Location: Cotton Ginning Research
Title: Discernment of lint trash in raw cottton using multivariate analysis of excitation-emission luminescence spectra Authors
Submitted to: Journal of Analytical and Molecular Techniques
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: April 29, 2014
Publication Date: May 8, 2014
Citation: Gamez, E., Fengshan, J., Ghale, S.B., Hughs, S.E., Anderson, D.M., Rayson, G.D. 2014. Discernment of lint trash in raw cottton using multivariate analysis of excitation-emission luminescence spectra. Journal of Analytical and Molecular Techniques. 1(2):1-9. Interpretive Summary: Being able to accurately discriminate and subsequently remove trash particles from cotton during the ginning process remains a challenge. The use fluorescence spectroscopy offers an innovative approach to identifying trash particles based on chemical structure rather than on particle size and shape. Investigative research into using a 12.5 pH buffered saline solution to extract compounds that will fluoresce appears to provide promising results when 295, 320 and 400 nm excitation wavelengths were used. The resulting emission wavelengths of 348, 435 and 473 nm suggests tryptophan and two other organic compounds appear responsible for the spectral signature differences found in several different plant parts of Pima and upland cotton varieties. The use of U-PCA procedures made it possible to detect these differences, however, some plant parts were easier to separate from the cotton fiber signatures than others. Stem fragments were the easiest material to separate from the cotton fibers in both cotton varieties.
Technical Abstract: Excitation-Emission luminescence spectra of basic (pH 12.5) phosphate buffer solution extracts were used to distinguish among botanical components of trash within seed cotton. All components were separated from whole plants removed from a field in southern New Mexico. Unfolded Principal Component Analysis (U-PCA) was applied to Rayleigh-corrected extract spectra from each plant component. This enabled distinction of seeds, stems, bracts and leaves as trash material. Sensitivity of U-PCA models to sample replicate selection was evaluated. This revealed significant sample dependence for U-PCA models using spectra from stems of both Pima and upland cotton varieties. Dependencies were also observed for seeds, burrs, walls, and fiber from Pima cotton. Spectra from fiber and seed samples of upland cotton also exhibited significant sample-selection sensitivity on U-PCA models. Application of parallel factor analysis (PARAFAC) further revealed three spectral factors enabling trash identification. Elucidated excitation and emission spectra of these factors exhibited excitation wavelengths of maximum intensity of 295, 320, and 400 nm and respective emission wavelengths of 348, 435, and 473 nm. These suggest the presence of 1) tryptophan-containing proteins or polypeptides, 2) NADH, periodic acid, or periodic acid-5’-phosphate, and 3) 6,7-dihydoxy or 7-hydroxy-6-methoxy coumarins as discriminating components.