Prep School

How Good Is That Fiberglass/Plywood Bond?

Coating experts will tell you that the coating is only as good as the surface preparation. You could have the world's most awesomest coating, like Unobtainium Enpregnated Epoxy, that requires no mixing, has infinite pot life, is unaffected by temperature and humidity, applies itself perfectly with no runs or sags, flows out to an auto showroom finish, cures completely the instant that you snap your fingers and is harder than the hubs of hell. But, if it doesn't stick to the surface, it is worthless. Surface preparation is very important if you want that coating to Bond, James Bond, to the surface.

I always hand sand and vacuum my wood surface before gluing with epoxy or laying-up fiberglass set in epoxy. The hand sanding removes any unseen contaminates and gives it a nice sanded wood grain look. The vacuuming removes the dust and wood particles. However, even after vacuuming, I notice that there is still a film of dust. How much does that last little bit of dust affect the strength of the bond between the epoxy and the plywood? I dunno, so I decided to find out.


Gotta Luv the Internet

Some internet searching revealed a little. But since the internet is either porn or BS, I was skeptical. Thanks to the internet, anyone with a big mouth and a modem can get published world-wide. I found a boat building forum where someone posted that you should always wipe the wood surface with tack cloth before applying epoxy.

Wait a minute!!!

I seem to recall learning over 20 years ago...

NEVER USE TACK CLOTH WITH EPOXY!!!

So I went back to my old early 1990's text book, The Epoxy Book by System Three Epoxy. And sure enough, right there on page 12.

From page 12 of The Epoxy Book

Me Own Little Experiment

Instead of being uninformed and posting BS, I decided to find out for meeself. I thought, "I'll prepare some test swatches, like I did with wood stains when building the Sea Scull. I'll do my own little pull tests. And while I'm at it, some of the test swatches will be placed on a surface wiped with tack cloth."

So I cut a little piece of plywood, hand sanded and vacuumed it, as usual. Since I am also interested in the effect of pre-coating (applying unthickened epoxy and letting it cure before applying fiberglass set in epoxy) I divided the piece of plywood into two halves, pre-coated and not pre-coated. Then each half was divided into three sections:

1. Hand sanded, vacuumed and wiped with tack cloth, 
2. Hand sanded, vacuumed and wiped with a dry paper towel and, 
3. Hand sanded and vacuumed.

Notice how the plywood gets darker with additional dust removal. But, the tack
cloth section is a bit too dark. That is probably residue left by the tack cloth

Then I pre-coated half of the plywood piece.

Pre-coated plywood sample

Before I get too far along, please note that this test was conducted using West System 105 epoxy resin and West System 207 epoxy hardener. The 105/207 system is blush free and there was no sanding of epoxy between coats. The only sanding was of the plywood surface.

After the pre-coat cured for a day, I applied 10 ounce fiberglass. But first I masked the outer 1/8-inch with wax paper so epoxy could not bond to the exposed edge grain of the plywood. This would affect the bond and skew the results of the pull test. 

Plywood masked to eliminate edge effects.

10 oz fiberglass saturated with unthickened epoxy

After the fiberglass epoxy cured, I divided the plywood into 18 samples so I would have 3 samples of each scenario.

Plywood divided into 18 samples

I made a clamp to grip the un-epoxied fiberglass at the end of each sample. Each sample was secured in my vice. Then, I pulled the fiberglass until it failed and recorded the pull force with a scale. This is a highly precise experiment using only the most accurate fisherman's scale in my tackle box.

Recording pull force with a fish scale.

So what are the results?
The Envelope Please...

Well, interestingly, or as interesting as pulling fiberglass off of plywood can be, I first observed that there were two modes of failure. The first was just what I wanted to test, the fiberglass disbonding from the plywood. That is, the fiberglass peeled off of the plywood like so...

Fiberglass disbonding from plywood as expected.
Note the disbonding of both samples (above and below the vise jaws).

The second failure mode was the fiberglass tearing at the start of its bond with the plywood. My test method bent the fiberglass at 90° and in some cases the fiberglass tore before any disbonding. The epoxy/plywood bond was greater than the bent fiberglass strength. Thus these torn samples provided no bond strength results.

Fiberglass broke at the start of its bond with the plywood
with no disbondment

The Moment You Have All Been Waiting For...

Data Please

For the numerically inclined...

For the graphically inclined...

Conclusions

First of all, please realize that due to the hokey testing method, the numerical values have no absolute meaning. There are actual ways to determine bond strength but I didn't have the wherewithal to pull off (pun intended) a real test.

However, the numbers are good for relative meaning. In other words, the numbers tell me which is stronger and which is weaker. And that is all I wanted to know. I don't really care how strong the bond is, I just want to know what surface preparation is weakest and what is strongest.

The first thing that jumps out is, pre-coating had the same failure load in all three surface preparation methods and the failure load is about as high as any. Looks like pre-coating is a good idea. Without pre-coating, 10-ounce fiberglass absorbs so much of the epoxy that the wood does not get enough epoxy. This is known to epoxy-smiths as "starving the lay up". The result is a less-than-optimal bond between the fiberglass and the wood. So, when working with 10-ounce fiberglass, I pre-coat the wood with unthickened epoxy and let the epoxy cure before applying fiberglass. This allows the wood to soak up as much epoxy as it wants and seals the surface so the wood will not suck future epoxy out of the fiberglass. With 4 ounce fiberglass, I do not pre-coat because the 4 ounce fiberglass is not as thirsty as 10 ounce fiberglass. 

The use of tack cloth provided no additional bond when pre-coating, and as expected, greatly reduced the bond when not pre-coating. Well, what do you expect when you leave a residue of sticky wax on the wood surface? Epoxy hates wax! The use of tack cloth appears to offer no advantage and can offer a significant disadvantage. The use of tack cloth is a rookie mistake. 

Hey! Isn't that odd? A reputable published source was right and some zero on the internet was full of baloney. What a revelation!!!

Surprisingly, wiping the surface with dry paper towels reduced the bond strength somewhat. Maybe the paper towels leave a residue or microscopic paper pieces behind. Also,  wiping the surface with dry paper towels rakes up a bunch of checks in the plywood surface and if you do not wear leather gloves when wiping, you will be picking microscopic splinters out of your palm. And how do I know this? ...The use of dry paper towels is probably a bad idea. 

Just sanding and vacuuming, without pre-coating, resulted in the highest failure load. Although the failure load was not tremendously higher, sanding and vacuuming, without wiping, works just fine.

From here on out, I'll just sand and vacuum. I'll pre-coat when using 10 ounce fiberglass, but no pre-coating when using 4 ounce fiberglass.




Don't Make Me Blush!

So here's another question...how important is it to remove the blush? Well, I use blush-free epoxy so it doesn't matter to me. But, for many years when blush-free epoxies did not exist, I suffered the agony of constant blush removal. So, for grins, I decided to look into blushing epoxies.

Blush, or technically amine blush, is a wax that forms on the surface of epoxy. Here is an excerpt from System Three's Epoxy Book.

From page 4 of System Three's The Epoxy Book

And from West Systems User Manual...

From page 6 of the West System User Manual

The blush forms near the end of the curing process so it is best to wait a while before removing it. If you try to remove it too soon, some more blush will likely form after your first try at removal. More from System Three's Epoxy Book

From page 11  of System Three's The Epoxy Book

However, System Three's Epoxy Book suggests that maybe that blush is not so bad afterall...

From Page 10 of System Three's The Epoxy Book

Enough of the research and speculation. Here is another experiment.

I decided to test the bond strength of four scenarios as follows:

Primary bond: The wood base is coated and the specimen is glued all in one session with no curing between coating and gluing. The joint is allowed to cure at least 72 hours

Secondary bond - 72 hours: The wood base is coated and allowed to cure at least 72 hours. Then, the base is sanded and the specimen is glued to the base and allowed to cure at least 72 more hours.

Secondary bond - tack-free: The wood base is coated and allowed to cure only until tack free (about 2 hours) then sanded and left to cure at least 72 hours. Then, with no further sanding, the specimen is glued to the base and allowed to cure for at least 72 more hours.

Secondary bond - no sanding: The wood base is coated and allowed to cure at least 72 hours. Then, the specimen is glued to the base with no sanding and allowed to cure at least 72 more hours.

Theoretically, the Primary bond should be strongest. The Secondary bond - no sanding should be weakest since the blush was not removed at all. The Secondary bond - 72 hours and Secondary bond - tack-free should be somewhere in between the Primary bond and the Secondary bond - no sanding, with Secondary bond - tack-free should be weaker than Secondary bond - 72 hours since some blush should have formed after the initial tack-free sanding.

I used West System 105 resin with West System 206 hardener. I prepared a wood base and 12 specimens (3 specimens for each scenario).

I coated the base and immediately glued the 3 specimens for the  Primary bond scenario.

I waited about 2 hours for the wood base coating to be tack-free then sanded the area where the three specimens for the Secondary bond - tack-free scenario would be glued. This was a gooey mess, clogging and ruining the sand paper.

I waited at least 72 hours then sanded the area where the three specimens for the Secondary bond - 72 hours scenario would be glued. No gooey mess, no clogging and ruining of sand paper.

Then I glued the all specimens for the Secondary bond - tack-free, Secondary bond - 72 hours and Secondary bond - no sanding scenarios.

After at least 72 hours, I conducted pull tests with a weight scale to measure pull force.

Tensile pull test

My intent was to measure the failure forces of the bond between the base and the specimens in purely tensile loading. However, after my first attempt, I was unable to pull hard enough to fail the specimen in purely tensile loading. So I switched to pulling to create bending stress.

Bending pull test

The Moment You Have All Been Waiting For...

Data Please

For the numerically inclined...

For the graphically inclined...

Conclusions

As expected, the Primary bond scenario exhibited much higher bond strength than the other 3 scenarios.

I was quite surprised by the strength of the Secondary bond - no sanding scenario. The strength is fairly high. I think it indicates that very little blush formed. This is not surprising since the curing occurred at 65°F in very dry air. However, I was not surprised that it was the weakest of the four scenarios.

I was not surprised by the Secondary bond - tack-free and Secondary bond - 72 hours scenarios. Both should have strengths between Primary bond and Secondary bond - no sanding. 

A primary bond is preferred whenever possible. But there are many instances in stitch and glue boat building where a primary bond is not achievable and a secondary bond is required. Since other temperature/humidity/epoxy resin/epoxy hardener combinations could result in drastically different results, I will always sand between coats when using blushing epoxies in secondary bonds and I will allow at least 24 hours of curing before sanding.

Easier yet, I simply avoid blushing epoxies all together, making this whole discussion moot!!!