This article may be helpful. I dont have access to the whole thing, but it looks to be relevant.
I did find someone bullet point the conclusions though:
● The impedance spectra for the samples that were anodized in the R&D tank mixed acid were very stable during 14 days exposure.
● The impedance spectra for the type III samples suggest that oxide layers are more complex than those for the other samples. The impedance spectra did not change significantly during the 14 days exposure time which suggests that these anodized surfaces were very corrosion resistant.
● The impedance spectra for the production tank samples were very stable for an exposure time of 365 days showing that these surfaces had very high corrosion resistance.
● The very high and stable values of the pore resistance Rpo for all three processes indicate that the newly developed anodizing and sealing processes were very effective.
Therefore im guessing that either the anodized layer was either too thin, or not properly sealed. There is probably not much else you can do to the yoyo to prevent this, but maybe someone in the field can answer the reason why their yoyos have not stood the test of time?