Are you evaporating these materials, or depositing by another method?  Typically, these materials adhere well to one another with heated evaporation.  Is the cracking actual separation of the layers (delamination), or is the cracking through the layers (tensile stress failure)?  How are you cleaning the glass, if this is an adhesion failure?

Yes, the material evaporating using e-beam method. Im using high vacuum chamber to do this process. So, you suggest to turn on the heater for previous layer for adhesion purpose to ZrO2?

FYI, while ZrO2 evoporation, O2 gas has been used. Is it effect the strength of the layer?

Im just think adhesion is the cause to this crack, but maybe it cause by others. The crack only happen when through high temp test(80c).

I already ask someone about plasma cleaning on the glass. But it will effect the surface of the glass. So, plasma cleaning is not an ideal solution for this problem.

Based on your description, I expect it is a tensile stress failure.  If the film is cracked, it has been pulled apart in different areas (in tension).  If the film is peeling off the substrate, that would be adhesion (delamination). 

To change the stress state of a tensile film, you need to make it more dense.  This can often be accomplished by heating the substrate (something like 200 deg. C), depositing more normal to the substrate surface, depositing with a lower O2 gas pressure, or using some form of an energetic process (ion assist, plasma assist, sputtering, etc.).  Depending on how your system is configured, I would expect substrate heating would be the easiest.  However, a more energetic process may be necessary, depending on your final film thickness and testing environment (particularly low humidity, such as N2 purge or vacuum).  Hope this helps.