Author
 |
Archibald, Douglas |
|
AKIN, DANNY |
|
Submitted to: Nordic Conference on Flax and Hemp Processing
Publication Type: Abstract Only Publication Acceptance Date: 3/4/1998 Publication Date: N/A Citation: N/A Interpretive Summary: Technical Abstract: Objective measurement of the progress of retting is important for achieving the optimal flax product in the least amount of time. NIR reflectance can probe the molecular composition of stems and involves minimal sample preparation. This study aims to identify NIRR spectral markers for retting which are unaffected by product color, moisture content, and sample packing gdifferences. 'Retted' samples were generated by incubating Ariane flax st specimens (n=24) for times ranging from 1 to 37 hours in a solution contain the Flaxzyme(TM) enzyme mixture. NIRR of air-dried stem specimens was obtained under three different relative humidity conditions. These samples were also evaluated by conventional methods including the Fried's test. Factor-based partial-least squares regression (PLSR) was able to predict th time of the retting treatment from the unprocessed spectral data (n=75), provided that 3 or 4 PLSR factors were used, and the size and position of t tspectral analysis range was optimized. The first two factors correct the slope and offset differences due to scattering, while the third factor typically corrected for differences in the levels of sample hydration. Optimal spectral windows and their tentative structural assignments are: I. 1722-1872nm (wax and cellulose, C-H first overtone, RMSEC=1.1hrs.); II. 1994-2032nm (cellulose, C-O-H combination, RMSEC=1.3hrs.); and III. 2322-2464nm (wax and cellulose, C-H C-C combinations, RMSEC=1.1hrs.). The NIRR method measures the end-point more precisely than the Fried's test and adequate for monitoring the enzymatic transformation, which is considered complete after 8 hours. |
