Falanx, I'm happy to be proved wrong by someone who actually works in this field. Always willing to learn.

The information regarding possible surface stresses came from my interpretation of several articles on shot peening of Ti alloys, some of which showed measureable reductions in fatigue strength of anodised Ti alloy. See for instance this one, esp fig 2.

I don't understand where your 30 nm figure comes from, I thought the colour effect needed a quarter wavelength to occur, that's a minimal thickness of 80 nm.

I know how hard it is to get anything including paint to stick to Ti. I understood that an agent which modified the surface activity (eg silane) was the best bet, on the grounds that etching Ti wasn't easy.