Skip to main content
ARS Home » Research » Publications at this Location » Publication #267046
Title: Predicting cotton stelometer fiber strength by fourier transform infrared spectroscopy

Author
item Liu, Yongliang
item Gamble, Gary
item Thibodeaux, Devron

Submitted to: Applied Engineering in Agriculture
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 5/7/2012
Publication Date: 8/14/2012
Citation: Liu, Y., Gamble, G.R., Thibodeaux, D.P. 2012. Predicting cotton stelometer fiber strength by fourier transform infrared spectroscopy. Applied Engineering in Agriculture. 28(4):585-591.

Interpretive Summary: The strength of cotton fibers is one of several important end-use characteristics. In routine programs, it has been mostly assessed by automation-oriented high volume instrument (HVI) system. An alternative method for cotton strength is near infrared (NIR) spectroscopy. Although previous NIR models have suggested a promising in accurately and reliably predicting HVI strength, in the latest research we have reported a much improved NIR model for HVI strength with a proposal of applying the pre-screening procedure to determine appropriate calibration samples. As a different and complementary approach, the present study was involved with partial least squares (PLS) analysis on mid-infrared (IR) spectra and cotton Stelometer strength. The model performance from the 1800-800 cm-1 IR region was nearly equivalent to that from the full 3600-600 cm-1 region. Considering the heterogeneous distribution of strength in native fibers and different sampling spots between IR spectral and reference measurement, a 90% confidence interval was applied to exclude outlier samples from the calibration and validation sets. The recalibrated model revealed the feasibility of the IR technique for the quantitative determination of cotton Stelometer strength. Of most interest is that the capability of IR model for Stelometer strength is in good agreement with the NIR model for HVI strength, verifying the potential of NIR technique for robust and reliable measurement of cotton strength.

Technical Abstract: The strength of cotton fibers is one of several important end-use characteristics. In routine programs, it has been mostly assessed by automation-oriented high volume instrument (HVI) system. An alternative method for cotton strength is near infrared (NIR) spectroscopy. Although previous NIR models have suggested a promising in the prediction of HVI strength, in the latest research we have reported a much improved NIR model for HVI strength with a proposal of applying the pre-screening procedure to determine appropriate calibration samples. As a different and complementary approach, the present study was involved with partial least squares (PLS) analysis on mid-infrared (IR) spectra and cotton Stelometer strength. The model performance from the 1800-800 cm-1 IR region was nearly equivalent to that from the full 3600-600 cm-1 region. Considering the heterogeneous distribution of strength in native fibers and different sampling spots between IR spectral and reference measurement, a 90% confidence interval was applied to exclude outlier samples from the calibration and validation sets. The recalibrated model revealed the feasibility of the IR technique for the quantitative determination of cotton Stelometer strength. Of most interest is that the capability of IR model for Stelometer strength is in good agreement with the NIR model for HVI strength, verifying the potential of NIR technique for robust and reliable measurement of cotton strength.