Hi Shyama,

From my point of view, if you put the same quantity on your plate,  you should first normalize them. Good luck, 

You can also weigh the samples before and after exposure to water.

I am not quite sure what you want to do. If you want to compare the water content... you should normalise to one of the other epoxy peaks and then examine  the OH variation.

First of all, the thicness of the film is important since ATR-FTIR shows  only surface functional groups. Secondly there should be a correlation between  peak absorbans and consantration of a funtional group (OH). 

ATR has a penetration depth of some microns, depending on the frequency - it is not true that it only sees surface functional groups

That is correct. If the film thicness above some microns then the measurenents may not be correct.

The chemometrics tools could be very interesting to help you. If you could send me the experimental file collected by the instrument-software, I mean, export the twenty file in " *.dat " format (or ASCII), I make some attempt to apply chemometric methods and send back you a text with the found correlations. 

I am really thankful to everyone who has replied. Yes, I am measuring the mass gain as well. I will try to use the chemometric tools. Thanks a ton to everyone!

You are welcome. If you want, send me the data in *.data form. I have full interest to help you. I can send back an example with the result of chemometric apply.

best regards.

As of my knowledge, we should not average. Generally in atr-ftir, we take 20 scans. It is always preferable to increase the number of scans say 200. 

For the fundamental understanding of ATR-FTIR, please find attached book.

Measure your sample at different locations (where you can specify the FTIR instrument to run several spectra in one run, e.g. 32 scans, which are automatically averaged into one spectrum), but since you’re doing ATR-FTIR you should not make an average of the spectra taken at different locations with different contact variations between the sample and the ATR crystal. Relatively hard surfaces complicate accurate quantitative measurements (height of absorbance peaks) due to varying contact with the ATR crystal for the different samples. Air between sample and ATR crystal results in a weaker absorbance signal. Unless other conditions indicate otherwise (e.g. inhomogenities, impurities, etc.), the FTIR curves with the largest absorbance peaks represent the most correct measurements on one and the same sample, and hence these curves should be chosen as they are assumed to be the most correct ones. Qualitative measurements (location of absorbance peaks at wave numbers) do not represent a problem as long as the contact area is large enough to ensure a sufficient strong measurement signal. Various details about the ATR-FTIR technique which may be helpful may be found in the following article (may be requested through Research Gate): B. P. Jelle, T.-N. Nilsen, P. J. Hovde and A. Gustavsen, ”Accelerated Climate Aging of Building Materials and their Characterization by Fourier Transform Infrared Radiation Analysis”, Journal of Building Physics, 36, 99-112, 2012.

Do you see any systematic variation between spectra of the same sample? For example, does a water-exposed sample gain or lose intensity in the OH stretch region at 3300 cm-1? If you see this peak changing in the same direction over time (meaning from the first spectrum you took to the last), then averaging them is a bad idea. If the peaks (all the peaks, not just my example) are all of relatively the same height, varying only with noise, then averaging them would be a way to reduce the noise, if only by a factor of 2 (noise varies with the square root of the number of scans).