EPA-AA-TEB-81-22
Evaluation of Mutagenic Characteristics of
Diesel Gaseous Hydrocarbons
by
Thomas J. Penninga
July, 1981
Test and Evaluation Branch
Emission Control Technology Division
Office of Mobile Source Air Pollution Control
U.S. Environmental Protection Agency
-------
Background
The organic materials extracted from diesel particulate produce a
positive mutagenic response in the Ames Salmonella test. Since a small
amount of high molecular weight hydrocarbon material passes through
particulate filters in a gaseous form, there has been a persistent
concern that these gaseous organics may also be mutagenic.
Several attempts at developing a test procedure to collect the gaseous
organics behind particulate filters have failed due to inadequacy of the
hydrocarbon absorbing media. This study evaluates previously exposed
diesel particulate filters as a collection medium after the filters have
been baked at about 1000°F to remove the particulate-bound organics.
Test Procedure
The test procedure involves collection of diesel particulate on
fluorocarbon coated glass fiber filters, baking off the particle-bound
hydrocarbons, and then exposing the baked filter to diesel gaseous
hydrocarbons.
A. Collection System
The particulate collection system consisted of a 20" by 20"
fiberglass filter commonly used to collect samples for subsequent
Ames Test analysis. The filter was positioned as shown below in a
fine wire mesh filter holder designated the primary filter location.
Exhaus t in
#2 aawcup
filterholder
#*-6ackup Primary filter
filterholder holder j i
i I
Bypass
Damper
Gaseous Hydrocarbon Collection System
Diluted diesel exhaust was pulled through the filter. The bypass
damper was adjusted to maintain a maximum filter temperature of
125°F. This resulted in approximately 1/3 of the flow passing
through the filter and 2/3 bypassed. The exhaust gas was pulled
through the filter via a constant volume sampler. The CVS was not
calibrated so total flow through the filter was not measured.
Therefore, revertent/mile calculations are not possible. However,
filter to filter comparisons can be made. The exhaust from an
Oldsmobile diesel was collected for the backup filters using the
Highway Fuel Economy Test (HFET) driving schedule. The filters were
removed and weighed. These weights minus the original filter weight
gave the net filter loading.
-------
The primary filter holder is followed by two identical filter
holders. When the previously collected filters had been baked at
1000°F for one hour they were put into the #1 and #2 back-up filter
holders. A double thickness clean filter was installed in the
primary filter holder. A test sequence, FTP or HFET, was then run
drawing exhaust gas through all 3 filter holders in series as shown
above. The filters were removed and weighed. The difference between
the post-test weights and the after-baking weight was the weight of
absorbed gaseous hydrocarbons and any diesel particulate which passed
through the double primary filter.
Preliminary testing using all 3 filter holders with clean filters
verified a very minimal weight gain on the #1 and #2 backup filters.
This weight gain was on the order of .05 grams. The actual results
of this system checkout are given below:
System Checkout Test Results
Filter Position Cycle Before Wt. (gms) After wt. (gms) weight gain (gms)
#1 Backup HFET 19.68 19.73 +.05
#2 Backup HFET 19.55 19.61 +.06
The fact that the #2 backup filter actually had a larger weight gain
than the #1 backup filter indicates that the weight gain may not be
from diesel particulate but from absorption of gaseous hydrocarbons
on the clean filter. Such absorbed hydrocarbons would be driven off
during the baking process. The primary filter was a single filter.
All subsequent tests were run with a double primary filter. With the
double primary filter and with the baking procedure, it was felt that
the .05 - .06 gram weight gain would not significantly affect the
results of the experiment.
B. Baking System
A problem encountered with baking the filters was the readiness with
which the diesel particulate oxidized. To avoid this oxidation the
following system was used. The 20"x20" filter was placed on the
inner circumference of a 6 inch diameter glass tube which was sealed
on one end. A tight fitting glass top was installed on top of the
tube. In the top was a small hole through which a stainless steel
nitrogen purge line and a J type thermocouple were positioned. This
container was then placed in a 24 inch diameter ceramic kiln. The
purge and thermocouple lines were attached to an N2 cylinder and a
0-2000°F J type thermocouple recorder, respectively. The oven was
then heated and controlled by the operator at 1000°F for one hour.
About two hours were needed to raise the oven to 1000°F, bake for one
hour, and cool down the oven.
C. Testing Samples
The actual test samples were divided into five groups.
1. Group 1 - Not Baked
-------
This group consisting of sample numbers 320, 330, &nd 340 was
not baked. After collection, the filters were weighed and sent
for extraction of organics. These filters served as a basis for
determining the effectiveness of the baking procedure.
2. Group 2 - Baked not Exposed
This group, consisting of sample numbers 350, 360, and 370, was
baked and sent directly for extraction. These filters compared
to Group 1, showed the feasibility of driving off the
extractable organics by baking. They also gave a baseline of
extractable organics for the filters exposed co gaseous
hydrocarbon.
3. Group 3 - Oldsmobile Results
This group, consisting of sample numbers 380, 390, 400, and 410,
were the baked filters put into backup filter holders #1 and #2
for either one Federal Test Procedure (FTP) or one HFET cycle.
The exhaust gas was drawn through all 3 filter holders during
the complete test sequence using an Oldsmobile Diesel as the
test vehicle.
4. Group 4 - Mercedes Benz Results
This group, consisting of sample numbers 420, 430, 440, and 450,
is identical to group 3 but run using a Mercedes Benz as the
test vehicle.
5. Group 5 - Background Air Results
This group, consisting of sample numbers 460, 470, 480 and 490,
was the same as groups 3 and 4 except no test vehicle was used.
Background air was drawn through the filters for the length of
either an FTP or an HFET sequence. These samples were taken to
verify that the extracted organics were due to the vehicle
exhaust and not to ambient air hydrocarbons.
Test Results
The test result analysis will be divided into four sections covering
filter weights, extractable organic, BAP, and Ames Analysis.
A. Filter Weights
The filters were weighed in a clean condition with the initial
particulate loading, after baking, and after exposure to the gaseous
diesel exhaust. The filter weights are given below:
-------
Group 1
Group 2
Group 3
Group 4
Group 5
Table I
Filter Weights
All Weights in Grams
Initial
Weight
Fresh Filter Weight Baked Loss Exposed Absorbed
Filter Loaded from Filter from Filter Weight
Filter No. Weight Weight Loading Weight Baking Weight Gain
520 19.62 21.96 2.34 ~N7A~ - N/A
330 19.56 20.69 1.13 N/A - N/A
340 19.55 20.47 .92 N/A - N/A
350
360
370
19.87
19.67
19.59
21.31
20.82
20.77
,44
,15
1.11
20.74
20.29
20.32
.57
.53
.38
N/A
N/A
N/A
380-IF* 19.73
390-2F* 19.82
400-1H 19.66
410-2H 19.62
420-IF 19.50
430-2F 19.46
440-1H 19.48
450-2H 20.21
460-1F 20.23
470-2F 20.24
480-1H 20.18
490-2H 20.10
22.18
20.73
21.27
21.24
20.83
20.88
20.82
22.07
2.45
.91
1.61
1.62
1.33
1.42
1.34
1.86
21.49
20.15
20.79
20.83
20.55
20.58
20.58
21.72
.69
.58
.48
.41
.28
.30
.27
.35
21.62
20.35
21.05
21.04
20.69
20.69
20.73
21.94
.13
.20
.26
.21
.14
.11
.18
.22
22.31
21.30
22.60
22.56
2.08
1.06
2.42
2.46
21.97
21.03
22.23
22.14
.34
.27
.37
.42
21.98
21.06
22.31
22.02
.01
.03
.08
-.12
*The number indicates the backup filter location, either #1 backup or #2
backup. The letter indicates the test sequence used during the exposure
of this filter; F = FTP, H = HFET.
Graphical presentation of the filter weights is given in Figures 1-4.
The weight gains compared to the 100% Dilution Air samples indicate
clearly that the filters exposed to the gaseous exhaust did achieve a
positive weight gain.
B. Ex tractable Organic Weight Analysis
The extractable organic weights are given below.
percentage of extractable organic to weight gain.
Also calculated is the
-------
Table II
Extractable Organics (all weights in grams)
Group 1
Un-Baked
Group 2
Baked but
Not Exposed
Group 3
Oldsmobile
'Group 4
Mercedes
Benz
Group 5
Background
Air
Filter No.
320
330
340
350
360
370
380
390
400
410
420
430
440
450
460
470
480
490
Filter Weight Gain Ext. Organic Wt.
.33005
.14995
.17259
.0030
.0052
.00140
% Ext/Wt. Gain
.13
.20
.26
.21
.14
.11
.18
.22
.01
.03
.08
.12
.16215
.04655
.12038
.08805
.05866
.04838
.06837
.06070
.00562
.00815
.00660
.00506
124.7%
23.3%
46.3%
41.9%
41.9%
43.9%
37.9%
27.6%
56.2%
27.2%
8.25%
4.22%
These results are not easily understood. It appears that filter #380 did
not have all of the extractable organic baked off since the extractable
weight was higher than the total weight gain. Therefore, results from
that filter are questionable. The percent extractable/weight-gained
figures average 37.5% for the exposed filters (other than filter #380).
The reason for this low percentage is probably due to desorbtion of light
hydrocarbons between the time of weighing and the time of extraction even
though after weighing, the filters were refrigerated at approximately 0°F
to minimize this desorbtion. The percentages for the 100% Dilution Air
are wide spread. However, considering the very low filter weight gains,
the percentage figures for this group are not considered to be very
meaningful.
The data does indicate that the baking procedure was very effective at
driving off extractable organics (other than filter #380). Filter
numbers 350, 360, and 370 indicate almost no extractable organic. The
four filters exposed to 100% dilution air show that very little
extractable organic was adsorbed from background air. These four filters
will serve well as a control for comparison with other filters.
Unfortunately, by baking the filters and driving off the extractable
organics, no Ames analysis can be performed on these control filters.
Therefore, the results discussed later are based on the assumption of no
activity from the control filters. The small amount of extractable
organic taken from the control filters was used on Strain TA98 and tend
to confirm this assumption.
-------
7
C. BAP Results
The BAP results are given below:
Table IV
BAP Results (all weights in ng.)
Filter // BAP (ng)
Group 1 320 1252
Unbaked 330 528
340 "U2
Group 2 350 BMD
Baked 360 16
Not Exposed 370 20
Group 3 380 48
Oldsmobile 390 96
400 BMD
410 BMD
Group 4 420 BMD
Mercedes 430 BMD
Benz 440 BMD
450 40
Group 5 460 BMD
Background 470 40
Air 480 360
Air 490 BMD
*BMD = Below measurable detection.
These results indicate normal diesel particulate BAP levels for filters
320, 330, and 340. The baked filters, other than filter number 480, show
very little BAP. Filter #480 is considered to be a testing anomaly.
D. Ames Analysis
The Ames test is normally run using five different strains of bacteria.
Due to limited extractable organic, some strains were not run. Each
strain, if possible, was run both with and without metabolic activation.
The analysis is done in culture plates, and is expressed in
revertants/microgram extractable organic. This result indicates the
mutagenicity of the extractable organic.
The Ames results then do not take into account the different extraction
rates. The actual Ames results and the dates the samples were run are
given below. Also given are the revertants/filter. Since identical
tests were run for each set of filters, a. direct comparison of
revertants/filter is possible.
-------
8
Table V
Strain TA-98
Non Activated
Activated
Filter #
Group 1
Unbaked
Group 2
Baked
Not Exposed
Group 3
Oldsmobile
Group 4
Mercedes
Benz
Group 5
Background
Air
320
330
340
350
360
370
380
390
400
410
420
430
440
450
460
470
480
490
Rev. /ug .ext.org .
2-27-81/3-24-81
14.1/13.5
10.2/11.9
5.8/3.3
0.3
0.3
1.4
18.9/36.8
6.4/12.5
17.9/11.0
8.5/8.0
10.1/14.7
.8/.3
28.3/18.6
1.6/2.5
0.5
0.2
2.0
0.1
Rev. per filter*
2-27-81/3-24-81
4653/4455
1514/1784
1001/569
0.9
0.15
1.96
3065/5967
298/582
2155/1324
748/704
592/862
38.7/14.5
1935/1272
97.1/151
2.81
1.63
13.2
.51
Rev./ug. ext.org.
2-27-81/3-24-8
8.1/7.0
9.0/7.8
9.8/7.2
—
2.4/.S
.5/.6
.6/.6
.6/.3
1.4/.6
.47.1
4.1/2.5
.47.1
^
Rev. /Filter*
2-27-81/3-24-81
2673/2310
1350/1170
1691/1243
_
389/130
23.3/27.9
72.2/72.2
52.8/26.4
82.1/35.2
19.4/4.84
283/171
24.3/6.07
-
* times 10-
Table VI
Strain TA-100
Non Activated
Activated
Filter #
Group 1
Unbaked
Group 3
Oldsmobile
Group 4
Mercedes
Benz
320
330
340
380
390
400
410
420
430
440
450
Rev./ug. ext.org .
2-27-81/4-10-81
27.0/17.7
47.3/20.7
5.9/12.2
3.9/44.7
14.8
59.1/20.5
•18.7/8.7
14.9
2.6
2.5
1.4
Rev. per filter*
2-27-31/4-10-81
8911/5776
7093/3104
1018/2106
632/7248
689
7114/2468
1646/766
874
126
171
85
Rev./ug. ext.org.
2-27-81/4-10-81
8.1/13.5
9.0/9.0
6.2/6.0
21.9/11.7
4.3/4.4
6.1/3.3
4.3
24.7
3.1
Rev. /Filter*
2-27-81/4-10-81
2673/4456
1349/1349
1070/1035
3551/1897
518/530
537/290
252
1689
188
* times 10-
-------
Table VII
Strain TA-1535
Non Activated
Activated
Group 1
Unbaked
Group 2
Old smo bile
* times 103
Group 1
Unbaked
Group 3
Old smo bile
* times 103
Group 1
Unbaked
Gro up 3
Old smo bile
Filter
320
330
340
380
Filter
320
330
340
380
Filter
320
330
340
380
400
Re v./ug
# 3-6-81
0.1
0.6
0.0
0.8
Re v./ug
# 3/6/81
1.9
2.7
1.1
1.7
.ext.org. Rev. /filter*
3-6-81
33.0
90.0
0.0
130
Table VIII
Strain TA-1537
Non Activated
.ext.org. Rev. /filter*
3/6/81
627
405
190
275.
Table IX
Strain TA-1538
Non Activated
Rev. /ug. ext .org. Rev. /filter*
# 3-6-81/4-10-81 3-6-81/4-10-81
6.3/8.3
7.3/3.4
2.5/2.1
.6/5.5
2.5/1.4
2079/2739
1094/510
431/315
97/892
301/168
Re v./ug. ext
3-6-81
0.3
0.1
0.3
0.0
Rev. /ug. ext
3/6/81
3.1
2.5
1.7
.4
.org. Rev. /Filter*
3-6-81
99.0
15.0
51.8
0.0
Activated
.org. Rev. /Filter*
3/6/81
1023.
375
193
64.8
Activated
Re v./ug. ext .org. Rev. /Filter*
3-6-81/4-1-/81 3-6-81/4-10-81
10.1/7.7
10.7/9.6
2.7/2.8
2.5/1.6
.8/1.3
3334/2541
1604/1440
465/483
405/259
96/156
times
103
Analysis of Ames Results
Since all strains were not run for all filters due to lack of extractable
organics, the analysis is not complete. The following comments on the
Ames data can be made:
-------
10
1. The diesel exhaust samples do not appear to be sensitive to strains
TA-1535 and TA-1537 either activated or non-activated. The results
will then be based on strains TA-98, TA-100, and TA-1538.
2. The three unbaked samples 320, 330, and 340 showed wide variability
in all three strains. The other samples appear to be less widely
distributed.
3. The #1 backup filters have significantly higher activity than the #2
backup filters for the three strains. This is consistent with the
theory that more of the gaseous hydrocarbons were removed by the #1
backup filter leaving less to be adsorbed by the #2 backup.
4. The activated samples were less mutagenic than the non-activated
samples for strains TA-98 and TA-100. This indicates the presence of
direct-acting mutagens. Direct acting mutagens do not require
metabolic activation in order to obtain a positive response.
(Indirect-acting mutagens require metabolic activation - they are
metabolically converted to an active mutagenic form.) The activity
was about even for strain TA-1538.
5. The activity of the control samples on strain TA-98 confirm the
assumption that very little activity was due to dilution air or
organics which survived the baking process.
6. The ratio of the total number of revertents from the backup filters
compared to the Group 1 filters is an indication of the amount of
mutagenic activity which would not have been accounted for on the
primary filter. Percentages are given on the next page (Table X) for
Group //3 and #4 filters compared to the average of Group #1 filters.
These results indicate that materials that are mutagenic passed
through the primary filters and were adsorbed on the backup filters.
In some cases, the backup filters were more reactive than the primary
filters. These Ames results indicate that results based on primary
filters only may be very conservative estimates of the true activity
of the diesel exhaust.
7. The FTP and HFET results showed varied behavior. The Oldsmobile data
indicated higher activity during the FTP than for the HFET cycle for
strain TA-98 and vice versa for strain TA-100. The Mercedes
indicated exactly the opposite results with higher activity during
the HFET for strain TA-98 and lower activity during the FTP for
strain TA-100.
8. The activity on the Oldsmobile was greater than that noted for the
Mercedes filters for both strains TA-98 and TA-100.
9. There are several problems with the data. If the experiments were
rerun, several more precautions would be taken. These problems are
noted below. The reason these problems were not addressed is that
until the extraction and Ames results were completed the feasibility
of the testing procedure was not known.
a. Lack of extractable organic was the largest problem. More
filters should have been taken to allow full Ames testing on all
strains.
-------
Table X
Ratio of Ames Test Activities; Backup Filter/Primary Filter
Strain TA 98 Strain TA-100 Strain TA-1538
Non-Activated Activated Non-Activated Activated Non-Activated Activated
Group 3
Oldsmobile
Group 4
Mercedes
Benz
380
390
400
410
420
430
440
450
2-27-81
128%
12.5%
90.2%
31.3%
24.8%
1.6%
81.0%
4.1%
3/24/81
263%
25.6%
58.4%
31.0%
38.0%
0.6%
56.0%
6.6%
2/27/81
20.4%
1.2%
3.8%
2.8%
4.3%
1.0%
14.8%
1.3%
3/24/81
6.8%
1.5%
3.8%
1.4%
1.8%
0.2%
8.9%
0.3%
2/27/81
11.1%
12.1%
125%
29.0%
15.4%
2.2%
3.0%
1.4%
4/10/81
198%
-
67.4%
20.9%
_
-
—
-
2/27/81
209%
-
30.5%
31.6%
14.8%
-
99.5%
11.1%
4/10/81
83.2%
-
23.2%
12.7%
_
-
-
-
3/6/81 4/10/81
8.1% 75.1%
- -
25.1% 14.1%
- -
-
— —
- -
3/6/81 4/10/81
22.4% 17.4%
-
5.3% 10.5%
- -
_ _
-
_ —
- -
-------
12
b. No Mercedes unbaked filters were taken and sent for analysis.
Thus, the ratio of backup filter activity to primary filter
activity is based on Oldsmobile primary filters.
c. Unloaded filters should have been baked and inserted into filter
holders #1 and #2 for an FTP and for an HFET test cycle. The
filters should have then been sent for analysis.
Conclusions
1. The data indicate that gaseous materials present in diesel exhaust
are mutagenic based on the Ames test results. The data also appear to
indicate the presence of direct acting mutagens. Both the gas phase and
particle-bound HC are more mutagenic in the Ames test without metabolic
activation.
2. The assessment of the mutagenic activity of diesel exhaust using only
the particulate organic may be overly conservative if significant
activity is indeed present in the gas phase as these experiments indicate.
3. While this method of collection shows promise, further development
work is needed to make it a routine collection procedure.
-------
ER5EDU5 HYDRDCRRBDN STUDY
DLDBMDBILE 20 X 20 FILTERB
in
z
LD
Z
in
Z
D
J
J
C
h
D
t ,,
S
—
t-
H
pi.
OADED
h-
Pt
H
^J
H
^
1
pq
PS
B
Kj
H
PM
s
CO
o
@
• . •
P4
W
H
•J
H
I
jpq
@
H
.J
M
Pn
uauvu
t-
—
Pi
W
H
hJ
H
h
AKED
PC
3
H
^
H
P>4
Q
W
CO
O
W
.UVU'
CQJ
L!
tl.H
FTP //I Backup
#0380
FTP #2 Backup
#0390
HFET //I Backup
//0400
HFET #2 Backup
#0410
FILTER BRMPLE NUMBERS
Figure 1
2U.U
CB 1/23/BI
-------
Ul
GflEEDUS HYDRDCflRBDN STUDY
MERCEDES 300D 20 X 20 FILTER5
23.0
LH
Z
ID
Z
EE
D
J
J
D
h
22.B
21.H
20.0
19.B
IB.0
•
1 BLANK FILTER 1
1 LOADED FILTER 1
I BAKED FILTER 1
I EXPOSED FILTER I
1 BLANK FILTER!
1 LOADED FILTER 1
••«
h-
1-
pq
1 EXPOSED FILTER 1
I BLANK FILTER |
I LOADED FILTER
| BAKED FILTER
| EXPOSED FILTER |
I BLANK FILTER t
1 LOADED FILTER t
1 BAKED FILTER 1
I EXPOSED FILTER I
0.0
FTP //I Backup
#0420
FTP //2 Backup HFET //I Backup HFET //2 Backup
//0430 //0440 //0450
TJp
FILTER 5RMPLE NUMBERS
Figure 2
-------
LJ1
Z
IE
LD
7i
6
IE
a
J
J
a:
h
a
23.0
22. B
21.0
20.0
19.0
IB.
GR5EDU5 HYD'RDCflRBDN 5TUDY
Fl I R 20 X 20 F I LTER5
'
•
PBLANK FILTER
[LOADED FILTER
H
to
pq
[EXPOSED FILTER
["BLANK FILTER
—
1 LOAD ED FILTER
^f~
^
H
pq
v^
xl
I
[BLANK FILTER
(LOADED FILTER
I BAKED FILTER
»•
1 EXPOSED FILTER
r
!
!
1
f
|H
Irt
ILOADED FILTER
:
| BAKED FILTER
•^
LEXi'UbiiU FILTER
0.0 FTP #1 Backup FTP #2 Backup HFET #1 Backup HFET #2 Backup 20 0
#0460 #0470 #0480 #0490 £B 2/Z/B 1
(Jl
FILTER 5RMPLE NUMBERS
Figure 3
-------
0.3Z
RD5DRBED NE EHT ER N
Ul
z
HI 0.2S
—7
0.20
LJ
ffl
E 0-I£
a
Ul
^j 0.10
E
h
I 0.05
m
LJ
~<[ 0.00
o
oo
CO
=tt=
0
CTi
cn
=»=
o
o
«*
=«fe
o
1-1
1
-*
=st=
o
ro
<•
=*fc
o
-*
-*
=«=
O
-tf
^
o
=«:
o
l~»
«*
=!fc
"~
O
00
<•
=*!
Oldsmobile Mercedes Benz Backgrou
E R D LJ P NUMBER
Figure 4
o
o\
=«=
id Air
FiJP 7-27-B 1
_
------- |