Linear and Nonlinear Calibration Methods for Predicting Mechanical Properties of Polypropylene Pellets Using Raman...

Area of Science:

• Materials Science
• Analytical Chemistry
• Polymer Science

Background:

• Traditional mechanical property testing of polypropylene (PP) can be time-consuming and destructive.
• Quantifying mechanical properties of industrial PP pellets requires efficient and reliable methods.

Purpose of the Study:

• To develop a nondestructive and faster methodology for quantifying mechanical properties of polypropylene (PP) pellets.
• To establish robust multivariate calibration models using Raman spectroscopy data.

Main Methods:

• Raman spectroscopy was employed to collect spectral data from various PP samples.
• Multivariate calibration models, including Partial Least Square Regression (PLS) and Artificial Neural Network (ANN), were developed.
• Genetic algorithms (GA) were utilized for wavelength selection and variable reduction to enhance model robustness.

Main Results:

• Models were built by correlating Raman spectra with ASTM-determined mechanical properties (Young's modulus, tensile strength, elongation at yield, flexural modulus).
• The combination of GA and ANN (GA-ANN) yielded the best multivariate calibration models.
• Excellent model performance was achieved with low relative standard errors (%RSE): 0.17–0.41% for training and 0.42–0.88% for validation.

Conclusions:

• Raman spectroscopy, coupled with advanced chemometric methods like GA-ANN, provides a highly accurate and efficient approach for quantifying PP mechanical properties.
• This nondestructive technique offers a significant improvement over conventional testing for industrial applications.
• The developed methodology enables rapid quality control and material characterization of polypropylene pellets.