Not a complete answer - think of this as added notes:
Glass + EVA + Crystalline Silicon + EVA + backsheet:
20-25 year lifetimes are typically attained from good quality panels from reputable manufacturers in real-world conditions with panels permanently mounted in their service position.
Real and available panels are getting 20++ years in those conditions.
I have an old rather degraded but still running BP Solar panel that is > 30 years old.
All EVAs are not created equal and I have seen EVA advertised on Alibaba for "In China use only".
A manufacturer I dealt with in China would use only German EVA and identified it as a major advantage in achieving longer lifetimes than his competitors.
The maker of the best small PV panels that I have found uses Chinese EVA.
Original PV6100 introductory paper - the product has quietly vanished from the market but the requirements and reasons are still valid. 2008 - Australian National University & Dow Corning joint paper.
NREL - types of encapsulant materials and differences
NREL 2008 - accelerated UV test methods for PV encapsulants
ebay ad with useful information
Silicone rubbers are useful but "tricky" for PV use.
Dow Corning introduced their magic PV6100 SR a few years back and it quietly vanished from the market over about 2 years - suggesting that it had unexpected priblems.
Dow Corning Sylgard 184 is available but very niche.
PV Encapsulants do not work as most people think.
Their main task is NOT to keep water out but
To ensure that the corrodable material is exposed to water vapour and not liquid water. As wv has ~ 1000 x lower water density than liquid water the corrosion rates are proportionally slower all else being equal.
Minimising dissolved water % in the SR helps address point 1.
Creating voidless contact with encapsulated materials addresses point 1.
Also, materials should be "low modulus' aka stretchy and rubber like to allow thermal and other stresses to be minimised. SRs can be substantially better than EVA in this respect.
Not form corrosive degradation products.
Not form transparancy degradation products.
Not degrade :-) - or, degrade as little as possible in high UV environments. The Si-Si bonds in the main chains have bond energies somewhat above the quanta energy in most terrestial UV - unlike eg epoxy resins C-C bonds which are able to be broken by UV quanta energy.
Low cost is always nice.
Water levels in a glass+ EVA panel equalise to ambient levels in a few months.
SR's are typically 10x MORE water permeable than EVA. What is needed is low DISSOLVED water %, no voids and low modulus (stretchy). Suitably designed SR's CAN be better than EVA in these respects.
Available solar encapsulant SRs include
GE RTV 615
Dow Sylgard 184 & 164
QSIL 216 & 220
I have only seen results for Sylgard 184 which is well spoken of. Others may be as good but I have no information. QSil is said to use Dow silicone products in its formulations