I've been reading a lot about the "boat pox" that many fiberglass hulls suffer from. There is no shortage of ideas as to what causes this issue, but very little in the way of science, which is publicly testing ideas by systematic observation and controlled experiment.
Fiberglass hulls typically consist of three major layers:
Blisters are bubbles in the gelcoat that are typically filled with a thin acidic fluid that smells like uncured polyester resin, which is the plastic component of fiberglass. Blisters appear in different forms. The most typical is a larger number of small blisters, but some boats have fewer (up to a couple of hundred) larger blisters.
Here are some observations drawn from various sources across the Web:
The last point is crucial, and is what made everything fall into place for me. Or at least give me an idea.
Here is a general rule of materials science: on a molecular level pretty much everything is porous, and therefore permeable. That is, the molecules that make up the polyester resin component of fiberglass are big enough that there are microscopic spaces between them. Smaller atoms and molecules--water, for example--can migrate through those spaces, eventually saturating them. Only in rare cases can this process be absolutely stopped. For example, most metals are impervious to anything larger than helium. But that doesn't mean they won't still get saturated with hydrogen, given enough time.
The second general rule is: once the molecules of a fluid saturate a solid, chemistry is bound to occur. Noble gases aside, almost everything is at least slightly reactive with everything else, and water is pretty much a universal solvent.
From these facts we can conclude that the inter-polymer spaces of every fiberglass hull everywhere that has been exposed to a reasonable amount of water--inside or out--is saturated with H2O molecules. It would be very interesting to perform an experiment of putting a piece of cured fiberglass resin in water and seeing if it expands over a year or two, or even less. My bet is that it would, but only if there is no mat or cloth present. In the presence of mat or cloth--glass fiber reinforcement--the effect would be to increase the strain in the glass fibers, not to increase the volume of the polyester.
This is a test I can perform.
Gelcoat is non-reinforced polyester, bonded to fiber-reinforced polyester underneath. As the boat sits in the water, therefore, the gelcoat and fiberglass underneath will absorb water, and the bond between the gelcoat and the fiberglass will experience increased strain, because it is only that bond that is preventing the gelcoat from increasing in volume.
Where the underlying mat is poorly saturated with resin during laying up, or where the gelcoat's inner surface has been compromised or contaminated before laying up, the bond between the two layers will be weaker, and the gelcoat can separate from the mat, producing a blister.
The interior of the blister creates a captive space where the hydrolized polymer--which is present throughout the resin matrix regardless--is free to come out. By this hypothesis the paint on the surface of the boat keeps the polymer from slowly being leached away by this process of hydrolization, which must be going on all the time.
This is another test I can perform: create a fibreglass block with an interior void, and let it sit in water for a good long time, and see what happens to the void. By hypothesis, it should become slowly filled with the same clear fluid as we see in blisters.
Furthermore, this changes the evaluation of the effects of blisters on hull integrity. The usual story is that blistering is a progressive process: blisters form under the gelcoat and the acidic hydrolized polymer fluid that fills them "eats away" at the fiberglass underneath. But if the polymer is entirely saturated with water there is a low level of hydrolized polymer everywhere that is kept in solid solution by the pressure of the surrounding matrix. On that basis gelcoat blisters are largely cosmetic.
Finally, the cure of blisters may not be to apply a "barrier coat" of epoxy or the like. The lower permeability of epoxy resins will slow the saturation by water and possibly shift the equilibrium value to something low enough for the problem not to recur. But barrier coats are typically applied after full-hull repairs in any case, which could well solve the underlying problem of poor bonding between the gelcoat and the underlying fiber-reinforced resin.
There may be an alternative fix, which suits the bias I bring from my surgical experience, which is to always preserver as much of the underlying anatomy as possible (total knees are a rare exception to this rule). Instead of grinding out the blister and filling with epoxy, it may be possible to puncture, flush and flood the blister with low-viscosity polyester resin that will bond the existing gelcoat to the underlying mat with sufficient strength to prevent the blister from reappearing. This would only work in cases with a smaller number of larger blisters, rather than the more frequent cases of thousands of small blisters covering the whole hull, but it would be fast, cheap and easy, and not introduce any alien materials into the mix.