An analytical strategy based on Fourier transform infrared spectroscopy, principal component analysis and linear discriminant analysis to suggest the botanical origin of resins from Bursera. Application to archaeological Aztec Samples

Abstract

Bursera species are the source of oleoresins that have been used by pre-Columbian American cultures as adhesives, raw materials for molding figurines, ritual offerings, among other uses. Spread along different museum collections all over the world, pre-Columbian artefacts contain these resins. The preservation and understanding of the technology of fabrication of these pieces constitute a major concern for conservators, historians and archaeologists. Few studies have so far dealt with the chemical composition and the botanical origin of Mexican copal, owing maybe to the difficulty on the procuration of resins from known botanical origin. In this work, fresh resins from six Mexican Bursera species, namely B. bipinnata, B. excelsa, B. grandifolia, B. laxiflora, B. penicillata and B. stenophylla, were analyzed by Fourier-transformed infrared spectroscopy (FTIR). Main spectral band positions were selected for chemometric analysis using principal component analysis (PCA), based on the loading plot of chemometric analysis. Sample distribution patterns were investigated with PCA. Score plots revealed a sample agglomeration with good differentiation in 5 out of the 6 species. This method was validated by linear discriminant analysis (LDA) with a 95.2% of global positive recognition for certified origin species. To compare the efficiency of this approach, high performance liquid chromatography coupled to diode array detection (HPLC-DAD) and FTIR results were coupled to PCA and LDA, for the same set of samples. “FTIR showed 94.4% of samples correctly assigned on the confusion matrix and 91% on the cross validation one. HPLC-LDA showed 100% of correct assignment in the confusion matrix and 95% on the cross validation one. These results are encouraging, as FTIR is much faster and less expensive than chromatographic techniques and it could more readily be available in conservation laboratories. Finally, an application to the identification of the botanical origin of four archaeological Aztec copal samples was performed and the model suggested an origin on B. bipinnata/B. stenophylla for these archaeological samples.