Submitted to: Journal of Cotton Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/2/2011
Publication Date: 8/1/2011
Citation: Fortier, C.A., Rodgers III, J.E., Foulk, J.A. 2011. Investigation of the impact of instrumental and software applications on cotton and botanical trash identification by ultraviolet-visible spectroscopy. Journal of Cotton Science. 15:170-178.
Interpretive Summary: Botanical cotton trash can become comingled with cotton fiber (lint) during the harvesting, ginning, and processing of cotton. Currently, conventional methods such as the High Volume Instrument (HVITM) and “classing” methods yield indirect qualitative information regarding the specific identification of botanical cotton trash types. Specific cotton quality measurements which offer individual trash type information may lead to improved removal of this foreign matter. A feasibility study was performed to classify botanical cotton trash types using Ultraviolet-Visible (UV-Vis) spectroscopy and Near-Infrared (NIR) spectroscopy. Both methods have been shown to identify cotton trash types with NIR yielding a much higher accuracy rating (98%) compared to UV-Vis spectroscopy (67%). These methods were compared using instrumentation, and chemometric software packages.
Technical Abstract: Given the worldwide production, usage, and manufacturing of cotton, protocols that could identify botanical cotton trash components which can become comingled with cotton could be advantageous for quality assessment prior to spinning. Conventional methods such as the High Volume Instrument (HVITM) or Shirley Analyzer do not classify or yield specific trash component information. A program was implemented 1.) to determine the efficacy of the UV-Vis spectroscopy technique to identify cotton trash types and 2.) to compare these identification results to the results of the Fourier-Transform Near-Infrared (FT-NIR) technique to identify cotton and individual cotton trash components. Chemometric routines involve pre-processing methods and evaluation of specific spectral wavelengths to enhance spectral differences among individual pure samples of cotton trash components and cotton fiber. The chemometric software package, Unscrambler, afforded a 67% correct botanical trash identification. The utility of this method to correctly identify cotton fiber and cotton trash components was compared to the FT-NIR spectrometer. Overall, a higher percentage of correct identifications (98%) were observed using the FT-NIR spectrometer coupled with the OPUS IDENT software package. When comparing OPUS IDENT and Unscrambler software packages for sample identification by uploading NIR data into Unscrambler, the FT-NIR identification results with Unscrambler were still significantly superior to the UV-Vis identification results. Thus, the FT-NIR technique proved to be a better technique than the UV-Vis technique at identifying cotton trash types.