Several factors that affect the discoloration rate of the ethylene-vinyl acetate ( EVA) copolymer encapsulants used in crystalline-Si photovoltaic ( PV) modules upon accelerated exposure have been investigated primarily by employing UV-visible spectrophotometry, spectrocolorimetry, and fluorescence analysis. A variety of film samples including the two typical ( unprimed) EVA formulations, A9918 and 15295, were studied. The films were laminated, cured, and exposed to either a concentrated 1-kW Xe or an enhanced-UV light source. The results indicate that the extent of EVA discoloration can be affected by factors of two general categories: chemical and physical. In the chemical category, the degradative factors include (1) EVA formulation, (2) presence and concentration of curing-generated, UV-excitable chromophores that depend on the type of curing agent used, (3) loss rate of the UV absorber, Cyasorb UV 531 TM, (4) curing agent and curing conditions, and (5) photobleaching reactions due to diffusion of air into the laminated films. In the physical category, the factors involve (6) UV light intensity, (7) UV-filtering effect of glass superstrates, (8) gas permeability of polymeric superstrates, (9) film thickness, and (10) lamination-delamination (maybe chemical and/or mechanical effect, too). Photodecomposition of the Cyasorb was first verified in cyclohexane solutions and then in Elvax 150 TM ( EVX) films (the copolymer without any additives and curing agent). Cyasorb decomposition rates in cyclohexane solutions are exponentially proportional to the light intensity, but can be greatly reduced by a free-radical scavenger, Tinuvin 770 TM, and furthermore by an antioxidant, Naugard P TM. The discoloration rate of EVA increases with increasing loss of Cyasorb UV 531 and is faster for the EVA A9918 films that have a greater concentration of UV-excitable chromophores generated from a slower curing than for the EVA 15295 films that are fast cured. In general, the loss rate of the UV absorber and the rate of discoloration from light yellow to brown follow a sigmoidal pattern. A reasonably good correlation for changes in transmittance at 420 nm, yellowness index, and fluorescence peak area (or intensity ratio) is obtained as the extent of EVA discoloration progresses. No discoloration was observed for the laminated EVX films that contain no stabilizers and curing-generated chromophores. The discoloration rate of both types of EVA can be largely reduced by UV-filtering glass superstrates that remove UV< 320nm. Photobleaching reactions are responsible for the non-discoloration of unlaminated EVA, the visually clear perimeter around the edges of laminated samples, and the EVA films laminated with gas-permeable polymer film superstrates. Delamination of EVA films from the top glass superstrate was observed after prolonged UV exposure.