不同塑料基体木塑复合材料(WPC)的识别及主要组分的定量分析对于废弃WPC产品的分类回收、 高效再利用, 以及产品生产过程中的质量控制、 产品销售和使用过程中规范市场秩序和维护消费者合法权益, 具有重要意义。 建立不同塑料基体WPC的主要组分的通用定量分析模型, 有助于降低检测成本, 扩大模型的适用范围。 然而。 目前国内外关于不同塑料基体的WPC定性识别研究, 尚未与WPC主要组分的定量分析相联系, 未能构建完整的技术体系。 WPC主要组分定量分析研究尚局限在单一塑料基体WPC的定量分析模型。 针对此种情况, 分别以聚乙烯(PE)和聚丙烯(PP)为增强体, 杉木为生物质填料, 加入一定量的添加剂后, 采用挤出成型法分别制备了20个不同杉木/PE配比和20个不同杉木/PP配比的WPC样品。 采用溴化钾压片法获取了40个WPC样品的红外光谱, 利用多变量统计软件对光谱数据先进行一阶导数处理, 再进行变量标准化。 利用主成分分析法(PCA)对杉木/PE和杉木/PP两种复合材料进行了分类, 由于PP和PE化学结构的差异明显, 两种复合材料在二维主成分空间中呈带状分布, 每种WPC样品处于相对独立空间, 分类正确率达100%。 利用偏最小二乘法(PLS)建立了两种复合材料通用定量分析模型, 木粉和塑料的校正模型的决定系数 R²分别为0.984和0.985, 校正标准偏差SEC分别为1.034%和1.206%; 木粉和塑料的预测模型的 R²均为0.956, 交互验证标准偏差SECV分别为1.779%和1.792%; RPD值分别为4.83和4.85。 为更客观准确地检验模型的预测能力, 随机选取10个样品对所建通用定量分析模型进行外部验证。 结果显示, 模型预测准确性高, 木粉含量的预测相对偏差在±8%以内, 塑料含量的预测相对偏差在±7%以内。 建立了一套PE基和PP基WPC快速准确的识别方法和通用定量分析模型, 为红外光谱法应用于WPC生产、 质检及回收再利用过程中的定性识别和定量分析奠定了技术基础。

Qualitative and quantitative analysis of different types of wood plastic composites (WPC) made of different plastics is important for waste WPC products classifying, recycling and quality controlin the production process, standardizing market order, protecting the legitimate rights and interests of the consumers during sales and use. Establishing a mixed model used for quantitative analysis of WPCmade of different plastics can reduce the costs and improve model applicability. However, the current studies on qualitative analysis of WPC made of different plastics do not address the quantitative analysis of WPC. Therefore, the complete technical systemcan not be established. There have been no studies concerning the quantitative analysis of WPC made of different plastics. For this purpose, in this study, Polyethylene (PE)and polypropylene (PP) were used as matrix materials, respectively. Chinese fir powders were used as filler, and some chemical regents were added. Then 20 Chinese fir/PE and 20 Chinese fir/PP composites were manufactured by extrusion moulding. FTIR spectral data of 40 WPC samples were obtained by potassium bromide pressed-disk technique. First derivatives and Standard Normal Variate(SNV)were used to preprocess the spectral data by The Unscrambler version 9.2. And the FTIR spectral data were analyzed by principal component analysis (PCA). Results showed that the WPC samples could be grouped according to their plastic matrixes, and the correct rate was 100% due to the differences between PE and PP. Partial least square regression (PLSR) models were developed to predict both wood flour and plastic contents in two types of WPC based the above FTIR spectra. Results indicated that for wood and plastic calibration, the coefficients of determination ( R²) were 0.984 and 0.985, respectively; the standard errors of calibration (SEC) were 1.034% and 1.026%, respectively. For both wood and plastic validation, the R² values were 0.956; the standard errors of cross validation (SECV) were 1.779% and 1.792%, respectively; the ratios of performance to deviation (RPD) were 4.83 and 4.85, respectively; The current model was used to predict the contents of wood and plastic in ten WPCs samples that were randomly selected for external validation. Results show that theaccuracy of the model is high, the relative prediction deviations for wood flour contents were lower than ±8%, and plastic contents were lower than ±7%. A rapid and accurate identification and determination method applied for PE-based WPC and PP-based WPC was established, whichlays the foundation for FTIR’s use in the manufacturing, quality control and recycling.