Study on Raman multi-peak fitting and structure quantitative analysis of PAN-based carbon fibers

To ascertain the best multi-peak fitting model for first- and second-order Raman spectral of both high-strength carbon fibers (T series) and high-strength and high-modulus carbon fibers (MJ series), three statistical evaluation methods including R2, Akaike information criterion (AIC) and Bayesian information criterion (BIC) values were calculated. And by establishing the relationship between spectral parameters and mechanical properties, the inconsistencies of fitting results under different peak fitting models were discussed. Results indicated that all fitting models of MJ series showed higher statistical consistency compared with that of T series. A five-peak model consisting of D, G, A, I and D′ peak and a four-peak model including 2I, G′, D+G and 2D′ peak with Voigtian function was recommended for first- and second-order Raman spectral of MJ series. But for T series, the above five-peak model with Lorentzian function might be recommended for first-order spectral. Importantly, for MJ series, it was found that the relationship between ID/IG value and tensile strength or modulus was almost the same as that between full width at half maximum (FWHM) of G′ peak and tensile strength or modulus. It was considered that the second-order spectral parameter of FWHM of G′ peak was helpful to supplement the first-order spectral parameter of ID/IG value, both of which were related to disorder features of carbon fibers.