Zhou Jinfeng, Wang Rongwu, Wu Xiongying, Xu Bugao
1 Department of Merchandising and Digital Retailing, University of North Texas, Denton, TX, USA.
2 Key Laboratory of Textile Science & Technology Ministry of Education College of Textiles, Donghua University, China.
Appl Spectrosc. 2017 Oct;71(10):2367-2376. doi: 10.1177/0003702817713480. Epub 2017 Jun 6.
Cashmere and wool are two protein fibers with analogous geometrical attributes, but distinct physical properties. Due to its scarcity and unique features, cashmere is a much more expensive fiber than wool. In the textile production, cashmere is often intentionally blended with fine wool in order to reduce the material cost. To identify the fiber contents of a wool-cashmere blend is important to quality control and product classification. The goal of this study is to develop a reliable method for estimating fiber contents in wool-cashmere blends based on near-infrared (NIR) spectroscopy. In this study, we prepared two sets of cashmere-wool blends by using either whole fibers or fiber snippets in 11 different blend ratios of the two fibers and collected the NIR spectra of all the 22 samples. Of the 11 samples in each set, six were used as a subset for calibration and five as a subset for validation. By referencing the NIR band assignment to chemical bonds in protein, we identified six characteristic wavelength bands where the NIR absorbance powers of the two fibers were significantly different. We then performed the chemometric analysis with two multilinear regression (MLR) equations to predict the cashmere content (CC) in a blended sample. The experiment with these samples demonstrated that the predicted CCs from the MLR models were consistent with the CCs given in the preparations of the two sample sets (whole fiber or snippet), and the errors of the predicted CCs could be limited to 0.5% if the testing was performed over at least 25 locations. The MLR models seem to be reliable and accurate enough for estimating the cashmere content in a wool-cashmere blend and have potential to be used for tackling the cashmere adulteration problem.
羊绒和羊毛是两种具有相似几何特性但物理性质不同的蛋白质纤维。由于其稀缺性和独特特性，羊绒是一种比羊毛昂贵得多的纤维。在纺织品生产中，羊绒常被有意与细羊毛混纺以降低材料成本。识别羊毛 - 羊绒混纺物中的纤维含量对于质量控制和产品分类很重要。本研究的目的是开发一种基于近红外（NIR）光谱法估算羊毛 - 羊绒混纺物中纤维含量的可靠方法。在本研究中，我们使用整根纤维或纤维片段制备了两组不同纤维比例（共11种）的羊绒 - 羊毛混纺物，并收集了所有22个样品的近红外光谱。每组的11个样品中，6个用作校准子集，5个用作验证子集。通过参考近红外波段与蛋白质中化学键的对应关系，我们确定了六个特征波长带，在这些波段中两种纤维的近红外吸收能力有显著差异。然后我们用两个多元线性回归（MLR）方程进行化学计量分析，以预测混纺样品中的羊绒含量（CC）。对这些样品的实验表明，MLR模型预测的CC与两组样品（整根纤维或片段）制备中给出的CC一致，并且如果在至少25个位置进行测试，预测CC的误差可限制在0.5%以内。MLR模型似乎足够可靠和准确，可用于估算羊毛 - 羊绒混纺物中的羊绒含量，并有可能用于解决羊绒掺假问题。
