I was wrong. Or at least my expectations were wrong. Here's what changed my mind completely:
"Two-liter PET bottles produced using a commercial grade of PET were used for this study. This resin did not contain any ultraviolet absorbing additives. These virgin bottles were crushed, stacked five to six deep, and placed in uncovered open sided crates to afford maximum exposure to the elements. These crates of bottles were then placed on the roof of Plastic Technologies, Inc. building in early January. Every three months, the bottles in the crates were agitated so that those on the bottom had a chance over time to move to the top or outside edges. Another set of bottles was stored indoors approximately 18-inches under a fluorescent light source. This light source was left on continuously, exposing the bottles for two months. A third set of bottles was stored and protected from light exposure for one year for use as a control."I can't believe it. Sure, they ran the mandatory expose-it-to-a-continuous-UV-light-source-to-scorch-it-beyond-all-reason, but they also had material exposed to REAL WORLD CONDITIONS. Industrial researchers got right what so many academic researchers have gotten wrong!
The results of the testing were pretty interesting as well. While the sunlight did yellow the PET a little, putting the exposed PET through an extruder (once again, duplicating REAL WORLD CONDITIONS) drastically increase the amount of yellow. While the authors did not offer an explanation, I would guess that the sunlight initiated a degradation reaction (perhaps along the lines of a Photo-Fries reaction) and the additional thermal cycle really allowed for the reaction to run wild.
My hats off to Dr. Schloss and Ms. Brown for getting the research right. You can simulate real world conditions all you want, but nothing beats using the real world conditions. Is it really that difficult of a concept?