Detection of Fibrilated Polymeric Contaminants in Wool and Cotton Yarns

Abstract

The contamination of natural fiber yarns with polymeric fibrils is a common problem in the textile industry. This work demonstrates that near-infrared spectroscopy in the 2250 to 2400 nm region is a viable technique for the detection of hydrocarbon-based polymeric contamination in both wool and cotton yarn samples. Both high-density polyethylene and polypropylene fibrils typical in size to those found in industry could be detected irrespective of polymer and yarn color. A principal component analysis model based on first-order derivative spectra was developed in which yarns contaminated with polymeric fibrils could easily be separated from pure wool yarns. The spectra obtained from the contaminated regions of the yarns were found to cluster well with spectra obtained from the pure polymeric materials. With the use of soft independent modeling for class analogy (SIMCA), a single class model for wool was developed on the basis of the raw data. This model easily discriminated between pure wool and polymer-containing samples, but the distinction between the different polymer types themselves was poor. This separation was enhanced, however, when the model was based on first-derivative spectra.