I’m like you and I’ve never seen any evidence of this, just heard that you should avoid bubbles. I can’t really imagine a tiny bubble being an issue, but a large bubble would probably be more problematic.
I had a problem last week when raster engraving aluminum signs. The laser is supposed to burn away the black dye in the anodized stock and leave a white area. 10mA at 100 mm/s is ample.
Suddenly long strokes with the laser on started to fade at the end. Long runs with short strokes were OK.
I tried moving the print area on the bed, thinking it was a mirror alignment problem. No difference.
At the output end in the tube there was a large bubble. Purging the bubble fixed the problem.
Wild guess: could the uneven cooling bend the tube so the mirrors didn’t align?
Interesting. I suspect that could be what happened. With a large bubble is certainly conceivable that you could get some distortion due to localized over heating. The other concern you might run into is how overheating near the mirror might effect the seal on the tube end mirror.