|
Results of BoatU.S. Sponsored Fuel and Fiberglass Gas Tank Tests
Soon after the Long Island Sound area of New York switched to
gas mixed with ethanol, BoatU.S. began receiving calls from
boaters who owned older gas-powered boats with fiberglass fuel
tanks. The tanks, it seemed, were being slowly dissolved by
ethanol; black stuff was building up on valves and intakes,
destroying engines and some fuel tanks were becoming soft and
even weeping fuel. Since then, BoatU.S. has sponsored several
tests to find out exactly what the problem is. Here are the
results of what we have learned so far. We’ve highlighted the
important aspects in each case.
Tests of the black material from an intake valve:
“A portion of the black material was scraped from the intake
valve, pressed flat, and transferred to a potassium bromide
crystal plate. The sample was then analyzed by Fourier Transform
Infrared Spectroscopy (FTIR), which produced an infrared
spectrum. As indicated on the spectrum, the peaks at 2800-3000
cm-1 are due to Carbon-Hydrogen absorptions; the strong
absorption at 1730 cm-1 can be from Esters or Ketones. The small
sharp peaks at 1460 and 1370 cm-1 are indicative of
Hydrocarbons. Finally, the large broad peak seen from 1000-1200
cm-1 is due at least in part from Carbon-Oxygen single bond
absorbers. It can be said that this spectrum is consistent
with the presence of polyester, whoever this cannot be
unequivocally proven from this spectrum alone.”
Test of fuel from an affected boat:
EPA 8260 fuel analysis of a sample of fuel taken from a 1968
Bertram 31 located on Long Island Sound indicated 736,426
µg/L of styrene, a component of polyester resin.
Test of sections of fiberglass fuel tanks from a 1967 and
1970 Bertram:
Testing Summary:
To date, the testing done by IMS, LLC indicates that the two
fuel tank samples have undergone some aggressive degradation
(40% of their strength). The bottoms of both tanks have lost
more strength than the tops. The older tank (1967) was laminated
to a much higher level of quality in terms of entrapped air and
fiber roll out. The mode of property reduction in the newer tank
(1970) appears to be both resin softening and loss of adhesion
between fiber and resin. This is evidenced by a moderate loss in
both strength and stiffness. The older tank has lost nearly a
similar amount of strength but has retained all its original
stiffness. This indicates some resin degradation has occurred
but no loss of the fiber/resin interface’s integrity has
occurred.
Both tanks were produced using a fire retardant resin system
although we feel the base resins for each tank are of a
different type. Both tanks have absorbed in the range of 4.2%
fuel into their volume over time (tank bottoms). The top of the
newer tank has also absorbed over 4% fuel. The top of the older
tank has absorbed 2.2% fuel over time.
Tests of Fiberglass Fuel Tank Samples
We recently analyzed intake valves having heavy, black deposits
under the crowns. The valves we have received to date had been
taken from gasoline engines in older yachts that had recently
changed over to gasoline containing ethanol. Bent pushrods and
bent valves have been reported with heavy engine damage when
pistons impacted the valves.
We removed some of the black sludge from under an intake valve
crown. We soaked the sludge in ethanol and, after taking the
extract down to dryness, we obtained the infrared absorption
spectrum. We found that the material is di-iso octyl phthalate.
We ran a series of experiments using straight gasoline and
gasoline with 10% ethanol on fiberglass coupons and coupons of
filler taken from one of the fuel tanks of the vessel from which
the valves had been taken. Shortly before engine failure that
vessel had changed over to gasoline having 10% ethanol. The
results can be summarized as follows:
With both the straight gasoline and the gasoline having 10%
ethanol, analysis by Gas Chromatography Mass Spectrometry (GCMS)
shows that the fuel's lightest fractions were absorbed into both
fiberglass and filler. Noting the very high flammability and
volatility of these light organic compounds, boaters needs to be
alert to possible outgassing and fire/explosion hazards.
GCMS shows that the gasoline having 10% ethanol picked up four
very heavy molecules from the fiberglass and two from the
filler. The molecular weights of these molecules were in the
range of 281 to 379. The straight gasoline did not pick up these
molecules. Evaporating the straight gasoline we were left with a
thin film. Evaporating the samples that had picked up the heavy
molecules we were left with heavy, brown sludge. Infrared
spectroscopy showed molecular similarities between the sludge,
and the material taken from under the intake valve crowns.
This is what we believe is happening:
Polyester resins, gel coats and fillers commonly incorporate
phthalates. In even the best resins and layups a small
proportions of these phthalates remain unreacted. There are
several water soluble molecules that are found in these
materials and they play a central role in blister formation and
delamination. Phthalates are only sparingly soluble in water,
however many are readily dissolved by ethanol.
Whereas gasoline free from ethanol never picks up phthalates,
when ethanol was introduced the very small ethanol molecules
diffused into the fiberglass, filler and gel coat materials
where they dissolved unreacted phthalates. Having been dissolved
by smaller molecules, and almost certainly accelerated by
osmotic pressure, some portion diffused back to the surface and
was dispersed in the gasoline. Based on our GCMS results to date
there are some other, presently unidentified, large molecules
that were also leached out by the ethanol and similarly
transferred into the gasoline.
This internal solution and diffusion back to the surface is the
process of leaching.
Since they are in solution, the phthalates and the other
heavy dissolved molecules are able to pass through the fuel line
filters. When the gasoline with ethanol evaporates in the
carburetor the heavy molecules do not evaporate but come out of
solution and are carried along in the air-fuel mix as an
aerosol. When the droplets impinge on throttle plates and on the
walls of the induction system they can collect as reported by
Chuck Fort at BoatU.S. We do not presently know if after
impingement the films are immobile or if they are able to
migrate through the induction system towards the intake valves.
Some of the molecules that impinge on the hot valve stems and
under the crowns decompose to leave carbon powder and ash.
Others, such as the phthalates that in general have exceptional
high temperature stability, remain intact or undergo only
partial decomposition and then act as the binder that holds
together the carbon particles and ash as the observed, black
sludge.
Frederick G. Hochgraf,
Senior Scientist
NH Materials Laboratory
Fiberglass Tank and Residue Analysis
Analysis of a piece of fiberglass tank (sample 1) and residue
from tank (sample 2). Samples were examined using
stereomicroscope and Fourier Transform Infrared Spectroscopy (FTIR).
Sample 1 showed that the inside section of the tank has begun to
erode. Portions of the resin material are flaking off in sheets,
exposing the actual fiberglass webbing. Additionally, the resin
on the inside of the tank had tiny granular particles adhering
to the sheets that were delaminating from the inside surface.
Examination of the residue (sample 2) showed sheets of
resin-like material, with a similar appearance as the material
flaking off the tank. The sheets observed in sample 2 also had
tiny granular particles adhering to the surface as viewed with
the aid of the microscope.
Analysis of the resin and granular material from sample 1 and
the residue and granular material in sample 2 showed they had a
similar composition. This indicates that the residue in
sample 2 is coming from the erosion of the tank material (sample
1).
Other independent test results:
Chemical Resistance Data From A Leading Epoxy Supplier.
The test was made using the company's most resistant epoxy and
exposing fiberglass lab samples to 10% ethanol gas and regular
unleaded gas as well as diesel and aviation gas.
The results for the ethanol gas showed a 10% loss in hardness
and a 10-15% loss of compressive strength over a 16 week period.
It is likely that the loss of hardness and strength would
continue to fall at a similar rate over time. The unleaded gas,
diesel, and aviation gas tests showed virtually no change.
Link to an informal test done by amateur boat builder:
http://egyptian.net/~raymacke/Cbnskif27.htm
http://egyptian.net/~raymacke/Cbnskif36.htm |