| Since the20th century, the rapid development of paper industry contributes to a large number of felled timbers. Bamboos replaced timber as essential processed materials of paper industry, board and spinning, for its ability of rapid growth, excellent mechanical property and process ability. The content and distribution of chemical composition of bamboo is the beginning of excellent mechanical property and the important basis of processing technic of bamboo. What’more, the content of chemical composition,which relates to the physical and mechanical property and processing property of bamboo, has a vital role in ensureing the reasonable cutting age. Therefore, the study on content and distribution of chemical composition of bamboo are the key problems to the utilization of bamboo.This study focus on bamboo, which is widely panted and high economic valued. The discrepancy of chemical composition among differential area, year and location were studied by multi-factor analysis of variance (MANOVA), and the chemical composition changed with these three factors were analyzed. Near-infrared spectroscopy technology were used to build a rapid detective model for the detection of hemicellulose, cellulose and lignin content of bamboo. In the end, the distribution of chemical composition of bamboo in the cell, which is very important to the study of macroscopic mechanical properties of bamboo, was analyzed by Confocal Raman Microscopy. The main search result were as follows:(1) The discrepancy of chemical composition among differential area, year and location were studied, and the change rule of chemical composition changed with these three factors were analyzed. The result of ANOVO showed that the discrepancy of hemicellulose of bamboo was non-significant in different place of a city,while between two citys the discrepancy was significant. The discrepancy of cellulose was significant, on the contrast the discrepancy of lignin was non-significant; With the increase of the year, the content of hemicellulose and cellulose trend to gradually decreasing, while the content of lignin varies with different year. Compared with the under, middle and upper positions of bamboo, the bamboo series existed the highest content of hemicellulose and lignin, while the lowest content of cellulose.(2) Elimination of uninformative variable(UVE) and genetic algorithm(GA) were separately combined to successive projections algorithm(SPA) to be UVE-SPA and GA-SPA. These two methods were applied to select characteristic wavelength in the rapid detection model of hemicellulose, cellulose and lignin of bamboo.(3) Near-infrared spectroscopy technology was applied to building the quantitative relationship between the content of hemicellulose and spectral information, and this study Achieved a rapid measurement the content of hemicellulose of bamboo, nine kinds of pretreatment methods were compared to build a partial least squares (PLS) model of full bands, and SPA, UVE-SPA and GA-SPA methods were used to extract characteristic wavelength to build SPA-PLSã€SPA-multiple linear regression (MLR) and SPA-least squares support vector machine (LS-SVM) model. The results of these models showed that the optimal analytical model was UVE-SPA-LS-SVM (1-Wt) model, and the determination coefficients(Rp2) was0.9485,the root mean square error (RMSEP) was0.0059. These results showed that it’s feasible to predict the content of hemicellulose of bamboo by Near-infrared spectroscopy technology, and the prediction accuracy was satisfied.(4) Smoothing algorithm,GA-SPA for selecting characteristic wavelength and LS-SVM methods were combined to achieve a high-precision rapid measurement of cellulose of bamboo. the determination coefficients(Rp2) was0.9310,the root mean square error (RMSEP) was0.0078.(5) Near-infrared spectroscopy technology was applied to achieve a high-precision rapid measurement of lignin of bamboo, the optimal analytical model was SPA-LS-SVM(raw) model, and the determination coefficients(Rp2) was0.8825,the root mean square error (RMSEP) was0.0065. These results showed that it’s feasible to predict the content of lignin of bamboo by Near-infrared spectroscopy technology, and the prediction accuracy was satisfied.(6) Confocal raman microscopy spectra combined PLS and MLR methods were applied to search the important bands for the differentiation of arenchyma cell (PCA), transition tissue between parenchyma cell and fiber cell (TC) and fibre cell (FC), and three Raman bands of1636,1319and1095cm-1, verified to the characteristic peaks of pectin, hemicellulose and lignin,were fonud finally. Subsequently, these three bands were used to chemical imaging analysis.The imaging result showed that a rather large microfibril angle between cellulose fibrils and fibre axis,and hemicellulose and cellulose have similar distribution in the raman chemical image,. The cell corners (CC) and compound middle lamella (CML) were heavily lignified, and a gradual decrease of lignification from the outer layer to the inner layer of the three cells indicate that lignification was first occurred at the CC and CML... |
