I could be wrong, but I suspect your deposited SiO2 layer actually is a SiO2.n(H2O) porous one. When dried, its volume tends to shrink, with pore collapse, which causes both an internal stress (plastically resolved by fracture) and also stresses the underlying and on-top layers.
Possibly, your deposition method is not the best choice, or you may try adding annealing steps before and after the SiO2 deposition (that is, before depositing the top Titania layer)
Okay you have mentioned the method, I didn't notice.. Apology.. I have experience in oxidation furnace and PECVD.. I guess I won't be able to assist you..