Skip to main content
ARS Home » Research » Publications at this Location » Publication #159495
Title: PREDICTION OF SHIVE CONTENT IN PILOT PLANT PROCESSED FAX BY NIR REFLECTANCE SPECTROSCOPY.

Author
item SOHN, MI RYEONG - USDA-FAS
item BARTON II, FRANKLIN
item MORRISON III, WILEY
item AKIN, DANNY

Submitted to: Near Infrared Spectroscopy Journal
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 9/6/2004
Publication Date: 11/5/2004
Citation: Sohn, M., Barton II, F.E., Morrison III, W.H., Akin, D.E. 2004. Prediction of shive content in pilot plant processed flax by NIR reflectance spectroscopy. Near Infrared Spectroscopy Journal. 12(4):251-258.

Interpretive Summary: In a previous study we reported the possibility of development of calibration model for determining shive or trash content in flax fiber using near infrared spectroscopy. To monitor the efficiency of processing of the fiber, a new calibration model was developed using a sample set with a range of shive expected in the various stages of fiber processing. The model estimates shive content from retted flax fiber what were physically ground to a fine texture ("ground"). To predict samples that one would expect to see on a processing line ("as-is"), we considered a bias and slope correction of the model. The results suggest that the developed NIR model can be used to evaluate the quality of flax fiber during processing.

Technical Abstract: Shive is the main contaminant in flax fiber and affects fiber quality. A calibration model for measurement of the shive percentage remaining in flax during process was developed using near-infrared spectroscopy. The best model was developed from multiplicative scatter correction after second derivative treatment of the spectral data using "ground" materials. The model was applied to a dew- and enzyme-retted flax after various stages of processing at a flax fiber pilot plant. Two sample types, "ground" and "as-is", were used to estimate the shive percentage. A slope and bias coefficients was obtained from linear regression between values estimated by "ground" and "as-is" samples, and an individual slope/bias correction was required for both retted flax. The shive content estimated from "as-is" flax samples tended to gradually decrease with the degree of processing. For enzyme-retted flax, the values ranged from about 25% after first processing step to a value of almost 0% after last processing step. For dew-retted flax, more shive remained in all steps compared to the enzyme-retted flax, showing a range of 56-7% shive. The results indicate the possibility of measurement of shive in flax on a processing line.