Are there removable/soluble potting compounds that are moisture-proof yet removable without damaging electronic components? I've been able to plunge-mill potted devices to get near enough to components or test points to take measurements. A potting compound that is soluble in some sort of solvent would make failure analysis and rework easier.
The ability of a material to mechanically protect components, keep water out, and withstand environmental conditions is in direct opposition to being able to easily remove it. Ultimately every compound is soluble in something (water, alcohol, acetone, benzene, piranha bath, aqua regia, HF...), but the problem is that components are too.
Depending on your application, you may consider a thin conformal coating, potting with a low melting point compound such as paraffin wax, or placing the circuit in an environmentally sealed enclosure. You may also redesign the circuit itself to have self-test capability, bring test points out to a connector, fail less often, or be cheap enough to throw away when it breaks.
May I suggest a conformal coat instead. For conformal coat, there are three main types.
1. Silicone conformal coat keeps contaminants off the circuit board whilst letting water through, effectively filtering the water. Pure water is a poor conductor so it will not affect the performance of the circuit. This can be removed using most alcohol based solvents, we have some electrolube cleaner that works well.
2. Polyurethane based conformal coat works in a similar way to silicon based CC, but is somewhat more robust. The removal of this is aided by solvents, but scraping with tools (dental picks etc.) is normally required before rework can take place.
3. Parylene CC is entirely watertight and keeps the water off the circuit board completely, but is a pain to remove and almost impossible to touch up after rework.
What you want (although few people know this :-) ) is not something that keeps water out but something which adheres superbly without voids to the surface AND which has a very low dissolved water content. It can pass water vapor but must not retain much water in itself. EVERYTHING let water in . Coatings that work keep it as vapor and keep liquid water away from the corrode-able items. Water vapor is OK.
This is the principle behind Dow Corning PV6100. It is targeted at PV panels BUT would probably do a superb job of doing what you want. You may find it hard to buy. DC are trying to keep its use 'closed shop' it seems.
PV6100 is 10+ times more water permeable than industry standard EVA which is used for almost all crystalline silion PV to glass laminating - but DC claim significantly superior anti corrosion performance due mainly to its low dissolved water ability.
A similar but older and not quite as good material is Sylgard 184
In both cases coating depth can be under 1 mm - but more OK. Remove like silicone rubber (as they ARE silicone rubber). If an area is scraped clear for rework (easy enough) the area can be resealed with new material.
Both compounds are clear (about 92% light transmission).