Study Of Near-infrared Spectroscopy On The Chemical Components Character Of Populus×Euramericana

The research uses five different families'Populus×euramericana of different kind of soil from three regions, which are Shunyi, Heze and Langfang, analying the chemical constituents, fiber morphology and structure, and wood basic density. Contrast with each data, in order to find the influence of different soil on Populus xeuramericana's chemical composition. Gathering near-infrared spectrogram of each sample, build the NIR calibration models with OPUS soft. The main conclusions are as follows:(1) The range of all strains of Populus xeuramericana's chemical composition is as follows:The holocellulose content is from 76.49% to 85.40%; The lignin content is from 20.82% to 23.89%; The pentosan content is from 20.84% to 25.06%; The phenethyl alcohol extract content is from 0.75% to 2.40%; The ash content is from 0.47% to 0.81%. The results show that the chemical components character of Populus xeuramericana conforms to the requirements of raw material used in pulping and papermaking. According to the compare of the different soil conditions at the same tree age, the 6 years old Populus xeuramericana cv.74/76'growing in Shunyi loam soil has the best chemical components character. The holocellulose content is 82.11% and the lignin conotent is 22.22%. The 6 years old Populus xeuramericana cv.'74/76'growing in Shunyi loam soil has the worst chemical components character. The holocellulose content is as low as 76.49% and the lignin conotent is 22.97%, higher than others. These characters show that Populus xeuramericana should be good to make pulp and paper, especially to make newsprint and printing paper.(2) The range of the fiber length of Populus xeuramericana in three regions is from 0.88mm to 1.18mm, and the width is from 19.14um to 25.04μm. The data's distribution is centralized. Most length-width ratio is above 45. The Populusxeuramericana's wood basic density is between 0.28g/cm3-0.42 g/cm3, and the crystallinity in wood cellulose is around 20%. Considering the chenmical composition, fiber morphology and trees' growth condition, the Populus xeuramericana cv.'74/76'and Populus xeuramericana cl.'Zhonglin46'should be promoted to plant. Their holocellulose contents are above 80% and lignin comtents are under 22%. Their fiber length is also big, and the length-width ratios are above 37.(3) Contrast with the chemical composition of the same year-old Populus xeuramericana in different regions. The light loam soil is the most suitable for the Populus×euramericana's growth as serving as pulping and papermaking. And the sandy loam soil is better. This shows Populus xeuramericana can be planted in sandy region.(4) The RMSECV of the calibration model of ash content is very small, only 0.0264, and R2 is 94.58%. This shows that the ash content has the very good relevance to the near-infrared spectrum data. This model may apply to fast predict the ash content of unkown Populus xeuramericana sample. The RMSECV of the calibration of holocelloluse and pentosan are both above 0.34. Although these are not as good as the model of ash content, R2 is high,91.25% and 89.85%, respectively. These show that the two models has a good relevance to the near-infrared spectrum data. The two models may apply to fast predict the corresponding content of unkown Populus×euramericana sample. The RMSECV of the calibration model of lignin is a little bigger as 0.364. Its R2 only achieves 82.35%. This shows that the lignin content and the near-infrared spectrum data's relevance is relatively bad. May improve the model later.(5) The R2 of the calibration model of fiber length only achieves 72.17%. This possibly is because the Populus×euramericana samples were in fast-growing period. Thus their fiber morphology were not steady. The RMSECV of the calibration model of fiber length is 0.0326, which shows the relevance is strong. The R2 of the calibration model of the crystallinity in wood cellulose is 89.03%, which shows the ideal predict ability of the model builded with the NIR technology and PLS method. The model reliability is high as RMSECV is 0.212.(6) When use the traditional method to measure the chemical content of holocellulose and lignin, jump-operation's steps are too many during the process.This is easy to cause the firsthand datas' accuracy low, thus the RMSECV of the model to be big. The accuracy of the model is not good. But the model predict effection does not affect to the industry application.