Identification of cellulose textile fibers

Area of Science:

• Materials Science
• Analytical Chemistry
• Textile Engineering

Background:

• Accurate textile fiber identification is vital for effective waste textile recycling.
• Current methods often focus on broad fiber categories, insufficient for chemical recycling needs.
• Detailed fiber composition and polymer properties are essential for advanced recycling.

Purpose of the Study:

• To develop and validate a non-destructive method for classifying natural and regenerated cellulose fibers.
• To assess the utility of near-infrared imaging spectroscopy and chemometrics for this classification task.
• To explore the potential for identifying fibers based on polymer chain length for recycling applications.

Main Methods:

• Near-infrared (NIR) imaging spectroscopy was employed for data acquisition.
• Chemometric models were developed and trained using spectral data from textile samples.
• Classifiers were validated using cross-validation and testing on known fiber compositions (cotton, viscose, lyocell).

Main Results:

• Classifiers achieved 100% true positive rates in model cross-validation.
• An average of 8-9 out of 10 test set pixels were correctly identified for cotton, viscose, and lyocell.
• The study demonstrated the potential to identify fibers based on polymer chain length.

Conclusions:

• Near-infrared imaging spectroscopy with chemometrics offers a highly accurate method for classifying cellulose fibers.
• This technique can support monitoring and controlling fiber dosing and dope viscosity in chemical recycling.
• Identifying fibers by polymer chain length opens new avenues for estimating dope viscosity and advancing cellulose modification research.