I think cracking will result due to the large lattice mismatch between Si and PZT. Maybe you have to prepare a buffer layer with increasing lattice constants towards PZT. Sometimes gradient mixtures of Ca --> BaF2 layer are used. This paper reports such layer to deposit PbSe onto Si (Lattice mismatch ~20%).

10.1063/1.96239

You can try to sputter PZT using RF magnetron sputtering.

thanks Benjamin for useful info.

@ Jea, but I have the option of sol-gel only (due to the process advantages).

I was wondering if there is some polymer agent like PEG or other which can help to depsoit a thicker film and relieve the thermal stressess as well.

~5-10 um I think flexible substrate may be possible

Thanks Bin, Can I use Stainless steel? and what should be optimum thickness?

I'm so sorry, This is also my future topic, now I am just trapped with 1um PZT film on silicon base --！