Characterizing Oxidatively Damaged Cotton Fabrics

Abstract
Treating cotton fabrics with electrogenerated oxygen at a Pt anode provides a lab oratory method for simulating the "catalytic damage" phenomenon. Samples of cotton voile were treated in this manner using different current and temperature conditions to damage the fabric. Diffuse reflectance infrared spectra of a series of these samples treated at 80°C were then analyzed using fuzzy clustering, an unsupervised classification method. The resulting classes were assigned a number (0 to 4) nominally indicating the extent of fabric damage as measured by this spectroscopic technique. This series of spectra together with the damage class numbers was then used to construct a cali bration set by principal components regression. The calibration was successfully val idated with a similar set of samples and also used to predict the relative fabric damage on other oxidatively affected voile samples.