I agree that the suggestion by Peter using a different wavelength laser is a very good suggestion.  However, I like to know whether your data were collected using the transmission mode set up that light passes through the entire sample thickness, or that your set up is reflect type, where reflected signals from the opaque sample is being analyzed. (in such cases, if the penetration length of the excitation source is less than the film thickness, then you will not detect the signals from the substrate, ad PL signals are entirely due to the film region.   

@ Peter, I tried measuring at two different wavelength(480nm and 532 nm), All gives emission peak at 620nm from all substrate. Peak some time moves with excitation wavelength but not disappear. 

@peter here I attache the file. I do not understand why all peaks at same position. Is it due to instrument calibration problem or common defects in all substrate.