[Investigation of fibrous cultural materials by infrared spectroscopy]

Cultural fibrous material includes both important categories, i. e. textile and paper, consisting of precious cultural materials in museum, such as costume, painting, and manuscript. In recent years more and more connoisseur and conservator's concerns are, through nondestructive method, the authenticity and the ageing identification of these cultural relics especially made from fragile materials. In this research, we used attenuated total reflection infrared spectroscopy to identify five traditional textile fibers, alongside cotton, linen, wool, mulberry silk and tussah silk, and another five paper fibers alongside straw, wheat straw, long qisong, Chinese alpine rush and mulberry bar, which are commonly used for making Chinese traditional xuan paper. The research result showed that the animal fiber (wool, mulberry silk and tussah silk) and plant fiber (cotton and linen) were easier to be distinguished by comparing the peaks at 3 280 cm-1 belonging to NH stretching vibration and a serious peaks related to amide I to amide III. In the spectrum of wool, the peak at 1 076 cm-1 was assigned to the S-O stretching vibration absorption of cystine in wool structure and can be used to tell wool from silk. The spectrum of mulberry silk and tussah silk seems somewhat difficult to be identified, as well as the spectrum of cotton and linen. Five rural paper fibers all have obvious characteristic peaks at 3 330, 2 900 cm-1 which are related to OH and CH stretching vibration. In the fingerprint wavenumber range of 1 600 - 800 cm, the similar peaks also appeared at 1 370, 1 320 cm-1 and 1 162, 1 050 cm-1, both group peaks respectively are related to CH and CO vibration in the structure of cellulose and hemicellulose in paper fibers. Although there is more similarity of the infrared spectroscopy of these 5 paper fibers, some tiny difference in absorbance also can be found at 3 300 cm-1 and in the fingerprint range at 1 332, 1 203, and 1 050 cm-1 which are related to C-O-C vibration in cellulose. Moreover, in order to explore direct and simple method to identify different materials with similar spectrum,. the principal component analysis (PCA) was applied to separate cotton and linen, mulberry silk and tussah silk, as well as five paper fibers. To eliminate and reduce the spectral scattering caused by sample uneven surface roughness, the multiplicative scatter correction (MSC) has been applied based on total spectral data. The result showed that the score plot using the first two principal components can effectively categorize both group textiles of cotton and linen, as well as mulberry silk and tussah silk, and they have similar chemical structure. For five paper fibers, the PCA was applied in different spectral range (918-550, 1 280-918, 1 700-1 280 and 3 800-2 800 cm-1), and the best result appeared in the range from 3 800 to 2 800 cm-1, in which the five paper fibers can be well categorized. This research showed that infrared spectroscopy combined with principal component analysis has great potential advantage on identifying fibrous materials with similar structure.