Fuel permeation rates after changing fuel.
continues to be a widely discussed topic in automotive engineering
circles. Two events resulting from the 1991 Clean Air Act
are keeping this subject at the forefront of engineers' and
materials scientists' minds. First, new enhanced evaporative
emissions limits are in the middle of their phase which
started in 1995 and will be 100% complete by 1999. These new
limits require that no vehicle emit more than 2.0 grams of
hydrocarbons in a diurnal test run by either the
Environmental Protection Agency (EPA) or the California Air Resources
Board (CARB) method. The EPA method is a two day
diurnal test cycled from 22 [degrees] C to 36 [degrees] C to 22 [degrees] C,
CARB diurnal test is cycled from 18 [degrees] C to 40 [degrees] C to 18
Second, in addition to this "enhanced evap" requirement in
1998, the phase in for on board refueling vapor recovery
(ORVR) starts at a 40% level for passenger cars in 1998, and
by 2003 all cars and light trucks will be equipped with this
system. ORVR may require larger charcoal canisters plus
considerable new vapor-vent lines, all of which will add
additional sources for hydrocarbon emissions. In spite of this
enlarged vapor recovery equipment, the 2.0 gram
hydrocarbon evaporative per EPA and CARB methods must be
maintained both as the vehicle is built and after 100,000 miles
(160,000 kilometers) of service.
The requirement that the vehicle maintain its less than 2.0
gram hydrocarbon evaporative emission rate after 100M
miles of service leads to some interesting questions. The
permeation rates of elastomers and plastics to various fuels has
been the subject of previous work, including work done by
MacLachlan (ref. 1) in 1979, Goldsberry, Chilous and Will
(ref. 2) in 1991, Stahl and Steven (ref. 4) in 1992, Hopf, Reis
and Gray in 1994 (ref. 5) and Fuller and Stevens (ref. 6) in
1996. The fuel permeation rates of elastomers such as nitrile,
epichlorohydrin and fluoroelastomers of various fluorine
contents, as well as plastics such as nylon 12 and
fluoroplastics, are well documented in these works. However, one real
world element of fuel permeation that has not been discussed
widely is the situation that will be encountered after up to
70,000 miles when the EPA and CARB tests could be rerun.
In this "in use" case, it is quite possible, even probable, that a
vehicle will have been running for an extended time on a
"reformulated" or "oxygenated" gasoline. These gasolines
can contain alcohols and ethers in low percentages such as
10-15% methyl-t-butyl ether (MTBE), 10% ethanol, and
even blends of 5% methanol with 2.5% of a cosolvent such
In the fuel permeation work referenced above, the
permeation rates were measured on one type of fuel. The potential
of switching or changing of the fuel part way through the test
has not been published. That changing or switching of the
fuel and the resulting patterns of fuel permeation behavior is
the focus of this article.
Permeation testing was conducted by two weight loss
* Thwing Albert Cup Method -- ASTM E96-66
* Fuel hose testing -- modified GM9061 P "fill and plug
It was desired to see if these two permeation test methods
would show similar trends when similar test specimens were
used. The Thwing Albert Cup method is well documented in
previous SAE work (refs. 1, 3 and 4). In this method, a cup is
filled with 100 ml of fuel, then a 76.2 mm diaphragm is
placed on top and secured metal flange ring. The exposed
surface area is 37 [cm.sup.2].
The fuel hose permeation testing was done by a modified
GM9061 P method. A longer length of hose was used for this
testing; in this case, a 91.4 cm length of hose having a
nominal ID of 7.9 mm. The hose was filled 75% by volume,
which was approximately 32 ml of fuel.
Both cup and hose testing were done in an oven @ 40 [degrees] C
to increase the severity of the test.
The materials tested included nitrile (NBR),
fluoroelastomers (FKM). and fluoroelastomer/fluoroplastic barrier
systems and are shown in table 1. The fluoroplastic barrier is
noted as FEP and had a thickness of 0.05 mm. The FKM cup
specimens were 0-76 mm thick, which is typical of the
thickness of FKM in FKM veneered fuel hose.
Table 1-cup and hose testing materials
1. NBR - 41% ACN
2. FKM - 66% fluorine
3. FKM - 68% fluorine
4. FKM - 70% fluorine
5. FKM/FEP laminate
1. FKM (66%F)/CSM/CSM(*)
2. FKM (68%F)/NBR/ECO(*)
3. FKM (70%F)/FEP/NBR/ECO(*)
4. FKM (70%F)/FEP/EAM/EAM(**)
5. FKM (68%F)/FEP/EAM/EAM(**)
(*) -- commercially made hose
(**) -- hose made on DuPont Dow prototype extrusion line
When evaluating via the cup or hose method, testing was
done for 336 hours at 40 [degrees] C. Weight losses were noted within
the first 24 hours, then every two to three days thereafter.
After 336 hours, the test specimens were broken down, fresh
fuel was placed inside, and the specimens returned for
another 336 hours at 40 [degrees] C. The testing sequence used was:
336 hrs. @ 40 [degrees] C [right arrow] Refill [right arrow] 336 hrs. @ 40
[degrees] C Fuel C [right arrow] Fuel C (control)
CE-10 [right arrow] Fuel C
CMTBE-15 [right arrow] Fuel C
The test sequence was designed to simulate an
oxygenated fuel such as 90/10 Fuel C/ethanol (CE-10), or 85/15 Fuel
C/MTBE (CMTBE-15) being used in a vehicle, then the
vehicle being refilled with a non-oxygenated fuel. The
question was, what would happen to the permeation rates of the
specimens that had been conditioned in the oxygenated fuel,
then switched to 100% Fuel C?
The results of the tests are shown in figures 1 through 10.
The first five figures are for the cup testing, and figures 6
through 10 show the hose results.
The cup results with 100% Fuel C and CE-10 are shown
in figure 1.41% ACN nitriles are still used in filler neck
hose, so it was desired to document their performance. The
results show that at 40 [degrees] C, the 41% ACN-NBR had a peak
permeation rate of ~400 g/[m.sup.2]/day in 90% Fuel C/10%
ethanol. The results show a drop off in permeation rate after
48 hours. This drop, especially the drop in the oxygenated
fuel, may be in part attributed to the "speciating" of the fuel.
By speciating, it is meant that the fuel permeates as separate
species. In this example, Fuel C has isooctane, toluene and in
the case of E10, ethanol. Once the fuel composition changes
and becomes less aggressive (less ethanol and toluene), the
permeation rate drops. Another concern here is that the total
volume of the fuel is only 100 ml. and a large weight loss is
occurring during this 336 hours.
[FIGURE 1, GRAPH OMITTED]
The 41% ACN-NBR was not run in the second phase of
testing. The 76 mm diaphragms were swelled enough so that
they could not be reassembled without damaging the
diaphragm and affecting the seal of the test specimen.
Testing was discontinued on the NBR at this point.
Figure 2 looks at a FKM dipolymer of 66% fluorine in a
fuel hose compound that is not postcured. The results for all
three fuels is shown: Fuel C, CE10 and CMTBE-15. The
permeation rate of the Fuel C at ~10/[m.sup.2]/day and the
CMTBE-15 at ~50g/[m.sup.2]/day seem stable at 336 hours at 40 [degrees] C, but
permeation rate of the CE-10 peaks at 120 hours and drops.
Fuel speciating may be occurring here too.
[FIGURE 2, GRAPH OMITTED]
After 336 hours, the fuel in all three cups was changed to
Fuel C. It is interesting to note that the two FKM diaphragms
exposed to oxygenated fuels CE-10 and CMTBE-15
continue to show a higher permeation rate than the Fuel C control,
but both do drop in permeation rate. The diaphragm exposed
to CE-10 seems to move toward the Fuel C control faster
than does the CMTBE-15 diaphragm. This is curious because
the CE-10 showed a higher permeation rate than CMTBE-15
in oxygenated fuel.
Figure 3, FKM (68% F). looks at a 68% fluorine FKM in
the same three fuels. As expected, the 68% fluorine FKM
shows an overall lower permeation rate than does the 66%
fluorine FKM with permeation rates of ~11 g/[m.sup.2]/day in Fuel
C, 50 g/[m.sup.2]/day in CE-10 and 34 g/[m.sup.2]/day in CMTBE-15
after 336 hours at 40 [degrees] C. The same drop in CE-10 permeation
rate is seen after 120 hours as was seen in figure 2, so fuel
speciating may be happening.
[FIGURE 3, GRAPH OMITTED]
After 336 hours, when all the cups are refilled with Fuel
C, a similar trend as is seen in figure 2 is noted. The CE-10
and CMTBE- 15 continue permeation at a higher rate than the
Fuel C, and it takes longer for the CMTBE-15 conditioned
diaphragm to return close to the Fuel C control than it does
for the CE-10 conditioned diaphragm.
Figure 4, FKM (70%F), charts the permeation rate of the
three fuels through a 70% fluorine FKM. As expected, the
70% fluorine FKM demonstrates an overall lower permeation
rate profile than does the 68% fluorine FKM in the first
336 hours of the test. After 336 hours at 40'C, permeation
rates of 6 g/[m.sup.2]/day for Fuel C, 25 g/[m.sup.2]/day for CE- 10 and 19
g/[m.sup.2]/day for CMTBE-15 are seen with the 70% fluorine
[FIGURE 4, GRAPH OMITTED]
When the fuel in the cups is changed to 100% Fuel C in
the second 336 hour phase of the test, a similar pattern as was
seen in figures 2 and 3 is seen here in figure 4. Both the
CE-10 and 10 the CMTBE-15 conditioned diaphragms continue to
permeate at a higher rate than does the Fuel C control.
Gradually, the permeation rates of the CE-10 and
CMTBE-15 conditioned diaphragms come close to the Fuel C control
after 552 hours, but the CMTBE-15 conditioned diaphragm
takes longer to recover to the Fuel C control.
Figure 5 shows the permeation rate of the three fuels to a
laminate of 70% fluorine FKM with 0.05 mm FEP
[FIGURE 5, GRAPH OMITTED]
The results show a different permeation profile than is
seen with any of the fluoroelastomers. The permeation rates
were quite low, below 3 g/[m.sup.2]/day, under any conditions, and
different fuels or fuel changes did not to seem to significantly
affect the fuel permeation rate of these laminates.
A summary of the average permeation rates as determined
by the cup method is shown in figure 6. These data look at
the original permeation rate in Fuel C, excluding data before
48 hours to allow the cup to come to equilibrium, and
compares this permeation rate to the cups where the rubber
diaphragm was conditioned 336 hours @ 40 [degrees] C in Fuel C.
CE-10 and CMTBE-15. All perm testing on figure 6 was run
in 100% Fuel C, but the diaphragms that were preconditioned
in other fuels, especially the ones seeing oxygenated fuel,
showed higher average permeation rates. In this case, the
CMTBE-15 fuel seemed to have a greater effect than the
[FIGURE 6, GRAPH OMITTED]
Figure 7 shows the phase of testing that used hose tested
by the modified GM9061P method. The hose tested in figure
7 is based on a 66% fluorine FKM (~0.7 mm veneer) with a
CSM interlayer, a fabric braid and a EAM cover. The
oxygenated fuels CE-10 and CMTBE-15 seem to come to a peak
permeation rate faster than does the Fuel C control with max
permeation rates of 18 g/[m.sup.2]/day for Fuel C, 50 g/[m.sup.2]/day for
CE-10 and 25 /[m.sup.2]/day for the CMTBE-15. Fuel speciating
giving lower than expected results may once again be
occurring in this test, especially in the CE-10 fuel.
[FIGURE 7, GRAPH OMITTED]
After 336 hours, the fuel in all the hoses was changed to
100% Fuel C. A similar trend, as was noted in figure 2 (FKM
66%F diaphragm), is noted with the CE-10 and CMTBE-15
conditioned hose permeating at a higher rate than the Fuel C
control, although the magnitude of the difference is not as
great. In this case, the CE-10 conditioned hose does not
closely approach the permeation rate of the Fuel C control
until 672 hours into the test.
Figure 8 shows the permeation data for the hose based on
68% fluorine FKM (~0.7 mm veneer), with a NBR
interlayer, fabric braid and a ECO cover. The peak permeation rate
after 336 hours at 40 [degrees] C was 16 g/[m.sup.2]/day for Fuel C, 40
g/[m.sup.2]/day for CE-10 and 22 /[m.sup.2/day for CMTBE-15. Fuel
speciating in the CE-10 fuel once again is a concern based on
the shape of the curve for this fuel. One other observation is,
that while this 68% fluorine FKM veneer hose does show
slightly lower permeation rates than is seen with the 66%
fluorine FKM veneered hose in figure 7, the magnitude of the
difference is not as great as is seen between the 66%F and
68%F FKM diaphragms in figures 2 and 3. It is speculated
that the differences in the interlayer and cover compounds on
the hose may be having an effect here. It was noted by
Edmonson and Balzer (ref. 3), that hoses made with a CSM
interlayer and cover had lower fuel-methanol permeation
than did hoses made with a ECO interlayer and cover when
the same FKM veneer was used. We may be seeing a similar
trend here with ethanol and MTBE blended fuels rather than
[FIGURE 8, GRAPH OMITTED]
After 336 hours, all the hoses were refilled with 100%
Fuel C, and the results show that the CE-10 and CMTBE-15
conditioned hoses continued to permeate at a higher rate than
the Fuel C control, but in this case, the CE-10 conditioned
hose did recover toward the Fuel C control faster, reaching
close to the Fuel C control permeation rate by 528 hours.
Figure 9 looks at a barrier hose made of 70@/( fluorine
FKM/FEP barrier/NBR interlayer/fabric braid/ECO cover. A
very different pattern is seen here compared to the hoses in
figures 7 and 8. The permeation rates are quite low and the
data seem to have no distinct peak. Based on the data curve.
it seems that fuel speciating is not so much of an issue here.
The permeation rates after 336 hours at 40'C of this hose is
1.5 g/[m.sup.2]/day in Fuel C. 2.0 g/[m.sup.2]/day in CE-10 and 1.6
g/[m.sup.2]/day in CMTBE-15.
[FIGURE 9, GRAPH OMITTED]
After 336 hours, the fuel in these hoses was changed to
100% Fuel C. In this case, no change in permeation rate for
any of the conditioned hoses was seen. The permeation rate
was still quite low with a rate varying between 1.5 to 3.5
g/[m.sup.2]/day throughout the entire 672 hour test.
Figure 10 looks at a hose made in a prototyping line. The
hose has a 0.4 mm veneer of 70% fluorine FKM, a 0.05 mm
FEP barrier, an interlayer and cover of 150 [degrees] C heat resistant
EAM and an aramid fiber braid. The results of the first 336
hours of permeation testing look quite similar to what was
seen on figure 9; the permeation rates are very low and no
distinct peaks in permeation are seen. After 336 hours, the
permeation rate for this hose was 1.6 g/[m.sup.2]/day in Fuel C, 1.9
g/[m.sup.2/day in CE-10 and 1.4 g/[m.sup.2]/day in CMTBE-15.
[FIGURE 10, GRAPH OMITTED]
After changing out to 100% Fuel C in all the hoses, no
significant change in permeation rate was seen through 672
hours at 40 [degrees] C.
In figure 11, data are seen on another prototype hose
made by DuPont Dow Elastomers. This hose uses exactly the
same construction as the hose in figure 10, except the 0.4 mm
FKM veneer is based on a 68% fluorine FKM. The
permeation rates are quite similar to those seen in figures 9 and 10,
with results for this hose of 2.0 g/[m.sup.2]/day in Fuel C, 3.0
g/[m.sup.2]/day in CE-10 and 1.8 g/[m.sup.2]/day in CMTBE-15 after the
336 hours at 40 [degrees] C phase of the test was completed.
[FIGURE 11, GRAPH OMITTED]
Once again, all the hoses were drained and refilled with
100% Fuel C. The permeation rates remained quite low and
did not vary much through the entire 672 hours of the test.
Figure 12 shows a summary of all hose average
permeation data much as figure 6 did for the cup data. This chart
looks at the hoses that were originally exposed to Fuel C
(data after 48 hrs. up to 336 hrs.) and compares them to the
hoses that were conditioned in Fuel C, CE-10 and
CMTBE-15, then rerun in 100% Fuel C. The results here are
interesting in that in this case the hose preconditioned in CE-10
seems to permeate at the highest rate with the CMTBE-15
having a lesser effect. This is different than figure 6 where
the CMTBE-15 had the greater effect. This difference
between hose and diaphragm may be due to the outerlayers
of the hose being a better barrier to CMBTE-15 than CE-10.
This hose effect was discussed in figure 8, and was noted by
Edmonson and Balzer (ref. 3).
[FIGURE 12, GRAPH OMITTED]
The trend noted in figure 12 supports what was seen in
figure 6 in that the hoses preconditioned with oxygenated
fuel show higher Fuel C permeation rates than hoses conditioned
with Fuel C only. These data support the theory that
just because the fuel handling components on a car pass an
initial SHED hydrocarbons emission test in a
non-oxygenated test fuel, the same components may not pass after
extended field service in oxygenated fuel in spite of the fact that the
test on the used vehicle will be run by draining the
oxygenated fuel out of the car and refueling with the non-oxygenated
test fuel. An extra degree of permeation resistance robustness
may be needed to insure SHED compliance after 160,000
kilometers of field service.
Based on the cup permeation data and the hose permeation
data, the following conclusions can be made:
1. When a FKM diaphragm, simulating a veneer in fuel
hose, having fluorine contents of 66, 68 and 70% is
conditioned in oxygenated fuel, then switched to a non-oxygenated
fuel like Fuel C, it:
* Permeates like the oxygenated fuel is still in the hose
for 48-72 hours.
* Gradually comes back toward the non-oxygenated
fuel permeation rate. This may take 336 hours or more to
* CMTBE-15 preconditioned FKM maintains a higher
permeation rate after being switched back to Fuel C than do
E10 conditioned specimens.
2. Hose data indicate that preconditioning in oxygenated
fuels, then switching back to Fuel C, does increase
permeation rate as compared to hose which has only seen Fuel C,
but in this case the CE-10 seems to have a larger effect than
3. Barrier hoses and laminates using the FKM/FEP
* Show significantly lower permeation rates than 70%
fluorine FKM, and regular FKM veneered hose.
* Are not significantly affected in permeation rate by
preconditioning in oxygenated fuel then switching back to
regular Fuel C.
It is further noted that fuel speciating may be occurring in
these types of "set volume of fuel" permeation tests even
with low permeating materials such as fluoroelastomers.
These phenomena had not been seen in fluoroelastomers
tested at 23 [degrees] C in previous work, but the test condition of 40
[degrees] C is
significantly more rigorous.
Further studies on changing fuel with a sufficient volume
of fuel available would be recommended. The new
recirculating fuel test method J 1737 would be an excellent way to
[1.] J.D. Maclachlan -- SAE Paper 790657, "Automotive fuel
permeation resistance -- a comparison of elastomeric
materials, June 11-15, 1979.
[2.] Goldsberry, Chillous and Will -- SAE Paper 910104,
"Fluoropolymer resins: Permeation of automotive fuels,"
Feb. 25-March 1, 1991.
[4.] W.M. Stahl and RD. Stevens -- SAE Paper 920163,
"Fuel-alcohol permeation rates of fluoroelastomers, fluoroplastics
and other fuel resistant materials," February 24-28, 1992.
[5.] Hopf, Reis and E. Gray, SAE Paper 940165,
"Development of multilayered thermoplastic fuel lines with
improved barrier properties," Feb. 28 -- March 3,1994.
[6.] R.E. Fuller and R.D. Stevens -- SAE Paper 960140,
"Unique low permeation elastomeric laminates for fuel
hose," February 26-29, 1996.
[7.] D. Kulp, M. Masten and M. Harrigan -- Detroit Rubber
Group Paper, Oct. 1995, Ford Report On Fuel Systems
|Printer friendly Cite/link Email Feedback|
|Author:||Fuller, Robert E.|
|Date:||Feb 1, 1998|
|Previous Article:||HNBR and long term serviceability in automotive lubricants: structure property relationships.|
|Next Article:||Elastomer blend approach to extend heat life of natural rubber based engine mounts.|