This is the penultimate article in the rapid prototyping series for now, and the last of our discussions on shop experiments. In the next article, we will look at a couple of opportunities for further education for those interested in learning more about composites and experimental design.
Last time we discussed some of the basic properties concerning epoxies. Based on that information, we conducted a set of tests designed to determine which epoxy would be best for a particular application, in this case the construction of a composite fuel tank.
Which Epoxy to Choose?
Like strength, chemical resistance of epoxy systems is strongly tied (among other factors) to the degree of crosslinking achieved in the cured part. For this reason, when constructing composite fuel tanks, the biggest advantage you can give yourself is to (a) choose a system with a proven history of success when used for fuel tanks, and (b) post cure the tank to the highest permissible temperature.
The SR-1 race plane that has been the basis of this series has integral wing fuel tanks. One proven method of sealing such tanks is to use a novolac sealer like Rhino (Jeffco) 9700, which forms a thick (.05-.15 inch, depending on application technique and number of coats) surface coat for sealing the inside of the tank. This is the factory specified method used on Lancairs, for example.
Close-up of typical coupons (left) and coupon in a test jar (right). Labeling nomenclature indicates epoxy brand (J=Jeffco, M=MGS, Z=E-Z Poxy), post cure temp (3=30 C/86 F, 5=50 C/122 F, 8=80 C/176 F), environment (V=avgas, E=ethanolated mogas, A=acetone). Each test had two coupons. Thus, the coupon on the left is E-Z Poxy, 50 C post cure, avgas exposure, coupon 1 of 2.
Unfortunately this method is somewhat heavy and thus not ideal for the SR-1, which is a severely weight-critical aircraft. However, Rutan-style aircraft builders (Long-EZs, etc.) have successfully built wing strake fuel tanks without sealer for years. Based on personal and forum discussions on this topic, I decided to seal my tanks by applying (over the existing sandwich skin) a skim layer of neat resin, followed by another skim layer of resin mixed with cabosil (to the consistency of petroleum jelly), followed by a final layer of Hexcel PrimeTex 284 carbon fiber with an approximately 40/60 fiber/resin ratio. (If you have read previous articles in this series, you will recognize 40/60 as being a very resin-rich layup.) PrimeTex 284 differs from bog-standard 284 in having very low open area, thus minimizing pinholes.
The next question was which epoxy would be most fuel resistant? Now, I have no intention of ever putting alcohol-laced mogas into the tank (a major reason for tank failures, as alcohol attacks epoxy). But while some tanks are fuel proof simply by virtue of having walls thick enough so that any chemical degradation never has a chance to make a hole through them, the thinness of my sealing layer (about .020 inch) meant that I wanted to assure that I had the most chemically resistant epoxy coating possible. I therefore decided to test three different epoxies for chemical resistance.
The test matrix consists of three epoxies, three post cure treatments, and three chemical environments, giving 27 test permutations, with two coupons per test (see Figure 1), for a total of 54 coupons tested over a period of 18 months.
According to various sources, E-Z Poxy E-Z 10 resin with E-Z 87 slow hardener is the epoxy of choice for fuel tanks for Rutan aircraft builders. According to EAA technical counselor and former Shell “Answer Man” for epoxy resins, Gary Hunter, E-Z 87 is the only aromatic amine hardener available for homebuilders, as compared to aliphatic and cycloaliphatic amines, which are more common in commercially available systems. While from a health and safety standpoint aromatic amines are nastier than the aliphatic/cycloaliphatic amines, they are superior for chemical resistance.
According to Gary, MGS 285/287 uses the aforementioned cycloaliphatic amine hardener, and is what I use generally for SR-1 parts, so I also wanted to test that system (it is also what the underlying wing skin is made with). Finally, I chose Jeffco 1307/3176, as it is a less expensive system that is popular among homebuilders.
Neither aviation gas (left) nor mogas test coupons (middle) showed any obvious physical degradation. Acetone coupons (right) showed moderate to severe degradation.
Jeffco acetone test coupons clearly show the effect of post curing in providing chemical resistance protection. (Left) Room-temperature post cure, (Middle) 122 F post cure, (Right) 176 F post cure.
All coupons were initially cured for 24 hours at room temp (approximately 85 F), followed by no post cure or a 24-hour post cure at 122 F or 176 F. Coupons were nominally 4.0×1.5x 1/8 inch, and were tested for Barcol hardness and weighed before being placed in a 16-ounce glass jar filled with either 100LL, 93 octane E10 auto gas, or acetone. The percentage of ethanol in the E10 mogas was 8.4%, as determined by a water admixture test.
Although one would not normally expect to have parts subjected to long-term immersion in a solvent like acetone, it was included as a means of examining the performance of coupons in a strong solvent environment, as a sort of proxy for a longer-term test. Whether this can be used to infer/extrapolate performance in a less harsh but longer-term environment is definitely debatable, and possibly specious. As Peter Meszaros of Airheart Distributing (North American distributors of MGS epoxy) points out by way of illustration, a paper towel soaked in acetone holds up just fine, but one soaked in water quickly disintegrates. So the acetone tests should be viewed in that light, but the results were nonetheless interesting and confirm Gary’s exoneration to builders that there is no substitute for post curing to achieve chemical resistance.
The test began June 19, 2016 and ended January 9, 2018. At the end of the test, coupons were removed and allowed to dry, then weighed.
None of the coupons immersed in either 100LL or 93 octane E10 mogas showed any visible signs of deterioration. No particular difference among coupons based on brand or post cure was discernible. Almost all coupons, however, did gain weight over the period of the test, with weight gain from avgas being least and acetone being greatest. As can be seen in Figure 3, post curing reduced the amount of weight gain.
Coupons immersed in acetone showed degradation ranging from slight to severe. Figure 4 shows a graphic representation of these coupons, with degradation being given a number value based on a subjective assessment of the degree of degradation, with 1 representing no degradation and 4 representing severe degradation. E-Z Poxy and MGS performed similarly, but surprisingly Jeffco outperformed both. That said, differences in post curing had a larger impact on degradation than differences in epoxy brand. This again appears to confirm Gary’s position that the most important consideration when constructing parts for chemical resistance is achieving a good post cure.
Figure 4: Higher post cure temperatures reduced the amount of coupon degradation from acetone (1 represents no degradation and 4 represents severe degradation).
It should be noted that with only two coupons per sample, these results lack any statistical significance. The results should be interpreted as observed trends.
That wraps up our look into epoxy. Next time we’ll finish up the series with a look at continuing education classes by Abaris Technology of Reno, Nevada, and Micro-Measurements of Raleigh, North Carolina. Until then, happy building.
Acknowledgements: Many thanks to Gary Hunter and Peter Meszaros for answering my questions while I was preparing these epoxy articles. Klaus Savier of Lightspeed Engineering provided access to his post curing oven for the coupon tests as well as his Barcol hardness tester. Ricardo Stary, the 2016 SR-1 Project intern, fabricated the coupons.