The increase in the background may be due to the fluorescence. In certain temperature, you may induce defects that cause the fluorescence. It also may be stress or strain. You also may check if peaks are slightly shifted.

I agree with Dr. Sergey Mamedov

I anticipate, the main cause here is fluorescence. It uplifts the entire background and sometimes this even eats up the Raman peaks! This is why samples having fluorescence or luminescence properties are so difficult for Raman spectroscopy. Quantum dots are most common examples. Fluorescence can have much higher intensity than inelastic scattering. In fact, you may find similar results if you increase the laser power, that promotes local heating.

Sergey Mamedov Souvik Bhattacharjee - if the sample does not absorb, it does not fluoresce. Mukesh Pandey - you did not specify what wavelength you are using, but I presume it is in the visible, where PVA does not absorb.

The large broadband background in Raman is very often due to scattering which enters the spectrometer as stray light, and therefore some morphological changes in PVA at elevated temperatures which increase the scattering could more likely explain the observed changes in Raman background.

This can also happen if you burn a hole in it.

Dear Dr. Hugh J Byrne

I think, you are right. I did not notice that, PVA can not absorb visible light earlier. So, I discussed a common reason for this. But, in this case your explanation seems correct to me. The scattering is the reason.

However, I think, effect of temperature on polymers like PVA is not just morphological. There are other changes as well (e.g., mechanical and vibrational), which is likely promoting a sudden hike in scattering here.

If the spectrum suddenly and irreversibly changes to that of amorphous carbon, then there is probably a hole burnt in the sample.

Thanks to all of you for your responses.

I used 532 nm laser light for temperature-dependent Raman characterization. I found a sudden upshift and also a change of linear Raman spectrum of PVA into a wavey spectrum. The intensity of Raman peak signals coming out of graphene deposited over it remains the same under thermal treatment but the background (PVA) spectrum suddenly shows the changes in shape and position on the intensity axis.

Both fluorescence and scattering result in baseline change. However, as Dr. Byrne said, PVA does not have noticeable absorption near the laser excitation frequency used. Impurity cannot be considered the reason as the same sample was used at or below 200oC analyses. So, the only possibility of fluorescence is the absorbing species generation caused by the thermal degradation (it can be extremely small amount). PVA is stable up to close to 300oC. Therefore, brief heating at 200oC is unlikely the cause of thermal degradation of fluorescing species.

This leads to the second possibility of increased scattering. The crystallization peak as observed by DSC is slightly above 200oC. This is under dynamic heating conditions. Thus, it would crystallize readily using isothermal heating at 200oC. Water that is absorbed in the sample mostly evaporate above 150oC. Therefore, the scattering will be rather efficient upon heating the sample at 200oC. Depending on how long the heating took place, the degree of crystallinity can reach almost 60%. Thus, the baseline increase caused by the increased scattering efficiency is likely.

If you take commercial transparent films made of any polymer, you will find plenty of fluorescence. Some of them may fluoresce, and some of them are not. It depends on additives. You don’t see strong absorption but keep in mind the efficiency of Raman scattering and fluorescence, which is a few orders of magnitude stronger compare with the Raman. PP, PE, PS, and others from different vendors show significant differences in fluorescence.

To Hugh J Byrne: The edge filter should block stray light. I may come close to the elastic line with a standard edge filter to 45-50 cm-1 with no problem, and the first fingerprint region is much further. So, scattered light will not contribute to the background.

The burning is another story, and it will contribute to the fluorescence background, but I assume that the laser power does not initiate it.

Hatsuo Ishida sir, thanks for your reply. I find your explanation quite matching with my observation up to a good extent.

The heating @200 degrees Celcius surely affects the crystallinity of PVA film. (seems thermal destruction since it does not regain the original condition when cooled down to room temperature).

Sergey Mamedov - please read the paper I attached to the discussion earlier.

Hugh J Byrne - Please, read my papers about nano-powders. The size of the nano-powder ranged from 8 nm to 50-40 nm. I did not see significant scattering background in the spectra of TiO2 for the Raman line at 143 cm-1. However, I saw very similar video images for catalysis samples based on SiO2 (powder) or zeolite + metal oxides, but no considerable background was detected. So, it may be sample-dependent, and your case is rare. But what happens if you close the confocal hole to 50 or 30 microns on your HR800? Did you see changes in the background relative to the signal? Besides, MPLAN x100 has a significant chromatic aberration for 785 nm excitation, and you may try the NIR objective to get better confocal performance.

Large backgrounds for biological tissues (especially for FFPP and subsequently dewaxed). This is often erroneously mistaken for fluorescence, even at 785nm.

the background is preferentially reduced by closing the confocal hole (see for example the study in Article Raman Microscopy: Complement or Competitor?