Thanks Jiri Vohlidal and Ahmed Mourtada Elseman for your reply,
The use of software for deconvolution is not an issue. I'm using origin and OMNIC, but as in this 1000 to 1300 cm-1 region, a large number of CFX peaks are present including C-O peak.
The issue is to find peaks locations as different database shows same points for C-O and CFx peaks and in previous work, no one deals it simultaneously.
Firstly, try to subtract LDPE polymer contribution from the spectra of both plasma coatings as thoroughly as possible. Probably, you'll get a cleaner difference spectrum for poly-C4F8 coating since the latter doesn't contain aliphatic C-H contribution.
Then separately deconvolve the remaining poly-AA and poly-C4F8 absorption bands to extract the corresponding components.
I guess it is reasonable to start from deconvolving the band in the poly-C4F8 coating spectrum. Most probably, you otta deconvolve the composite bands in the neat coatings spectra only within judgeously chosen comparatively short wavenumber intervals: say 3600-3000 cm-1 for OH, 1800-1500 cm-1 for C=O+COOH, and 1500-800 cm-1 for C-O or C4F8. This should help to obtain more precise results (bands parameters) for the bands of interest to be deconvolved.
Finally, in the same manner try to deconvolve composite absorption bands of interest in the spectra of your mixed (AA/C4F8) plasma polymer coatings. This time, while deconvolving rather complex absorbance in the range of 1500-800 cm-1 you can use the knowledge obtained earlier (bands parameters) of the C4F8-bands. E.g., as a first approximation, you can now fix the deconvolved C4F8-peaks positions to find possible C-O contribution in the spectra of mixed coatings.
Good Luck!
P.S. If you are not happy with the deconvolution programs in your hands, try to find Systat's "PeakFit", which is very friendly and precise.