known functional relationship between porosity, mean pore
diameter, and mean center-to-center pore spacing (39):
E = Effective porosity (%) derived as 4.88% for a clean trap
MPD = Mean Pore Diameter given as 12>im (91) for a clean trap
B = Mean Pore Spacing (center-to-center) same units as MPD
(E/100)
the value of 51.73 micrometers was obtained, which was used in all computations
since it should remain constant as the trap is loaded. The pores are much more
widely spaced than originally thought.
Table 3.2.25 gives the porosity and mean pore diameter predicted for each
particulate filter taken during the Corning trap tests, assuring that the
measured pressure drop was accurately estimated by the model-predicted pressure
drop. Each filter corresponds to a one-hour sampling period, and the traps were
regenerated in an oven between each mode to establish baseline conditions. The
porosity and mean pore diameter are seen to decrease for all modes but those for
which continuous regeneration was observed (modes 9 and 10).
A parametric study of the membrane filter theory was implemented by
approximating two standard conditions: EPA mode 4 and EPA mode 9 with the
Caterpillar 3208. Material properties and flow conditions chosen for these
conditions are shown in Table 3.2.26.
Figures 3.2.61 and 3.2.62 show the filtering efficiency for both sets of
flow conditions. The total filtering efficiency as well as individual mechanism
filtering efficiencies are plotted as a function of particle diameter. The
diffusion mechanism, which, if 100% effective for particles smaller than about
0.4 micrometers, is slighty more efficient for SC/1, perhaps due to lower
exhaust viscosity. Higher gas velocities cause slightly higher efficiencies for
the inertial impaction mechanism for SC/2. The direct interception mechanism is
predicted to be the same for both mechanisms because it depends only upon mean
-------
199
Table 3.2.25 Predicted Properties for the Corning Trap Based on the
Membrane Filter Theory Model
EPA
Mode
3
3
3
3
4
4
4
4
5
5
5
9
9
9
10
10
10
11
11
11
Filter
Number
3-T-l (a)
3-T-2
3-T-3
3-T-4
4-T-l
4-T-2
4-T-3
4-T-4
5-T-l
5-T-2
5-T-3
9-T-l
9-T-2
9-T-3
• 10-T-l
10-T-2
10-T-3
11-T-l
ll-T-2
ll-T-3
Model
Predicted
Porosity
(percent)
3.40
3.28
3.18
3.11
3.55
3.43
3.33
3.23
4.12
3.94
3.70
4.42
4.47
4.50
3.20
3.04
3.33
3.33
2.72
2.35
Mean
Pore
Diameter
(micrometers)
10.02
9.84
9.69
9.58
10.23
10.06
9.91
9.76
11.03
10.78
10.45
11.42
11.48
11.52
9.72
9.74
9.91
9.91
8.96
8.33
(a) Each filter represents a one hour time period in the total sample time for
the mode. Therefore the predicted porosities and mean pore diameters are the
average predicted values for that one hour period.
-------
200
Table 3.2.26 Standard Material and Flow Conditions Chosen for Parametric
Study of Corning Trap using the Membrane Filter Theory Based
Model Developed.
Parameter
SC/1
•v EPA Mode 4
SC/2
^ EPA Mode 9
Porosity
(percent)
Pore Spacing
(micrometers)
Mean Pore Diameter
(micrometers)
Filter Wall
Thickness
(micrometers)
Total Filter
Area (m )
Total Exhaust
Flow (m /min)
Exhaust Temp.
(deg K)
Exhaust^ Viscosity
(N-s/m")
Exhaust Density
(kg/in )
Exhaust Molecular
Weight (kg/kg-mole)
3.500
51.73
10.16
431.8
4.75795
16.39
639
.273(10 4)
.648
28.80
3.500
51.73
10.16
431.8
4.75795
34.44
901
.379(10 4)
.379
28.82
-------
201
pore diameter. Overall, the filtering efficiencies display a concave upward
"droop" with a minimum of about 35% for both sets of standard conditions; but
this minimum is shifted slightly to larger particle diameters for SC/1. The
input particle size distribution is necessary for predicting which set of
standard conditions would produce the better overall particle size filtration
efficiency.
Figures 3.2.63 through 3.2.66 show independent effects of four trap
properties using the flow conditions of SC/1. Although these plots are
informative, one must have information on the particle size distribution to
predict the optimum filter properties.
The best particle size distribution that could be found for this
application was a log-normal distribution presented by Vuk et al (65) . The data
given by Vuk et al were based on Anderson inertial impactor results from the raw
exhaust of Caterpillar 3150 Engine. The average of their data over the 13 mode
cycle was taken as a starting point and the two parameters needed to define the
log-normal distribution were determined using the mathematical procedures of
Reference 77. The mass median diameter used was 0.075 micrometers with a
standard geometric deviation of 8.10 micrometers.
Because a cumulative frequency is needed for a filter input model and the
-x2
lognormal distribution is expressed by a non-integrable function (e ).
The log-normal distribution was transformed into a cumulative distribution by
integration of a fifth order polynomial fit of the normal distribution given in
Reference 78; this is plotted in Figure 3.2.67. The model was modified to
integrate the predicted efficiency using the cumulative distribution of Figure
3.2.67 to get overall efficiency. The model was then applied to the predicted
porosities for the actual filter data (Table 3.2.25) and the actual flow
conditions. The predicted overall efficiencies are compared with the actual
-------
202
LU
(J
&.
UJ
CL
CJ
z
UJ
t~t
o
I— «
LJL
U_
UJ
• CD
UJ
I I [tillI f 11)11
FILTERING MECHRNISM EFFICIENCIES VS.
PRRTICLE DIflMETER FOR THE CORNING TRflP
CORNING TRRP BEING MODELLED flS H
MEMBRflNE TYPE FILTER WITH FILTERING
MEDIUtl PROPERTIES HNO EXHRUST FLOH
CONOIT10NS flS SHOWN BELOW.
Porosity = 3.500%
Pore Spacing = 52 \m
Wall Thickness = 432 pm
Filter Area = 4.758 sq meters
Exhaust Flow = 16.39
Exhaust Temperature = 639 K _
Viscosity = .273(10~4) N-s/m
Density = 0.648 kg/m3
NR = dpart/dpore
E = Diffusion Mech. Efficiency
Ey = Irnpaction Mech. Efficiency
= Interception Mech. Efficiency
EQ = Overall Efficiency
PRRTICLE DIRMETER (MICROMETERS)
FIG.3.2.61FILTERING MECHRNISMS FOR CORNING TRRP
MODELLED RS R MEMBRRNE TYPE FILTER
INCLUDING: DIFFUSION, INERTIRL
IMPRCTION, RND DIRECT INTERCEPTION
-------
203
LJ
CJ
&L
LJ
CL
S-/
U
~z.
LU
I—I
U
I—I
LL
LJ_
LU
CD
LU
U_
FILTERING MECHRNISM EFFICIENCIES
PRRTICLE DIflMETER FOR THE CORNING TRRP
CORNING TRflP BEINB MODELLED RS fl
MEMBRflNE TYPE FILTER WITH FILTERING
MEDIUM PROPERTIES FIND EXHRUST FLOW
CONDITIONS flS SHOWN BELOW.
Porosity = 3.500%
Pore Spacing = 52 ym
Kail Thickness = 432 ym
Filter Area = 4.758 sq meters
Exhaust Flow = 34.44
Exhaust Temperature = 901 K
Viscosity = .379(10-4) N-s/m
Density = 0.379 kg/m
N_ = d ./d
R part pore
E = Diffusion Mech. Efficiency
E = Impaction Mech. Efficiency
E = Interception Mech. Effici
E = Overall Efficiency
E + (l-EI)(ED4-ERy(1 V)
10
10
10
10
PnRTICLE DIRMETER (MICROMETERS)
FIG.3.2.62 FILTERING MECHRNISMS FOR CORNING T
MODELLED RS R MEMBRflNE TYPE FIL~FR
INCLUDING: DIFFUSION, INERTIRL
IMPRCTION. RND DIRECT INTFRrEPTT''M
RRP
-------
FILTERING EFFICIENCY (PERCENT)
20 .0
40 .0
60 .0
80 .0
100 .0
~n
i—i
CD
LO
*
hO
U)
~n co n: ~o
>-< —1 O 3D
r~ cz a TO
—i a m 3D
m ^c r~ 3
TO >—< m m
•-I ~z. m —i
Z CD O TO
CD i-"
-H 3D CD
3 X CO
mm co
a 3D —i
i—ii—i CZ
c: z IK o
is "n m -<
r~ 2
~~Q a 33 ~n
o m TO o
70 Z 3D 70
O O Z
co m m n
>-< o
-H O —I 70
-< ~n -< z
~0 HH
o m m z
z ~n CD
~n m
rn m i—i —i
m o r~ TO
~n —i —i 3D
>-H i—i m TJ
n < TO
m
n
"D
3D
n
r~
m
o
t—i
m
—i
m
70
s-\
12
t—H
n
o
z
CO
o
o
-------
FILTERING EFFICIENCY (PERCENT)
20 .0
40 .0
60 .0
80 .0
100 .0
O CO
rn —i
z c
-H CD
m -<
70 >—i
Z
Tl CD
O
70 -H
m IE
m
CO
~o I-H
ID Z
n ~n
O
a
m
~n
i — i
CD
•
OJ
•
to
•
c^
-ts
~D
H3
/D
ZD
3
m
m
a 73
> — i
ID n
CO
CO
ID —I
C
3 a
m -<
z ci co ~n
CD m TO o
Z ID 70
o n z
z m m o
o
m o —i TO
t-< -n -< z
r~ TI i—i
-H n m z
mm CD
m m r~ TO
m TO —i ID
m i m ~o
>-< -H 70
n o
TJ
ID
70
n
i—
m
a
i—i
33
IZ
m
—i
m
70
n
70
o
z
CO
CD
I
i
m
KJ
o
Ln
o
-------
206
UJ
u
en
LU
Q_
CJ
Z
LLi
i—i
CJ
u_
UJ
CD
LJ
I—
_J
i—i
LL
CD
CD
CD
00
CD
CO
CD
•
CD
CD
•
CD
C\J
THE INFLUENCE OF WRLL THICKNESS ON THE
EFFICIENCY OF R CORNING CERRMIC TRflP
CORNING TRRP BEING MODELLED flS fl
MEMBRRNE T\PE FILTER WITH DIFFERENT WRLL
THICKNESSES flND FLOW CONDITIONS RS SHOWN
Wall
AP =
Wall
AP -
Wall
AP =
Wall
AP -
Wall
AP =
Thickness
3.61 kPa
Thickness
5.42 kPa
Thickness
7.22 kPa
Thickness
10.83 kPa
Thickness
14.44 kPa
Effective Porosity = 3.500 %
Pore Spacing - 52 ym „
Filter Area = 4.758 -m -.
Exhaust Flowrate = 16.39 m /min
Exhaust Temperature = 639 K _.
Exhaust Viscosity = 0.273(103 )
Exhaust Density = 0.648 kg/m
N-sec/m'
10
-2
10
-1
10
10
PRRTICLE DIRMETER (MICROMETERS)
FIG. 3.2.65PRRRMETRIC STUDY FOR CORNING TRRP
MODELLED RS R MEMBRRNE TYPE FILTER
STUDYING THE INFLUENCE OF FILTERING
MEDIUM THICKNESS ON FILTER EFFICIENCY
-------
207
UJ
u
LY
LJ
CL
CJ
~2L
UJ
I — I
CJ
I — I
L_
u_
UJ
CD
UJ
h-
_J
i — i
LJ_
o
•
o
o
CD
•
O
00
o
CO
o
•d-
o
C\J
TTTTI
THE INFLUENCE OF FILTER flREfl ON FILTER
EFFICIENCY OF R CORNING CERflMIC TRflP
CORNING TRRP BEING MODELLED flS H
MEMBRflNE TYPE FILTER WITH DIFFERENT
FILTER HREflS RND FLOW CONDITIONS SHOWN
Filter Area = 2.379 m
AP = 14.44 kPa
Filter Area - 3.568 m
AP = 9.63 kPa
Filter Area = 4.758 m
AP - 7.22 kPa
Filter Area - 7.137 m
AP = 4.81 kPa
Filter Area - 9.516 m
AP = 3.61 kPa
Effective Porosity = 3.500
Pore Spacing = 52 jjm
Wall Thickness - 432 ym
Exhaust Flowrate = 16.39 m /min
Exhaust Temperature = 639 K
Exhaust Viscosity = 0.273(10 ) N-sec/m"
Exhaust Density = 0.648 kg/m
ID'2
ID'1
10
10
PflRTICLE DIflMETER (MICRONS)
FIG.3.2.66PflRflMETRIC STUDY FOR CORNING TRflP
MODELLED RS R MEMBRflNE TYPE FILTER
STUDYING THE INFLUENCE OF TOTRL FILTER
MEDIUM RRER.ON FILTERING EFFICIENCY
-------
208
o
o
LU
I—
LJ
z:
(T
i—i
Q
~2L
CE
CO
CO
CO
CO
CE
z:
CD
UJ
CJ
LY
ill
Q_
o
00
CD
•
CD
CD
O
0
•
CD
CM
~H I I II II l| I 1—I MINI 1 1—|—TT
LOG-NORMflL PRRTICLE SIZE DISTRIBUTION
HRSS MEDIF1N DIHMETER = 0.075 MICROMETER
STD. GEOMETRIC DEVIflTION = 8.1033
BRSED ON DflTfl BY VUK ET fit (4O)
FOR 13 MODE CYCLE flVERRGE
OF R CRTERPILLRR 3150
J L
10
10
-1
10
10
PRRTICLE DIflMETER (MICROMETERS)
FIG.3.2.67 |_OG-NORMflL PflRTICLE SIZE DISTRIBUTION
flSSUMED FOR INTEGRflTION OF THE MEMBRRNE
FILTER THEORY PREDICTED EFFICIENCY TO
RPPROXIMRTE THE OVERRLL EFFICIENCY
-------
209
observed efficiencies in Table 3.2.27. The model predicts between 89 and 92
percent efficiency through the whole range of engine conditions. Obviously the
original purpose of the model (explaining the load dependent efficiency at 1680
rpm) was not satisfied. Some inadequacies in the model which may account for
the discrepancies are: i) The model assumes a sticking probability of 100% (i.e.
if a particle contacts a surface of the ceramic it will stick and not become
re-entrained). This is probably not realistic and it is possible that the
variables which affect the sticking probability (unknown are load-related. 2)
The particulate bulk density was assumed to be 1 gram/cubic centimeter which is
probably also a function of load. 3) A large fraction of the particles are
actually not in the continuum (or Stokes) regime. Many particles are in the
slip flow regime (where corrections can be applied) but many are also in the
transition region between slip and free molecule flow. The various size regimes
and the percent by mass of particles in these regimes are given in Fig. 3.2.68
for SC/1 and SC/2.It is likely that the theory is inadequate for many of the
very small particles and the diffusion mechanism may be somewhate overestimated.
4) The particle size distribution used was an average over a wide range of
conditions when in fact it is actually a function of load. There are some
discrepancies among studies of load dependence of particle size but generally
particle size distributions are shifted smaller with increased load. This would
yield higher efficiency for increased load as was observed in this study. 5)
There are many discrepancies between the idealized representation of the porous
ceramic and the true structure (such as variable pore diameters and lengths).
These variations could result in inertial effects which might be responsible for
a load dependent efficiency (higher velocities at higher loads resulting in
greater inertial impaction).
The model as it exists now is not of any quantitative use because of the
inadequacies outlined above but it should be useful for predicting general
-------
210
Table 3.2.27 Comparison of Model Predicted Efficiencies with
Actual Measured Efficiencies of the Corning Trap
on the Caterpillar 3208
EPA
Mode
Filter
Number
Model
Predicted
Efficiency
(percent)
Average
Predicted
Efficiency
(percent)
Average
Measured
Efficiency
(percent)
3
3
3
3
4
4
4
4
5
5
5
9
9
9
10
10
10
11
11
11
3-T-l
3-T-2
3-T-3
3-T-4
4-T-l
4-T-2
4-T-3
4-T-4
5-T-l
5-T-2
5-T-3
9-T-l
9-T-2
9-T-3
10-T-l
10-T-2
10-T-3
11-T-l
ll-T-2
ll-T-3
91.42
91.38
91.34
91.32
91.35
91.31
91.28
91.25
91.48
91.43
91.36
90.50
90.50
90.50
90.14
90.08
90.15
90.13
89.89
89.74
91.37
91.30
91.42
90.50
90.12
89.92
34.5
68.1
97.0
98.7
98.8
94.7
-------
211
UJ
CD
UJ
LU
M
i — i
CO
CO
CO
CE
>-
DO
LU
(_)
(Y.
LlJ
Q_
o
•
o
o
o
•
o
oo
CD
•
O
CD
O
•
O
0
•
o
I I I I I I
i r
I I I I
PRRTICLE SIZE REGIMES FOR CRT 3208
EPfl MODES 4 RND 9 .
PERCENT BY MflSS IN SIZE REGIME f)S
LRBELLED RND DETERMINED BY KNUDSEN
NUMBERS.
Size Regimes:
Free Molecule
Transition
Slip Flow
Continuum
Mode 4
• Mode 9 —
J L
J I
10
-2
10
-1
10
10
PRRTICLE DIRMETER (MICROMETERS)
FIG.3.2.68PRRTICLE SIZE REGIMES FOR CRT 3208
FREE MOLECULE, TRRNSITION, SLIP FLOW,
RND CONTINUUM REGIMES FOR TWO ENGINE
CONDITIONS: EPR MODE 4 RND 9 .
-------
212
trends. A parametric study of the effects of certain variables on trap
performance was carried out. The parameters varied independently were:
porosity, pore spacing, filter thickness, total filter flow area, exhaust
temperature, mass median particle diameter, and standard geometric deviation.
The variables predicted were overall filtering efficiency, pressure drop, and a
composite variable called filter quality. Filter quality considers both
efficiency and pressure drop together such that high filter quality is
desirable. The definition of filter quality is given at the bottom of the
results of the parametric study, Table 3.2.28. The lack of precision of the
model dictated the presentation of directions of change rather than magnitudes.
It can be seen that all combinations of changes in efficiency and pressure drop
are present but the filter quality parameter decides if the net effect is
beneficial. Based on filter quality alone, the following conditions are
desirable in the construction and operation of a porous ceramic trap: high
porosity, large pore spacing (i.e., large pores relative to the 12 micrometers
mean pore diameter used in the Corning material EX-47), thin cell walls, large
filter area (i.e., largest trap practical), high exhaust temperature (i.e., close
to engine), small mass median particle diameter, and standard geometric deviation
(i.e., close to the engine).
The last step in the model development taken was the prediction of the
output particle size distribution, given the assumed log-normal input. The
output particle size distribution was expected to have a mass median diameter
somewhere in the low efficiency region (0.4 to 10 micrometers) with a much
smaller standard geometric deviation. The result of this final step is given in
Fig. 3.2.69 and appears to match the expectation. The particle size
distribution is not strictly log-normal because it does not extend to infinity
in the negative and positive directions. However, it can be approximated as a
log normal distribution with a very small standard geometric deviation. The
-------
213
Table 3.2.28 Results of Parametric Study using the Corning Trap-
Membrane Filter Theory-Based Computer Model
Change in
Dependent Variable
Increase In
Independent
Variable
Overall
Efficiency
Pressure
Drop
Filter
Quality
Effective
Porosity *
Mean Pore
Spacing ^
Filter (wall)
Thickness *
Total Filter
Flow Area *
Exhaust
Temperature ^
Mass Median
Particle Dia. *
Standard Geom.
Dia. Deviation^
Filter Quality is a composite of the effects of efficiency and pressure
drop relating the desirability of high efficiency and low pressure drop in
one parameter.
Filter Quality =
E = overall filtering efficiency (fraction)
AP = pressure drop across filter
-------
214
o
o
UJ
h-
LJ
z:
CE
i — i
Q
CE
X
CO
CO
UJ
CO
CO
CE
CD
UJ
0
Qi
UJ
QL
o
oo
o
CD
o
•=»-
o
CNJ
TT
1
1—I Mill
PREDICTED OUTPUT PflRTICLE SIZE
DISTRIBUTION FOR THE CORNING -TRflP
PREDICTED FPON THF CORNING TRflP
MEMGRflNE FILTFR THCOi.'Y MODEL
fiSSUMINO ft LQG-NORMRL INPUT
PfiRTICLC ZIZF DISFRIDUTION RS
SHOWN mO THE TRflP fiND FLOW
VRRIRBLES SHOWN BELOW:
Porosity = 3.500%
Pore Spacing = 52 ym
Pore Diameter = 10.16 pm
Wall Thickness = 432 pm
.Filter Area = 4.758 sq meters
Exhaust Flow = 16.39 m3/rnin
Temperature = 639K .
Viscosity = .273(10 )^
Density = .648 kg/m3 m
Input to
Trap
I I
io-2
10-1
10
10
PRRTICLE DIRMETER (MICROMETERS)
FIG.3.2.69PREDICTED PRRTICLE SIZE DISTRIBUTION
OUTPUT FROM THE CORNING TRRP FOR THE
RSSUMED LOG-NORM INPUT SIZE DISTRIBUTION
MRSS MEDIRN DIR = 0.075 MICROMETERS
-------
UJ
I—
UJ
51
tr
^—i
Q
CE
1C
CO
CO
LJ
CO
CO
(T
5-
m
o
LZ.
UJ
Q_
PREDICTED OUTPUT PRRTICLE SIZE
DISTRIBUTION FOR THE CORNING TRRP
PRFDICTFn FROM THF. CORNING TRf)P
MI Hrii:i)ic FILTER THEORY MODEL
ROOMING fl mG-NORMni. INPUT
PfikTICLE fc!7L m&Tk'ILtUTION flS
SHOWN Finn THt IK'HP flND FLOW
VRRIflBLEb SHOWN BELOW:
10
Input Mass
Median Dia
Output Mass
0.075 ym Median Dia = 1.344 ym
PRRTICLE DIRMETER (MICROMETERS)
FIG.3.2.70PREDICTED PRRTICLE SIZE DISTRIBUTION
OUTPUT FPOM THE CORNING TRRP FOR THE
RSSUMED LOG-NORM INPUT SIZE DISTRIBUTION
MF15S
MED I FIN DIR = 0 .075
MICROMETER'S
-------
216
parameters of the output distribution are derived from the diameters
corresponding to the 50 and 84 cumulative mass percent. The mass median
diameter is the diameter at 50% while the standard geometric deviation is the
diameter at 84% divided by the mass median diameter = 1.34 micrometers and
standard geometric deviation = 1.90 micrometers. The shift in mass median
diameter is shown in Fig. 3.2.70. This observation is significant in terms of
health effects because it appears that a large portion of the very fine
particulates (which are capable of entering and being retained by the lungs) are
effectively removed by the Corning trap.
Adsorption and SOF/Solids Ratio
Figure 3.2.71 plots the ratio of SOF/SOLID as a function of raw exhaust
hydrocarbon concentration. Of the three aftertreatment devices thoroughly
investigated during this project, the Corning traps produced the highest
SOF/SOLID ratio. Despite the extreme reduction in solids, the exhaust
hydrocarbons were sufficiently high to produce an SOF/SOLID ratio much higher
than that of the port catalysts at modes 3 and 4, which are the two rightmost
points in each 1680 rpm curve. The PTX and port catalysts always had enough
solids emissions to establish an SOF/SOLIDS ratio below that of the baseline.
Also, the mode 11 data point (rightmost point on the 2800 rpm curve) of the
Corning trap lies above that of the baseline or either of the other
aftertreatment devices, despite the extreme reduction in solids for the Corning
trap tests at this mode. This finding supports the hypothesis of Clerc (79) who
developed a semi-empirical computer model for predicting SOF concentrations
based upon raw exhaust opacity, hydrocarbon concentration, and BET adsorption
theory. The SOF/solids ratio is the fundamental adsorption parameter which
should be a function of hydrocarbon concentration and filter temperature. Clerc
saw the greatest difference between condensed hydrocarbon (SOF) as measured in
the dilution tunnel and the SOF predicted by the semi-empirical model for mode 3
-------
217
o
(—i
_j
o
CO
CO
z:
CE
CD
LL
O
CO
CO
z:
CE
CD
O
CO
o
CO
o
1 r
I ' ' ' ' I '
SOF/SOL VS. HYDROCRRBON CONCENTRRTION
CRT 3208 BRSELINE (NO CONTROLS FIND WITH
EXHHUST PORT CHTRLYSTS. PTX CflTRLYSTS.
flND CORNING TRflPS. DILUTION TUNNEL OP-
ERRTED RT fl DILUTION RRTIO OF 15:1 WITH
FILTER TEMPERRTURE RLLOWED TO VHRY WITH
flMBIENT TEMPERRTURE RND RflW EXHRUST
INJECTION TEMPERRTURE CHHNGES.
1680 rpm-
PTX catalyst
1680 rpm-trap
'1680 rpm-baseline
1680 rpm^port catalyst —
00 rpm-
baseline
2800 rpm-PTX catalyst
1
0 .0
100 .0
200 .0
300 .0
RflN EXHRUST HYDROCRRBON CONC (PPM)
FIG.3.2.71SOLUBLE ORGRNIC FRRCTION TO SOLID RRTIO
CRT 3208 BRSELINE (NO CONTROL) RND NITH
EXHRUST PORT CRTRLYSTS RND NITH CORNING
TRRPS TESTED RT EPR MODES 3,4,5,9,10,11
-------
218
on the Caterpillar engine, with lesser differences at modes 4 and 11. This
difference was attributed to condensed high-molecular-weight hydrocarbons.
Dilution tunnel mixture (sample filter) temperature is not accounted for in
the plot of Fig. 3.2.71 and could explain some of the other differences. Clerc
showed that ambient temperature (which influences filter temperature) affects
SOF/SOLID significantly (a 10°C ambient temperature decrease can cause a 70%
increase in SOF). The effect of ambient temperature could be significant in
these comparisons because of the different times of the year during which the
tests were conducted. Clearly one needs either ambient condition
controllability or a method for correcting to a standard ambient condition.
Clerc recommends a constant filter temperature (filter temperature determined by
running the engine at the various speeds and loads with dilution ratio = 15:1,
and correcting the observed filter temperature vs. load curve as the basis for
setting dilution ratio in all future experiments) as the best way of achieving
repeatability between different ambient condition test days.
-------
219
Johnson Matthey Cylindrical Mesh Trap
The fourth aftertreatment device tested was a trap designed by
Johnson-Matthey Corporation. Table 3.2.29 gives specifications for this trap,
which contains a proprietary catalyst and a washcoat formulated to minimize
sulfate emission. The mesh material in this trap has been tested in the
modified manifold of a 2-liter passenger car diesel with 52-76% reduction in
emissions during a 50,000-mile test period (35), but has not received testing on
a larger engine until now.
Table 3.2.29 Cylindrical Mesh Trap Specifications
Manufacturer Johnson-Matthey
Catalytic Agent Not specified
Substrate Material 310 Stainless Steel ribbon,
knitted into mesh
Substrate Geometry Compressed blocks of mesh
with radially graduated
bulk density and surface-to
volume ratio; designed to
trap gradually smaller
particles in radial flow
direction
Washcoat Not specified
Trap dimensions 35cm dia x 101 cm long
Table 3.2.30 compares BSFC and emissions data for the baseline engine as
well as for the Johnson-Matthey trap. Due to difficulties in obtaining stable
filter weights, the particulate data is missing for many of these tests.
The first mode run after receipt of the trap was mode 11. Although only
one test was obtained, the trap displayed higher fuel consumption and NO and HC
emissions; solid particulate was also increased as was sulfate, but SOF was
greatly reduced. Modes 9, 4, 10, and 5 were then run in that order.
-------
Table 3.2.30 - Comparison of Fuel Consumption and Emissions for the Johnson-Matthey Mesh Trap
with Baseline Values, Before and After Reactivating Mesh Material.
All values are Brake-Specific, in g/kw-hr, ± 1 Standard Deviation
EPA MODE
Mesh Trap
Speed rpm
Load, N-m
BS FC
BS NO
X
BS N02
BS NO
BS HC
BS TPM
BS SOF
BS SO.
4
BS SOLIDS
Opacity(%)
3
W/0
1680
160
287
0.00
10.241
0.03
1.16±
0.22
5.91
0.16
1.92±
0.21
1.0471
0.031
0.9321
0.030
0.0311
0.001
0.084
0.3
W,as rec W, react
1680
160
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
1680
160
284
0.00
10.101
0.00
0.311
0.00
6.381
0.00
0.901
0.00
(a)
(a)
(a)
(a)
0.2
4
W/0 W,as rec
1680
320
2311
0.00
10.391
0.13
0.851
0.05
6.221
0.07
0.771
0.02
0.4821
0.03
0.3771
0.024
0.0361
0.002
0.069
0.91
0.05
1680
320
2291
2
9.331
0.21
2.171
0.57
4.671
0.24
0.051
0.02
(a)
(a)
(a)
(a)
1.11
0.39
W, react
1680
320
2261
2
9.601
0.09
2.391
0.01
4.701
0.05
0.051
0.00
0.8521
0.014
0.0131
0.001
0.6641
0.005
0.177
0.71
0.0
5
W/0
1680
480
2171
2
9.691
0.08
0.361
0.01
6.081
0.05
0.251
0.00
0.2411
0.005
0.0381
0.001
0.0311
0.002
0.172
1.91
0.0
W,as rec
1680
480
2161
2
9.651
0.19
1.801
0.20
5.131
0.01
0.041
0.00
(a)
(a)
(a)
(a)
4.51
0.35
9
W, react W/0
1680
480
N.A.
N.A.
N.A.
N.A.
N.A.
N.A
N.A.
N.A
N.A.
N.A.
2800
399
2161
0
8.411
0.08
0.2501
0.033
5.321
0.05
0.261
0.01
0.7951
0.064
0.0661
0.007
0.0391
0.009
0.7301
0.051
4.91
0.2
W,as rec
2800
399
2511
5
7.811
0.09
0.1731
0.067
4.981
0.08
0.041
0.00
2.1211,..
0.174 (b)
0.0121
0.015
1.8421
0.080
0.7701
0.210
5.61
0.29
(a) Particulate data notavailable due to insufficient filter ammoniation and/or insufficient sampling times
N.A. Not available
(b) Particulate data at Mode 9 may be inaccurate, as sulfate values steadily increased with time and
filters may have not been completely ammoniated
-------
Table 3.2.30 - Comparison of Fuel Consumption and Emissions for the Johnson-Matthey Mesh Trap
With Baseline Values, Before and After Reactivating Mesh Material.
All values are Brake-Specific, in g/kw-hr, ± 1 Standard Deviation
EPA MODE
Mesh Trap
Speed rpm
9
10
11
W.reactivitated W/0 W,as rec W, reactivated W/0 W,as rec
2800
Load , N-m
BS
BS
BS
BS
BS
BS
BS
BS
BS
FC
NO
X
N02
NO
HC
TPM
SOF
S°4
SOLIDS
N
N
N
N
N
N
N
N
N
399
.A.
.A.
.A.
.A.
.A.
.A.
.A.
.A.
.A.
2800
9
0
0
5
1
0
0
0
0
0
0
0
0
0
Opacity(%)
266
282
0
.02
0
.261
.026
.72+
0
.3051
.052
.792±
.016
.1331
.021
.0331
.005
.6261
.031
3.81
0.84
2800
266
2781
2.191
8.491
0.170
1.4161
0.296
4.611
0.301
0.100
0.055
(a)
(a)
(a)
(a)
1.81
0.05
2800
266
8
1
4
0
3
0
o
0
2
0
0
0
2751
0
.661
0
.841
0
.451
0
.151
0
.171
.025
.0431
.018
.3431
.001
.7931
.042
5.21
0.0
2800
133
370
0
10.091
0.1011
0.651
0.026
6.151
0.062
2.9261
0.117
1.8641
0.019
0.9691
0.107
0.8571
0.111
0.0381
0.004
3.21
0.10
2800
133
383
(c)
9.78
(c)
2.19
(c)
4.45
(c)
0.18
(c)
2.967
(c)
0.067
(c)
2.467
(c)
0.428
(c)
1.4
(c)
W, reactivated
2800
133
3911
0
10.811
0
2.581
0
5 . 361
0
0.151
0
2.298
0.106
-.3961
0.052
1.7331
0.066
0.1891
0.069
2.01
0.0
(c) One test value, only
-------
222
Hydrocarbon emissions were reduced at all modes, but NO was increased at all
modes but mode 9. The most notable increase over baseline emissions was the
sulfate, which dominated the total particulate to the extent that SOF and solids
were difficult to measure. A series of six filters run at mode 9 gave steadily
increasing sulfate levels, with 71 to 94 percent of the total particulate
analyzed as sulfate.
Since sulfate conversion had been shown to be a problem in the development
of this trap (79), Johnson-Matthey agreed to reactivate the trap mesh material
with a washcoat designed to minimize this problem. After return of the trap to
MTU, a re-run of mode 11 showed that the sulfate conversion had been reduced
slightly but not to baseline levels. Subsequent running of modes 4 and 10
showed that the particulate was not now completely dominated by sulfate, but
that sulfate (as well as solids) were still higher than baseline levels. Results
from mode 3 as of this writing showed particulate values well below baseline
emissions, but sulfate predominance made TPM measurements difficult; modes 5 and
9 have not been rerun since catalyst reactivation. It appears as if higher load
conditions eliminate the trap's ability to reduce sulfate emissions.
The regeneration characteristics of the Johnson-Matthey trap were not
completely investigated; however, modes 4 and 10 showed no increase in trap
pressure drop after two hours of running, while mode 3 showed a pressure drop
increase of less than 0.5kPa over 2 hours. Mode 11 showed a pressure drop
increase of about 0.6 kPa per hour, based on a 30 minute run, and modes 5 and 9
were essentially stable over 30 minutes.
-------
223
CUMMINS ADIABATIC ENGINE ' - ALTERNATE FUEL SAMPLES
This work by Michigan Tech involved collection of filter samples by Cummins
Engine Company followed by extraction and mutagenicity testing on SOF obtained
from the first build of a single-cylinder experimental adiabatic engine. This
engine configuration was the non-adiabatic build of the engine. Two other
builds of the engine were carried out on a DOE Cummins contract and the results
are not reported here. Specifications of this engine are given in Table
3.3.1.
Table 3.3.1 Cummins Engine Specifications
Manufacturer and Model
Type
Bore x Stroke
Displacement
Compression Ratio
Rated power
Rated torque
Cummins NTC-400
Single-cylinder, direct injection, turbocharged
140mm x 153 mm (5.52in x 6.02 in)
2.35 liter (143.4 in )
14.5:1
49.8KW (67.0hp) @ 1900 rpm
344 N-M (254Ft-lbf)
A mini-dilution tunnel system similar to the system described by MacDonald et
al. was used for particle collection. Dilution ratios were established
to dilute exhaust gases to 130°F. For modes 6 and 8, dilution ratio was
typically 15:1, while mode 9 utilized a 10:1 dilution ratio and mode 11
used a 6.5:1 dilution ratio. Particulate samples were collected on 110 mm round
Pallflex filters, and were extracted with DCM to give SOF samples, which
were bioassayed as described in the experimental section of this report.
Table 3.3.2 lists the fuel properties of the No. 2 diesel fuel, shale
fuel, and solvent-refined coal fuel used in this work. The most notable
differences between these fuels are the low cetane number, H/C ratio, and per-
cent saturates as well as the high viscosity, percent aromatics and percent
nitrogen of the coal-derived SRC-II fuel.
-------
224
Table 3.3.2 Fuels Analysis: Cummins Experimental Engine
FUEL PROPERTY
NO. 2
SRC II
MIDDLE
DISTILATE
SHALE
PHYSICAL AND CHEMICAL
PROPERTIES
Energy Content
BTU/LB Gross
19,200
17,102
19,365
Gravity 60 °F
Specific Gravity
60/60, ° F
Viscosity, cP
Pour Point,0 F
Flash Point, °F
Cetane Number
Hydrogen/Carbon Molar Ratio
% Aromatics
% Saturates
% Olefins
Elemental Analysis
% Sulfur
% Nitrogen
33
.86
2.85
-35
167
45-50
1.75
36
62
20
.29
.4
12.4
.983
3.87
-65
175
22
1.24
79.1
14.4
6.5
.31
1.0
38.2
.834
2.72
-5
178
46
1.86
30.5
67.7
1.8
.07
.024
-------
225
Sample collection was carried out at EPA modes 6, 8, 10, and 11. Due to
unstable running conditions for SRC-II at mode 11, this fuel was run at 75% load
and rated speed (mode 9) instead.
Effects of Operating Conditions on Ames Bioassay Results
Table 3.3.3 gives dose-response statistics from the thirteen samples
assayed. Because of sample limitations, only three samples (mode 8, SRC-II
fuel; mode 11, shale fuel; and mode 6, No. 2 fuel) showed evidence of toxicity,
and a uniform dose-range analysis was not performed. All samples were assayed
without metabolic activation, and those with sufficient concentration were also
assayed with activation (S-9). The only sample demonstrating increased
mutagenicity with S-9 was the mode 6 (peak torque) sample obtained with No.2
diesel fuel. This sample showed considerably higher activity than any other
sample tested, but no explanation for this high activity can be offered.
Comparing fuels, it appears that SRC-II SOF is not appreciably more
mutagenic than the shale fuel, and is similar to No. 2 fuel SOF if the mode 6
data point for No. 2 fuel is disregarded. However, it should be stressed that
the effects of toxicity cannot always be discounted, and a toxic effect may
suppress mutagenic activity at lower doses before manifesting itself at higher
doses.
Comparing modes, the mode 8 (rated speed at full load) samples showed the
greatest biological activity if the extremely active No. 2 fuel, mode 6 sample
is disregarded. This mode also demonstrated slightly higher activity at the
retarded timing of 24°BTC than at 27°BTC for SRC-II.
-------
226
Table 3.3.3 Specific Activity Calculations for Cummins Test Data
(a)
Sample
Fuel, Mode, Timing,0
SRC-II,
SRC-II,
SRC-II,
SRC-II,
SRC-II,
SRC-II,
SRC-II,
SF, 6,
SF, 8,
SF, 10,
SF, 11,
SF, 6,
SF, 8,
SF, 10,
SF, 11,
No. 2, 6
No. 2, 8
No. 2, 10
No. 2, 11
No. 2, 6
No. 2, 10
No. 2, 11
6, 29°
8, 27°
8, 24°
9, 27°
10, 27°
8, 27°
8, 24°
27°
27°
27°
27°
27°
27°
27°
27°
, 27°
, 27°
, 27°
, 27°
, 27°
, 27°
, 27°
s-'c <
— 1
3
4
3
1
+ 3
+ 3
2
5
0
1
+ 2
+ 2
+ 0
+ 0
29
3
2
0
+ 37
+ 2
+ 1
.43
.05
.19
.54
.63
.19
.79
.04
.99
.61
.38
.51
.95
.70
.70
.09
.59
.97
.82
.43
.05
.02
Cl
18
4
4
4
9
9
9
9
9
75
18
18
18
37
75
1
9
18
18
4
18
18
.75
.69
.69
.69
.38
.38
.38
.38
.38
.00
.75
.75
.75
.50
.00
.17
.38
.85
.75
.69
.75
.75
C2
300.
150.
300.
300.
600.
75.
150.
600.
600.
1200.
600.
300.
300.
600.
1200.
37.
300.
300.
1200.
37.
300.
300.
00
00
00
00
00
00
00
00
00
00
00
00
00
00
00
50
00
00
00
50
00
00
b
334
294
685
578
534
232
443
668
699
573
533
587
690
328
654
702
692
694
539
1360
478
237
r
.94
.94
.98
.96
.96
.95
.91
.96
.98
.97
.96
.87
.95
.93
.95
.95
.97
.96
.96
.98
.92
.94
(a)
(b)
(c)
Using Ames S^. typhimurium tester strain TA100
Fuel type (SRC-II - Solvent-Refined Coal-Derived Fuel, SF-Shale Fuel,
No.2-Standard Lab Fuel); Mode; Timing in °BTC)
Microsomal activation system
SL
( ' = Specific Activity (rev/jig)= - log
VC1
1Q
C. = Minimum dose (yg/plate) on linear portion of dose-response curve
C9 = Maximum dose (jig/plate) on linear portion of dose response curve
b = Slope of dose-response curve in linear-log format, determined
statistically
r = Coefficient of correlation
-------
227
4
CONCLUSIONS
MACK ENDT-676 FUEL INJECTION PARAMETER STUDY
1) The effect of load at intermediate speed increased NO, NO ,
2.
smoke, and solids emissions, while reducing BSFC and SOF emissions,
primarily due to increased equivalence ratio and temperatures at higher
load. At rated speed, increased mixing rates and temperatures above the
solid-formation range kept solids emissions from increasing.
2) Increased engine speed reduced NO and NO- while increasing SOF
and BSFC and made solids emissions less load-sensitive, all probably due
to increased mixing rates and decreased residence times.
3) Increased fuel injection rates (at the same injection timing)
reduced solids, SOF, and BSFC while increasing NO and NO, primarily due
to increased fuel-air mixing rates. The reduction in pumping efficiency
with the ultra-high-rate shuttle pump increased BFSC. Excessive fuel
accumulation during the premixed combustion period led to increased SOF
with this pump. The overall biological activity and percentage of
transitional, oxygenated, and acidic subfractions in the SOF generally
increased with increased injection rates, because increased amounts of
fuel in lean-limit regions underwent thermal decomposition. Reduced SOF
at high rates failed to offset the increased SA; SOF increased with ultra-
high rates. The result was a net increase in BSSA with increased injection
rates at constant timing.
4) Retarded injection timing (at higher injection rates) increased solid
particulates and BSFC, but decreased SOF, NO, and NO-, as a result of shorter
ignition delays and reduced peak cycle temperatures. Ultra-high rates
-------
228
showed similar trends, for the same reasons and also due to a high-density
air charge (which reduced over-penetration). Retarded timing at high rates
increased acidic and decreased ether-insoluble subfractions in the SOF.
The acidic, transitional, and oxygenated subfractions did not always decrease
as expected from the amount of fuel undergoing thermal decomposition. This
was true for mode 3 conditions for both high and ultra-high injection rates.
At mode 9, retarded timings caused an increase in transitional and oxygenated
subfractions, suggesting that monatomic oxygen is more available for production
of these species due to lower combustion temperatures and reduced pyrolysis
of N to monatomic nitrogen.
5) Larger injector nozzle sac volume, as expected, increased HC, solids,
and SOF with minimal reductions in NO . The increase in SOF was much less than
x
the increase in solids or HC, and the composition of the SOF with large sac-volume
nozzles showed extensive modification.
6) The combination parameter BSFC-BSTPM-BSNO was improved by both increased
~ ~~~~~ X
injection rates and retarded timings, except for extremely retarded timings
at the ultra-high rate. Moderately increased injection rate was more effective
than injection timing in reducing the BS solids and BSNO_, but greatly increased
injection rates made most emissions more sensitive to injection timing. All
emissions and the BSSA decreased with increased rates and retarded timings,
but BSFC increased with the ultra-high-rate shuttle pump.
7) The biological activity of the particulate is related to at least
three factors: 1) exhaust temperature, 2) relative amounts of acidic, oxygenated,
and transitional subfractions, and 3) exhaust N0? concentration. Whether NO
concentration is an influential parameter in the real world or is merely a
contributor to artifactual mutagenic compounds is unanswered.
8) Use of the ideal injection characteristics hypothesized could result
in 1) reducing solids by as much as 70% and reducing SOF, 2) maintaining the BSFC
-------
229
at levels similar to the 3.8-inch bowl, high-rate system, and 3) achieving
BSNO emissions of about 10 g/kw-hr. Although the overall rate of injection
x
is important, the instantaneous rate (especially the beginning and ending) of
injection is more important in controlling solids and SOF formation. The use
of high-sulfur fuels with the ideal injection characteristics may result in the
majority of the particulate being sulfate. This conclusion is warranted in
light of the insensitivity of exhaust sulfate levels to injection parameter
modifications.
-------
230
CATERPILLAR 3208 AFTERTREATMENT DEVICE STUDY
The conclusions from the Caterpillar 3208 engine study will be subdivided
according to control device tested, and the final section will summarize the
advantages and disadvantages of all devices.
Englehard PTX Oxidation Catalysts
1) The effect of the oxidation catalysts on NO emissions was
x
negligible for all fuels and modes tested, but NO emissions increased at
the expense of NO for all modes and fuels except mode 3 with No. 1 and
No. 2 fuels. Mode 3 showed a decrease in NO , probably due to a net catalytic
reduction of NO to NO. Values of NO were highest for the No. 2 fuel, with
£• X
No. 1 and shale fuel giving equivalent NO emissions both with and without
X
the catalysts (mode 4 only).
2) The catalysts decreased both HC and SOF emissions for all fuels and modes
tested. Percent reduction of HC was generally lower for No. 1 fuel than for
No. 2 fuel.
3) The catalysts increased sulfate emissions for all fuels and modes tested.
Conversion of fuel sulfur to sulfate ranged from 6% at mode 3 with No. 1 fuel
to 89% at mode 9 with No. 2 fuel; the uncatalyzed exhaust typically contained
less than 3% of the fuel sulfur as sulfate. Shale fuel sulfate emissions were
extremely low due to the very low fuel sulfur content.
4) The catalysts increased solid particulates for all fuels and modes except
mode 3 with both No. 1 and No. 2 fuels; solids emissions using shale fuel
increased, but not significantly. Dehydrogenation of exhaust hydrocarbons on
the catalyst are thought to account for this solids increase. A decrease in
percentage paraffins in the catalyzed SOF supports the idea that dehydrogenation
is occurring.
5) The catalysts increased total particulates more for the No. 2 than the
-------
231
No. 1 fuel, due to higher fuel sulfur and higher aromatics content
(leading to more solids).
6) Biological activity of the SOF is increased by use of the PTX catalyst
and by the use of No. 2 fuel. Shale fuel gave SOF of higher SA,
with and without the catalyst, than either of the other fuels (mode 4 only).
BSSA was increased by use of the catalysts except at mode 3.
Johnson-Matthey Close-Coupled Port Catalysts
1) The exhaust port catalysts usually increased NO slightly, although
NO emissions were significantly higher than the baseline engine at the high-
load modes (5 and 9). A significant reduction of N02 back to NO took place
at the low-load modes (3 and 11), while NO levels for intermediate-load modes
(4 and 10) were not significantly different from baseline conditions.
2) The port catalysts decreased HC and SOF at all modes tested. Shale
fuel increased HC emissions over No. 2 fuel (mode 4 only).
3) The port catalysts increased sulfate emissions at all modes tested,
although the percent conversion of fuel sulfur to sulfate was lower than for
the PTX catalyst (ranging from 2% at mode 3 to 32% at mode 5). For both catalysts,
molar conversion of sulfur to sulfate peaked at near 500°C, which is close to
the temperature used in manufacturing sulfuric acid.
4) The port catalysts did not significantly change solid particulates
except for mode 11 (a large increase). Slight increases were noted at modes
3, 9, and 10, mode 4 was essentially unchanged, and mode 5 showed decreased
solids emissions. For the intermediate-speed modes, shale fuel showed a
reduction in solids at mode 3, little change at mode 4, and a large increase
at mode 5. This fuel-dependent behavior is probably due to more stable aromatics
in the shale fuel, leading to a predominance of dehydrogenation (to solids) at the
-------
232
same conditions for which oxidation (to CO and HO) is predominant using
the No. 2 fuel.
5) The port catalysts increased total particulate at rated-speed modes
(due largely to increases in sulfate), and had mixed results at intermediate
speed (due to lower sulfate and variable competition between dehydrogenation
and oxidation).
6) The port catalysts increased SA at all intermediate-speed
modes and decreased it at all rated-speed modes. The port catalysts
reduced BSSA at all modes with greater reductions at rated speed.
Corning Particulate Trap
1) The Corning traps decreased N00 emissions with little change in
£
total NO and NO emissions. The reduction in N0_ may be due to either
X ji
increased back pressure or reaction with trapped particulate, but is more
likely due to reaction of NO with water of combustion to form nitric acid.
Mixing conditions in the traps could increase residence time sufficiently
to favor the conversion of NO to nitric acid.
2) The Corning traps decreased HC emissions considerably at rated
speed with smaller reductions at intermediate speed.
3) The Corning traps reduced sulfates significantly at all speeds and
loads, probably via sulfate trapping, lower conversions of SO to sulfate
at lower particulate levels (36), or lower ambient humidity (36). Since
modes 9 and 10 showed continuous regeneration, the first possibility may
be less important.
4) The Corning traps decreased solids emissions significantly,
especially at rated-speed modes. The solids in particulate matter should, after
all, be removed preferentially by the trap.
5) The Corning traps reduced total particulate, especially at
rated-speed modes. Using porous wall filter theory, this control device
-------
233
should be capable of completely trapping particulate less than 0.4 pm and greater
than 10 \im with lesser trapping efficiencies for particles of intermediate size.
The approximate particle input size distribution for the Caterpillar 3208
yields a predicted trapping efficiency quite different from that measured
for 2 of the 6 modes tested. This difference was ascribed to particulate
and trap characteristics not in the model. The deviation of the SOF/solid
ratio from that expected from adsorption theory at certain modes can be
explained by condensation of SOF onto the particulate after the traps.
6) The Corning traps increased SA of the SOF at all modes
but mode 3. BSSA were drastically reduced for most modes, but very high
SA of certain SOF subfractions suggests that production of certain
compounds could negate the advantages of the trap as a particulate
control device.
7) The Corning traps increased exhaust back pressure with time. The
increase was mode-dependent with rated-speed modes showing the most difference;
a trap temperature of about 500°C (achieved at modes 9 and 10) was necessary
for continuous trap regeneration. These traps, without regeneration hardware,
would require continuous operation at high-load, high-speed conditions.
Johnson-Matthey Trap
1) The Johnson-Matthey trap increased BSFC and N0? emissions at most
modes with little change in NO emissions and a large reduction in HC emissions.
X
Total particulate increased, due largely to higher sulfate emissions which
dominated the particulate at most modes. Reactivation of the mesh trap
material with a washcoat designed to minimize sulfate conversion (which
Johnson-Matthey [J-M] believed was left off in the new trap) improved sulfate
emissions slightly, but reduction of sulfate to baseline levels was not
achieved.
-------
234
2) The pressure drop across the Johnson-Matthey trap increased with time
at mode 3 with a larger increase at mode 11. For modes 4, 5, 9, and 10,
pressure drop increased only slightly during brief test periods.
Summary of All Af tertreatment Devices Tested
Table 4.1 gives the BSFC and emissions changes for all four aftertreatment
devices. Disregarding time-dependent behavior, particulate traps are more
effective overall in reducing the greater number of emissions. In this study,
Corning traps are best overall except for their substantial fuel consumption
penalty and SOF biological activity.
An important concern in a comparison such as this is the time-dependence
of trap emissions. The Johnson-Matthey trap, particularly, was not investigated
thoroughly enough to establish all pertinent specifics for regeneration.
An effective and economical means of regeneration under highway or field
conditions is needed, if continuous regeneration under normal operating
conditions is not possible.
Also important is the relative change in particulate and gaseous emissions
with a given control device. While all four devices were effective in reducing
HC and SOF, only the Corning traps showed a net decrease in sulfate and NO
emissions.
Consideration for widespread use of any control device must include
health effects assessment. Although the Corning traps were the most effective
for overall emissions reduction, they produced a particulate SOF with subfractions
of extremely high biological activity (measured as SA). The mode-averaged HC were
reduced less with the Corning trap than with the other devices, and this
gaseous phase could be as biologically active as the particulate phase.
The gaseous phase HC could be significantly increased by the Corning traps.
Short-term bioassay and inhalation studies should be done before widespread
use of these or any other traps.
-------
Table 4.1 Comparison of Four Aftertreatment Devices for Diesel Emissions Control,
No. 2 fuel, six-mode weighted average (equal weighting)
Percent Change (from Baseline)
Brake-Specific Fuel Consumption and Emissions
DEVICE
Brake-
Specific
Fuel
Consumption
Particulate Emissions
Gaseous Emissions
Solid Soluble Sulfate
Total Fraction Organic Fraction
Fraction
NO
NO.
NO
HC
Englehard PTX
Oxidation
Catalysts
-0.8
+254
+162
-75
+3626
+2
+398
-6
-83
Johnson-
Mat they
Exhaust Port
Catalysts
+3.3
+41
+17
-49
+861
+16
+45 +17
-43
Corning
Ceramic
+5.0
-63
-82
-73
-63
-2
-51 +0.4
-31
Johnson-
Mat they
Mesh Trap
. ,
^ '
+0.3
+140
+36
-74
+3068
+1
+371 -19
-92
(a) Values are averages of consecutive 7-hour tests and do not consider the time-dependent behavior of the Corning traps.
(b) Values are equally weighted between modes 4, 10, and 11 only and do not consider any time-dependent behavior.
K3
LO
Ul
-------
236
CUMMINS ADIABATIC ENGINE - ALTERNATE FUEL SAMPLES
Fuel change in the Cummins experimental engine in its cooled configura-
tion slightly increased the SA of the particulate SOF from the use of
No. 2 and solvent-refind, coal-derived fuel above that of shale fuel
except at mode 8. Mode 6 SOF using No. 2 fuel showed much higher SA
than SOF from other fuels and modes and was the only SOF whose SA was
appreciably increased with metabolic activation. In this engine, there
was a general increase in SA with increasing speed and load.
-------
237
5
RECOMMENDATIONS FOR FUTURE RESEARCH
IN-CYLINDER MODIFICATIONS
1) The APE-6G pump should be further tested with additional instrumenta-
tion, specifically: pressure transducers in the injection lines and cylinder
to monitor fuel injection pressure and rate, and a photomultiplier window
in a cylinder for observing combustion temperature. Testing at selected modes
and timings with a different-rate cam and the additional temperature and pressure
information should provide a data base for an engine simulation program.
Including chemical and biological information in this study would allow
the effects of proposed injection/combustion systems to be screened without
extensive testing.
2) The effect of spray geometry should be investigated using a larger
piston bowl diameter and smaller orifice nozzles, which may limit the spray
over-penetration anticipated from the use of higher injection pressures. Lower
sac-volume nozzles (ideally, nozzles with plungers covering orifice holes)
should be utilized to see if particulate can be reduced substantially.
3) The ideal injection characteristics discussed in this report should
be investigated. This would be possible by obtaining the Universal Fuel
Injection System or UFIS (80). Test procedures outlined in the first suggestion
should be followed.
AFTERTREATMENT MODIFICATIONS
1) An uncatalyzed monolith similar to the PTX catalysts should be investigated.
The results with the Corning traps suggest that SOF and HC reductions may be achieved
without a metal catalyst, and the increase in solids seen with the catalyzed PTX
could be reduced if platinum catalyst were excluded from the monolith. Regeneration
-------
238
would be unnecessary with the uncatalyzed monoliths, and the effectiveness
of the platinum catalyst would be determined
2) A ceramic, uncatalyzed port catalyst or other close-coupled trap should
be investigated to determine the effect of increased temperature. Higher
operating temperatures may enable continuous regeneration, and a smaller filter
flow area may not degrade operating conditions at higher temperatures.
3) Limited testing of catalyzed Corning traps should determine whether the
Pt catalyst improves trap effectiveness. Higher exhaust temperatures and
continuous regeneration at more modes should result, but the Pt catalyst might
cause dehydrogenation of hydrocarbons to solid and thereby negate the effect
of higher temperatures.
4) Diesel particulate size distributions should be determined for those
engines and running conditions expected to receive widespread trap applications.
This information is necessary to determine the most effective pore structure
for particulate traps. They would also help to determine the weaknesses of the
membrane filter model and provide information on other particle retention
factors.
GENERAL RECOMMENDATIONS
1) A chemical fractionation scheme should be developed to separate
particulate SOF into classes of known chemical characteristics without
modification. The scheme should be automated and be able to handle enough
material to provide subfractions for biological testing. No present scheme
is ideal in all these respects.
2) The effect of sampling time on SOF biological activity should be
investigated. Use of a variety of nitroreductase-deficient strains of
Salmonella typhimurium should disclose nitroaromatic compounds and
establish how nitration (artifact formation) of polynuclear aromatic
hydrocarbons relates to biological activity.
-------
239
3) Widespread use of the Ames bioassay requires standardization of methods
for dose-response data analysis. A standard method should be relatively simple,
statistically sound, and usable with samples of widely varying mutagenic activities.
Current analysis methods do not meet these requirements.
4) A closer look at the gaseous phase of diesel emissions is warranted,
especially regarding biological activity. Unregulated gaseous emissions,
especially nitric acid and vapor-phase organic nitro compounds, should be
measured, since they may actually be more bioavailable than particulate, due to
diffusion across cell membranes.
5) The rate constants for formation of mutagenic precursors in diesels
should be determined experimentally. The nitro-aromatics are prime candidates:
aromatic hydrocarbons should be exposed to various levels of NO and nitric acid
with and without exhaust. Polymer or carbon traps may be useful to investigate
volatile compounds in similar experiments.
-------
240
REFERENCES
1) South Coast Technology, Inc., 1979, An Initial Assessment of the Literature
on the Measurement, Control, Transport, Transformation, and Health Effects
of Unregulated Diesel Engine Emissions. Final Report, DOT Contract No.
DOT-HS-7-01790.
2) National Research Council of the U.S., 1981, Health Effects of Exposure to
Diesel Exhaust. National Academy Press, Washington, D.C.
3) Aaronson, A. E., and R. A. Matula, 1971, "Diesel Odor and Formation of
Aromatic Hydrocarbons during the Heterogeneous Combustion of Pure Cetane in
a Single-Cylinder Diesel Engine," Thirteenth International Symposium on
Combustion, p. 421.
4) Ames, B. N.; J. McCann, and E. Yamasaki, 1975, "Methods for Detecting
Carcinogens and Mutagens with the Salmonella Mammalian-Microsome
Mutagenicity Test," Mutation Research, 31, p. 347.
5) Hollstein, M.; J. McCann, F. A. Angelsanto, and W. W. Nichols, 1979, "Short
Term Tests for Carcinogens and Mutagens," Mutation Research, 65, p. 133.
6) Wright, A. S., 1980, "The Role of Metabolism in Chemical Mutagenesis and
Chemical Carcinogenesis," Mutation Research, 75, p. 215.
7) Brooks, A. L.; C. R. Clark, and R. E. Royer, 1980, "Mutagenic
Characterization of Particulate Materials in Diesel Exhaust," Inhalation
Toxicology Research Institute Annual Report, Lovelace Biomedical and
Environmental Research Institute.
8) Wang, Y. Y.; M. S. Lee, C. M. King, and P. 0. Warner, 1980, "Evidence for
Nitroaromatics as Direct-Acting Mutagens of Airborne Particulates,"
Chemosphere, 9, p. 83.
9) Wei, E. T.; Y. Y. Wang, and S. M. Rappaport, 1980, "Diesel Emissions and
the Ames Test: A Commentary," J. Air Pollution Control Assoc., 30, p. 267.
10) Ames, B. N., 1979, "Identifying Environmental Chemicals Causing Mutations
and Cancer," Science 204, 587.
11) Schuetzle, D.; F. S. C. Lee, T. J. Prater, and S. B. Tejada, 1980, "The
Identification of Polynuclear Aromatic Hydrocarbon Derivatives in Mutagenic
Fractions of Diesel Particulate Extract," Presented at the 10th Annual
Symposium on the Analytical Chemistry of Pollutants, Dortmund, Germany.
12) Funkenbusch, E. F.; D. G. Leddy, and J. H. Johnson, 1979, "The
Characterization of the Soluble Organic Fraction of Diesel Particulate
Matter," SAE Paper No. 790418.
13) Mentser, M., and A. G. Sharkey, Jr., 1977, "Characterization of Diesel
Exhaust Particulate," Pittsburgh Energy Research Center Publication Number
RI-77/5.
-------
241
14) Frisch, L. E.; J. H. Johnson, and D. G. Leddy, 1979, "Effect of Fuels and
Dilution Ratio on Diesel Particulate Emissions," SAE Paper No.790417.
15) Schuetzle, D., L. M. Skewes, G. E. Fischer, S. P. Levine, and R. A. Gorse,
Jr., 1981, "Determination of Sulfate in Diesel Particulate," Analyt. Chem.,
53, p. 837.
16) Aoyagi, Y.; T. Kamimoto, Y. Matsui, and S. Matsuoka, 1980, "A Gas Sampling
Study on the Formation Processes of Soot and NO in a DI Diesel Engine," SAE
Paper 800254 (SAE Special Publication P-86).
17) Nielsen, T., 1981, "Nitro Derivatives of Polynuclear Aromatics: Formation
Presence, and Transformation in Stack and Exhaust Gases, and in the
Atmosphere," Ris
-------
242
28) Society of Automotive Engineers, 1979, "Measurement of Carbon Dioxide,
Carbon Monoxide, and Oxides of Nitrogen in Diesel Exhaust—SAEJ177," SAE
Handbook, Part 2.
29) Matthews, R. D.; R. F. Sawyer, and R. W. Schafer, 1977, "Interferences in
Chemiluminescent Measurements of NO and NO Emissions from Combustion
Systems," Environ. Science and Tech., 11, p. 1092.
30) EPA Draft Recommended Practice for Measurement of Gaseous and Particulate
Emissions from Heavy-Duty Engines under Transient Conditions, 1979.
31) Hunter, G.; J. Scholl, F. Kibbler, S. Bagley, D. Leddy, D. Abata, and J.
Johnson, 1981, "The Effects of Fuels on Diesel Oxidation Catalyst
Performance and the Physical, Chemical, and Biological Character of Diesel
Particulate Emissions," Society of Automotive Engineers, Publication
SP-484, SAE Paper No. 810263.
32) Hillenbrand, L. J., and D. A. Trayser, 1981, "A Concept for Catalyzed
Ignition of Diesel Soot," SAE Paper No. 811236.
33) Wade, W. R.; J. E. White, and J. J. Florek, 1981, "Diesel Particulate Trap
Regeneration Techniques," SAE Paper No. 810118.
34) Mogaka, Z. N.; V. W. Wong, and S. M. Shahed, 1982, "Performance and
Regeneration Characteristics of a Cellular Ceramic Diesel Particulate
Trap," SAE Paper No. 820272.
35) Enga, B. E.; M. F. Buchman, and I. E. Lichtenstein, 1982, "Catalytic
Control of Diesel Particulate," SAE Paper No. 820184.
36) Oh, S. H.; J. S. MacDonald, G. L. Vaneman, and L. L. Hegedus, 1981,
"Mathematical Modeling of Fibrous Filters for Diesel Particulates—Theory
and Experiment," SAE Paper No. 810113.
37) MacDonald, J. S., and G. L. Vaneman, 1981, "Experimental Evaluation of
Fibrous Filters for Trapping Diesel Exhaust Particulates," SAE Paper No.
810956.
38) Howitt, J. S.; and M. R. Montierth, 1981, "Cellular Ceramic Diesel
Particulate Filter SAE Paper No. 810114.
39) Orr, C., 1977, Filtration—Principles and Practices, Part I, New York and
Basel: Marcel Dekker, Inc.
40) Orr, C., 1966, Particulate Technology, New York: The MacMillan Company.
41) Davies, C. N., 1973, Air Filtration, London and New York: Academic Press.
42) Davies, C. N. (Editor), 1966, Aerosol Science, London and New York:
Academic Press.
-------
243
43) Lipkea, W. H.; J. H. Johnson, and C. T. Vuk, 1978, "The Physical and
Chemical Character of Diesel Particulate Emissions—Measurement Techniques
and Fundamental Considerations," SAE Paper No. 780108, SAE Publication
SP-430.
44) EPA Recommended Practice for Measurement of Exhaust Sulfate from Light-Duty
Vehicles and Trucks, 1977.
45) Lee, F. S. C.; T. M. Harvey, T. J. Prater, M. C. Paputa, and D. Schuetzle,
1980, "Chemical Analysis of Diesel Particulate Matter and an Evaluation of
Artifact Formation," ASTM STP-721, American Society for Testing and
Materials, p. 92.
46) Killough, P.; J. Watson, and R. Bradow, 1979, "Filter Type, High Volume
Particulate Sampler for Automotive Diesel Emission Studies," Environmental
Protection Agency Report on Contract No. 68-02-2566.
47) Huisingh, J.; R. Bradow, R. Jungers, L. Claxton, R. Zweidinger, S. Tejada,
J. Bumgarner, F. Duffield, M. Waters, V. F. Simmon, C. Hare, C. Rodriguez,
and L. Snow, 1978, "Application of a Mutagenicity Bioassay to Monitoring
Light-Duty Diesel Particle Emissions," in; Waters, M.D.; S. Nelnow, J. L.
Huisingh, S. S. Sandhu, and L. Claxton (eds.), Applications of Short-term
Bioassays in the Fractionation and Analysis of Complex Environmental
Mixtures, New York: Plenum Press, pp. 383-418.
48) Khatri, N. J.; J. H. Johnson, and D. G. Leddy, 1978, "The Characterization
of the Hydrocarbon and Sulfate Fraction of Diesel Particulate Matter," SAE
Publication No. 780111, Society of Automotive Engineers Transactions, V.
87.
49) "Current Methods to Measure Atmospheric Nitric Acid and Nitrate Artifacts,"
1979, EPA Report No. 600/2-79-051.
50) Ames, B. N.; J. McCann, and E. Yamasaki, 1975, "Methods for Detecting
Carcinogens and Mutagens with the Salmonella/Mammalian-Microsome
Mutagenicity Test," Mutation Research, V. 31, p. 374.
51) Belser, W.; S. D. Shaffer, R. D. Bliss, P. M. Hynds, L. Yamamoto, J. N.
Pitts, and J. A. Winer, 1981, "A Standardized Procedure for Quantification
of the Ames Salmonella/Mammalian-Microsome Mutagenicity Test," Environ.
Mut., V. 3, p. 221.
52) Vois, J. R., and R. E. Vanderpoel, 1977, "The Shuttle Distributor for a
Diesel Fuel Injection Pump," SAE Transactions Volume 86, SAE Paper No.
770083.
53) Kido, H.; Y. Wakuri, S. Ono, and E. Marase, 1980, "Prediction of
In-Cylinder Gas Motion in Engines by an Energy Method," SAE Special
Publication SD-468, SAE Paper No. 800985.
54) J. Kimberly, 1980, Personal Communication. American Bosch Diesel Products.
-------
244
55) Parker, R. F., 1976, "Future Fuel Injection System Requirements of Diesel
Engines for Mobile Power," SAE Transactions, V. 85, SAE Paper No. 760125.
56) Obert, E. 0., 1973, Internal Combustion Engines and Air Pollution, New
York, Intext Educational.
57) J. Hansel, Englehard Industries Division, 1979, Personal Communication.
58) AMOCO, 200 E. Randolph Dr., Chicago, IL.
59) Chevron Research Company, 576 Standard Ave., Richmond, CA.
60) "Diesel Fuel Oils—1974," Petroleum Products Survey, No. 87, Mineral
Industry Surveys, Bureau of Mines, U.S. Dept. of Interior.
61) Culberson, S. F., and P. D. Rolniak, 1981, "Shale Oil—Likely Prospect for
Refining," Technol. Oil and Gas J., V. 43.
62) Lin, Chieh-Shen, and D. E. Foster, 1981, "A Study of Fuel Nitrogen
Conversion, Performance, and Emission Characteristics of Blended SRC-II in
a High-Speed Diesel Engine," SAE Paper No. 810251, SAE Publication SP-480.
63) Tuteja, A. D., and D. W. Clar, 1980, "Comparative Performance and Emission
Characteristics of Petroleum, Oil Shale, and Tar Sands Derived Diesel
Fuels," SAE Paper No. 800331.
64) Muller, K., 1975, "Influence of Engine Design and Fuel Quality on Diesel
Emissions—Part II, Fuel Quality," V. 6, Proceedings of the Ninth World
Petroleum Congress, Tokyo.
65) Vuk, C. T.; M. A. Jones, and J. H. Johnson, 1976, "The Measurement and
Analysis of the Physical Character of Diesel Particulate Emissions," SAE
Paper No. 760131.
66) M. Skaates, MTU, 1980, Personal Communication.
67) Goodger, E. M., 1975, "Hydrocarbon Fuels," New York, NY, Halsted Press.
68) Marke, L. F., and D. A. Hahn, 1932, "The Catalytic Oxidation of Organic
Compounds in the Vapor Phase," The Chemical Catalog Co., New York, NY.
69) Latimer, W. M., and J. H. Hildebrand, 1951, "Reference Book of Inorganic
Chemistry, 3rd Ed.," New York, NY: The MacMillan Co.
70) Wang, Y. Y., S. M. Rappaport, R. F. Sawyer, F. E. Talcott, and F. T. Wei,
1978, "Direct-Acting Mutagens in Automobile Exhaust," Cancer Letters, V. 5,
p. 39.
71) Allen, J. R., and J. M. Learn, 1971, "Regan 2: A Multiple Linear
Regression Analysis," Academic Computing Center, University of
Wisconsin-Madison.
-------
245
72) Coates, P., 1981, 3M Company, Technical Ceramic Resources, Personal
Communication.
73) Gibson, T.L.; A. I. Ricci, and R. L. Williams, 1980, "Measurement of
Polynuclear Aromatic Hydrocarbons, Their Derivatives, and Their Reactivity
in Diesel Automobile Exhaust," General Motors Research Laboratories,
Publication GMR-3478, 11 pp.
74) Ahmad, T., 1980 "Effect of Particulate on the Measurement of Oxides of
Nitrogen in Diesel Exhaust," SAE Transactions, v. 89, SAE Paper No. 800189.
75) Dean, David L. , 1982, "Processing Pollution into Product"', J. Chemical
Education, v. 59 p. 639.
76) Roberson, J. A., and C. T. Crowe, 1975, "Engineering Fluid Mechanics,"
Boston" Houghton Mifflin Co.
77) Larsen, R. I., 1971 "A Mathematical Model for Relating Air Quality
Measurements to Air Quality Standards," U.S. Government Printing Office,
Office of Air Programs Publication No. AP-89.
78) Abraiaowitz and Stegun, 1964, "Handbook of Mathematical Functions," National
Bureau of Standards.
79) Clerc, J. C., 1982, "A Computer Model for Predicting Temperature, Solids,
Concentration, and Adsorbed Hydrocarbon Concentration for Dilute Diesel
Partieulate Emissions," M.S. Thesis, Michigan Technological University.
80) Kimberly, J. A. ; and R. A. DiDomenico, 1977, "UFIS - A New Diesel
Injection System." SAE Paper 770084.
81) Campbell, J.; J. Scholl, F. Kibbler, S. Bagley, D. Leddy, D. Abata, and J.
Johnson, 1981, "The Effect of Fuel Injection Rate and Timing on the
Physical, Chemical, and Biological Character of Diesel Particulate
Emissions," Society of Automotive Engineers, Publication SP-495, SAE Paper
No. 810946.
82) Stead, A. G.; V. Hasselblad, J. Creason, and L. Claxton, 1980, "Modelling
the Ames Test," Mutation Research, V. 85, p. 13.
83) Myers, L. E., N. H. Sexton, L. I. Southerland, and T. J. Wolff, 1981,
"Regression Analysis of Ames Test Data," Environ. Mutagen., V. 3, p. 575.
84) Dent, J. C., 1980, "Turbulent Mixing Rate—Its Effect on Smoke and
Hydrocarbon Emissions from Diesel Engines, Society of Automotive Engineers
Special Publication P-86, SAE Paper No. 800092.
85) Kibbler F. 1982, "Quantification of Ames Salmonella Microsome System
Variables", M.S. Thesis, Michigan Technological University.
86) Sokal, R. R., and F. J. Rolf, 1981, Biometry, W. H. Freeman & Co., San
Francisco.
-------
246
87) Zar, J. H., 1974, Biostatistical Analysis, Prentice-Hall Inc., Englewood
Cliffs, NJ.
88) Schuetzle, D., and J. M. Perez, 1981, "A CRC Cooperative Comparison of
Extraction and HPLC Techniques for Diesel Particulate Emissions, "Presented
at the 74th Annual Meeting of the Air Pollution Control Association,
Philadelphia.
89) Snyder, L. R., and J. J. Kirkland, 1979, Introduction to Modern Liquid
Chromatography, 2nd Edition. New York: Wiley-Interscience, p. 106.
-------
APPENDICES
-------
APPENDIX A
-------
247
MEAN
TABLE A-l
MACK ENDT-676 PARTICULATE CONTROL STUDY C.V.
SEE EXPERIMENTAL TEST MATRIX (TABLE 3.1.3) FOR EXPLANATION OF TEST CODES
MODE ' ~~ —— EPA MODE 3
TEST NUMBER J3B 3.8-3 G-12-3 G-17-3 G-19-3 G-22-3 S-8-3 S-12-3 S-14-3 S-17^
AMBIENT TEMP. 29 32 28 23 22 24 27 22 29 25
(deg. C) 0% 0% 2% 15% 3% 0% 2% 3% 0% 4%
SPEC. HUM.
(a 1 o )
V&H20 &Air
BSFC
(kg/kw-hr)
4.75
0%
0.233
0%
16.41
1%
0.231
2%
3.36
9%
0.240
0%
5.62
28%
0.231
1%
2.11
15
0.288
2%
4.06
7%
0.233
0%
2.15
41%
0.290
0%
2.21
26%
0.254
0%
3.94
8%
0.247
0%
2.59
16%
0.242
3%
N0x 705. 607. 380. 713. 716. 976. 545. 612. 542. 860.
(ppm, corr.)
, v
(ppm, corr.)
TPM
1% 7% 1% 3% 0% 1% 4% 1% 3% 1%
RqNO 10.13 9.42 5.42 10.71 10.12 13.63 7.99 9.35 7.75 12.11
(g/kw-hr) 1% 1% 2% 4% 4% 1% 5% 0% 3% 5%
81.0 26.1 7.2 55.4 40.3 86.5 78.5 60.2 31.8 49.0
8% 31% 6% 11% 5% 14% 16% 4% 11% 1%
BSNO? 1.19 0.40 0.10 0.83 0.57 1.21 1.15 0.92 0.45 0.69
(g/kw-hr) 8% 28% 10% 6% 7% 14% 17% 4% 11% 3%
NO 624. 581. 373. 658. 675. 889. 466. 552 510. 810.
(ppm, corr.) 1% 7% 1% 4% 1% 2% 3% 1% 3% 1%
BSNO 5.95 5.88 3.47 6.44 6.23 8.10 4.46 5.50 4.76 7.45
(g/kw-hr) 0% 1% 2% 5% 4% 1% 3% 1% 2% 5%
(mg/m3, std.)
47.9 71.1 65.0 48.3 51.6 46.9 33.1 28.2 35.5 54.6
5% 18% 7% 1% 1% 0% 8% 4% 8% 6%
-------
TABLE A-l (cont'd)
MACK ENDT-676 P ARTICULATE CONTROL STUDY
SEE EXPERIMENTAL TEST MATRIX (TABLE 3.1.3) FOR EXPLANATION
MODE
TEST NUMBER
BSTPM
(g/kw-hr)
S04
(mg/m , std.)
BSSOT
(g/kw-hr)
SOF
(mg/m3, std.)
BSSOF
(g/kw-hr)
INJECTION
DURATION (°CA)
RATE PARAMETER
(mm3/°CA)
3B
0.41
5%
9.9
39%
0.08
41%
21.2
12%
0.18
12%
9.0
5.6
3.8-3
0.58
14%
3.7
14%
0.03
10%
10.1
33%
0.08
30%
10.0
4.2
G-12-3
0.55
6%
6.2
12%
0.05
12%
3.5
15%
0.03
13%
9.4
6.1
EPA
G-17-3
0.40
4%
6.3
3%
0.05
6%
7.2
2%
0.06
3%
9.1
6.0
MODE 3
G-19-3
0.43
5%
5.0
33%
0.04
38%
6.1
8%
0.05
6%
9.0
5.9
G-22-3
0.39
1%
6.0
7%
0.05
6%
15.1
4%
0.12
4%
9.0
6.0
OF TEST CODES
S-8-3
0.29
9%
4.4
8%
0.04
8%
15.1
8%
0.13
92%
3.0
17.7
S-12-3
0.26
4%
5.2
3%
0.05
4%
11.1
5%
0.10
5%
3.0
17.7
S-14-3
0.30
9%
5.9
5%
0.05
6%
12.1
14%
0.10
14%
3.0
17.7
248
MEAN
C.V.
S-17-3
0.46
3%
6.1
2%
0.05
0%
14.5
2%
0.12
3%
3.0
17.7
PEAK INJECTION 40.0 41.4 46.5 46.5 46.5 46.5 65.5 65.5 65.5 65.5
PRESSURE (mPa)
-------
MEAN 249
TABLE A-l (cont'd)
MACK ENDT-676 PARTICULATE CONTROL STUDY C.V.
SEE EXPERIMENTAL TEST MATRIX (TABLE 3.1.3) FOR EXPLANATION OF TEST CODES
MODE
TEST NUMBER
AMBIENT TEMP.
(deg. C)
SPEC. HUM.
BSFC
(kg/kw-hr)
NOX
(ppm, corr.)
BSNO
(g/kw-hr)
NO 2
(ppm, corr.)
BSN02
(g/kw-hr)
NO
(ppm, corr.)
BSNO
(g/kw-hr)
TPM
(mg/m-^, std.)
EPA MODE 4
4B
33
0%
5.47
0%
0.209
1278
0%
10.74
0%
133.7
4%
1.12
4%
1144.
1%
6.27
0%
84.0
6%
3.8-4
20
10%
6.65
2%
0.213
1%
1061.0
1%
9.52
1%
31.3
19%
0.28
21%
1030.
1%
6.03
0%
79.1
3%
G-12-4
18
3%
1.77
36%
0.236
10%
746.0
1%
7.04
11%
15.6
24%
0.15
33%
730.
2%
4.49
10%
71.4
12%
G-17-4
24
7%
6.68
10%
0.209
0%
1284
0%
10.88
1%
22.5
44%
0.19
42%
1261.
1%
6.97
2%
63.0
7%
G-19-4
22
3
1.56
0%
0.214
0%
1444
1%
11.71
3%
93.3
2%
0.76
1%
1351.
1%
7.15
3%
65.9
1%
G-22-4
19
0%
3.08
11%
0.213
0%
1768
1%
14.46
1%
100.6
1%
0.82
1%
1668.
1%
8.89
0%
57.0
4%
S-17-4
31
3%
10.16
4%
0.212
0%
1396
1%
12.27
3%
69.1
27%
0.61
26%
1327.
2%
7.61
4%
107.5
2%
-------
TABLE A-l (cont'd)
MACK ENDT - 676 PARTICULATE CONTROL STUDY
SEE EXPERIMENTAL TEST MATRIX FOR EXPLANATION OF TEST
MODE
TEST NUMBER
BSTPM
(g/kw-hr)
(mg/m , std.)
BSSO^
(g/kw-hr)
SOF
(mg/m3, std.)
BSSOF
(g/kw-hr)
INJECTION
DURATION (°CA)
RATE PARAMETER
(mm /°CA)
CODES
250
MEAN
C.V.
EPA MODE 4
4B
0.40
6%
7.5
2%
0.04
3%
10.8
14%
0.05
14%
15.0
4.2
3.8-4
0.39
3%
5.2
6%
0.03
3%
6.9
12%
0.03
13%
16.0
4.2
G-12-4
0.39
13%
7.0
5%
0.04
13%
1.2
67%
0.01
50%
13.4
5.4
G-17-4
0.30
6%
7.3
3%
0.04
3%
2.1
23%
0.01
20%
12.8
5.3
G-19-4
0.31
4%
6.6
32%
0.03
37%
3.3
8%
0.02
5%
13.0
5.3
G-22-4
0.28
4%
8.4
5%
0.04
5%
7.3
14%
0.04
13%
12.6
5.3
S-17-4
0.50
3%
9.4
5%
.04
5%
5.0
16%
0.02
13%
PEAK INJECTION 46.5 48.3 51.7 51.7 51.7 51.7
PRESSURE (mPa)
-------
251
TABLE A-l (cont'd)
MACK ENDT - 676 PARTICULATE CONTROL STUDY
SEE EXPERIMENTAL TEST MATRIX FOR EXPLANATION OF TEST CODES
MODE
TEST NUMBER
AMBIENT TEMP.
(deg. C)
SPEC. HUM.
(8H20/gAir)
BSFC
(kg/kw-hr)
NOx
(ppm, corr.)
BSNO
(g/kw-hr)
NO 2
(ppm, corr.)
BSNO,
(g/kw-hr)
NO
(ppm, corr.)
BSNO
(g/kw-hr)
TPM
(me/nr*. std.}
MEAN
C.V.
EPA MODE 5
5B
28
0%
10.64
0%
0.212
0%
1247
5%
8.67
6%
55.9
63%
0.39
62%
1190.
6%
5.40
5%
155.7
6%
.3.8-5
21
0%
6.72
5%
0.216
0%
1170
2%
7.90
3%
33.3
17%
0.22
18%
1136.
2%
5.01
3%
166.2
9%
G-12-5
18
3%
2.77
12%
0.222
0%
974.
1%
6.52
2%
14.1
30%
0.09
33%
960.
1%
4.19
1%
108.5
6%
G2-12-5
14
8%
3.97
15%
0.228
1%
743
7%
4.61
18%
20.6
53%
0.13
53%
722.
6%
2.91
16%
8.62
9%
G-17-5
7
9%
4.40
0%
0.211
0%
1670
2%
9.10
25%
93.6
20%
0.50
18%
1576.
3%
5.61
26%
81.9
11%
G2-17-5
31
8%
5.16
0%
0.216
0%
1070
0%
8.21
2%
32.5
0%
0.25
2%
1037
0%
5.19
2%
179.4
7%
-------
TABLE A-L (cont'd)
MACK ENDT - 676 PARTICULATE CONTROL STUDY
SEE EXPERIMENTAL TEST MATRIX FOR EXPLANATION OF TEST CODES
MODE
TEST NUMBER
BSTPM
(g/kw-hr)
SO^
(mg/m , std.)
BSSO^
(g/kw-hr)
SOF
(mg/m3, std.)
BSSOF
(g/kw-hr)
INJECTION
DURATION (°CA)
RATE PARAMETER
(mm /°CA)
252
MEAN
C.V.
EPA MODE 5
5B
0.62
2%
4.6
42%
0.02
35%
6.65
5%
0.028
18%
19.0
4.0
3.8-5
0.62
8%
3.6
3%
0.01
10%
4.7
12%
0.02
10%
22.0
3.7
G-12-5
0.42
7%
6.2
22%
0.02
25%
1.2
73%
0.00
0%
18.0
4.9
G2-12-5
0.51
17%
0.376
14%
0.02-2,
25%
0.22
12%
0.013
15%
18.0
4.9
G-17-5
0.26
17%
8.3
13%
0.03
17%
0.5
80%
0.00
0%
17.0
4.7
G2-17-5
0.78
5%
5.4
6.5%
0.023
4%
3.87
19%
0.017
16%
17.0
4.7
PEAK INJECTION 51.9 58.6 51.7 51.7 51.7 51.7
PRESSURE (mPa)
-------
TABLE A-l (cont'd)
SEE
MODE
TEST NUMBER
AMBIENT TEMP.
(deg. C)
SPEC. HUM.
(e /P. )
OTT f\ * O A -I «• '
novj "• -*- i
BSFC
(kg/kw-hr)
NOx
(ppm, corr.)
BSNO
(g/kw-hr)
N02
(ppm, corr.)
BSNO-
(g/kw-hr)
NO
(ppm, corr.)
BSNO
(g/kw-hr)
MACK ENDT-676
P ARTICULATE CONTROL
EXPERIMENTAL TEST MATRIX
EPA
G-19-5
19
0%
1.3
89%
0.216
1809
0%
11.50
2%
121.8
1%
0.77
1%
1687.
0%
6.99
2%
MODE 5
G-22-5
15
7.5%
3.3
15%
0.211
2078
0%
13.23
0%
115.2
12%
0.73
12%
1963.
1%
8.15
1%
STUDY
FOR EXPLANATION OF
G2-22-5
29
2%
10.8
0%
0.208
1751
0%
14.36
0%
26.7
9%
0.22
8%
1724.
0%
9.22
0%
S-12-5
33
6%
12.7
6%
0.230
813
1%
5.44
6%
21.9
15%
0.14
15%
798.
1%
3.45
6%
TEST CODES
S-14-5
32
5%
12.1
4%
0.231
900
1%
5.76
12%
32.8
29%
0.23
25%
864.
2%
3.61
12%
253
MEAN
C.V.
S-17-5
35
1%
10.9
10%
0.212
1723
1%
10.87
2%
83.0
6%
0.53
8%
1639.
1%
6.74
1%
TPM
131.6
73.3
187.1
183.5
193.3
210.1
, std.)
-------
TABLE A-l (cont'd)
MACK ENDT-676 PARTICULATE CONTROL STUDY
SEE EXPERIMENTAL TEST MATRIX FOR EXPLANATION OF TEST CODES
MODE
TEST NUMBER
BSTPM
(g/kw-hr)
SOA
(mg/m , std.)
BSSO=
(g/kw-hr)
SOF
o
(mg/mj, std.)
BSSOF
(g/kw-hr)
INJECTION
DURATION (°CA)
RATE PARAMETER
(imn /°CA)
G-19-5
0.50
4%
3.3
25%
0.01
30%
4.0
6%
0.02
5%
17.0
4.8
EPA MODE
G-22-5
0.28
5%
7.6
4%
0.03
3%
2.4
21%
0.01
20%
17.0
4.8
5
G2-22-5
0.81
3%
6.0
4%
0.03
4%
2.3
89%
0.01
87%
17.0
4.8
S-12-5
0.66
9%
8.6
11%
0.03
10%
4.9
46%
0.02
44%
S-14-5
0.65
3%
9.7
7%
0.03
15%
5.6
31%
0.02
42%
254
MEAN
C.V.
S-17-5
0.70
2%
7.4
11%
0.02
12%
3.6
32%
0.01
33%
PEAK INJECTION 51.7 51.7 51.7
PRESSURE (mPa)
-------
TABLE A-l (cont'd)
MACK ENDT - 676 PARTICULATE CONTROL STUDY
SEE EXPERIMENTAL TEST MATRIX FOR EXPLANATION OF TEST CODES
MODE
TEST NUMBER
AMBIENT TEMP.
(deg. C)
SPEC. HUM.
BSFC
(kg/kw-hr)
NOX
(ppm, corr.)
BSNO
(g/kw-hr)
NO 2
(ppm, corr.)
BSNO
(g/kw-hr)
NO
(ppm, corr.)
BSNO
(g/kw-hr)
TPM
(mg/nr5, std.)
9B
31
0%
9.45
0%
0.252
0%
586.1
2%
6.14
4%
24.7
40%
0.26
38%
561.4
2%
3.84
5%
50.7
9%
3.8-9
26
0%
18.08
4%
0.261
2%
503.1
1%
6.05
8%
12.9
11%
0.16
13%
490.2
1%
3.85
8%
152.3
9%
EPA MODE 9
G-12-9
7
14%
2.46
10%
0.285
0%
395.0
0%
4.34
1%
10.1
71%
0.11
73%
384.9
2%
2.76
2%
56.6
3%
G-17-9
9
7%
4.62
6%
0.255
0%
665.7
1%
7.65
1%
45.9
5%
0.53
6%
619.8
1%
4.65
1%
62.8
1%
G-19-9
15
7.5%
0.0
0%
0.262
0%
728.3
0%
7.66
2%
37.9
6%
0.40
5%
690.4
0%
4.74
2%
56.7
3%
255
MEAN
C.V.
G-22-9
20
3%
3.93
7%
0.248
0%
897.7
4%
9.46
4%
62.9
28%
0.66
27%
834.8
4%
5.74
4%
59.1
6%
-------
TABLE A-l (cont'd)
MACK ENDT -
SEE EXPERIMENTAL TEST
MODE
TEST NUMBER
BSTPM
(g/kw-hr)
soA
(mg/m , std.)
BSSO=
(g/kw-hr)
SOF
(mg/m3, std.)
BSSOF
(g/kw-hr)
INJECTION
DURATION (°CA)
RATE PARAMETER
(mm /°CA)
676 PARTICULATE CONTROL STUDY
MATRIX FOR EXPLANATION OF TEST
CODES
256
MEAN
C.V.
EPA MODE 9
9B
0.30
10%
3.0
7%
0.02
5%
7.2
13%
0.04
15%
25.0
3.0
3.8-9
0.89
10%
3.1
5%
0.02
5%
8.1
19%
0.05
18%
28.0
2.9
G-12-9
0.36
4%
7.4
2%
0.05
20%
5.8
16%
0.04
15%
22.0
4.2
G-17-9
0.39
1%
7.2
1%
0.04
0%
12.0
5%
0.07
6%
20.0
4.1
G-19-9
0.36
3%
9.0
28%
0.06
25%
9.5
17%
0.06
15%
20.0
3.8
G-22-9
0.36
6%
6.2
4%
0.04
5%
17.8
9%
0.11
9%
19.0
4.0
PEAK INJECTION 77.6 75.8 82.7 82.7 82.7 82.7
PRESSURE (mPa)
-------
TABLE A-l (cont'd) 257
MACK ENDT - 676 PARTICULATE CONTROL STUDY
SEE EXPERIMENTAL TEST MATRIX FOR EXPLANATION OF TEST CODES
MODE
TEST NUMBER
AMBIENT TEMP.
(deg. C)
SPEC. HUM.
(gH20/'gAir)
BSFC
(kg/kw-hr)
NOx
(ppm, corr.)
BSNO
(g/kw-hr)
NO 2
(ppm, corr.)
BSN02
(g/kw-hr)
NO
(ppm, corr.)
BSNO
(g/kw-hr)
TPM
fmcrlm-J o f A \
MEAN
C.V.
EPA MODE 10
10B
35
0%
4.22
0%
0.269
1%
426.8
1%
5.65
1%
17.5
15%
0.23
17%
409.3
1%
3.53
0%
56.9
3%
3.8-10
26
2%
4.27
35%
0.276
1%
398.9
3%
5.45
3%
7.3
25%
0.10
30%
391.6
4%
3.49
3%
129.3
8%
G-12-10
19
3%
1.98
31%
0.299
0%
307.3
0%
4.59
1%
2.9
14%
. 0.04
25%
304.4
0%
2.96
1%
56.4
2%
G-17-10
21
0%
7.31
3%
0.268
0%
447.5
0%
6.59
0%
9.6
40%
0.14
43%
437.9
1%
4.20
1%
68.6
3%
G-19-10
16
7%
1.39
4%
0.278
7%
515.3
1%
6.89
5%
20.2
4%
0.27
7%
495.1
2%
4.32
5%
68.9
4%
G-22-10
28
6%
2.87
5%
0.259
0%
669.6
2%
8.02
2%
48.5
4%
0.58
2%
621.0
2%
4.85
2%
74.9
7%
-------
TABLE A-l (cont'd)
MACK ENDT -
SEE EXPERIMENTAL TEST
MODE
TEST NUMBER
BSTPM
(g/kw-hr)
S04
(mg/m , std.)
BSSO"
(g/kw-hr)
SOF
(mg/m , std.)
BSSOF
(g/kw-hr)
INJECTION
DURATION (°CA)
RATE PARAMETER-
Cmni /"CAI
676 PARTICULATE CONTROL STUDY
MATRIX FOR EXPLANATION OF TEST CODES
258
MEAN
C.V.
EPA MODE 10
10B
0.45
3%
5.1
14%
0.04
15%
10.6
5%
0.08
5%
18.0
3.6
3.8-10
1.23
37%
2.0
89%
0.02
700%
14.6
10%
0.115
9%
20.0
3.6
G-12-10
0.50
2%
5.6
38%
0.05
38%
3.7
36%
0.03
40%
17.0
4.6
G-17-10
0.56
3%
5.7
1%
0.05
0%
12.0
14%
0.10
14%
16.0
4.4
G-19-10
0.56
5%
5.7
14%
0.05
16%
12.2
6%
0.10
9%
16.0
4.4
G-22-10
0.57
8%
5.6
4%
0.04
5%
27.6
12%
0.21
12%
15.0
4.2
PEAK INJECTION 56.9 68.9 68.9 68.9 68.9 68.9
PRESSURE (mPa)
-------
TABLE A-l (cont'd) 259
MACK ENDT - 676 PARTICULATE CONTROL STUDY
SEE EXPERIMENTAL TEST MATRIX FOR EXPLANATION OF TEST CODES
MODE
TEST NUMBER
AMBIENT TEMP.
(deg. C)
SPEC. HUM.
'V/W
BSFC
(kg/kw-hr)
NOx
(ppm, corr.)
BSNO
(g/kw-hr)
NO 2
(ppm, corr.)
BSNO 2
(g/kw-hr)
NO
(ppm, corr.)
BSNO
(g/kw-hr)
TPM
fmcj/m-^ std-^
EPA MODE 1]
11B
35
0%
2.74
0%
0.337
255.8
0%
6.65
1%
1.8
28%
0.05
20%
254.0
0%
4.31
1%
54.4
5%
3.8-11
26
2%
3.62
15%
0.356
1%
236.5
2%
5.68
1%
6.7
70%
0.16
69%
229.8
3%
3.63
2%
172.2
10%
G-12-11
25
26%
2.72
74%
0.372
2%
179.3
3%
4.41
2%
2.3
52%
0.06
50%
177.0
3%
2.84
2%
82.2
2%
[
G2-12-11
19
11%
1.72
31%
0 .395
1%
131
2%
4.41
6%
3.1
22%
0.10
20%
128
2%
4.31
6%
84.3
3%
G-17-11
30
2%
5.36
3%
0.337
0%
284.0
1%
6.56
0%
8.4
38%
0.19
37%
275.6
1%
4.15
1%
106.2
6%
MEAN
C.V.
G2-17-11
23
2%
9 .38
0%
.357
0%
208
1%
7.04
1%
0.6
83%
0.02
100%
207
1%
4.58
0%
94.9
0.5%
-------
TABLE A-l (cont'd)
MACK ENDT -
SEE EXPERIMENTAL TEST
MODE
TEST NUMBER
BSTPM
(g/kw-hr)
s°4 3
(mg/m , std.)
BSSO"
(g/kw-hr)
SOF
o
(mg/mj, std.)
BSSOF
(g/kw-hr)
INJECTION
DURATION (°CA)
RATE PARAMETER
(mm3 /°CA)
676 PARTICULATE CONTROL STUDY
MATRIX FOR EXPLANATION OF TEST CODES
260
MEAN
C.V.
EPA MODE 11
11B
0.78
6%
4.0
30%
0.06
28%
17.6
10%
0.25
108%
13.0
4.2
3.8-11
2.44
10%
2.3
13%
0.03
13%
20.03
10%
0.28
9%
14.0
4.2
G-12-11
1.21
7%
3.9
3%
0.06
2%
5.7
39%
0.09
41%
13.1
5.1
G2-12-11
1.48
3%
4.6
15%
.082
17%
7.7
13%
0,14
14%
13.1
5.1
G-17-11
1.44
7%
5.5
6%
0.07
7%
17.7
9%
0.24
9%
12.0
4.7
G2-17-11
1.69
1%
5.0
3%
0.09
2%
11.0
22%
0.20
22%
12.0
4.7
PEAK INJECTION 55.2 55.2 65.5 65.5 65.5 65.5
PRESSURE (mPa)
-------
TABLE A-l (cont'd) 261
MACK ENDT - 676 PARTICULATE CONTROL STUDY
MEAN
SEE EXPERIMENTAL TEST MATRIX FOR EXPLANATION OF TEST CODES
C.V.
MODE
TEST NUMBER
AMBIENT TEMP.
(deg. C)
SPEC. HUM.
BSFC
(kg/kw-hr)
(ppm, corr.)
BSNO
(g/kw-hr)
N02
(ppm, corr.)
BSN02
(g/kw-hr)
NO
(ppm, corr.)
BSNO
(g/kw-hr)
TPM
(mg/m-*, std.)
G-19-11
23
0%
1.66
12%
0.346
0%
302.
1%
7.07
1%
11.8
8%
0.28
7%
290.
1%
4.43
1%
115.8
4%
G-22-1
30
8%
2.92
6%
0.324
2%
373.
5%
8.27
14%
15.7
16%
0.35
23%
357.
6%
5.16
14%
88.0
2%
EPA MODE 11
1 G2-22-11
29
2%
12.73
0%
0.324
0
319
1%
10.1
1%
19.4
0.5%
0 .62
1.6%
300.
1%
6.18
1%
82.3
3%
G3-22-11
30
4%
9.50
0%
0.338
0%
295.
1%
9.46
2%
1.9
0%
0.06
0%
293.
1%
6.13
2%
107.9
3%
S-12-11
31
13%
13.68
10%
0.396
2%
282.
1%
7.42
4%
35.4
1%
0.93
4%
246.
1%
4.23
4%
36.0
9%
-------
TABLE A-l (ccmt'd)
MACK ENDT - 676 PARTICULATE CONTROL STUDY
SEE EXPERIMENTAL TEST MATRIX FOR EXPLANATION OF TEST
MODE
TEST NUMBER
BSTPM
(g/kw-hr)
(mg/m , std.)
(g/kw-hr)
SOF
o
(mg/m-3, std.)
BSSOF
(g/kw-hr)
INJECTION
DURATION (°CA)
RATE PARAMETER -
(ran /°CA)
G-19-11
1.63
4%
5.05
10%
0.080
9%
17.2
3%
0.24
3%
12.0
4.8
G-22-1
1.20
6%
4.60
9%
0.060
8%
28.5
3%
0.39
6%
11.5
4.7
EPA MODE 11
1 G2-22-11
1.37
4%
5.70
4%
0.095
4%
15.3
2%
0.25
3%
11.5
4.7
C3-22-11
1.82
3%
4.11
4%
0.069
4%
11.6
8%
0.20
8%
262
MEAN
CODES C.V.
S-12-11
0.50
11%
4.34
8%
0.060
5%
18.1
14%
0.25
16%
PEAK INJECTION 65.5 65.5 65.5
PRESSURE (mPa)
-------
TABLE A-2
263
CATERPILLAR 3208 PARTICULATE CONTROL STUDY
ENGLEHARD PTX CATALYST EVALUATION
INTERMEDIATE SPEED TEST SUMMARY
MODE
FUEL
PTX CATALYST
AMBIENT TEMP.
(deg. C)
BAROMETRIC PRES.
(kPa)
SPEC. HUM.
So/8Air}
SPEED
(rpm)
LOAD
(N-m)
BMEP, Corr.
(kPa)
BSFC
(kR/kw-hr)
EPA MODE 3
NO.
W/0
29
0%
99
0%
4.5
11%
1680
160
— —•
193
0%
.286
1%
2
W
30
1%
100
0%
3.0
6%
1680
__
160
— —
191
1%
.293
1%
NO.l
W/0
34
2%
99
0%
13.2
5%
1680
_ —
160
"
197
0%
.280
4%
W
28
3%
100
0%
9.1
5%
1680
— • —
160
""
193
0%
.280
2%
NO.
W/0
31
0%
101
0%
2.0
10%
1680
320
"
385
0%
.233
0%
EPA MODE 4
2
W
32
1%
100
0%
2.7
29%
1680
— -*•
320
""
381
0%
.236
1%
NO.
W/0
34
4%
100
0%
6.7
9%
1680
__
320
"
389
0%
.243
6%
1
W
31
3%
100
0%
9.5
6%
1680
— —
320
"
390
0%
.235
3%
SHALE
W/0
29
2%
98
0%
4.3
6%
1680
"~^
320
"
388
0%
.229 .
0%
MEAN
C.V.
EPA MODE 5
W
30
3%
98
0%
5.1
4%
1680
»_
320
389
0%
231
0%
NO.
W/0
31
0%
9*
0%
3.1
13%
1680
_«
480
580
0%
.217
1%
2
W
31
3%
100
0%
2.4
25%
1680
— ~-
480
571
0%
.214
1%
NO.
W/0
33
3%
98
0%
10.8
2%
1680
""""""
480
594
0%
.219
0%
1
W
31
0%
99
0%
10.4
3%
1680
"
480
585
0%
.215
1%
-------
264
TABLE A-2 (cont'd)
CATERPILLAR 3208 PARTICULATE CONTROL STUDY
ENGLEHARD PTX CATALYST EVALUATION
INTERMEDIATE SPEED TEST SUMMARY
MODE
FUEL
PTX CATALYST
AIR/FUEL RATIO
^Alr'^Fuel*
MASS FLOW AIR
(kg/hr)
MASS FLOW EXH.
(kg/hr)
MASS FLOW EXH.
(m /min., act.)
EXH. TEMP.
(deg. C)
EXH. 02 CONG.
(%)
NO CONC.
/ x N
(ppm, corr.)
N00 CONC.
, 2 ,
(ppm, corr.)
NO CONC.
Cnnm. corr.")
EPA MODE 3
NO.
W/0
67
1%
536
0%
544
0%
13.9
1%
252
0%
16.0
1%
432
2%
28
10%
405
2%
2
W
67
1%
547
0%
555
0%
14.2
0%
258
1%
16.1
1%
429
1%
18
1]%
411
1%
NO.l
W/0
64
3%
514
1%
522
1%
13.2
1%
246
1%
15.5
1%
352
2%
29
7%
352
2%
W
69
1%
539
0%
546
0%
14.0
1%
254
0%
16.0
1%
398
2%
15
13%
398
2%
NO.
W/0
41
2%
530
0%
543
0%
16.8
1%
365
0%
13.1
0%
805
1%
16
12%
788
1%
EPA MODE 4
2
W
41
2%
535
0%
546
0%
17.0
1%
372
1%
13.1
0%
765
1%
246
2%
519
1%
NO.
W/0
37
5%
514
1%
529
1%
16.1
0%
356
0%
12.2
3%
673
1%
12
8%
673
1%
1
W
40
3%
529
0%
542
0%
16.9
0%
367
0%
12.6
1%
701
2%
209
6%
701
2%
SHALE
W/0
40
0%
522
0%
534
0%
16.2
0%
346
0%
12.9
0%
688
2%
40
10%
647
2%
MEAN
C.V.
EPA MODE 5
W
40
1%
518
0%
531
0%
16.4
0%
358
0%
12.7
0%
691
2%
234
4%
455
1%
NO.
W/0
30
3%
539
2%
558
2%
20.4
2%
499
0%
10.2
2%
1119
1%
17
29%
1102
1%
2
W
31
0%
554
1%
571
0%
20.6
1%
481
0%
10.8
2%
1019
2%
105
24%
914
5%
NO.
W/0
26
0%
495
1%
514
1%
19.3
0%
489
0%
8.6
1%
1027
1%
14
71%
1027
1%
1
W
29
3%
523
1%
541
0%
20.4
0%
500
1%
9.6
1%
1062
4%
114
7%
1062
4%
-------
TABLE A-2 (cont'd)
CATERPILLAR 3208 PARTICULATE CONTROL STUDY
ENGLEHARD PTX CATALYST EVALUATION
RATED SPEED TEST SUMMARY
MODE
FUEL
PTX CATALYST
N02 CONC.
(ppm, act.)
HC CONC.
(ppm C)
EPA MODE 3
NO.
W/0
30
10%
210
5%
2
W
21
10%
55
0%
NO.
W/0
29
7%
220
0%
1
W
15
13%
54
4%
EPA MODE
NO. 2
W/0 W
19 286
11% 1%
150 25
7% 0%
NO.
W/0
13
8%
85
8%
4
1
W
215
7%
29
21%
265
MEAN
C.V.
EPA MODE
SHALE
W/0
45
9%
152
0%
W
262
4%
52
0%
NO.
W/0
19
32%
97
3%
2
W
120
21%
18
17%
5
NO.
W/0
15
67%
52
4%
1
W
115
9%
13
23%
BSNO
(g/kw-hr)
15.5 16.2 10.2 13.0 14.9 14.2 11.2 11.3 10.7 10.7 13.6 12.9 10.1 11.2
1% 1% 1% 2% 1% 1% 2% 3% 2% 2% 4% 3% 1% 4%
BSN02
(g/kw-hr)
1.0 0.7 0.9 0.5 0.3 4.6 0.2 3.4 0.6 3.6 0.2 1.3 0.2 1.2
10% 14% 11% 20% 0% 2% 0% 6% 10% 4% 50% 23% 50% 8%
BSNO
(g/kw-hr)
9.5 10.2 6.1 8.2 9.5 6.3 7.2 5.2 6.6 4.6 8.7 7.6 6.5 6.5
2% 1% 2% 2% 1% 2% 1% 2% 2% 1% 5% 5% 0% 6%
BSHC
(g/kw-hr)
2.05 0.25 2.30 0.60 0.73 0.11 0.50 0.20 0.71 0.24 0.31 0.05 0.20 0.10
5% 0% 0% 0% 7% 0% 0% 0% 0% 0% 3% 20% 0% 0%
EXH. OPACITY
0.3 0.4 0.4 0.3 0.5 1.2 0.5 0.7 0.4 0.8 1.8 1.9 1.3 1.7
0% 0% 0% 0% 0% 0% 20% 14% 0% 0% 0% 0% 0% 6%
-------
266
TABLE A-2 (cont'd)
CATERPILLAR 3208 PARTICULATE CONTROL STUDY
ENCLEHARD PTX CATALYST EVALUATION
INTERMEDIATE SPEED TEST SUMMARY
MODE
FUEL
PTX CATALYST
N02 @ FILTER FACE
(ppm)
TEMP. @ FILTER FA(
(dego C)
TPM CONC.
(mg/m , act.)
TPM CQNC.
(mg/m , std.)
BSTPM
(g/kw-hr)
TPM E.F.
(g/D
SOLID3CONC.
(mg/m , std.)
EPA MODE 3
NO.
W/0
2.0
10%
39
0%
3.3
0%
53
6%
0.89
4%
2.10
5%
7
"
2
W
1.4
7%
40
3%
2.1
5%
34
0%
0.57
0%
1.40
7%
5
"
NO.
W/0
2.0
10%
44
2%
3.2
9%
51
8%
0.80
9%
1.92
7%
6
"
1
W
1.0
20%
41
2%
1.2
0%
20
5%
0.33
3%
0.80
0%
4
"
NO.
W/0
1.3
15%
48
0%
2.5
0%
40
3%
0.33
3%
1.00
0%
10
"
EPA
2
W
19.2
1%
47
4%
12.7
1%
204
2%
1.70
2%
5.00
2%
70
"
MODE
NO.
W/0
0.9
11%
50
2%
2.1
10%
35
14%
0.28
14%
0.77
13%
9
"
4
1
W
14.7
7%
49
4%
2.6
4%
42
5%
0.35
6%
1.00
0%
15
"
SHAL
W/0
2.8
17%
44
3%
3.3
2%
59
9%
0.48
9%
1.46
9%
8
""
E
W
17.4
3%
47
1%
1.2
3%
20
2%
0.16
2%
0.49
2%
9
""
EPA
NO.
W/0
1.3
31%
51
0%
2.9
3%
47
0%
0.27
4%
0.80
13%
35
"
MEAN
C.V.
MODE
2
W
8.4
24%
51
2%
22.2
3%
343
2%
1.99
3%
6.50
5%
87
""
5
NO.
W/0
1.0
60%
50
2%
1.9
5%
32
3%
0.16
6%
0.50
0%
26
"
1
W
7.8
6%
48
0%
4.2
7%
69
7%
0.37
5%
1.17
8%
30
"
SOF CQNC 2'8 1'1 2'8 °'9 1>6' °'4 1<6 °'4 2'8 °'5 °'4 °'1 °'3 °'1
(mg/m act.) 0% 9% 11% 0% 0% 25% 13% 25% 2% 4% 25% 100% 33% 100%
qm? TONf 45 18 45 15 26 6 26 7 50 8 6252
, / 3 . , , 4% 11% 9% 7% 0% 0% 15% 14% 8% 6% 17% 50% 20% 50%
(mg/m , std.)
-------
TABLE A- 2 (cont1
MODE
FUEL
PTX CATALYST
BSSOF
(g/kw-hr)
SOF E.F.
(g/D
SO^ CQNC.
(mg/m , act.)
SOT CQNC.
(mg/m , std.)
BSSOT
(g/kw-hr)
S0~ E.F.
(g/D
CONV. FUEL
s ->• so"
4
DEVICE OPER.
TEMP.
(deg. C)
DEVICE SPACE
VELOCITY
(device vol. /sec.)
'd)
CATERPILLAR 3208
ENGLEHARD PTX
INTERMEDIATE
EPA MODE 3
NO. 2
W/0 W
0.76 0.31
4% 16%
1.53 0.61
4% 15%
0.1 0.7
0% 14%
1.0 11.0
0% 9%
0.02 0.18
50% 11%
0.04 0.42
25% 14%
0.83 7.82
19% 13%
269
0%
167
0%
NO.l
W/0
0.71
10%
1.54
10%
0.0
0%
0
0%
0.01
100%
0.02
50%
2.01
55%
W
0.24
4%
0.47
9%
0.1
0%
1
0%
0.02
0%
0.05
20%
5.6
14%
262
0%
167
1%
PARTICULATE CONTROL
CATALYST EVALUATION
SPEED TEST SUMMARY
EPA
NO. 2
W/0 W
0.21 0.5
5% 0%
0.72 0.16
1% 6%
0.2 7.6
50% 8%
4.0 128
25% 3%
0.03 1.08
33% 5%
0.10 3.13
20% 4%
1.92 58.0
22% 4%
383
0%
200
1%
MODE
NO.l
W/0
0.21
14%
0.73
15%
0.0
0%
0
0%
0.00
0%
0.01
0%
1.1
0%
4
W
0.06
17%
0.20
10%
1.2
25%
20
25%
0.17
24%
0.48
21%
58.9
22%
STUDY
SHALE
W/0
0.41
8%
1.24
8%
0.04
12%
0.8
12%
0.006
15%
0.02
13%
N.A.
267
MEAN
C.V.
W
0.61
6%
0.18
6%
0.2
11%
2.9
10%
0.02
9%
0.07
10%
N.A.
376
1%
203
1%
EPA
NO. 2
W/0
0.03
33%
0.13
15%
0.4
25%
5.7
37%
0.03
66%
0.10
40%
1.89
41%
MODE
W
0.01
0%
0.04
50%
16.5
5%
254
6%
1.48
6%
4.80
6%
89.0
6%
498
0%
242
1%
5
NO.l
W/0 W
0.03 0.01
33% 100%
0.12 0.03
33% 67%
0.0 2.3
N.A. 17%
1.0 37
N.A. 16%
0.00 0.20
N.A. 20%
0.01 0.64
N.A. 17%
1.3 78.8
N.A. 18%
513
0%
245
1%
-------
268
TABLE A-2 (cont'd)
CATERPILLAR 3208 P ARTICULATE CONTROL
ENGLEHARD PTX CATALYST EVALUATION
RATED SPEED TEST SUMMARY
MODE
FUEL
PTX CATALYST
AMBIENT TEMP.
(deg. C)
BAROMETRIC PRES.
(kPa)
SPEC. HUM.
(8H20/8Air)
SPEED
(rpm)
LOAD
(N-m)
BMEP, Corr.
(kPa)
BSFC
(kg/kw-hr)
AIR-FUEL RATIO
Cks . . /kg , )
EPA MODE 9
NO.
W/0
37
3%
99
0%
6.2
6%
2800
399
489
0%
.255
0%
24
4%
2
W
22
5%
99
0%
7.4
4%
2800
399
479
0%
.271
3%
25
4%
EPA MODE 10
NO.
W/0
25
4%
102
0%
2.1
10%
2800
266
308
0%
.296
2%
36
3%
2
W
27
0%
99
0%
4.3
28%
2800
266
318
0%
.285
1%
35
3%
STUDY MEAN
C.V.
EPA MODE 11
NO. 2
W/0
32
3%
99
0%
9.8
29%
2800
133
162
1%
.379
1%
50
2%
W
27
0%
99
0%
2.3
13%
2800
133
159
0%
.342
1%
52
2%
Air 6Fuel
-------
TABLE A- 2 (cont'd)
MODE
FUEL
PTX CATALYST
MASS FLOW AIR
(kg/hr)
MASS FLOW EXH.
(kg/hr)
MASS FLOW EXH.
(m3/min. , act. )
EXH. TEMP.
(deg. C)
EXH. 0? CONG.
(%)
NO CONG.
(ppm, corr.)
N02 CONG.
(ppm, corr.)
NO CONG.
(ppm, corr.)
269
CATERPILLAR 3208 PARTICULATE CONTROL STUDY MEAN
ENGLEI1ARD PTX CATALYST EVALUATION
RATFD SPEED TEST SUMMARY C.V.
EPA MODE 9
NO.
W/0
738
1%
768
1%
34.1
1%
642
0%
8.0
1%
904
3%
7
71%
897
3%
2
W
803
0%
834
0%
38.3
1%
666
1%
8.5
4%
1061
2%
16
69%
1045
2%
EPA MODE 10
NO.
W/0
795
0%
817
0%
29.0
0%
475
1%
12.1
1%
627
1%
13
8%
614
1%
2
W
111
0%
799
0%
29.7
0%
491
0%
11.9
1%
657
3%
23
13%
634
3%
EPA MODE 11
NO. 2
W/0
754
2%
769
2%
24.3
2%
377
1%
14.0
0%
432
5%
13
15%
419
6%
W
780
1%
795
1%
25.1
1%
374
1%
14.7
0%
352
1%
48
6%
304
2%
-------
TABLE A~2 (cont'd) 270
CATERPILLAR 3208 PARTICULATE CONTROL STUDY MEAN
ENGLEHARD PTX CATALYST EVALUATION
RATED SPEED TEST SUMMARY C.V.
MODE
FUEL
PTX CATALYST
NO 2 CONC.
(ppm, act.)
HC CONC.
(ppm C)
BSNO
(g/kw-hr)
BSNO 2
(g/kw-hr)
BSNO
(g/kw-hr)
BSHC
(g/kw-hr)
EXH. OPACITY
EPA MODE 9
NO.
W/0
8
75%
113
13%
10.2
3%
0.1
100%
6.6
3%
0.38
13%
8.6
1%
2
W
16
69%
25
20%
13.0
2%
0.2
50%
8.3
1%
0.08
25%
8.8
6%
EPA MODE 10
NO.
W/0
14
14%
320
0%
12.4
1%
0.3
0%
7.9
1%
1.61
0%
3.8
3%
2
W
26
12%
40
0%
12.3
2%
0.4
25%
7.7
3%
0.21
0%
4.4
2%
EPA MODE 11
NO. 2
W/0
14
21%
310
3%
13.8
7%
0.4
25%
8.7
7%
3.14
3%
4.9
2%
W
56
7%
58
5%
13.8
1%
1.9
5%
7.8
1%
0.68
6%
4.8
0%
-------
TABLE A-2 (cont'd) 271
CATERPILLAR 3208 PARTICULATE CONTROL STUDY MEAN
ENGLEHARD PTX CATALYST EVALUATION
RATED SPEED TEST SUMMARY C.V.
MODE EPA MODE 9
FUEL
PTX CATALYST
NO.
W/0
2
W
EPA MODE 10
NO.
W/0
2
W
EPA MODE 11
NO. 2
W/0
W
NO @ FILTER FACE
(ppm)
TEMP. @ FILTER FACE
(deg* C)
TPM CONG.
(mg/m , act.)
TPM CQNC.
(mg/m , std.)
BSTPM
(g/kw-hr)
TPM E.F.*
(g/D
SOLID3CONC.
(mg/m , std.)
SOF CQNC.
(mg/m , act.)
SOF CQjNC.
(mg/m , std.)
0.5
80%
60
2%
11.6
3%
187
5%
1.04
5%
2.8
7%
170
0.4
25%
7
29%
1.1
64%
52
2%
25.3
11%
397
12%
2.46
12%
6.3
11%
285
0.1
100%
2
50%
1.0
10%
50
0%
5.8
5%
92
2%
0.84
2%
2.0
5%
65
1.4
0%
23
4%
1.7
12%
51
0%
23.4
5%
368
4%
3.27
4%
8.0
4%
206
0.2
100%
4
50%
0.9
22%
49
2%
8.5
5%
138
5%
2.32
5%
4.2
5%
78
3.6
3%
58
5%
3.9
8%
47
2%
15.5
3%
235
2%
4.15
2%
7.4
3%
84
0.6
33%
10
30%
*E.F. = Emission Factor
-------
TABLE A-2 (cont'd) 272
CATERPILLAR 3208 PARTICULATE CONTROL STUDY MEAN
ENGLEHARD PTX CATALYST EVALUATION
RATED SPEED TEST SUMMARY C.V.
MODE
FUEL
PTX CATALYST
BSSOF
(g/kw-hr)
SOF E.F.
(g/D
SO^ CQ^NC.
(mg/m , act.)
SOT CQNC.
,4,3
(mg/m , std.)
BSSO~
(g/kw-hr)
SO" E.F.
(g/1)
CONV. FUEL
s -> soT
f/\ 4
U)
DEVICE OPER.
TEMP.
(deg. C)
DEVICE SPACE
VELOCITY
(device vol. /sec.)
EPA MODE 9
NO.
W/0
0.03
33%
0.14
29%
0.6
83%
10.0
82%
0.06
67%
0.15
80%
2.80
82%
—
2
W
0.01
100%
0.03
67%
7.0
24%
110
24%
0.68
24%
1.74
26%
32.2
26%
695
1%
450
1%
EPA MODE 10
NO.
W/0
0.21
10%
0.53
4%
0.3
33%
4.5
33%
0.04
50%
0.91
55%
1.69
54%
—
2
W
0.03
67%
0.09
56%
10.1
17%
158
17%
1.41
17%
3.44
17%
63.7
17%
507
0%
349
0%
EPA MODE 11
NO. 2
W/0
0.98
6%
1.58
4%
0.1
100%
2.0
50%
0.03
67%
0.06
50%
1.09
43%
—
W
0.17
29%
0.27
30%
9.3
4%
141
6%
2.48
6%
4.40
6%
81.5
6%
391
1%
295
1%
-------
273
.,.„, ,, . _ CftTERPILlAR 3208 PARTICULATE CONTROL STUDY MEAN
TABLE A- 3
INTERMEDIATE SPEED TEST SUMMARY C.V.
PARAMETER
AMBIENT' TbMPERATURE
(deg C)
BAROMETRIC PRESSURE
(kPa)
SPECIFIC HUMIDITY
(g H20/kg AIR)
SPEED
(rpm)
LOAD
(N-m)
CORRECTED POWER
(kW)
CORRECTED BMEP
(kPa)
BSFC
(kgAw-hr)
AIR-FUEL RATIO
(kg AIRAg FUEL)
MASS FIOW AIR
(kg/hr)
MASS FIOW FUEL
(kg/hr)
MASS FIOW EXHAUST
(kg/hr)
TOL. FIOW EXH. (ACTUAL)
(cubic msters/min)
DENSITY EXHAUST
(kg/cubic ireter)
MDL WT. EXHAUST
( kg/no le)
EXHAUST TEMP (MANIFOLD)
(deg C)
EXHAUST VISCOSITY
(kg/m-sec)
EXH. OXYGEN CONC.
(percent)
NOx CONC. (corr)
(ppm)
ND2 CONC. (oorr)
(ppm)
ND CONC. (corr)
(ppm)
ND2 CONC. (actual)
(ppm)
HC ODNC.
(ppm C)
EPA MODE 3
BASELINE PORT TRAP
ND CONTROL CATALYST (UNCAT. )
31.7
0%
99.6
0%
5.94
0%
1680.
0%
160.
0%
28.14
0%
193.0
0%
.2874
0%
63.13
0%
510.6
0%
8.089
0%
518.7
0%
12.96
0%
.667
0%
28.75
0%
242.9
0%
.268-004
0%
15.72
0%
341.3
0%
38.7
18%
302.6
2%
43.2
38%
212.
10%
29.4
0%
97.6
0%
7.47
0%
1680.
0%
160.
0%
28.62
0%
196.4
0%
.2910
0*
61.23
0%
509.9
0%
8.328
0%
518.2
0%
14.04
0%
.615
0%
28.72
0%
275.0
0%
283-004
0%
15.52
0%
399.6
0%
22.7
6%
376.9
0%
24.2
6*
195.
0%
23.9
0%
99.7
0%
2.94
0%
*1680.
0%
160.
0%
27.54
0%
188.5
0%
.3007
0*
62.67
0%
519.1
0%
8.283
0%
527.4
0%
13.56
0%
.648
0%
28.80
0%
259.5
0%
.273-004
0%
15.78
0%
346.8
0%
37.9
6%
30S.9
0%
43.0
6*
230.
5*
EPA NODE 4
BASELINE PORT TRAP
tO CONTROL CATALYST (UNCAT.
29.4
0*
99.9
0%
8.21
0%
1680.
0%
320.
0%
56.00
0%
384.3
0%
.2307
0%
39.66
0%
512.5
0%
12.921
0%
525.4
0%
15.90
0%
.551
0%
28.72
0%
353.1
1%
.301-004
0%
12.73
0%
681.5
1%
56.0
5%
625.5
0%
58.7
5%
168.
2%
31.1
0%
98.2
0%
6.01
0%
1680.
0%
320.
0%
57.02
0%
390.9
0%
.2293
0%
38.87
0%
508.1
0%
13.072
0%
521.2
0%
16.75
0%
.518
0%
28.76
0%
381.9
0%
318-004
0%
12.64
0%
700.6
0%
40.0
6%
660.6
0%
43.8
6%
86.
0%
24.
0%
100.5
0%
2.12
0%
1680.
0%
320.
0%
54.67
0%
373.9
0%
.2427
0%
39.24
0%
520.6
0%
13.269
0%
533.9
0%
16.39
0%
.543
0%
28.82
0%
368.2
0%
.305-004
0%
12.82
0%
617.9
0%
24.1
6%
593.8
Ot
27.5
6*
US.
0*
EPA MODE 5
BASELINE PORT TRAP
NO CONTROL CATALYST (UNCAT. )
32.2
0%
99.2
0%
8.54
0%
1680.
0%
480.
0%
85.45
0%
584.8
0%
.2167
0%
26.71
0%
494.6
0%
18.519
0%
513.1
0%
18.89
0%
.453
0%
28.72
0%
483.2
0%
.337-004
0%
8.99
0%
986.0
0%
36.4
3%
949.7
0%
37.7
3%
85.
1%
23.1
1%
98.5
0%
3.68
20%
1680.
01
480.
0%
83.46
0%
571.8
0%
.2181
0%
28.03
0%
510.0
0%
18.199
0%
528.2
0%
19.44
0%
.453
0%
28.80
0%
480.0
0%
352-004
0%
9.65
0%
1179.1
1%
100.3
1%
1078.8
1%
110.0
2*
36.
0%
30.0
0%
100.3
0%
3.12
0%
1680.
0%
480.
0%
83.33
0%
570.0
0%
.2227
0%
27.43
0%
508.9
0%
18.553
0%
527.5
0%
19.42
0%
.453
0%
28.81
0%
494.8
0%
.340-004
0%
9.43
0%
897.0
0*
14.0
9%
883.0
0%
15.7
9%
58.
1*
-------
274
TUi||. CATERPILLAR 3208 PARTI CULATE CONTROL STUDY MFAN
(cont'd) INTERMEDIATE SPEED TEST SUMMARY C.V.
PARAMETER'
bs NOx
(gAW-hr)
bs N32
(gAW-hr)
bs NO
(gAW-hr)
bs HC
(gAW-hr)
EXHAUST OPACITY
(percent)
ND2 @ FILTER FACE
(ppm)
TEMP @ FILTER FACE
(
MASS FLOW SOLID
(g/hr)
CONC. SOLID
(rog/m**3 act)
CONC. SOLID
(nq/m**3 std)
bs SOLID
(g/kw-hr)
SOLID E.F.
(g/l)
MASS FLOW SOF
(g/hr)
CONC. SOF
(nq/m**3 act)
CONC. SOF
(irrj/m**3 std)
bs SOF
(q/kW-hr)
SOF E.F.
(q/1)
EPA MODE 3
BASELINE PORT TRAP
NO CONTROL CATALYST (UNCAT. )
10.24
0%
1.16
18%
5.92
2%
1.917
10%
.3
0%
2.81
17%
37.2
0%
29.461
3%
37.882
2%
66.419
2%
1.047
3%
2.519
2%
2.357
2%
3.031
1%
5.314
2%
.084
2%
.202
1%
26.230
3%
33.727
3%
59.134
3%
.932
3%
2.243
2%
12.02
0*
.68
6%
7.40
0%
1.766
0%
.3
0%
1.65
6%
40.0
0%
22.617
6%
26.844
6%
49.990
6%
.790
6%
1.878
6%
3.979
13%
4.723
13%
8.795
13%
.139
13%
.330
13%
17.431
5%
20.689
5%
38.528
5%
.609
5%
1.418
5*
10.73
0*
1.18
6%
6.26
0%
2.152
5%
.3
23%
3.00
5%
31.7
1%
28.617
6%
35.172
6%
63.638
6%
1.039
6%
2.390
'6%
1.528
4%
1.877
4%
3.397
4%
.055
4%
.128
4%
26.762
6%
32.892
7%
59.512
6%
.972
6%
2.2.15
7*
EPA MODE 4
BASELINE PORT TRAP
NO CONTROL CATALYST (UNCAT.)
10.39
1%
.85
5%
6.22
1%
.771
2%
.9
6%
3.81
5%
45.6
0%
26.974
6%
28.272
5%
60.161
6%
.482
6%
1.444
6%
3.875
12%
4.061
12%
8.643
12%
.069
12%
.207
12%
21.106
6%
22.121
5%
47.073
6%
.377
<>%
1.130
6*
10.56
0%
.60
6%
6.50
0%
.390
0%
.5
0%
2.99
5%
47.2
0*
26.902
2%
26.762
2%
59.559
2%
.472
2%
1.423
2%
4.228
4%
4.206
4%
9.359
4%
.074
4%
.224
4%
6.576
5%
6.542
5%
14.559
5%
.115
5%
.348
rj%
9.71
0%
.38
6%
6.08
0%
.591
0%
.1
0%
1.85
6%
37.2
2%
9.448
19%
9.618
20%
20.940
19%
.173
19%
.493
20%
1.223
5%
1.244
5%
2.710
5%
.022
5%
.064
5%
7.512
24%
7.648
25%
16.647
24%
.137
24%
.392
25%
EPA MODE 5
BASELINE PORT TRAP
NO CONTROL CATALYST (UNCAT.)
9.69
0%
.36
3%
6.08
0%
.253
1%
1.9
0%
2.50
3%
50.4
0%
20.594
2%
18.171
2%
46.689
2%
.241
2%
.769
2%
14.680
3%
12.953
3%
33.281
3%
.172
3%
.548
3%
3.235
5%
2.855
5%
7.335
5%
.038
5%
.121
5%
12.26
1%
1.04
1%
7.31
1%
.113
0%
1.4
0%
7.39
3%
41.1
0%
59.186
1%
50.751
1%
129.847
!%•
.709
1%
2.249
1%
9.331
13%
8.001
12%
20.471
13%
.112
12%
.355
12%
2.021
25%
1.733
25%
4.433
25%
.024
25%
.077
25%
9.15
0%
.14
9%
5.88
0%
.178
1%
.1
0%
1.07
9%
45.9
1%
1.269
17%
1.090
18%
2.841
17%
.015
17%
.047
18%
.453
26%
.389
26%
1.014
26%
.005
25%
.017
26i
.152
8%
.130
8%
.340
8%
.002
8%
.006
8!.
-------
275
TA»)LE A-3 CATERPILLAR 3208 PARTICUIATE CONTROL STUDY MEAN
(cont'd) INTERMEDIATE SPEED TEST SUMMARY C.V.
PARAMETER
MASS FIOW S04
(g/hr>
CDNC. SOt
(ng/m**3 act)
UONC. BJ4
(mg/T.**3 std)
bs SO4
(gAw-hr)
SO4 E.F.
(g/l)
OONV. FUEL S to S04
(percent)
DEVICE OPERATING TEMP.
(deg C)
DEVICE SPACE VELOCITY
(device volumes/second)
DEVICE RESIDENCE TIME
(msec)
TRAP PRESSURE DROP
(kPa)
TRAP LOADING PARAMETER
(dinensionless)
EPA MDDE 3
BASELINE PORT TRAP
ND CONTROL CATALYST (UNCAT. )
.874
4%
1.124
4%
i.971
4%
.031
4%
.075
4%
1.33
4%
242.9
0%
.0
0%
.00
0%
.000
0%
0.
0%
1.207
4%
1.432
4%
2.667
4%
.042
4%
.100
4%
1.79
4%
288.9
0%
404.0
0%
2.48
0%
.000
0%
0.
0%
.327
18%
.403
19%
.728
18%
.012
18%
.027
19%
.49
19%
258.5
0%
42.0
0%
23.83
0%
6.996
8%
1694.
7%
EPA MODE 4
BASELINE PORT TRAP
ND CONTROL CATALYST (UNCAT. )
1.993
6%
2.090
6%
4". 445
6%
.036
6%
.107
6%
1.90
6%
353.1
1%
.0
0%
.00
0%
.000
0%
0.
0%
16.099
1%
16.015
1%
35.640
1%
.282
1%
.852
1%
15.19
1%
406.9
0%
488.1
0%
2.05
0%
.000
0%
0.
0%
.714
6%
.726
5%
1.582
6%
.013
5%
.037
6%
.66
6%
365.9
0%
50.6
0%
19.75
0%
8.616
9%
1548.
8%
EPA MODE 5
BASELINE PORT TRAP
ND CONTROL CATALYST (UNCAT. )
2.679
5%
2.364
5%
6.073
5%
.031
5%
.100
5%
1.78
5%
483.2
0%
.0
0%
.00
0%
.000
0%
0.
0%
47.835
1%
41.017
1%
104.943
1%
.573
1%
1.818
1%
32.43
1%
530.4
0%
581.9
0%
1.72
0%
.000
0%
0.
0%
.664
17%
.570
18%
1.486
17%
.008
17%
.025
18%
• .44
18%
489.3
0%
59.8
0%
16.72
0%
8.565
11%
1170.
11%
-------
276
TMll.F A- I CATERPILLAR 3208 PAKTICULATK OONTHWL STUDY MEAN
(cont'd) RATED SPEED TEST SUMMARY C.V.
PARAMETER
AMBIFOT TEMPERATURE
(deg C)
BAROMLTRIC PRESSURE
(kPa)
SPECIFIC HUMIDITY
(g H20/kg AIR)
SPEED
(rpm)
IDAD
(N-m)
CDRRECrED POKER
(kW)
CORRECTED aMEP
CkPa)
BS^C
(kg/kW-hr)
AIR-FUEL RATIO
(kg AIR/kg FUEL)
MASS FIDK AIR
(kg/hr)
MASS FIDW FUEL
(kg/hr)
MASS FIDW EXHAUST
(kg/hr)
VOL. FLOW EXH. (ACTUAL)
(cubic rreters/min)
DEMSITY EXHAUST
(kg/cubic meter)
MDL KT. EXHAUST
( kg/no le)
EXHAUST TEMP (MANIFOLD)
(deg C)
EXHAUST VISCOSITY
( kg/m-sec )
EXH. OXYGEN CONC.
(percent)
NOx CONC. (corr)
(ppm)
ND2 CONC. (corr)
(ppm)
NO CONC. (corr)
(ppm)
N02 CONC. (actual)
(ppm)
HC CONC.
(prm C)
EPA (ODE 9
BASELINE PORT TRAP
NO CONTROL CATALYST (UNCAT. )
31.7
0%
98.5
0%
11.56
0%
2800.
0%
399.
0%
119.29
0%
490.8
0%
.2516
0%
25.34
0%
760.4
0%
30.009
0%
790.4
0%
33.38
0%
.395
0%
28.68
0%
587.2
0%
.367-004
0%
8.31
0%
768.7
1%
23.2
13%
745.5
1%
21.0
33*
80.
5?.
26.1
2%
100.1
0%
3.76
19%
2800.
0%
399.
0%
114.59
0%
470.9
0%
.2718
0%
25.97
0%
808.9
0%
31.147
0%
840.0
0%
36.52
0%
.383
0%
28.81
0%
630.9
0%
.397-004
0%
8.79
0%
857.6
1%
55.4
1%
802.2
1%
60.8
0%
35.
Or,
34.4
0%
99.0
0%
2.90
0%
2800.
0%
399.
0%
118.02
0%
484.5
0%
.2623
0%
24.31
0%
752.5
0%
30.955
0%
783.5
0%
34.44
•0%
.379
0%
28.82
0%
631.8
0%
.379-004
0%
8.02
1%
681.3
0%
9.6
28%
671.7
1%
10.7
28%
34.
2*
EPA MODE 10
BASELINE PORT TRAP
NO CONTROL CATALYST ( UNCAT. )
37.8
0%
101.2
0%
11.73
0%
2800.
0%
266.
0%
78.47
0%
323.1
0%
.2816
0%
34.36
0%
759.1
0%
22.096
0%
781.2
0%
28.02
0%
.465
0%
28.66
0%
477.2
0%
.338-004
0%
11.44
0%
564.0
0%
16.1
10%
548.0
03
16.2
10%
271.
4t
27.7
2*
98.7
0%
6.44
3%
2800.
0%
266.
0%
78.05
0%
320.9
0%
.2895
0%
34.52
0%
780.0
0%
22.595
0%
802.6
0%
30.23
0%
.443
0%
28.75
0%
497.7
0%
.358-004
0%
11.62
0%
592.7
0%
19.0
3%
573.7
0%
20.3
3%
125.
0%
33.3
0%
98.0
0%
4.11
0%
2800.
0%
266.
0%
79.34
0%
326.1
0%
.2879
0%
32.94
0%
752.5
0%
22.842
0%
775.3
0%
29.46
1%
.439
1%
28.79
0%
500.4
1%
.344-004
0%
11.26
0%
527.7
1%
7.4
6%
520.3
1%
8.3
61
10=;.
0»
EPA ("DDE 11
BASELINE PORT TRAP
NO CONTROL CATALYST (UNCAT.)
36.7
0%
97.6
0%
12.02
0%
2800.
0%
133.
0%
40.63
0%
167.1
0%
.3703
0%
48.81
0%
734.4
0%
15.047
0%
749.5
0%
23.67
0%
.528
0%
28.65
0%
363.7
0%
.305-004
0%
14.09
0%
328.0
1%
21.3
4%
306.7
1%
21.5
4%
316.
4*
31.1
0%
98.3
0%
5.35
0%
2800.
0%
133,
0%
39.39
0%
162.1
0%
.3988
0%
48.79
0%
766.6
0%
15.711
0%
782.3
0%
25.89
0%
.504
0%
28.76
0%
401.7
0%
.331-004
0%
14.28
0%
349.2
0%
11.6
0%
337.6
0%
13.0
0%
180.
0%
27.8
0%
99.9
0%
3.52
0%
2800.
0%
133.
0%
38.90
2%
157.8
0%
.4113
5%
49.32
3%
788.0
0%
15.992
3%
804.0
0%
25.71
2%
.522
2%
28.79
0%
390.4
4%
.300-004
6%
14.40
1%
321.2
3%
4.4
0%
316.8
3%
5.0
0%
206.
0*
-------
277
TAJil.K A-D CATERPILIAR 3208 PARTICUIATE COWITOL STUDY MEAN
(cont'd) RATED SPEED TEST SUMMARY C.V.
PARAMETER
bs M3x
(gAW-hr)
bs N02
(gAW-hr)
bs (O
{ gAW-hr)
bs HC
(gAW-hr)
EXHAUST OPACITY
(percent)
ND2 8 FILTER FACE
(ppn)
TEMP @ FILTER FACE
(deg C)
MASS FI£« TPM
(g/hr)
CONC. TPM
(mg/m**3 act)
CONC. TPM
(ng/m**3 std)
bs TPM
(gAW-hr)
TPM E.F.
(g/D
MASS FIDW SOLID
(g/hr)
erne. SOLID
-------
278
TABLE A- 3 CATERPILLAR 3208 PARTICUIATE CONTROL STODY MEAN
(cont'd) RATED SPEED TEST SUMMARY C.V.
PARAMETER
MASS FIDW SO4
(g/hr)
CONC. S04
(rag/is**3 act)
CONC. SO4
(n*3/m**3 std)
bs S04
(gAW-hr)
S04 E.F.
(g/D
COW. FUEL S to SO4
(percent)
DEVICE OPERATING TEMP.
(deg C)
DEVICE SPACE VELOCITY
(device volumes/second)
DEVICE RESIDENCE TIME
(msec)
TRAP PRESSURE DROP
(kPa)
TRAP LOADING PARAMETER
(dinensionless )
EPA M3DE 9
BASELINE PORT TRAP
ND CONTROL CATALYST (UNCAT. )
4.693
23%
2.343
23%
6.848
23%
.039
23%
.108
23%
1.93
23%
587.2
0%
.0
0%
.00
0%
.000
0%
0.
0%
45.257
3%
20.652
3%
63.426
3%
.395
3%
1.005
3%
17.93
3%
690.9
0%
1092.9
0%
.92
0%
.000
0%
0.
0%
1.418
2%
.686
?*
2.109
2%
.012
2%
.032
3%
.57
3%
627.9
0%
106.3
0%
9.40
0%
12.902
1%
889.
1%
EPA MODE 10
BASELINE PORT TRAP
NO CONTROL CATALYST (UNCAT. )
2.567
15%
1.528
15%
3.893
15%
.033
15%
.080
15%
1.43
15%
• 477. 2
0%
.0
0%
.00
0%
.000
0%
0.
0%
34.506
8%
19.025
8%
49.827
9%
.442
8%
1.056
8%
18.84
8%
545.5
0%
900.8
0%
1.11
0%
.000
0%
0.
0%
.700
10%
.397
11%
1.041
10%
.009
10%
.021
11%
.38
11%
498.1
1%
91.1
1%
10.98
1%
20.617
6%
1825.
4%
EPA MODE 11
BASELINE PORT TRAP
NO CONTOOL CATALYST (UNCAT. )
1.538
11%
1.083
11%
2.342
11%
.038
11%
.071
11%
1.26
11%
363.7
0%
.0
0%
.00
0%
.000
0%
0.
0%
12.788
2%
8.233
2%
18.873
2%
.325
2%
.563
2%
10.04
2%
445.0
0%
773.1
0%
1.29
0%
.000
0%
0.
0%
.918
10%
.597
12%
1.342
10%
.024
10%
.040
13%
.71
13%
344.7
17%
74.2
9%
13.56
10%
21.726
37%
2848.
53%
-------
Key
790
PORT
CATALYST
Total Particulate Matter (TPM)
Solid Fraction (SOL) of TPM
Soluble Organic Fraction (SOF) of TPM
Sulfate Fraction (SO4) of TPM
1.047
BASELINE ENGINE (NO CONTROL)
1.039
CORNING
TRAP
Fig. Al The Effect of Aftertreatment Devices on Caterpillar 3208 Particulate Emissions:
Johnson-Matthey Close-Coupled Exhaust Port Catalysts and Corning Ceramic Traps,
EPA Mode 3 (16.80' rpm and 25% load = 160 N-m)
(numbers indicate bsTPM in g/kw-hr)
-------
o
oo
CM
472
PORT
CATALYST
Key: Q
Total Particulate Matter (TPM)
Solid Fraction (SOL) of TPM
Soluble Organic Fraction (SOF) of TPM
Sulfate Fraction (SO4) of TPM
Control Device Indication
.482
BASELINE ENGINE (NO CONTROL)
.173
CORNING
TRAP
Fig. A2 The Effect of Aftertreatment Devices on Caterpillar 3208 Particulate Emissions;
Johnson-Matthey Close-Coupled Exhaust Port Catalysts and Corning Ceramic Traps
EPA Mode 4 (1680 rpm and 50% load = 320 N-m)
(numbers indicate bsTPM in g/kW-hr)
-------
oo
CSJ
.709
PORT
CATALYST
Key: Q Total Particulate Matter (TPM)
Solid Fraction (SOL) of TPM
Soluble Organic Fraction (SOF) of TPM
Sulfate Fraction (S04) of TPM
Control Device Indication
.241
BASELINE ENGINE (NO CONTROL)
.015
CORNING
TRAP
Fig. A3 The Effect of Aftertreatment Devices on Caterpillar 3208 Particulate Emissions:
Johnson-Matthey Close-Coupled Exhaust Port Catalysts and Corning Ceramic Traps,
EPA Mode 5 (1680 rpm and 75% load = 480 N-m)
(numbers indicate bsTPM in g/kW-hr)
-------
CNI
oo
OM
1.192
PORT
CATALYST
Key: Q
Total Particulate Matter (TPM)
Solid Fraction (SOL) of TPM
Soluble Organic Fraction (SOF) of TPM
Sulfate Fraction (SO4) of TPM
- Control Device Indication
.795
BASELINE ENGINE (NO CONTROL)
.022
CORNING
TRAP
Fig. A4 The Effect of Aftertreatment Devices on Caterpillar 3208 Particulate Emissions
Johnson-Matthey Close-Coupled Exhaust Port Catalysts and Corning Ceramic Traps,
EPA Mode 9 (2800 rpm and 75% load = 399 N-m)
(numbers indicate bsTPM in g/kW-hr)
-------
oo
CN
1.256
PORT
CATALYST
Key
Total Particulate Matter (TPM)
Solid Fraction (SOL) of TPM
Soluble Organic Fraction (SOF) of TPM
Sulfate Fraction (S04) of TPM
Control Device Indication
.792
BASELINE ENGINE (NO CONTROL)
.022
CORNING
TRAP
Fig. A5 The Effect of Aftertreatment Devices on Caterpillar 3208 Particulate Emissions:
Johnson-Matthey Close-Coupled Exhaust Port Catalysts and Corning Ceramic Traps.
EPA Mode 10 (2800 rpm and 50% load = 266 N-m)
(numbers indicate bsTPM in g/kW-hr)
-------
oo
(N
2.079
PORT
CATALYST
Key: Q
Total Particulate Matter (TPM)
Solid Fraction (SOL) of TPM
Soluble Organic Fraction (SOF) of TPM
Sulfate Fraction (S04) of TPM
Control Device Indication
1.864
BASELINE ENGINE (NO CONTROL)
139
CORNING
TRAP
Fig. A6 The Effect of Aftertreatment Devices on Caterpillar 3208 Particulate Emissions
Johnson-Matthey Close-Coupled Exhaust Port Catalysts and Corning Ceramic Traps
EPA Mode 11 (2800 rpm and 25% load = 133 N-m)
(numbers indicate bsTPM in g/kW-hr)
-------
285
APPENDIX B
AMES BIOASSAY VARIABILITY AND DATA QUANTIFICATION
Sensitivity Analysis Experiments
Preliminary research in this laboratory has shown that a number of
system variables can influence the Ames bioassay dose-response curve of
a sample. The observed dose-response curve, as depicted in Figure Bl.l,
is influenced by many variables in addition to the sample concentration.
o
Although a nominal value of 10 cells is used on each plate, this value
is influenced by harvesting time, incubation temperature, medium
variability, and strain idiosyncrasies. Small variations in nutrient
levels may cause wide variability in bacteria capable of reverting.
Space restrictions on the Petri plate limit the number of colonies;
resolution of automatic colony counters is well below the saturation
point of the agar surface. For meaningful laboratory or sample
comparisons, these system variables must be considered and controlled.
A series of pilot experiments was performed to enable us to
construct a model delineating the interactions of Ames test variables
including initial cell concentration, histidine concentration, and
presence of trace histidine from overnight growth broth. Although these
experiments were performed without rigid statistical analysis and offer
no proof of any theories set forth in this section, the data are still
valuable, as they comprise a solid base from which a model of the Ames
system can be obtained.
-------
NUMBER OF S. fyphimurium
> 11
3 M
« c
w i
9
"* W
X
VI
«
0
* .,
3 «
a
< ~
Q -,
Z. •
o a
tr •*
•o
M
IT
a
9
«
3
n
O
2
n
O
2
N3
00
ON
-------
287
The Ames genotypic strain TA98, a frame-shift mutagen indicator,
was used in these experiments. The standard Ames plate incorporation
assay (50) as modified by Belser et al. (51) was followed. A
direct-acting, known frameshift mutagen, 2-nitrofluorene (2-NF) was used
as the test compound, without S-9 mammalian microsomal activation. The
spontaneous mutation rate was estimated without addition of 2-NF. For
determining the influence of residual histidine in the overnight growth
broth, cells were washed and resuspended in sterile 0.001 M
tris-(hydroxy methyl) amino methane buffer (Tris) at pH 7. This was
accomplished by centrifuging the growth broth, discarding the supernate,
and resuspending the cells in Tris. This procedure was repeated three
times.
Figure B1.2 shows the effects of cell washing and initial cell and
histidine concentration on the number of spontaneous revertants per
plate. Unless exogenous histidine was added washed cells did not
? f\
spontaneously revert with cell inocula between 10 and 10 cells. With
7 8
10 cells, one spontaneous revertant formed, and with 10 cells, two
spontaneous revertants formed. Washed cells responded similarly to
unwashed cells when exogenous histidine was added.
Figure B1.2 also shows the dependence of spontaneous
revertants/plate on cell density and exogenous histidine concentration.
2
Histidine remaining in the overnight growth broth allowed 10 but not
3
10 cells to grow into countable colonies, and 0.5mM histidine allowed
3
growth of 10 cells into countable colonies. Due to colony counter
4
limitations, the number of cells counted with 10 cells/plate was
deceptive. At all three histidine concentrations, there were visible
colonies too small to be detected by the colony counter. The histidine
-------
288
O
CL
U>
to
o>
10'
LU
< 10*
0.
f
cc
IU
CO
r>
O
UJ
10'
EFFECT OF WASHING CELLS AND HISTIDINE
ON SPONTANEOUS MUTATION RATE
CELL TYPE
WASHED
UNWASHED
WASHED
UNWASHED
WASHED
UNWASHED
HISTIDINE
0-OmM
O.OmM
0.5mM
O.BmM
1.0m M
10m M
If
10* 10' 104 10' 10* 10'
INITIAL INOCULUM OF TA 98 (/PLATE)
10'
Fig. B1.2
SYSTEM VARIABLES INFLUENCING AMES DATA
-------
10V
10'
LLJ
<
CO
h-
h-
cn
HI
a
-------
10
290
UJ
DC
UJ
UJ
CL
CO
O)
.OmM HISTIDINE
ADDED TO
TOP AGAR
10J
SYMBOL
,5mM HISTIDINE
ADDED TO
TOP AGAR
1.0mM HISTIDINE
ADDED TO
TOP AGAR
1.0 1.8 3.2 5.6 10.0 1.0 1.8 3.2 5.6 10.0 1.0 1.8 3.2 5.6 10.0
2 - NITROFLUORENE mg/PLATE
Fig
Bi.4 Influence of histidine and cell concentration.
-------
291
concentration limited the number of spontaneous revertants between 10
g
and 10 cells/inoculum. Varying the histidine concentration affected
the number of revertants within a cell concentration, but the number of
revertants remained constant between concentrations. This supports the
thesis that a given histidine level allows a constant fraction of cells
4
to grow and be counted as spontaneous revertants above 10
cells/inoculum.
Figure B1.3 shows the effects of cell washing, histidine
concentration, and cell concentration on the induced revertant rate of
TA98 with 3ug/plate of 2-NF. Washed cells showed no revertants at any
concentration, indicated that the cellular DNA must be replicating (with
the histidine) before 2-NF can be effective. The influence of washing
and histidine concentration is more pronounced with induced mutations
than with spontaneous mutations.
Figure B1.4 shows the effect of histidine and initial cell inoculum
concentration of the dose-response curve of TA98 to 2-NF. Increasing
histidine concentration changes the y-intercept of the curves but has
little effect upon their slopes. Colony counter limitations make the
slopes of the l.OmM histidine curves appear lower at high 2-NF
concentrations. From 10 to 10 cells/inoculum, dose curves showed a
higher y-intercept for lower cell concentration. This indicates that at
lower inoculum concentration there was more histidine and sample per
bacterium, allowing a greater opportunity for reversion.
The model derived from these results is not yet complete, but can
be described by:
Log (Revertants) = a + bl x (Wash) + b2 x (Hist .5) +
b3 x (Hist 1.) + b4 x (102 cells) +
-------
292
b5 x (103 cells) + b6 x (104 cells) +
b7 x (105 cells) + b8 x (10 cells) +
b9 x (10? cells) + blO x (10 cells) + error.
This model utilizes the variables Wash, Hist .5, Hist 1., and seven
cell concentrations as indicator variables with values of 1 or 0. The
use of indicators allow one to quantify the effect of within-variable
changes on the regression sum of squares. For instance, if the cells
4
were not washed and l.OmM histidine were added to an inoculum of 10
cells, the model reduces to
Log (Revertants) = a + bl x (Wash) + (Hist 1.) +
4
b6 x (10 cells) + error
because all the variables were zero, except those specified.
Applying this model to the data of Figure B1.3 shows that washing
has little effect on the spontaneous revertant rate of TA98 with
addition of exogenous histidine. Raising the histidine concentration
had little effect on the histidine regression coefficient; likewise 10
Q
to 10 cells/inoculum showed insignificant regression coefficients, but
3 4
significant coefficients resulted from use of 10 or 10 cells/inoculum,
attributable to all the cells growing when histidine was added.
When the model was applied to induced revertant data, the results
were quite different. Washing had a greater influence on the level of
induced revertants than on the level of spontaneous revertants, and
doubling of histidine concentration significantly increased its
regression coefficient. The induced revertant data showed that all of
the cell concentrations had an equal and significant influence.
-------
293
This model enables us to conclude that:
1) Washing affects induced revertants more than spontaneous
revertants;
2) histidine concentration has a greater affect on induced
revertants than on spontaneous revertants; and
3) changes in cell inoculum concentration affect induced
revertants more than spontaneous revertants.
Quantification of Ames Data
Originally, the Ames test was developed as a qualitative model for
evaluating the mutagenic activity of both pure compounds and complex
mixtures. When the link between bacterial mutagenicity and mammalian
carcinogenicity was realized, the widespread use of the Ames test made
its results useful in the regulatory decision-making process. This
placed an increasing emphasis on quantifying Ames test data. Simple
positive or negative results were no longer sufficient.
A direct method of dose-response curve quantification is a simple
linear regression analysis for calculating the specific activity (SA) of
sample, as used by Hunter et al. (31) and Campbell et al. (81).
Analysis of the dose-response curve was done as follows:
1) Statistically determine the curve fit for the dose-response data by
least squares;
R = a + b (log1QC)
where R = Number of revertants/plate;
C = Dose concentration (jig), and
a and b are determined statistically
-------
294
2) the slope at any point is given by:
dR = 1 b,
dC JlnlO (C}
3) the average slope of the dose-response curve in Cartesian
coordinates is found from:
C2,
c - Cl - c2 -
where C. = minimum dose (rig) observed on linear portion of dose-response
curve and
C0 = maximum dose Oj^g) observed on linear portion of dose-response
curve; and
4) the value ^E is then the SA (specific activity) of the
dC bL
dose-response curve, in semi-log coordinates.
The brake-specific revertants are obtained by taking brake-specific SOF
(rate of SOF emissions per rate of work done by engine) and multiplying
this by SA :
bsREV = dR x bsSOF x 106/[4&x
dC < ;
-------
295
This linear regression method suffers from the inability to consistently
choose a linear portion of the dose curve, and some data are always
discarded.
Stead et al. (82) and Myers et al. (83) used non-linear regression
techniques to quantify Ames data. This method offers no bias in
choosing a linear portion of the dose curve, and little data are
discarded. However, when the functional relationship between variables
is not known, a non-linear analysis gives confusing and misleading
results. There is also no assurance that the calculated parameters are
the best estimates of the population parameters. For this reason, simple
linear regression methods should be preferred if certain statistical
assumptions can be met (84):
1) The independent variable, X (dosage for Ames data), is measured
without error;
2) For any given value of X, corresponding Y values are independent
and normally distributed;
3) Independent variable values are not autocorrelated;
4) The variance around the regression line is constant and independent
of the magnitude of X and Y;
5) There are no extraneous variables which influence the relationship
between X and Y; and ^
6) The expected value for the dependent variable, Y (revertants/plate
for Ames data), is described by the linear function:
Y = a + bX
-------
296
Unfortunately, none of these assumptions is rigorously met for data
obtained from the Ames test as implemented at MTU and elsewhere (85). A
transformation of the data, however, may enable some of these
assumptions to be met. Additionally, if a statistically sound method
for determining the beginning and ending of a dose-response curve's
mutagenic region is utilized, simple linear regression may be useful in
Ames data quantification.
A linear regression analysis provides a regression coefficient
which describes the "goodness of fit" of the observed data to a
calculated regression line. In addition, an analysis of variance table
provides sum of squares data known as the residual, or unexplained, sum
of squares:
degrees of sum of mean
Source of Variation freedom Squares square
Total
Regression
Residual
Within Groups
Deviation from Linearity
The residual (unexplained) sum of squares is divided into
c
within-group sum of squares (true experimental deviation of the data)
and deviation from linearity sum of squares (bias of the regression from
a simple linear model).
In addition to a non-transformed linear dose-revertant model, two
transformations of this model (y = a + bX) were examined for
within-group and deviation from linearity sum of squares as well as
-------
297
y-intercept and slope variability. The other transformations of this
model were the log dose-revertant transformation (colonies/plate
regressed on the common logarithm of concentration/plate) and the log
dose-log revertant transformation (common logarithm of colonies/plate
regressed on common logarithm of concentration/plate). Tables Bl.l thru
B1.4 of this appendix display the four aforementioned regression
statistics for every combination of three or more doses. The
dose-revertant model and its log dose-revertant, and log dose-log
revertant transformations are compared for data obtained with TA-98 and
2-NF assayed on July 8, 1981.
Tables Bl.l and B1.2 show that, given three or more doses, the
slope and y-intercept is most stable if the log dose-log revertant
transformation is used. Perhaps more importantly, Tables B1.3 and B1.4
show that the log dose-log revertant transformation offers the lowest
variation in both the sum of squares due to deviation from linearity,
and that due to random error in the data. Minimization of the deviation
from linearity sum of squares helps satisfy statistical assumption
number 4 pertaining to data subjected to linear regression analysis,
which asserts that the variance around the regression line should be
constant and independent of the magnitude of X and Y. Minimization of
the random error sum of squares gives a regression line that is less
sensitive to variation in choosing minimum and maximum doses on the
mutagenic portion of the dose-response curve. This eliminates much of
the bias in estimating the beginning of mutagenicity and the beginning
of toxicity; however, a sound statistical method should be employed to
choose these points consistently.
-------
Table Bl.l Comparison of the slopes from the three Linear models
describing Ames data collected on July 8, 1982.
298
LOG DOSE - LOG REVERTANT MODEL
.56
1.0
MAX. DOSE (yg)
1.8 3.2 5.6
10.
18.
10.
5.6
3.2
MIN. 1.8
DOSE 1.0
(lag) .56
.32
.18
9088
.6235
.8639 .7614
.8464 .7839
LOG DOSE - REVERTANT MODEL
.56
1.0
MAX. DOSE
1.8 3.2
5.6
10.
18.
DOSE - REVERTANT MODEL
.56
1.0
MAX. DOSE
1.8 3.2
5.6
10.
18.
32.
.9272
.8102
.8413
.8347
.9261
.8595
.8033
.8276
.8260
.6979
.8567
.8359
.8025
.8210
.8212
.4822
.5430
.6990
.7299
.7317
.7617
.7737
.0139
.2671
.3776
.5355
.5999
.6303
.6743
.6999
32.
MIN.
DOSE
(yg)
10.
5.6
3.2
1.8
1.0
.56
.32
.18 214.
355.
302. 336.
249. 289.
912.
678.
560.
463.
1267.
1018.
825.
698.
591.
1721.
1570.
1301.
1081.
923.
790.
1446.
1433.
1429.
1279.
1120.
989.
869.
57.
785.
992.
1114.
1081.
1002.
920.
835.
32.
MIN.
DOSE
fog)
10.
5.6
3.2
1.8
1.0
.56
.32 214.
.18 282. 223.
144.
165.
175.
215.
203.
204.
204.
159.
163.
165.
169.
172.
126.
140.
145.
147.
150.
152.
54.
67.
81.
88.
93.
97.
100.
5.
17.
27.
37.
43.
47.
51.
53.
-------
299
Table Bl.2 Comparison of the Y-intercepts from the three Linear models
describing Ames data collected on July 8, 1981.
LOG DOSE - LOG REVERTANT MODEL
MAX. DOSE
.56
1.0
1.8
3.2
5.6
10.
18.
MIN.
DOSE
(}4g)
10.
5.6
3.2
1.8
1.0
.56
.32
.18
1.518
LOG DOSE - REVERTANT MODEL
.56
1.0
MAX. DOSE
1.8 3.2 5.6
10.
18.
DOSE - REVERTANT MODEL
.56
1.0
MAX. DOSE (yg)
1.8 3.2 5.6
10.
18.
32.
1.530
1.545
1.749
1.599
1.576
1.421
1.578
1.539
1.546
1.395
1.500
1.585
1.550
1.552
1.784
1.493
1.530
1.585
1.556
1.556
2.170
2.043
1.731
1.671
1.668
1.616
1.596
3.174
2.582
2.333
1.990
1.856
1.795
1.710
1.663
32.
MIN.
DOSE
(pg)
10.
5.6
3.2
1.8
1.0
.56
.32
.18 -4.
-63.
-18.
-107.
-86.
-39.
-716.
-402.
-254.
-141.
-1233.
-840.
-549.
-370.
-227.
-1940.
-1664.
-1194.
-827.
-576.
-376.
-1487.
-1460.
-1452.
-1165.
-872.
-642.
-443.
1482.
-221.
-686.
-951.
-883.
-724.
-567.
-411.
32.
MIN.
DOSE
(wg)
10.
5.6
3.2
1.8
1.0
.56
.32
.18 4.
88.
30. 57.
23. 44.
-2.
29.
27.
25.
111.
89.
81.
66.
56.
284.
168.
135.
117.
97.
83.
798.
612.
430.
337.
211.
230.
196.
1608.
1167.
923.
707.
572.
479.
406.
352.
-------
300
Table B1.3 Comparison of the deviation from Linearity sum of squares
for Ames data collected on July 8, 1982.
LOG DOSE - LOG REVERTANT MODEL
MAX. DOSE (yg)
1.8 3.2 5.6
.56
1.0
10.
18.
MIN.
DOSE
(yg)
10.
5.6
3.2
1.8
1.0
.56
.32
.18
.0040
.0001
.0062 .0125
.0063 .0135
LOG DOSE - REVERTANT MODEL
.56
1.0
MAX. DOSE (yg)
1.8 3.2 5.6
10.
18.
MIN.
DOSE
(yg)
10.
5.6
3.2
1.8
1.0
.56
.32
.18
1049
433
4 677
1774 4772
32.
.0138
.0226
.0245
.0247
.0142
.0168
.0227
.0253
.0253
.0017
.0171
.0179
.0227
.0257
.0257
.0137
.0162
.0622
.0663
.0664
.0806
.0844
.0027
.0436
.0674
.1780
.2140
.2286
.2801
.3073
32.
32749
67742
94752
136129
3203
40074
109618
179995
280562
44777
58664
189769
399508
619826
898293
94470
94580
94613
191872
412859
687636
1062860
38281
376747
459602
525655
534986
634165
810818
1111451
DOSE - REVERTANT MODEL
.56
1.0
MAX. DOSE (yg)
1.8 3.2 5.6
10.
18.
32.
MIN.
DOSE.
()|8)
10.
5.6
3.2
1.8
1.0
.56
.32
.18 133
227
956 3070
1095 4533
7635
10533
10582
10639
31252
32525
33018
37190
40372
2017
40366
49002
54572
67204
77403
200988
304343
574505
726370
836463
949273
1037988
39503
606642
1091694
1900261
2470475
2913128
3318391
3642705
-------
Table B1.4 Comparison of the random error of the three models
describing Ames data collected on July 8, 1981.
301
LOG DOSE - LOG REVERTANT MODEL
MAX. DOSE (pg)
1.8 3.2 5.6
.56
1.0
MIN.
DOSE
(yg)
10.
5.6
3.2
1.8
1.0
.56
.32
.18 .0076
.0245
.0104 .0281
.0118 .0294
.0221
.0248
.0284
.0297
.0183
.0225
.0252
.0288
.0301
LOG DOSE - REVERTANT MODEL
.56
1.0
MAX. DOSE (yg)
1.8 3.2 5.6
10.
.0010
.0186
.0228
.0255
.0291
.0304
18.
.0019
.0022
.0198
.0239
.0267
.0302
.0316
10.
18.
DOSE - REVERTANT MODEL
.56
1.0
MAX. DOSE (pg)
1.8 3.2 5.6
10.
18.
32.
.0042
.0046
.0049
.0225
.0267
.0294
.0329
.0343
32.
MIN.
DOSE
(Tig)
10.
5.6
3.2
1.8
1.0
.56
.32
.18 573.
11747.
1799. 11893.
1823. 11917.
12165.
12567.
12713.
12738.
12931.
14181.
14583.
14729.
14754.
6541.
16635.
17885.
18287.
18433.
18459.
23755.
24575.
34669.
35920.
36322.
36468.
36493.
59181
61197
62018
72112
73363
73765
73911
73936
32.
MIN.
DOSE
(148)
10.
5.6
3.2
1.8
1.0
.56
.32
.18 573.
11747.
1799. 11893.
1823. 11917.
12165.
12567.
12713.
12738.
12931.
14181.
14583.
14729.
14754.
6541.
16635.
17885.
18287.
18433.
18458.
23755.
24575.
34669.
35920.
36322.
36468.
36493.
59181
61197
62018
72112
73363
73765
73911
73936
-------
302
The Student-Newman-Keuls (S-N-K) multiple comparison test (86) was
applied to the data examined in Tables Bl.l through B1.4, and six doses
were obtained which were neither significantly non-mutagenic nor toxic.
Figure B1.5 shows the log-log transform of the dose-response curve of
this data and that from the same test compound and bacterial strain
assayed on eight other dates. Figure B1.6 shows the linear portion of
eight of these curves, after applying the S-N-K multiple comparison test
to obtain the linear region.
To obtain an unbiased estimate of any sample's regression
coefficient, the data from several dates should be pooled after
eliminating any statistically significant outliers. An analysis of
covariance ( cx= 0.05) performed on the dose-response curves of Figure
B1.5 showed that the slopes of all these curves were not statistically
different, while the y-intercepts for all curves except that obtained
from the March 16, 1982 data, were not statistically different.
To demonstrate blocking of assay dates using diesel SOF samples,
two samples were analyzed simultaneously on four different dates.
Figure B1.7 shows the dose-response curve of these samples in log-log
format; sample 4-B is the SOF from the baseline Caterpillar 3208 engine
without exhaust aftertreatment, while sample 4-P is SOF from the same
engine with Johnson-Matthey close-coupled port catalysts installed. To
choose the mutagenic section of each curve, the S-N-K test yielded the
points between C and C , as listed in Table B1.5.
-------
303
UJ
J-
CE
_J
CL
\
CO
h-
2
tC
UJ
CO
CO
U_
O
Ul
DO
z:
ID
CD
O
CD
UJ
51
o
o
•
CO
o
LO
•
CO
o
o
•
OJ
o
ur>
r
2-NITROFLUORE:NE
BIOLOG1CRL nci IVlTf
WITHOUT S-3
REPERTED flSSflV DHTES
RS LISTED BELOW:
RSSRY
RSSRY
RSSRY
RSSflY
flSSRY
RSSRY
RSSflY
RSSflY
flSSRY
DflTE:
DfiTE:
DOTE:
DRTE:
DRTE:
DHTE:
ORTE:
DRTE:
DflTE:
110381
121181
102781
70881
11182
31182
32582
31682
33082
-3.00
1
I I I I I I I I I 1 I I 1 1 1 i
-2 .00
-1 .00
0 .00
1 .00
2 .00
LOG SflMPLE MRSS
A
LOG MICROGRRMS PER PLflTE
Fig. Bi.5 2-NF Dose-Response Curve with TA 98.
-------
304
UJ
h-
CE
_J
Q_
•*x
CO
LU
O
_J
O
O
u_
O
UJ
OQ
CD
O
CE
UJ
o
O
•
CO
o
to
•
CM
o
o
•
CO
O
LT5
r~"i—r
rr
~\—i—i—r
2-NITROFLUORENF
Tfl 98
WITHOUT S-9
RSSRY DflTE:
RSSflY DHTE:
ftSSRY OflTE:
flSSflY OflTE:
flSSflY DOTE:
flSSBY DflTE:
RSSftY DHTE:
RSSflv
110331
U 1181
102781
70881
1116?
32582
31682
33082
J 1 L I
J L
-1 .50
-1 .00
I I I I
.50
0 .00
.50
1 .00
LOG SflMPLE MPSS
LOG MICROGPPMS PER PLRTE
Fig. Bi.6 Linear section of 2~ NF curve .
-------
305
Table B1.5: Summary Statistics of Dose-Response curves for
Samples 4-B and 4-P chosen by the Student-Newman-Keuls test
with an alpha level of 0.001
Degrees of
Sample freedom C^ _Cg a. o^ %CVQ b_ a& %Cvb
4_P 18 37.5 1200 0.5779 0.0559 9.677 0.7862 0.0235 2.99
4-B 15 75 1200 0.3484 0.0719 20.61 0.7312 0.0285 3.90
The slopes (b-values) in this table appear to be similar and have low
variability, while y-intercepts (a-values) seem quite different, despite
their relatively higher variabilities. To test the equality or non-equality
of both values, an analysis of covariance was performed by calculating a
t-statistic for each group of data (87). Comparison of the calculated
t-statistic with the critical t-value for slope and y-intercept showed that
slopes, but not y-intercepts, were equal at the 0.05 significance level.
In practical terms, this means that the slopes of dose-response curves
may have little meaning when the log dose-log revertant transformation of a
simple linear model is used. Instead, one could use the y-intercept as a
predictor for the dose required to obtain a predefined mutagenic effect,
similar to the use of LD-50 values for mortality data. If one chooses the
spontaneous revertant level (shown to be stable for constant experimental
parameters), one can define the Order of Magnitude Mutagenic Potential
(OMMP) dose as the dose level exhibiting a 10-fold increase in mutagenicity
over background levels on the regression line described by the dose-response
curve. Application of this potential to the data for 4-B and 4-P diesel SOF
-------
306
discussed previously, the value giving ten-fold increase over background
revertants with 4-B SOF is 2.806 |lg/plate, while the same value for 4-P SOF
is 2.318 yg/plate. As seen in Figure B1.7, the leftward shift of the
dose-response curve for the catalyzed (4-P) SOF yields a lower figure for
the Order of Magnitude Mutagenic Potential.
This OMMP is derived from the y-intercept of the log-dose-log
revertant transformation of the linear regression model. Table B1.6
compares the standard SA (semi-log) parameter with the y-intercepts obtained
from this log-log transformation for three compounds and nine different
assay dates. The coefficient of variation was only slightly better for the
2-NF data, while it was either worse for the y-intercept value (baseline
engine SOF) or essentially the same (port catalyst SOF), for diesel SOF
samples. The improvement in variability expected by using the OMMP was not
realized; so a closer look at the definition of slope was made.
Table B1.6 Ames Data Parameter Comparisons on Repeat Assays, 3
Different Samples, 9 Different Assays Each
a b
Model Rev = a + b log1QX log Rev = a + b log1QX or Rev = 10 *X
Characteristic „
Activity Parameter SA. = ^~£°g 7T~ 10
GZ ~~ ^i *-• i
mean C.V.,% mean C.V.,%
2-NF 89.6 32 221 25
Baseline Engine Q,23;l 33 5.28 42
Port Catalyst 1.513 4° 8.71 39
-------
307
1 T
LU
J-
o:
_j
Q.
*-o
CO
LU
>
LU
CO
CD
(E
U.
O
LU
CO
2:
ID
CD
O
_J
z
,'GC
LU
o
O
•
oo
o
in
o
o
O
IT)
CRTERPILLRR 3208, MODE
SOF BIOLOGICni ACTIVITY
WITHOUT S-3
JOHNSON-MfiTTMEY CLOSE-COUF'l. fD
EXHflUST PORT
ftSSRY DflTE:
flssRY ORTE:
BSSflY ORTE:
RSSHY ORTE:
RSSRY DRTE-.
RSSRY OflTE:
fiSSflY OflTE:
RSSflY DfiTE:
121181
1?M81
1 1182
1 1132
30982
30982
31 182
31 13?
uncatalyxed
J I I I
I I [_
_J I 1 I
I I I
0 .00
1 .00
2 .00
3 .00
LOG SRMPLE MRSS
LOG MICROGRRMS PER PLRTE
Fig. B1.7
Two diesel SOP samples assayed on four dates.
-------
308
Figure B1.8 shows the wide variety of values for specific activity
obtainable from one assay depending upon the units of dose range chosen;
this demonstrates the sensitivity of an assay to non-uniform dose ranges.
The S-N-K multiple comparison test cannot be effectively utilized for
samples not reaching toxicity because of mass limitations, so a uniform dose
range may be a more accurate method to use for calculating specific
activity. The Caterpillar 3208 data base showed no SOF samples toxic at 600
yg/plate, while 18.75 yg/plate was always into the linear portion of the
responding region. Using these values as the C and C values for 4-B and
4-P SOF yielded nearly equal coefficients of variation for SA values (Table
B1.7) This shows that the SA figure can be made repeatable if replicate
assays demonstrate that major linear portions of different samples'
dose-response curves are confined to identical dose levels. This may not
always be the case, however. The optimum parameter for measuring s sample's
mutagenicity may not be the same for different groups of data.
Table B1.7 Results of 9 Repeat Assays, C = 18.75 yg/pl.
C2 = 600 yg/pl.
CATERPILLAR 3208, EPA MODE 4
SA
(rev/|ag) (rev/u g)
Baseline 0.317 0.041 12.8
Close-Coupled Catalyst 1.028 0.113 11.0
-------
'ECIFIC RCTIVITY (REV/MICROGRRM)
~n
i—i
CD
•
w
M
•
00
01 O
o
3D
o
~z.
o
o
CO
o
! | |
.20 .40
, | , , , ,
1 1 1
.SO
'I1'--
.80
1 ' '
1
i
.00
1 , , , , 1
m o ~n co
"O 3D 0 13
3D —1 70 m
n
Z2 co < .-H
O ro HD ~H
CD o 70 i— i
m co >-, n
o
-P- 00 CZ 3D
3D CO n
^ CO -H
•-i m < I-H
z r~ 3D <
— t i — i r~ i — i
i ' ' l ' '
m z cz — i
TO m m -<
ZS CO
. ^ HD
Z 0 CO
LO o ~n
"0 3D
rn n n
m o ^ -n
o z cz
- —i z
~Q
"D
m
TO
0
o
CO
m
s-^-
n
ro
^
t — i
Z
Z2
I 1
n
70
O
CD
7(3
X'
1 — *
0
i— »
o
(X")
( ^
CD
O
p^rno -ncn
013
Tom
013 J
rino?
con-^cp
o
i—
I Tl n
HH ZJ
?in5 mo.
O ^ Zi •*—
g^1 2 ID
a c^ 3 <-** i'—'i
v-j -n ^^v k >
°z-io CZ<
<..or; p-,^
rj;^ oo—i
m '-o r~ —S
c ^:D r~i
n
CZ
-^
O
—I
I—I
o
o
n
J
-i
CJ
_L L_
OJ
o
-------
310
It can be concluded that quantification of Ames test data is a very
difficult task, mainly due to the complex nature of the interaction between
sample and bacteria and also due to the fact that convenient implementation
of the Ames test and limitations in counting methodology interfere with many
of the statistical assumptions necessary for a reliable data analysis. However,
if certain practices are followed, the data are usually amenable to reliable,
statistically sound quantification. These practices include the following:
1) Use of the log dose-log revertant transformation of a simple linear
regression model, which reduces bias in estimating the regression line.
2) Use of a multiple comparison test such as the Student-Newman-Keuls
test, if the sample is sufficiently concentrated to allow demonstration of a
toxic effect. This reduces bias in estimating the extremes of mutagenic
potential. If toxicity is not reached, use of a uniform dose region for
the regression line may give more accurate results for samples with similar
activities. Comparison of samples with widely differing activities may not
be accurate unless sufficient sample is available to demonstrate toxicity.
3) Use of an SA parameter calculated from the log dose-log revertant
transformation of a simple linear model (85, 86), although it may not reduce
variability significantly. This transformation can be expressed as a power
function, since the equation upon which it is based (log R = a + b log C)
is equivalent to R = 10 *C , when antilogs of both sides are taken. The
advantage of this transformation is the inclusion of the y-intercept in
calculation of the specific activity parameter, since slopes of different
samples have been shown to be similar, while y-intercepts are quite different.
The SA value from this transformation (SA^ ) may be calculated as:
-------
311
SA • 10a (C2b -
J_»J_i
where a, b, GI, and C? are defined as before and SA^ is the average
specific activity (revertants/yg) between C and C .
Reproducibility of Ames Data
Intralaboratory and interlaboratory reproducibility is crucial to any
attempt to quantify Ames data. The intralaboratory reproducibility of Ames
data was examined at MTU by collecting numerous data sets over time, using
tester strain TA98. The first data set of TA 98 spontaneous revertants
(without S-9 activation) was collected from experiments performed between
March 31, 1980 and November 16, 1981. The statistics presented in Table
B1.8 confirm that no trends existed over this time period. This indicated
the spontaneous background reversion rate of TA 98 was consistent and
reproducible throughout this time period. The second data set is the
response of TA 98 to 10 yg/plate of 2-NF (without S-9 activiation). These
data were collected from various experiments between June 6, 1980 and
November 12, 1981. Table B1.9 gives the mean, standard deviation and 98%
confidence interval for the response of TA 98 to 10 jig/plate 2-NF over this
time period. Since no trends appear in the data, TA 98's response to 10
yg/plate 2-NF is considered to be stable.
-------
312
Table B1.8 Statistics describing TA 98 spontaneous revertants/plate as a
function of time.
TA 98 SPONTANEOUS REVERTANTS/PLATE STATISTICS
THE REGRESSION EQUATION IS
REVERTANTS = 27.8 - .0387 * DATE
R-SQUARED = .6 PERCENT
MEAN NUMBER OF REVERTANTS/PLATE M = 26.971
STD. DEV. =7.23
90% C.I. IS 26.051 u 1 27..S8
WITH ( 172- 2) = 170 DEGREES OF FREEDOM
ANALYSIS OF VARIANCE
DUE TO DF
REGRESSION 1
RESIDUAL 170
TOTAL 171
F .05(1) 1,170 = 3.84
ss
54.87
8873.98
8928.85
MS=SS/DF
54.87
52.20
CALC. F
1.05
-------
313
Table B1.9 Statistics describing revertants/plate induced by 2-NF as a
function of time.
INDUCED REVERTANTS/PLATE WITH 10 -pg/PLATE 2-NF
THE REGRESSION EQUATION IS
REVERTANTS = 1577. - 3.15 * 2NF CONG
MEAN NUMBER OF REVERTANTS/PLATE M = 1555.6
STD. DEV. = 366.0
R-SQUARED = .1 PERCENT
90% C.I. IS 1504.27 < u < 1608.69
WITH ( 37-2) =35 DEGREES OF FREEDOM
ANALYSIS OF VARIANCE
DUE TO DF
REGRESSION 1
RESIDUAL 35
TOTAL 36
F .05(1) 1,35 = 4.13
ss
4887
4813335
4818222
MS=SS/DF
4887
137524
CALC. F
0.036
-------
314
Interlaboratory variability was assessed at MTU by participation in two
sets of round-robin sample analyses with EPA-RTP and EG&G Mason
laboratories. Due to lack of dilution instructions for the sample and
positive controls in the first round robin, only spontaneous revertants and
2-nitrofluorene induced revertant statistics are available. Although five
tester strains were utilized for both round-robins, only the results of
strain TA100 will be discussed here.
Figure B1.9 shows the spontaneous revertant rate and Figure B1.10 shows
the revertants induced by 3 pg/plate of 2-NF, as assayed in the three
laboratories (in both tests). A nested analysis of variance showed that
neither set of data demonstrated any significant variability between
laboratories, at a 5% alpha level. There were significant differences for
different assay dates, at the same 5% alpha level, for spontaneous
revertants, but not for mutants induced by 3 pg of 2-NF.
For the second round robin, two diesel SOF samples generated at MTU
using the Caterpillar 3208 Engine at mode 4, with and without an oxidation
catalyst, were assayed using strain TA100 without S-9 activation. Figure
Bl.ll shows the dose-response curve in log dose-revertant format. Although
statistics were not applied to these data, inspection of the curves shows
similar trends. Some of the differences may have been due to differences in
automatic colony counter setting with the MTU counter being able to detect
smaller and therefore more colonies.
In summary, intralaboratory variability of Ames data was shown to be
less than random experimental error while interlaboratory variability was
shown to be less than experimental error in most cases. The number of
spontaneous revertants varied significantly between assay dates, but not
between laboratories. Interlaboratory experimental error was found to be
-------
315
~. 150
"O
e
I
3
a.
r 100
c
a
k
i
i 50
0
o
I
i
~ f-
T
i
T
1
T
i
T
i
laboratory: MTU MTU I G CM EGGM EGGM IGGM E8GM EGGM EPA
««*oyd«l«: 50480 111480 111680 51480 51480 51480 51180 51180 51680
lampU(n): 33 3 333 33 3
Fig. B1.9 TA 1OO Spontanoous R«vartanli/plat« In rhr«« Am«t laboralorUi.
-------
316
>• as oo
•
•a
•o
••
M
+ 1
c
0
0
£
• 1500-
*-.
o
a
o
2 500-
laboratory: MTU MTU
asiay date: 50180 111980
sample (n): 3 3
EPA EGGM
5018-0 111980
3 3
Fig. Bl.10 Responte of TA 1OO to
2"N F.
-------
317
0)
-d
•d
->-}
tn
0)
.*j
(0
r—«
a,
CO
0)
• I—I
a
o
r—I
O
u
o
o
o
(!)
a>
00
cd
t_
0)
2000 -
1500
MTU Catalyzed Mode 4 SOF
EG&G Catalyzed Mode 4 SOF
MTU Uncatalyzed Mode 4 SOF
EG&G Uncatalyzed Mode 4 SOF
Without S-9 Metabolic Activation
Fig. Bl.ll
2 34 5678^02 234 5678^03
Mode 4 SOF Concentration (jig/plate)
Interlaboratory comparison of two diesel samples.
-------
318
greater than the experimental random error, although similar trends were
seen in the data.
Ames Bioassay Results for Selected Data
The Caterpillar 3208 provided the greatest number of SOF samples for
repeat bioassays of any of the engine systems tested during this study. A
detailed analysis of Ames bioassay results from SOF collected using No.2
fuel and this engine, after replacement of its injection pump, was
performed.
Figures B1.12 through B1.20 plot the Ames dose-response curves of
twelve assay dates between July 1981 and April 1982, using SOF from the
Caterpillar 3208 engine in its baseline configuration as well as with
close-coupled port catalysts and Corning ceramic traps. Triangles placed
above and below data points designate standard deviation for all plate
replicates.
Figures B1.12 through B1.17 compare the three test conditions on a
modal basis, and show that the Corning traps yielded SOF with the highest
mutagenic activity at all modes but mode 3. Additionally the Corning trap
SOF exhibited signs of toxicity at all modes (including mode 3) and was
significantly toxic at 1200-|jg/plate for all modes but mode 3.
Figures B1.18 through B1.20 show the same data plotted as a function of
aftertreatment configuration. Mode 9 was the only mode exhibiting
significant toxicity for the baseline engine, while mode 10 displayed a
slight indication of toxicity. The port catalyst SOF was noticeably toxic
at modes 5 and 10, while the Corning trap SOF showed toxicity for all modes.
As discussed in the section on Ames bioassay variability and data
quantification, the mutagenic response is often marked by toxicity. The
most significant example of this is shown in Figure B1.21, which is a plot
-------
Average Number of TA 98 Colonies/Plate (ave. ± std. dev.)
-s
O
i-i»
O
o
o
H-
TO
CJ
ro
o
CD
O
CL
£ CO
(-"
£> C/J
o
O
o
CD
o>
o
o
OD
O
O
O
o
o
o
o
-------
C/}
p
3
3
Average Number of TA 98 Colonies/Plate (ave. ± std. dev.)
Ol
O
o
o
o
o
Ol
o
o
ro
o
o
o
en
o
o
CO
o
o
o
n
o?
CO
ro
o
CO
a
o
3
CD
CD
O
>
CO
CD
Tl
H-
TO
W
01
C/D
O
00
T)
O
• 1 1 1 1
1
1
1
*
M<
1
JH»
\\
~'«
-**1*
~ t \
~ 1 \
:0< ; ,
N, """--.. ""-•-...
> ^ t >'*• o _ 4 » '"'• • -q
f •% < » " o « » • <
— >• -^ •> > • <
till
1
'^ o n t
X ° o *
tr 3 w f
o 2. t '
(= P i >
<->• OQ O f
O
y z c (
1 3 "o r
10 ^ 0" 0"
cop.:
yi^ e
ro ^ ^ f
oT § o .
V 0 rl J
Op'
CT: (-"• o "
o ^ o ,.
> ^ £ C
fD ^«tf f
£• CL M c
< f" >•
P> ra w f
f™** ^^ ^
o' i <
3 1
0
• o *
1
33
0
3
j
u ~
3 -
3
3
C
s
r
_
3
3 I
—
j
3
3
1
3
-
r
_
™
—
-------
Average Number of TA 98 Colonies/Plate (ave ± std. dev.)
3
o:
* ~c
o o
> ra ^^
o I?-
tt1 CTQ
CD w
5 r
^~: i— '
— -O
K •—
i O
CO ,_,
ro 5,
0 2
CD a.
a
0> CO M-
^ 0 o
0
^i
£— i 0*3
c \
H- r»- f*
CO fl> O
CD ^-^ «
1 — ^
r»-
/-\
•<*>
>
•c
t""t
•—• . .
,__l *•"•
o
^ . .. 1^
CO
CD
ro
1— ' H*
Ol O Ol
O 0 O
3000
1 I
• 1 I
1
^ o o c
5 ° o* 0
tW ^^ »^f ^ ft
. •& 2 to 2
o 5. » 2
1 . fc^ 1 )-.
i- • & 3 i r
H* f* OQ O n
- 1'. 0 "
- h w JT C n
bU ' 3 *S. c
It '• «3 CO
', \ ^ w d- C
.«)*(»• ft) x~, ^ fl
. \\\ S" 3 0 a
• •frf<»^~'.. g_ p_ ^ J
1 \ *'\\ - «' 2. » c
— ^fc. — fc.*»J. lj* *^ ^
: " *\ "01vT--, > N £L
- X *-. '•-•••. ° CD x r
N. X. -•-... ~. CL co c
» i(t i p o.._«ns. < rr t
^N. '*fc»^'"'-.. r^" s^ ••
X. -»••.. >-'• ^
9 ~^~^ * ** **'"-*- 3 ,£>. "
: /^^* />Tr>-o , ^
L ^x ., """*--->... . 0 ,
% t « > ~~fO <
•- «• ^ ^ f _..- <>' 4
: , o---\:V""
L \*
1 !
ft
C
C
j^
0
{
3
^
5
j
.
3 J
5
3'
' _
i
-*-
) -
5 ~
3
3
1
3^
-
—
-
—
U)
ro
-------
03
I
Number of TA
-------
IS
1C
Summary <
U >~x -
0 -r] to
O
3 ^
<*___,! ^"*
~_j uQ
,_ £T ^
«o /5T °
CO O- co
»— '
(•*•
O
>
>-v
I— (-»
o
5
CD
TV 1
Average Number of TA 98 Colonies/Plate (ave. ± st<
i—*
CO ^ Oi CD O
O O O O O
3 O O O O O
— 1 1 1 1 1
• III 1
• S3
!-••
M ("^
cr
o
!•*• •« S-
: 1 \ M
~t ^.. «°
11 x- g;
_ «K« » •. •• nj
I \ \ ^J-
" ' » "''••-. a"
•t v -... 9-
I •, V * •" >
\ f^
w^ ^ ^\^ '^" * ••* ^
^\- "••••• r*
» 'o < » ^-^. , ,. ""••• , , O
; "\ * *<^^* '•""'" "
— "\ ^^^^^ • '
* °x* ' *s>.-£^"" "
s\ ^^^^^^^^
'•-•;:*-"-•*' ' * *-~^^*
- ""---'"' ^^^-^^.
„ . . .
5- > *
III 1
i. dev
0 G
2"* to
. 0)
3 1
OQ O
o
r-f rr
*^ ^^
^d tT c
M p.
_, v
§ °
p **
r s
^ £
fD <<"
(X M
X
1
• o
• o
•»
.)
t— »
fo
o
o
— 1
1
CO
(39
CO
JL
CD
_
fl) _
2. ~
3'
f6
3. .
?
P
o :
—
Q
|_J
rt-
O^
* i
-i
*
.
-
-
_;
i
OJ
K3
U)
-------
324
W
o
o
CO
05
SB
1200
0)
1000
800
600
400
200
0
Baseline engine with no control *-
Close-coupled port catalysts
Corning traps (uncatalyzed EX-47) .
Without S-9 Metabolic Activation
- I'"
t •••" • ____-§.-*
i. ; i T'l* Tl'i n i i i i i 1111 i i_ i i i 1111 i i i i i 1111
1° 101 10
10'
10'
Fig. Bl.17 Mode 11 SOF (Hg/plate)
Summary of Caterpillar 3208 Data from July 1981 to April 1982
-------
325
V
OH
W
E-
-------
326
2000
1500
a
£
*O
U
CO
1000
-------
w Average Number of TA 98 Colonies/Plate
&
3
3 i— * (— i fvj
5 01 o 01 o
!-, o o o o
V; £ O O O O O
- ; °
o r
a? M
re
t— ••
5T r/j ^
~" 0 ^
CO "^
o o
CD 0
t-J
o o
Q) rt>
1 1 1
0000
PL p. O. p-
(T) tB (D CD
^fr
-IjS >-i o co en
3^\
nV. DO CO CO CO
*Jv^ o o o o
"Trvr . ^ "^ *^d ^
• 1 ' \ '•'•
•* 4 ^H. 'rf^
:\\^iv-->. ^ G n •
i .*•••• iii
<-+• P r i ' s v • -
ft £• ^ |;*»s ^'dh,. o. 1
-5 <-*• ZO 1 ^x Jt^'^S-L ""' 1 '
S o" 1 *\ °^5v -•- °- s ^ G n »
•J ^~f 1 — \ x ^*X '•'•«.
^ \ ^x^ '^"^'-
«^ r *\ >e>^. '^'^--
£- 'CTV
^ OQ
VNX i-*-
\ NX~~"*'~-^
^ v& ©""~'°-_ *"'O
- \ N^ ~ ~ ~ ~ _ ^ "".
^ "^ Q L \^ '' ^^^ ~~~~~^~^^
CD p" w *\ ,. °. • y- x^~~ °~-
*~ fD~
o"
^ c
•^"•••'^^x. / y" ~~--^
:" ""J.'Q""~ "S^-^* - ^ "^°
."" - - ' ~~^~~' "^^X. ^^
»-'' — #"""" ^
""' ~ \ ''
^_ o F- *
to *"
CU
I I I
en
0
o
i
n
o o n
CL C1 ^ ~
CB (D »3
OQ
4^ CO T -
•o ~"
CO GO M
O O
""^ ^j ^ ~
3
O
! 7 <"*"
! N -Z
! S.
w
X
1
Ci -
_
-
~
-
—
I
-------
328
of all subtraction dose-response curves from the Caterpillar Engine between
July 1981 and April 1982, in log-log format. The Corning trap mode 4
oxygenated, ether insoluble, and basic subfractions (each assayed twice)
stand out as more mutagenic than any other samples run in this laboratory.
The shape of the theoretical toxicity curve may not allow direct observation
of a toxic effect for a composite sample, and the extremely active
oxygenated subfraction (toxic at 18.75 yg/plate), which constitutes only
6.5% of the SOF mass, does not raise the total SOF activity as
significantly as expected. This suggests that hazardous compounds may be
obscured if one observes the total SOF mutagenic activity.
The SOF subfractions obtained from the Caterpillar engine during this
study are ranked in decreasing order of specific activity in Table B1.10.
Despite the lack of specific activity data for three subfractions from the
PTX catalyst, it can be seen that the ether insoluble, acidic, and
oxygenated subfractions are the most active of the eight subfractions
tested. While more limited data are available from the Mack engine, the
general trend shows transitionals to be somewhat more and hexane insolubles
somewhat less active, in a relative sense. Acidic, ether insoluble, and
basic subfractions are still quite active, as shown in Figure 3.1.9 in the
discussion of the Mack engine emissions.
The effect of sample storage on mutagenic specific activity was
investigated using mode 11 SOF from the Caterpillar 3208 engine equipped
with port catalysts. The SOF was stored on four different filters and
extracted on three separate assay dates, with storage in either DMSO, on the
particulate filter, as a dry residue, or in dichloromethane. Although not
all storage conditions were utilized for all assay dates, a uniform dose
-------
329
LU
(-
cc
_J
QL
v»
Ul
Qi
oo
<7>
CC
li.
a
(Y.
u
oo
CD
O
GC
UJ
O
O
I 1 1 1 1 1 1 1 1—
CRTERPILLRR 3208, MODE
SOr SL'SFPq — ION SIOflCTIVITY
WITHOUT b-3
BHSE'-INE ^MINE (NO CONTRDl )
JOHN'?aN-M;iTTj£Y EX-SJ5T P35T C^TRLYST
flM3 COSMNG CISPMIC [=nRTICULRTE TRflPS
T r
O
O
Corning Trap 3/16/82
(Oxygenated) 3/25/82
Corning Trap 3/16/82
(Ether Insoluble) 3/25/82
- Corning Trap 3/16/82
. (Basic) 3/25/82...
o
o
1
-i
-I
-4 .0
-2 .0
0 .0
2 .0
4 .0
LOG SflMPLE MRSS
LOG MICROGRflMS PER PLRTE
FIG, BI.21 CATERPILLAR 3208 SOF SUBFRACTION DOSE RESPONSE CURVES
BASELINE, J-M PORT CATALYST, AND CORNING TRAP
REPEAT ASSAYS FOR EACH SUBFRACTION, EPA MODE 4
-------
330
Table B1.10
Ranking of SOF Subfractions According to Specific Activity
for the Baseline, PTX Catalyst, Port Catalyst, and Corning
Trap tests at Mode 4 with the Caterpillar 3208 and an Overall
Ranking Considering all Three. (Highest Specific Activity on
Top Decreasing Down)
Baseline
* PTX Port
Catalyst Catalyst
Corning
Trap
Overall
ACD
BIN
OXY
HXI
ARM
TRN
BAS
PRF
OXY ACD
ACD EIN
TRN OXY
BAS BAS
ARM HXI
TRN
ARM
PRF
EIN - Ether Insoluble
BAS - Basic
ACD - Acidic
PRF - Paraffin
ARM - Aromatic
TRN - Transitional
OXY - Oxygenated
HXI - Hexane Insoluble
OXY
EIN
BAS
ACD
HXI
TRN
ARM
PRF
EIN, ACD
OXY
HXI, BAS
TRN
ARM
PRF
* PTX Subfractions not assayed were EIN, PRF, HXI
-------
331
range analysis of the resultant Ames data showed no clear pattern of sample
gain or loss of mutagenic activity. This suggests that the practice of
holding samples for testing on the same date is acceptable.
Tables Bl.ll and B1.12 give SA data calculated using the log
dose-revertant and log dose-log revertant tranformation of the linear
regression model. The use of the power function (log-log) transformation
generally results in a coefficient of variation for SA values somewhat
higher than that for the semilog transformation. This is expected, as the
log-log tranformation utilizes the y-intercept as well as the slope
coefficient of the linear regression model, and y-intercepts of repeat
assays in log-log coordinates are more variable than slopes.
The results in Tables Bl.ll and B1.12 also show that the variability
due to repeat engine runs and extractions may not contribute significantly
to the total variability in SA, since C.V. values for all tests are not
consistently higher or lower than C.V. values for repeat tests of identical
samples.
-------
Table Bl.ll Summary of specific activity (SA) from model: Revertants = A + B * log (Sample Concentration)
SOF Concentration range: 18.75 - 600.00 pg/plate. Caterpillar 3208 data July 1981 to April 1982.
Assay
Date
B
Mode 3
P T
B
Mode 4
P T
B
Mode 5
P T
B
Mode 9
P T
B
Mode 10
P T
B
Mode 11
P T
7/10/81
0.564
1.117
1.639
0.460
0.252
0.597
11/03/81
0.367 1.156
12/11/81 0.358 0.456
0.332 0.835
1.176 1.690
0.588
0.637 0.171
0.449
0.344
1/11/82 0.452 0.578
0.276 1.096
0.369 0.997
1.224 1.912
0.837
0.827 0.184
0.443
0.396
3/09/82
0.377 0.989
0.341 1.167
3/11/82
0.287 0.963
0.299 0.930
3/16/82
0.348 0.295
1.380
1.402
2.005
1.442
1.217
(a)
3/16/82
0.360
3/25/82
0.456
1.418
3/30/82 0.426
0.409 0.267
1.798 0.670
1.411 1.269
2.234 0.413
1.499 0.529
1.199
(a)
3/30/82 0.355
4/13/82
"00
4/13/82
0.256
0.384
0.299
1.401 1.212
0.842
0.725
0.378
OJ
U)
ro
-------
Table Bl.ll Summary of specific activity (SA) from model: Revertants = A + B * log (Sample Concentration)
SOF Concentration range: 18.75 - 600.00 ug/plate. Caterpillar 3208 data July 1981 to April 1982.
Assay
Date
4/15/82
(a)
4/15/82
(a)
4/27/82
n
average
C.V.(%)
n
(a)
average
C.V.%
Mode 3 Mode 4 Mode 5 Mode 9 Mode 10 Mode 11
BPTBPT BPTBPTBPTBPT
0.415
0.314
0.446 1.817 0.717 1.449 0.483 0.683
332 13 93 33231233 2332
0.388 0.533 0.379 0.343 1.028 1.532 1.023 1.747 1.407 1.898 0.460 2.119 0.626 0.202 1.471 0.474 0.446 1.208
14 12 11 17 11 15 30 8 0.4 38 8 33 21 3 10 30 1
5 17 5 5 5 5 5
0.407 0.335 1.563 1.000 0.997 0.617 0.496
12 17 14 28 35 28 24
(a) Indicates repeat engine run (sample) and/or extraction (therefore includes
the variation of the engine, sample collection, and extraction process)
Two numbers in the same box indicate separate extractions from the same
engine run.
B = baseline engine with no control
P = with close-coupled port catalysts
T = with Corning traps (uncatalyzed EX-47)
CO
Co
Co
-------
Table B1.12 Summary of specific activity (SA) from model: log (Revertants)= a + b * log (Sample Concentration)
SOF Concentration range:18.75 - 600.00 g/plate. Caterpillar 3208 data July 1981 to April 1982.
Assay
Date
7/01/81
11/01/81
12/11/81
1/11/82
3/09/82
3/11/82
3/16/28
(a)
3/16/82
3/25/82
3/30/82
(a)
3/30/82
4/13/82
Mode 3
B P T B
0.643
0.
0.308 0.429 0.
0.411 0.537 0.
0.
0.
0.
0.
0.
0.318 0.
0.
0.390 0.375 0.
0.341 0.
0.
Mode 4 Mode 5 Mode 9 Mode 10 Mode 11
PT BPTBPTBPTBPT
283
307
266
330
293
275
256
272
277
304
24b
243
3M
1.
1.
0.
1.
0.
0.
1.
0.
0.
143 1.828 0.509 0.224 0.643
135
802 1.368 1.966 0.614 0.649 0.154 0.451 0.305
100 1.322 2.174 0.821 0.737 0.173 0.445 0.387
971
860
023
957
917
1.368 1.603 2.231 1.794 1.390
1.281
1.407 1.426 1.637 0.860 2.640 0.799 1.791 0.336 1.302
1.708 0.684 1.477 0.406 0.496
CO
Co
-------
Table B1.12 Summary of specific activity (SA) from model: log (Revertant^ = a + b * log (Sample Concentration)
SOF Concentration range:18.75 - 600.00 ug/plate. Caterpillar 3208 data July 1981 to April 1982.
Assay
Date
(a)
4/13/82
4/15/82
(a)
4/15/82
4/27/82
n
average
C.V.(%)
n
(a)
average
C.V.(%)
Mode 3
B P T
0.400
332
0.310 0.536 0.346
15 20 12
5
0.370
12
B
0.267
0.395
0.274
12
0.296
15
17
0.300
19
Mode 4
P T
1.803
9 3
0.990 1.352
12 5
5
1.513
15
Mode 5 Mode 9
B P T B P T
0.749 1.670
432312
1.216 1.989 1.620 0.765 0.509 2.436
26 9 1 17 12
5 5
1.110 1.088
33 42
Mode 10 Mode 11
B P T B P T
0.426 0.664
332332
0.528 0.184 1.793 0.411 0.445 1.346
10 20 0 18 40 5
4 5
0.603 0.478
30 25
(a) Indicates Repeat Engine Run (Sample) and/or B = baseline engine with no control
Extraction (therefore includes the variation P = with close-coupled port catalyst
of the engine, sample collection, and T = with Corning traps (uncatalyzed EX-47)
extraction process)
Two numbers in the same box indicate separate extraction from the same engine run.
-------
336
Table B^2- Mack ENDT-676 SOF: Biological Activity
Data for Tester Strain TA100 without S-9
Metabolic Activation
(b)
Testv ' C C2
(^g/plate) (pg/plate)
SA
SL
BSSA
(rev/yg) (rev/kw-hr)
3B
4B
5B
9B
10B
11B
3.803
3.8-9
G-19-3
G-19-9
G-12-13
G-12-9
G-17-3
G-17-9
G-22-3
G-22-9
S-8-3
S-12-3
S-14-3
S-17-3
75
(c)
NAV }
75
75
75
75
150
37.5
37.5
75
37.5
150
37.5
150
37.5
37.5
9.4
9.4
9.4
4.7
1200
NA
600
1200
600
1200
1200
1200
1200
1200
1200
1200
2400
1200
1200
1200
600
2400
600
2400
-1825
NA
0557
-1834
-687
-1136
-407
-631
-1226
-1199
-1096
-609
-1063
-880
-880
-541
-1584
-1690
-1915
-1730
1039
NA
436
1095
485
647
364
382
816
644
700
348
664
480
480
316
1072
944
1280
965
0.97
NA
0.98
0.97
0.95
0.98
0.98
0.95
0.95
0.98
0.92
0.95
0.95
0.98
0.98
0.95
0.99
0.99
1.0
0.98
1.112
NA
0.75
1.17
0.835
0.692
0.313
0.494
1.06
0.690
0.907
0.299
0.507
0.413
0.621
0.409
3.28
1.00
3.91
1.09
202,000
NA
21,000
46,800
67,000
173,000
25,000
25.000
53,000
41,000
27,000
12,000
30,000
29,000
75,000
45,000
426,400
100,000
391,000
130,800
(a) Calculations explained in Appendix B, "Quantification of Ames Data"
(b) An explanation of the test code is given in Tables 3.1.1 and 3.1.3
(c) Not Available
(d) SA was used rather than SA^ because data analysis was performed
before the SA^ parameter was developed
-------
Table B-3. Caterpillar 3208 With and Without Englehard PTX Oxidation Catalysts:
Biological Activity for SOF from Modes 3, 4 and 5 Using Ames Tester
Strain TA98 Without S-9 Metabolic Activation
No. 2 Fuel
No. 1 Fuel
Shale Fuel
EPA MODE
PTX
Catalyst
a
b
a
b
a
b
r
SA
Without
37.5
600.
-508.05
337.51
.96
.72
37.5
600.
-362.87
239.18
.94
.51
N.A.
With
37.5
600.
-340.42
232.53
.96
.50
37.5
600.
-141.02
98.99
.91
.21
N.A.
Without
37.5
600.
-374.15
247.48
.96
.53
37.5
600.
-196.09
139.19
.91
.30
37.5
600.
-714.56
461.08
.94
.99
With
37.5
600.
-2279.22
1533.40
.99
3.28
37.5
600.
-2431.43
1736.70
.99
3.72
Without
37.5
600.
-354.92
244.16
.97
.52
37.5
600.
•2311.40
1521.44
.98
3.26
37.5
600.
-334.72
248.54
.95
.53
N.A.
With
37.5
600.
-1585.79
1557.65
.98
3.33
37.5
600.
-1063.91
750.76
.99
1.61
N.A.
SA
SL
a
b
r
N.A.
= specific activity (REV/yg) (average slope of dose-response curve, semi-log transformation of linear
model)
minimum dose in linear portion of dose-response curve
maximum dose on linear portion of dose-response curve
response intercept of dose-response curve, determined statistically
slope of dose-response curve in linear-log format, determined statistically
coefficient of correlation
data not available
OJ
u>
Calculations are explained in Appendix B.
SA was used, rather than SA^ , because data analysis was performed before the SALL parameter was developed.
-------
Table B-4. Caterpillar 3208 With and Without Englehard PTX Oxidation Catalysts:
Biological Activity for Mode 4 SOF Subfractions, Using Ames
Tester Strain TA98 Without S-9 Metabolic Activation
Fraction
PTX
ACD* va'
ARO
BAS
EIN
Catalyst _ + _ + _ + _ +
#2
Fuel
#1
Fuel
Shale
Fuel
c..
1
c«
2
a
b
r
si
c
C0
2
a
b
r
SA
c..
1
°2
a
b
r
SA
18.75
600.
-404.0
306.0
, .93
18.75
2400.
-621.33
420.60
.95
SL *37
18.75
4800.
-882.6
564.23
.95
.28
18
300
-1186
1812
7
18
1200
-2019
1657
2
18
600
-1836
1552
4
.75
.2
.11
.99
.76
.75
.2
.0
.98
.53
.75
.6
.8
.99
.02
18.75
600.
-145.0
125.66
.94
.33
18.75
2400.
22.77
6.92
.78
<.01
18.75
2400.
-43.90
45.68
.88
.04
18.75
1200.
-45.20
49.98
.88
.08
18.75
1200.
20.17
20.4
.88
.03
18.75
2400.
-198.3
141.70
.93
.13
18.75
300.
-1.02
23.59
.98
.10
18.75
300.
18.12
18.60
.97
.08
18.75
600.
-112.04
99.09
.89
.26
18
300
-1058
813
4
75
-34
112
1
18
300
-808
840
2
.75
.6
.20
.95
3.48
.69
.73
.28
.91
.92
.75
.7
.8
.96
.74
18.75
600.
-246.4
198.18
.94
.51
18.75
600.
-17.75
34.93
.93
.09
18.75
2400.
-698.97
457.36
.94
.40
N.A.
18.75
600.
-1189.38
878.13
.94
2.27
18.75
1200.
-1000.3
727.74
.98
1.11
Symbols and calculations are explained in Appendix B.
(a) ACD-Acidic, ARO-Aromatic , BAS-Basic, EIN-Ether Insoluble
(b) SACTwas used, rather than SS , because data analysis was performed before the
oL LiL
parameter was developed
A^
LL
UJ
u>
00
-------
Table B-4 Continued
Fraction
PTX
Catalyst
#2 c,
Fuel
#1
Fuel
Shale
Fuel
c2
a
b
L S-, \
sib)
bASL
Cl
r
2
a
b
r
SASL
c
1
r
2
a
b
r
SASL
HIN
18.75
300.
-11.98
34.55
.97
.15
18.75
300.
-3.27
32.89
.99
.14
18.75
600.
-8.18
30.37
.91
.09
N.
18.
1200.
-1132.
796.
.
1.
18.
4800.
-584.
353.
.
*
A.
75
08
55
94
22
75
7
18
89
18
OXY
18.75 18.
1200.
-581.25
473.97
.98
.72
18.75
4800.
-114.70
105.69
.98
.05
18.75
2400.
-164.86
138.01
.97
.12
300.
-2076.
1986.
8.
18.
2400.
-1755.
1229.
.
1.
18.
1200.
-1289.
1271.
.
1.
75
5
85
99
51
75
6
55
97
08
75
2
11
99
94
-
N.A.
300.
19200.
22.39
4.15
.49
2.01
1200.
4800.
15.76
18.27
.90
<.01
PAR
N.A.
600.
9600.
.06
16.61
.84
2.01
1200.00
9600.
-46.06
25.58
.99
<.01
18.75
1200.00
-52.68
91.82
.99
.14
18.75
2400.
-21.88
40.30
.98
.04
18.75
1200.
-114.3
117.45
.94
.18
TRN
18.75
300.
-1634.5
1413.14
.99
6.05
9.375
150.
-95.0
184.36
.99
1.58
18.75
600.
-469.8
435.93
.99
1.13
Symbols and calculations are explained in Appendix B.
(a) HIN-Hexane Insoluble, OXY-Oxygenated, PAR-Paraf f inic, TRN-Transitional
(b) SA was used, rather than SA^ , because data analysis was performed before the
parameter was developed
A^
u>
-------
340
300
250
200
150
100
50
SHALE OIL CAT * *
CATERPILLAR 3208, MODE 4
AROMATIC FRACTION
CATALYZED AND UNCATALYZED
STRAIN TA98, WITHOUT S-9
UNC * —UNCATALYZED
CAT * *—CATALYZED
9- J5'~"~ n-—O" NO 1 FUEL
*i—0—o—o- UNC
3 10 30 100 300 1000 3000
AROMATIC FRACTION CONCENTRATION ( n g/PLATE)
Fig. B2 - Fuel effects on dose-response curves for
Ames tester strain TA98 without S-9
activation for aromatic fraction of
SOF, with and without PTX oxidation
catalysts
1200
— 1000
|
ui 800
_i
Q.
W 600
z
S 400
200
NO. 2 FUEL CAT.* *
CATERPILLAR 3208, MODE 4
BASIC FRACTION
CATALYZED AND UNCATALYZED
STRAIN TA98, WITHOUT S-9
UNC. * —UNCATALYZED
CAT.* *—CATALYZED
NO 1 FUEL UNC
NO 2 FUEL UNC
3 10 30 40 100 300
BASIC FRACTION CONCENTRATION (^g/PLATE)
Fig. B3 - Fuel effects on dose-response curves for
Ames tester strain TA98 without S-9 activa-
tion for basic fraction of SOF, with and
without PTX oxidation catalysts
1400
1200
^1000
800
600
111
CL
400
200
NO 1 FUEL CAT * *
CATERPILLAR 3208, MODE 4
ETHER INSOLUBLE FRACTION
CATALYZED AND UNCATALYZED
STRAIN TA98,WITHOUT S-9
UNC.*-UNCATALYZED
CAT * *-CATALYZED
3 10 30 100 300 1000 3000
ETHER INSOLUBLE FRACTION CONCENTRATION (^.g/PLATE)
1400
1200
1000
BOO
600
400
200
NO 1 FUEL CAT * *
CATERPILLAR 3208, MODE 4
HEXANE INSOLUBLE FRACTION
CATALYZED ANDUNCATALYZED
STRAIN TA98, WITHOUT S-9
UNC **—CATALYZED
CAT * *—CATALYZED
SHALE OIL CAT.
S
NO. 1 FUEL UNC *
NO 2 FUEL UNC
S SHALE OIL UNC
3 10 30 100 300 1000 3000
HEXANE INSOLUBLE FRACTION CONCENTRATION (Hg'PLATE)
Fig. B4 - Fuel effects on dose-response curves for
Ames tester strain TA98 without S-9
activation, for ether-insoluble fraction
of SOF, with and without PTX oxidation
catalysts
Fig. B5 - Fuel effects on dose -response curves for
Ames tester strain TA98 without S-9 act-
ivation for hexane-insoluble fraction
of SOF, with and without PTX oxidation
catalysts
-------
70
NO 1 fan CAT • •
SO
30
CATt«MUAR ttOI
PARAf FIN FRACTION
CATACY/fO ANOUNCArALtZCD
STRAIN TAM. WITHOUT S-9
UNC • -UNCATALYZFD
CAT * * -CATALYZED
3 to 30 tOO 300 1000
PARAFFIN FRACTION CONCENTRATION
3000 10,000
Fig. B6 - Fuel effects on dose-response curves for
Ames tester strain TA98 without S-9
activation tor prararrin traction of
SOF, with and without PTX oxidation
catalysts
341
2000
•z.
^ 1600
1200
800
400
i i
CATEnPllLAR 3JO« MODE 4
TRANSITIONAL FRACTION
CATALYnOANDUNCATALVZtD
STHAIN TA9H.WITHOUT S-l
UNC • —UHCATALYZEO
CAT » *— CATAL>/tD
HO }FUH CAI • *
3 10 30 100 300 1000 3000
TRANSITIONAL FRACTION CONCENTRATION (Mg/PLATE)
Fig. B7 - Fuel effects on dose-response curves
for Ames tester strain TA 98 without
S-9 activation for transitional fraction
of SOF, with and without PTX oxidation
catalysts
-------
342
APPENDIX C
NON-ROUTINE CHEMICAL INVESTIGATIONS
CRC Round Robin
In 1980, the Coordinating Research Council (CRC) Chemical
Charactization panel conducted an interlaboratory round robin to investigate
an automated (HPLC) method for diesel SOF characterization. MTU
participated in that round robin, and the results are presented here.
Table Cl.l shows retention time values and relative areas of compound
classes from Nissan and Cummins Engine particulate as determined by the
microparticulate silica gel HPLC method. Figures Cl.l and C1.2 are HPLC
chromatograms for these two extracts. Four replicate runs were taken, and
the best reproducibility was obtained on both samples for the 04 and a2 peaks.
The a-i and a2 peaks are polynuclear aromatic hydrocarbons, and the B peak (not
always seen in our runs) consists of larger (higher molecular weight)
polynuclear aromatic hydrocarbons. The Yi and Y2 peaks consist mainly of
quinone and hydroxy derivatives of polynuclear aromatic hydrocarbons, while
acid, acid anhydride, and aldehyde derivatives of polynuclear aromatic
hydrocarbons comprise the 6 peak.
Table Cl.2 is a pool of data collected by the CRC from ten laboratories
for this HPLC round robin (88). Since a 3 mm-i.d. HPLC column (not
available in time for MTU to participate) and solvent flow rate of 1.0
ml/min was prescribed in the experimental section of this round robin, the
data generated by MTU was not included in this data set. Retention times
were earlier for MTU analyses than the CRC round robin analyses, due to
-------
Table Cl.l: -Setention Time Values and Area Percentages of Compound
Classes from Nissan and Cummins Engine Extracts for Four Replicate
HPLC Runs using a Silica Gel Column, MTU Data only.
Retention Time, Minutes
Nissan Extract
Xia
%C
v
Cummins Extract
Xla
%C
v
6.43±0.82
12.7
1.83±0.26
14.4
9.69±1.26
13.0
6.9210.47
6.8
21.0310.50
2.4
21.6610.52
2.4
28.3410.85
3.0
28.5610.76
1.8
37.601.06
2.8
37.3610.69
1.8
66.1110.42*
0.6*
65.9210.05*
0.08*
Relative Peak Areas
Nissan Extract
Xla
%C
v
Cummins Extract
Xla
%C
v
1.00
0.05910.005
9.0
0.0110.013
134
0.02410.038
158
0.6710.073
10.8
1.00
12
0.69910.074
10.5
0.15010.045
30.6
0.44610.07
15.7
1.2410.26
20.7
*Standard deviation and %C for retention time of 6 peak was this low only because this peak required
acetonitrile injection to elute.
LO
-C-
U)
-------
344
Table C1.2
Summary of Coordinating Research Council HPLC Round-Robin Results (n = 10, ref.88,)
Retention Time (Minutes)
Sample
a
1
a,.
Nissan
Xlo 10.0±1.3 15.4±2.1 24.4+3.0 35.1±2.7 50.6±3.8 56.7±3.6
%C
v
Cummins
Xla
%c
v
13.0
9.311.
19.4
13.6
12.3
7.7
7.5
6.3
14.7+2.7 26.0+2.7 34.4+4.5 49.3+5.1 56.510.5
18.4 10.4 13.1 10.3 0.9
Relative Peak Areas
Nissan
XI CT
%C
v
Cummins
1.00
_
XI a
%C
v
0.21±0.06 0.7710.14 0.7010.11 0.6610.18
28.5 18.2 15.7 27.3
0.0710.01 0.1410.03
14.3 21.4
1.00 0.2010.05 1.4310.19
25.0 13.3
-------
345
-------
346
different column size and flow rate as well as initial solvent composition
which may have varied significantly from that prescribed by the round robin.
Solvent was not pre-mixed by MTU, but was formed by the gradient system of
the liquid chromatograph. Because of the imprecision of such systems at low
percentage composition of one solvent (89), the prescribed 5% hexane in DCM
composition may have been somewhat different from that actually delivered by
the HPLC pumps. Although retention times were less than those observed by
most investigators, relative peak areas (except for the $ peak) were similar
to those of other investigators. The discrepancy in g peak areas could also
be due to difference in solvent composition from that of others since this
was not always observed by MTU, and actually belongs to the a peaks in
chemical character.
Sephadex LH-20 Lipophilic Gel
Table C1.3 is the acid-base-neutral fractionation data for SOF from the
baseline Caterpillar engine run at mode 3 (25% rated load, 1680 RPM) using
No.l fuel and an air volume dilution ratio of 5:1. The same SOF was
fractionated using LH-20 lipophilic gel; results of this fractionation are
shown in Table C1.4. Recoveries of material in both cases were 90% or
better; relative standard deviation was less than 10% for major fractions
and total recovery. Dose-response curves for LH-20 separated fractions are
shown in Figure C1.3. The curve labeled "E" is unfractionated SOF, and that
labeled "BHT" is butylated hydroxytoluene, the solvent preservative in THF.
The majority of biological activity resides in fractions 4 thru 7, which
consititutes approximately 11% of the SOF mass.
-------
347
TABLE C1.3 ACID-BASE-NEUTRAL FRACTIONATION RESULTS
PERCENTAGES OF TOTAL SAMPLE
(AVERAGE SAMPLE SIZE 146 MG)
FRACTION NAME
ETHER INSOLUBLE
BASIC
ACIDIC
PARAFFIN
AROMATIC
TRANSITIONAL
OXYGENATED
HEXANE INSOLUBLE
1-3D5-1
5.1
3.4
3.4
65.6
0.8
4.2
11.2
0.9
1-3D5-2
1.2
0.6
5.0
59.7
4.5
6.3
13.5
0.3
1-3D5-3
3.0
1.3
9.0
61.3
4.6
4.3
11.9
0.4
X
3.1
1.8
5.8
62.2
3.3
4.9
12.2
0.5
a
1.6
1.2
2.4
2.5
1.8
1.0
1.0
0.3
PERCENT RECOVERY
94.6
83.4
91.5
90.0
5.5
-------
348
at
o:
u
a.
LC
UJ
>
UJ
c:
-1400
1200
-1000
•600
•600
-400
-200
FIGURE C1.3
CATERPILLAR ENGINE, MODE 3,
DILUTION 3ATIO 5:1, #l FUEL
(SEPHADEX LH-20 PREP.)
E - SOF (unfr&ctionated)
BHT-Butylated Hydroxytoluene
1-7- Fraction Numbers
I.I7 2345 4.687 9375 18.75 37.5 75 I50 300 600 I200 2400 4800 9SOO
I-3D5-LH20 AQS/PLATE '
FIGURE C1.4
SUBFRACTIONATION OF SAMPLE I-3D5
FRACTION 5
VYDAC 50ITPIO
CYANOPROFYL POLAR BONDED PHASE
10mm 1.0. HPLC COLUMN
6mj SAMPLE INJECTED
HEXANE- ISOPROPANOL GRADIENT
ATTENUATION 256
UV DETECTION, 350nm
ATTENUATION
2048
5 (0 15 20 25
TIME (MINUTES)
30 32
-------
349
Table C1.4 SEPHADEX LH-20 GEL CHROMATOGRAPHY RESULTS
PERCENTAGES OF TOTAL SAMPLES
AVERAGE SAMPLE SIZE 195 MG FOR TRIALS 1-3;
420 MG FOR TRIALS 4 and 5
FRACTION
1
2
3
4
5
6
7
1-3D5-1
4.9
57.1
25.1
3.7
2.9
6.1(a)
-2
5.2
52.7
16.1
4.1
3.3
3.4
0.8
-3
4.4
61.9
20.5
2.7
2.5
4.0
3.0
-4
6.3
54.2
25.0
3.1
2.8
3.7
0.8
-5
4.7
64.0
18.4
2.3
3.2
3.1
1.2
X
5.1
58.0
20.7
3.2
3.9
3.6(b)
1.4(b)
•:a
0.7
4.3
3.1
0.6
0.3
0.3(b)
0.9(b)
PERCENT
RECOVERY 99.5 84.5 97.8 93.3 97.3 94.5 5.4
(a) FRACTIONS 6 & 7 WERE COMBINED IN RUN #1.
(b) FOUR DETERMINATIONS
-------
350
Subfractional:ion of the most active LH-20 separated fraction (A5) from
this SOF was accomplished using a Vydac cyanopropyl polar bonded phase
semi-preparative HPLC column (10 mm I.D.). Figure C1.4 shows the resultant
chromatogram. Capillary gas chromatography of subfractions Bl and B2 of
this fraction is shown in Figures C1.5 and C1.6. These fractions were
subjected to GC/MS analysis using the Hewlett-Packard 5985 GC/MS - data
system; subfraction A5-B1 was also analyzed using the Kratos MS-50 GC/MS at
the Midwest Center for Mass Spectrometry in Lincoln, Nebraska. Due largely
to high background levels, no identifications of compounds in these
subfractions could be made.
-------
351
FIGURE C1.5
SAMPLE I-3D5
FRACTION A5-BI
SE-54, 30METER
FUSED SILICA CAPILLARY COLUMN
SPLITLF.SS INJECTION IN TOLUENE
ATTENUAT ON 8 (AFTER 6 MINUTES)
I20°C ISOTHERMAL, 5 MINUTES
3VMINUTE TO 260°; 10 MINUTE HQLD
T~
10
I
20
30 40
TIME (M/NUTES)
50
60
-------
352
CD
CM
CM
rr>
ID
FIGURE C1.6
SAMPLE I-305
FRACTION A5-B2
SE-54, 30 METER
FUSED SILICA CAPILLARY COLUMN
SPLITLESS INJECTION IN TOLUENE
ATTENUATION 8 (AFTER 6 MINUTES)
120°C ISOTHERMAL, 5 MINUTES
3*/MINUTE .TO 280°; 10 MINUTE HOLD
in
ro
K> 15 20 25 30 35
TIME (MINUTES)
40
45
55
-------
353
Sulfate Analysis By Ion Chromatography Versus the Barium Chloranilate Method
Table C1.5 shows comparative values for four pairs of filters collected
for sulfate analysis by the EPA barium chloranilate method, as well as by
ion Chromatography using a Dionex Model 10 Ion Chromatograph. Each pair of
filters was run for identical length of time in parallel sampling probes and
collected similar amounts of particulate and SOF from the Caterpillar 3208
engine with close-coupled port catalysts installed. Although the
reproducibility of each method is Rood, it appears from these data that the
barium chloranilate method gives a consistently higher sulfate value than ion
Chromatography. To investigate the possibility that extraction efficiency for
isopropanol-water 60:40 (used in the barium chloranilate method) is higher
than that of deionized water (used in ion Chromatography), a series of
samples were extracted using both deionized water and IPA: H 0, 60:40, to
Table C1.5 Comparison of the Barium Chloranilate Coloriraetric Method
with the Dionex Ion Chromatographic Method for Sulfate Analysis
(n = 4 determinations per filter)
Ion Chromatography
Sample mg S0~ mg TPM
4-P-3 5.1210.109 7.61
%C = 2.1
v
4-P-5 5.3810.157 7.27
2.9%
4-P-7 12.1410.209 16.92
1.7%
4-P-ll 12.0010.136 17.95
1.5%
Barium Chloranilate
67.3
74.0
71.8
66.8
Sample mg S0~
mg TPM
4-P-4 6.28±0.084 8.84
1.3%
4-P-6 6.42+0.23
3.5%
8.46
4-P-8 14.6010.73 17.37
5.0%
4-P-12 14.0810.54 18.67
3.8%
%SO~
71.1
75.9
84.0
75.4
-------
354
analyze by ion chromatography. Table C1.6 shows that filters extracted with
IPA: H.O, 60:40 are not extracted more efficiently than those extracted with
deionized water.
Table C106: Effect of Solvent on Sulfate Ion Analysis Using Dionex
Model 10 Ion Chromatograph. Caterpillar 3208, Close-
Coupled Port Catalyst, EPA Mode 4
Sample
20
22
15
14
Solvent
IPA:
IPA:
D.I.
D.I.
H
H
H
H
2
2
2
2
0
0
0
0
mg
6.
6.
8.
12.
TPM
24
13
75
70
mg SO,
4
3
6
8
*+
.05
.62
.16
.74
%SO,
65
59
70
68
*=r
.0
.0
.4
.8
Extraction volume as a possible factor in extraction efficiency was
investigated for the barium chloranilate method, with results shown in Table
C1.7. Smaller volumes than the 25 ml normally employed do not show
consistent effects on the percentage sulfate removable from filters.
Likewise, reducing the particulate loading demonstrates no consistent
solvent effect except perhaps the greater efficiency of the larger
extraction volume (Table C1.8).
Recoveries of sulfate solution added to aliquots of various filter
extracts are shown in Table C1.9. Recoveries ranged from 92.9 to 105.9%.
Similarly, recoveries for sulfate added as extract to filters whose original
sulfate loading was calculated from most similarly-loaded, non-spiked
filters varied between 97.0 and 102.1 percent (Table C1.10), well within the
statistical variability of the method.
-------
355
Table C1.7: Effect of Solvent Volume on Percentage Sulfate
Measured using the Barium Chloranilate Method. Caterpillar 3208,
Close-Coupled Port Catalysts, EPA Mode 4
Sample
21
24
13
17
Solvent
IPA: HO,
10 ml
IPA: HO,
10 ml
IPA: HO,
25 ml
IPA: HO,
25 ml
mg TPM
3.192
6.204
8.89
8.43
mg SO
-T
2.43
3.69
5.88
5.84
%so4
76.1
59.5
66.2
69.3
Table C1.8: Effects of Solvent Volume and Composition on
Percentage Sulfate Measured with Light Particulate Loadings,
Caterpiller 3208, Close-Coupled Port Catalyst, EPA Mode 4
Barium Chloranilate
Sample
32
30
Solvent
IPA: HO
10 ml
IPA: HO
25 ml
mg TPM
0.309
0.312
mg S0=
0.198
0.250
%so4
63.9
80.2
Ion Chromatography
27
29
IPA: HO
10 ml
D.I. HO
25 ml
0.328
0.401
0.203
0.306
61.9
76.2
-------
356
The precision and true chromatographic nature of ion chromatography, as
well as its simplicity of execution relative to the barium chloranilate
method convinced us to adopt the method in March, 1981.
Table C1.9: Recoveries of Sulfate Added to Aliquots of Filter Extract
Using Both Sulfate Analysis Methods. Caterpillar 3208, Close-
Coupled Port Catalysts, EPA Mode 4
Sample Method PPM SO? in Extract PPM S07 Added % Recovery
" tf
17 BCA 233.4 250 92.9
21 BCA 97.2 125 98.4
20 1C 162.0 250 105.9
14 1C 349.6 250 104.0
Table C1.10: Recoveries of Sulfate Added to Filters Whose Sulfate
Levels Were Estimated from the Loading of Other,
Similarly Loaded, Filters: Extracted With Standard
Sulfate Before Analysis, Caterpillar 3208, Close-Coupled
Port Catalysts, EPA Mode 4.
Sample Method PPM S0~ on filter mg SO Added % Recovery
23 BCA 3.13' 2.50 98.8
16 BCA 8.15 6.25 102.1
19 1C 3.11 2.50 97.0
18 1C 7.90 6.25 99.6
-------
357
TABLE Cr-,2
MACK ENDT-676 SOF CHEMICAL FRACTIONATION DATA
TEST (a
ACD (%
BSACD
ARM (%
BSARM
BAS (%
BSBAS
BIN (%
BSEIN
HIN (%
BSHIN
OXY (%
BSOXY
PRF (%
BSPRF
TRN (%
BSTRN
)
of SOF)
(mg/kw-hr)
of SOF)
(mg/kw-hr)
of SOF)
(mg/kw-hr)
of SOF)
(mg/kw-hr)
of SOF)
(mg/kw-hr)
of SOF)
(mg/kw-hr)
of SOF)
(mg/kw-hr)
of SOF)
(mg/kw-hr)
3B
5.7
10.3
3.5
6.35
2.7
4.90
13.3
2.41
4.2
7.62
13.6
24.7
49.2
89.3
7.8
14.2
G-19-3
7.
3.
4.
2.
2.
1.
5.
3.
2.
1.
19
9.
55
28
3.
1.
5
76
0
03
1
07
9
0
4
22
.4
86
.1
.0
8
93
G-12-3
14.4
4.26
3.8
1.13
2.4
0.71
6.1
1.81
3.2
0.95
13.4
3.97
52.3
15.5
4.5
1.33
G-17-3
9.
5.
2.
1.
4.
2.
7.
4.
3.
2.
13
8.
52
31
6.
3.
5
66
9
73
3
56
0
17
7
21
.9
29
.7
.4
1
64
G-22-3
3.5
4.39
2.8
3.52
0.7
0.88
8.0
10.05
1.7
2.13
12.3
15.4
65.9
82.7
5.2
6.53
G-19-9
8.6
5.19
4.5
2.71
2.1
1.27
15.5
9.35
3.3
1.99
26.1
15.7
36.0
21.7
3.3
1.99
G-12-9
8.
3.
3.
1.
1.
0.
6.
2.
3.
1.
27
10
43
16
5.
2.
3
06
6
33
7
62
3
32
2
18
.9
.3
.3
.0
7
10
(a)
A complete description of engine and fuel injection components corresponding
to the test numbers is given in Tables 3.1.2 and 3.1.3
-------
358
TABLE C-2 (cont'd)
MACK ENDT-676 CHEMICAL FRACTIONATION DATA
TEST(a)
ACD (%
BSACD
ARM (%
BSARM
BAS (%
BSBAS
EIN (%
BSEIN
HIN (%
BSHIN
of SOF)
(mg/kw-hr)
of SOF)
(mg/kw-hr)
of SOF)
(mg/kw-hr)
of SOF)
(mg/kw-hr)
of SOF)
(mg/kw-hr)
OXY (%of SOF)
BSOXY
PRF (%
BSPRF
TRN (%
BSTRN
(mg/kw-hr)
of SOF)
(mg/kw-hr)
of SOF)
(mg/kw-hr)
G-17-9
3.9
2.91
2.2
1.64
6.8
5.07
5.8
4.32
3.2
3.35
27.5
20.49
46.1
34.4
3.2
2.38
G-22-9
4.3
4.66
10.6(b)
-i 1 t^f**'
1.3
1.41
7.6
8.24
4.5
4.79
18.4
19.91
53.5
58.0
(b)
(b)
G2-12-5
10.6
1.38
6.2
1 0.81
0.95
0.12
4.5
0.58
4.4
1.07
5.0
0.65
64.0
8.32
0.41
0.05
G2-17-5
12
2.
7.
1.
0.
0.
7.
1.
8.
1.
6.
1.
62
10
1.
0.
.0
04
5
28
87
15
7
31
3
30
1
04
.8
.7
59
27
G2-22-5
23
3.
10
1.
2.
0.
8.
1.
7.
0.
7.
1.
49
7.
0.
0.
.9
56
.1
51
17
32
9
32
6
62
1
06
.4
35
28
04
G2-12-11 G2-17-11
19.7
27.0
4.1
5.6
2.32
3.20
9.3
12.7
4.2
16.3
16.8
23.1
34.1
46.8
3.77
5.20
17
33
4.
9.
1.
3.
7.
13
11
11
16
31
46
89
2.
4.
.2
.5
8
3
53
00
0
.7
.9
.5
.4
.9
.0
.7
41
70
A complete description of engine and fuel injection components corresponding
to the test numbers is given in Tables 3.1.2 and 3.1.3
^ ARM and TRN combined due to lack of distinct ARM band
-------
359
TABLE C-2 (cont'd)
MACK ENDT-676 CHEMICAL FRACTIONATION DATA
TEST(a)
ACD (%
BSACD
ARM (%
BSARM
BAS (%
BSBAS
EIN (%
BSEIN
HIN (%
BSHIN
OXY (%
BSOXY
PRF (%
BSPRF
TRN (%
BSTRN
of SOF)
(mg/kw-hr)
of SOF)
(mg/kw-hr)
of SOF)
(mg/kw-hr)
of SOF)
(mg/kw-hr)
of SOF)
(mg/kw-hr)
of SOF)
(mg/kw-hr)
of SOF)
(mg/kw-hr)
of SOF)
(mg/kw-hr)
G2-22-11 G3-22-11 S-8-3
14
.7
26.1
4.
7.
1.
2.
7.
13
5.
8.
12
21
53
95
1.
2.
0
0
6
9
5
.3
0
9
.4
.9
.8
.2
19
1
17.3
33.8
4.9
9.6
1.6
3.1
6.6
12.9
5.9
11.5
16.3
31.7
44.5
86.8
2.87
5.6
16.9
21.9
4.7
6.1
1.9
2.5
46.6
60.6
4.5
5.9
11.1
14.4
11.5
15.0
2.80
3.6
S-12-3
4.8
4.8
27.7
27.3
2.5
2.5
16.6
16.6
3.0
3.0
16.5
16.5
25.1
25.1
4.30
4.3
S-14-3
5.
5.
14
14
1.
1.
4.
4.
1.
1.
18
18
52
52
3.
3.
1
1
.1
.1
6
6
5
5
7
7
.1
.1
.4
.4
00
0
o_
6.
7.
5.
6.
2.
3.
11
13
1.
1.
16
20
46
55
9.
11
17-3
2
4
7
8
8
4
*
•
2
4
•
•
•
•
6
9
9
3
3
6
40
•
3
(a)
A complete description of engine and fuel injection components corresponding
to the test numbers is given in Tables 3.1.2 and 3.1.3
-------
TABLE C~3
360
CATERPILLAR 3208 SOF CHEMICAL FRACTIONATION
DATA: ENGLEHARD PTX CATALYST EVALUATION
MODE
FUEL
PTX CATALYST
ACD (% of SOF)
BSACD (mg/kw-hr)
ARM (% of SOF)
BSARM (mg/kw-hr)
BAS (% of SOF)
BSBAS (mg/kw-hr)
BIN (% of SOF)
BSEIN(mg/kw-hr)
HIN (% of SOF)
BSHIN (mg/kw-hr)
OXY (% of SOF)
BSOXY (mg/kw-hr)
PRF(% of SOF)
BSPRF (mg/kw-hr)
TRN(% of SOF)
BSTRN (mg/kw-hr)
EPA MODE 3
NO. 2
W/0 W
10.5
79.8
3.7
28.1
3.3
25.1
7.5
57.0
2.3
17.5
11.8
89.7
55.3
420.3
5.6
42.6
3.9
11.7
6.3
18.9
2.1
6.3
9.0
27.0
1.8
5.4
13.0
39.0
60.0
180.0
4.0
1.2
NO.l
W/0
11.0
82.9
3.1
23.4
1.9
14.3
3.1
23.4
3.7
27.9
13.1
98.8
61.6
464.4
2.4
18.1
W
11.5
34.2
3.5
10.4
2.4
7.1
4.6
13.7
0
0
15.9
47.2
59.2
175.8
2.9
8.6
NO. 2
W/0 W
4.5
9.5
4.2
8.8
1.0
2.1
9.4
10.7
0.6
1.3
12.2
25.6
63.3
132.9
4.9
10.3
EPA MODE 4
7.1
3.6
1.2
0.6
1.7
0.9
14.2
7.1
1.4
0.7
24.8
12.4
45.2
22.6
4.4
2.2
NO.l
W/0
8.5
18.2
5.5
11.8
1.9
4.1
3.1
6.6
0
0
13.0
27.8
60.8
130.1
5.8
12.6
W
13.4
7.5
2.9
1.6
2.9
1.6
7.9
4.4
6.7
3.8
13.0
7.3
46.7
26.2
6.6
3.7
SHALE
W/0
18.8
76.3
8.9
36.1
2.9
11.8
12.7
51.6
1.8
7.3
17.0
69.0
33.2
134.8
4.6
18.7
W
7.4
4.5
18.9
11.5
2.8
1.7
6.1
3.7
3.3
2.0
12.4
7.6
45.7
27.9
3.3
2.0
-------
TABLE C-3(continued)
361
CATERPILLAR 3208 SOF CHEMICAL FRACTIONATION DATA
ENGLEHARD PTX CATALYST EVALUATION
MODE
FUEL
PTX CATALYST
ACD (% of SOF)
BSACD (mg/kw-hr)
ARM (% of SOF)
BSARM (mg/kw-hr)
BAS (% of SOF)
BSBAS (mg/kw-hr)
EIN (% of SOF)
BSEIN(tng/kw-hr)
HlN (% of SOF)
BS3IN (mg/kw-hr)
OXY (% of SOF)
BSOXY (mg/kw-hr)
PRF(% of SOF)
BSPRF (mg/kw-hr)
TRN(% of SOF)
BTRN (mg/kw-hr)
EPA MODE 5
NO. 2
W/0
12.6
3.8
0.6
0.2
1.5
0.5
14.7
4.4
1.7
0.5
14.4
4.3
49.7
14.9
4.7
1.4
W
28.5
2.9
2.4
0.2
4.5
0.5
13.3
1.3
1.8
0.1
28.7
2.9
14.5
1.5
6.3
0.6
NO.l
W/0
10.4
2.7
2.1
0.5
2.3
0.6
4.9
1.3
1.3
0.3
14.1
3.7
55.3
14.4
9.6
2.5
W
12.5
1.1
8.4
0.8
1.3
0.1
25.2
2.3
1.2
0.1
23.8
2.1
22.8
2.1
6.1
0.5
EPA MODE 9
NO
W/0
21.1
8.4
1.6
0.6
1.8
0.7
3.0
1.2
5.4
2.2
53.5
21.4
10.6
4.2
3.1
1.2
.2
W
24.2
2.4
3.6
0.4
1.7
0.2
29.7
3.0
0.9
0.1
17.9
1.8
17.6
1.8
17.9
1.8
EPA MODE 10
NO.
W/0
9.6
23.0
3.6
8.6
2.6
6.2
17.8
42.7
3.7
8.9
20.1
48.2
34.6
83.0
8.0
19.2
2
W
26.9
8.1
3.2
1.0
4.6
1.4
2.3
0.7
4.8
1.4
23.8
7.1
29.7
8.9
3.2
1.0
EPA MODE
NO. 2
W/0
15.5
152.0
3.3
32.3
2.7
26.5
6.6
64.7
1.7
16.7
19.2
188.2
44.6
437.1
6.4
62.7
W
8.
14
4.
6.
1.
2.
11
18
4.
7.
24
38
39
63
5.
8.
11
9
.2
2
7
7
7
.8
.9
6
4
.1
.6
.6
.4
3
5
-------
362
Table C -4 CATERPILLAR 3208 SOF CHEMICAL FRACTIONATION DATA:JOHNSON-MATTHEY
CLOSE-COUPLED PORT CATALYST AND CORNING TRAP EVALUATION, NO.2 FUEL
EPA MODE
AFTERTREATMENT
ACD (% of SOF)
BSACD (mg/kw-hr)
ARM (% of SOF)
BSARM (mg/kw-hr)
BAS (% of SOF)
BSBAS (mg/kw-hr)
EIN (% of SOF)
BSEIN(mg/kw-hr)
HIN (% of SOF)
BSHIN (mg/kw-hr)
OXY (% of SOF)
BSOXY (mg/kw-hr)
PEP(% of SOF)
PRF (mg/kw-hr)
TRN(% of SOF)
BSTRN (mg/kw-hr)
NONE
5.0
46.
5.5
51.
0.8
7.
5.1
48.
3.1
28.
10.8
100.
66.7
622.
3.1
29.
3
PORT
CATALYST
6.1
37.
4.9
30.
1.3
8.
12.2
74.
5.2
32.
12.9
78.
54.1
330.
3.2
20.
TRAP
6.3
62.
3.6
35.
0.9
9.
7.7
77.
5.5
53.
7.5
73.
66.2
643.
2.0
20.
NONE
4.2
16.
2.4
9.
2.7
10.
8.6
32.
1.4
5.
16.7
63.
58.5
220.
5.7
21.
4
PORT
CATALYST
6.5
7.
8.7
10.
0.7
1.
15.1
17.
3.4
4.
13.1
15.
48.5
56.
4.0
5.
TRAP
9.4
13.
2.8
4.
0.5
1.
3.2
4.
1.0
1.
6.5
9.
73.0
100.
3.5
5.
5
PORT
NONE CATALYST TRA£
9.4
4.
3.4
1.
2.4
1.
6.9
3.
5.6
2.
16.4
6.
50.0
19.
5.9
2.
20.4
5.
10.5
3.
0.6
0
8.2
2.
1.3
0.
11.6
3.
44.3
11.
3.2
1.
(a) No SOF samples obtained due to extremely low levels.
-------
363
Table C-4 CATERPILLAR 3208 SOF CHEMICAL FRACTIONATION DATA-'JOHNSON-MATTHEY
CLOSE-COUPLED PORT CATALYST AND CORNING TRAP EVALUATION, NO.2 FUEL
EPA MODE
AFTERTREATMENT
ACD (% of SOF)
BSACD (mg/kw-hr)
ARM (% of SOF)
BSARM (mg/kw-hr)
BAS (% of SOF)
BSBAS (mg/kw-hr)
EIN (% of SOF)
BSEIN (mg/kw-hr)
HIN (% of SOF)
BSHIN (mg/kw-hr)
OXY (% of SOF)
BSOXY (mg/kw-hr)
PRF(% of SOF)
PRF (mg/kw-hr)
TRN(% of SOF)
BSTRN (mg/kw-hr)
NONE
18.0
5.
7.9
2.
1.7
0.
10.7
3.
3.9
1.
12.3
3.
41.6
11.
3.9
1.
9
PORT
CATALYST TRAP
7.7 -(a)
1.
8.1
1.
5.1
1.
37.7
4.
1.1
0.
14.3
2.
21.3
2.
5.3
1.
NONE
8.6
11.
0.5
1.
3.7
5.
11.6
15.
7.6
10.
19.9
27.
41.8
56.
6.4
9.
10
PORT
CATALYST
14.3
17
3.4
4.
1.2
1.
18.4
22.
4.8
6.
24.0
9.
30.7
37.
3.3
4.
11
PORT
TRAP NONE CATALYST
-(a) 7.0
67.
3.0
29.
1.1
10.
5.2
50.
2.4
23.
23.0
223.
51.7
500.
6.9
67.
6.4
18.
5.8
16.
0.4
1.
20.9
59.
5.0
14.
17.2
48.
39.2
110.
5.1
14.
TRAP
33.6
22.
1.0
1.
1.3
1.
10.0
7.
1.7
1.
7.6
5.
43.3
29.
1.4
1.
(a) No SOF samples obtained due to extremely low levels.
-------
364
Ld
CJ
C£
UJ
Q_
CO
2
O
CJ
cr
m
ID
CO
o
CO
CD
•
O
CO
LD
CO
O
•
O
CM
O
•
LO
O
•
O
o
•
LO
T
T
SOF SUBFRflCTIONS VS.
MODE 3
EIN-ETHER INSOLUBLE
TRN-TRRNSITIONflL
OXY-QXYGENflTED
HXI-HEXRNE INSOLUBLE
flRO-flROMflT I C
BflS-BRSIC
RCD-flCIDIC
PRF-PHRflFFIN
ENGINE TIMING
J I
OXY - Shuttle
EIN - Shuttle
TRN - Shuttle
ACD - Shuttle
ARM - Shuttle
PRF/10 - Shuttle
BAS - Shuttle
BIN - Shuttle
I I I 1 L
5 .0
10 .0
15 .0
20 .0
25 .0
ENGINE TIMING, DEGREES BTC
Fig. C2 - Trend in chemical subtractions as
a percent of total SOF with engine timing,
Mack ENDT-676 equipped with APS "ultra-
high rate" shuttle pump, mode 3
-------
365
LJ
CJ
CY
LL)
Q_
CO
Z
o
CJ
cr
LY
LJ_
m
ID
CO
L_
O
CO
CD
CO
CM
CD
CM
O
•
LO
CD
•
CD
CD
•
LO
SOF SUBFPflCTIONS VS.
rn
ENGINE TIMING
MODE 3
FIN-ETHER INSOLUBLE
TRN-TPRNSITIONRL
OXY-OXYGElIfiTED
HXI-HEXPNE INSOLUBLE
HRO-flROMflTIC
BflS-BRSIC
RCD-RCIDIC
PRF-PRRRFFIN
6G - ACD
6G - OXY
6G - EIN
6G - PRF/10
6G - TRN
6G - BAS
i i i i
5 .0
10 .0
15 .0
20 .0
25 .0
ENGINE TIMING, DEGREES BTC
Fig« C3 - Trend in chemical subfractions as
a percent of total SOF with engine timing,
Mack ENDT-676 equipped with APE-6G "high
rate" pump, mode 3
-------
366
"I "I" !"" "I T * I 1 - 'I "I
SOF SUPFPPCTIONS V
MODE 5
EIN-ETriEP INSOLUBLE RRO-RROMfiTIC
! PN- TPnMS IT I u| p,L BflS-Bfl? IC
.:,,i u.-VOHUfiTtlD ;j,CD fiCIDIC
MXI-HEXHNE IN50LUBLE fRF-PflPRFFIN
NGIhF
L_
LJJ
CJ
CY
LJ
Q_
CO
u
CE
LD
Z)
CO
O
CO
O
CD
•
CD
CD
LO
6G - ACD
6G - IIXI
6G
6G
6G
OXY
BIN
10 .0
15 .0
0 .0
25 .0
ENGINE TIMING, DEGREES BTC
Fig. C4 - Trend in chemical subtractions as
a percent of total SOF with engine timing,
Mack ENDT-676 equipped with APE-6G "high
rate" pump, mode 5
-------
367
UJ
U
ClL
LJ
CL-
O
U
CE
Q/
L_
CD
ID
01
LL
O
CO
CD
00 -
Li")
f\!
O
•
c^
CXI
CD
•
LO
CD
•
CD
CD
•
LO
c. n r C i • u; n r p| r;
MODE 9
EiM-EFHER INSXUBLE
rPN-Tc-RN0] I IONRL
L,,;; 0/YGENriiED
HXI-HEXRNE INSOLUBLE
6G - OXY
R '••_ !' ir
RRO-RROMRTIC
BRS-BR-K
PRF-PHRRFF1N
6G - ACD
5 .0
10 .0
15 .0
2Q .0
25 .0
ENGINE TIMING, DEGREES BTC
Fig. C5 - Trend in chemical subfractions as
a percent of total SOF with engine timing,
Mack ENDT-676 equipped with APE-6G "high
rate" pump, mode 9
-------
368
o
. «
o
CO
I
,
I - • I -I'l
:-.nr c.
MOPC: 1 1
IMI-rMLi' lfl.ti'U(LE
T(-M- fh'iH'.I] ! -I'-L
0..' --1. f OL! «M 1! n
nx]-,tLxiiNt IH.-OLUOLI:
I I "I • T -"I -T ~T " I '
. EN'C.INE TIMING.
10
Cl
LU
CJ
-------
o
oo
369
i 1 1 1 1 1 1 1 i 1 1 r~
SOF SUBFRflCTIONS VS. ENGINE LORD
ENGINE TIMING: 12 DEGREES BTC
EIN-ETHER INSOLUBLE
TRN-TRRNSITIONRL
OXY-OXYGENflTED
HXI-HEXHNE INSOLUBLE
RRO-RROMHTIC
BHS-8RSIC
flCD-RCIOIC
PRF-PRRRFFIN
LJ
O
C£
LJJ
Q,
CO
2
O
o
•
o
CM
CJ
CT
CK
U_
CD
ID
CO
u_
o
CO
ACD
OXY
o
•
o
EIN
PRF/10
TRN
A.RO
3X1
BAS
0.0
500 .0
1000 .0
LORD, N-M
Fig. C7 - Trend in chemical subtractions as
a percent of total SOF with engine load,
Mack Endt-676 equipped with APE-6G "high
rate" pump, modes 3 and 5, 12°BTC
-------
370
LiJ
CJ
C£
UJ
Q_
CO
•z.
o
CE
LY
LL
co
ID
CO
u_
o
CO
o
•
o
CO
o
•
o
CVJ
o
•
o
T
T
SOF SUBFRflCTIONS VS. ENGINE LORD
ENGINE TIMING: 17 DEGREES BTC
EIN-ETHER INSOLUBLE
TRN-TRfiNSITIONRL
OXY-OXYGENflTED
HXI-HEXflNE INSOLUBLE
flRO-flROMRTIC
ERS-BRSIC
nCD-RCIDIC
PRF-PHRflFFIN
OXY
0.0
500 .0
1000 .0
LORD, N-M
Figo C8 - Trend in chemical subfractions as
a percent of total SOF with engine load,
Mack ENDT-676 equipped with APE-6G "high
rate" pump, modes 3 and 5, 17°BTC
-------
371
CD
CO
T
T
SOF SUBFRflCTIONS VS. ENGINE LORD
ENGINE TIMING: 22 DEGREES BTC
EIN-ETHER INSOLUBLE RRO-flROMRTfC
TRN-TRflNSITIONnL 6RS-BRSIC
OXY-OXYGENRTED RCO-RCIDIC
HXI-HEXRNE INSOLUBLE PRF-PRRRFFIN
z:
UJ
o
on
LJ
OL
CO
•z.
O
o
CE
CD
ID
CO
LL
O
CO
O
CM
OXY
O
•
o
EIN
PRF/10
TRN
ACD
ARO
HXI
,BAS,
0.0
500 .0
1000 .0
LORD, N-M
Fig. C9 - Trend in chemical subfractions as
a percent of total SOF with engine load,
Mack ENDT-676 equipped with APE-6G "high
rate" pump, modes 3 and 5, 22°BTC
-------
372
Z
UJ
CJ
o:
u
CL
co
u
CE
&.
\JL
GO
ID
CO
LL,
o
CO
T
T
SOF SUBFRflCTIONS VS. ENGINE LORD
ENGINE TIMING: 12 DEGREES BTC
CD
•
O
oo
EIN-ETHER INSOLUBLE RRO-flROMRTIC
TRN-TRRNSITIONRL BRS-BFISIC
OXr-OXYGENRTED RCO-HCIDIC
HXI-HEXRNE INSOLUBLE PRF-PBRflFFIN
O
•
o
CM
ACD
OXY
O
O
HXI
EIN
ARQ
TRN
PRF/10
BAS
i
1
1
0.0
250 .0
500 .0
750 .0
LORD, N-M
Fig. CIO - Trend in chemical subtractions as
a percent of total SOF with engine load,
Mack ENDT-676 equipped with APE-6G "high
rate" pump, modes 11 and 9, 12°BTC
-------
373
o
oo
UJ
O
(Y.
LJ
0_
00
•z.
o
o
cr
(x
LJL
CO
Z)
o
tn
o
•
o
SUBFRRCTIONS
DEGREES
EIN-ETHER INSOLUBLE
TRN-TRfiNSITIONflL
OXY-OXYGENRTED
HXI-HEXRNE INSOLUBLE
flPO-HROMflTIC
BflS-BflSIC
RCO-flCIDIC
PRF-PflRRFFIN
' I 1 '
0.0
250 .0
500 .0
750 .0
LORD, N-M
Fig. Cll - Trend in chemical subfractions as
a percent of total SOF with engine load,
Mack ENDT-676 equipped with APE-6G "high
rate" pump, modes 11 and 9, 17°BTC
-------
374
I I T I II I I IIIII IT^
SOF SUBFRRCTICNS VS. ENGINE LORD
ENGINE TIMING: 22 DEGREES BTC
EIN-ETHER INSOLUBLE
TRN-TRRNSITIONflL
OXY-OXYGENfiTEO
HXI-HEXHNE INSOLUBLE
flRO-flRGMFITIC
BflS-BRSIC
HCD-fiCIDIC
PRF-PRRRFFIN
o
00
h-
z
LU
UJ
Q_
CO
Z
o
o
cr
C£
LJL
QQ
ID
CO
U_
O
CO
o
-------