Machine Learning for Infrared Spectral Classification of Polyvinyl butyrals with Identical Chemical Groups: An Example for Undergraduate Chemistry Classes

Machine learning (ML) is extensively applied in chemistry, particularly in vibrational spectroscopy. However, few teaching examples effectively demonstrate the capabilities of ML in classifying polymeric materials, exhibiting subtle spectral differences that elude visual discrimination. This study presents a teaching example specifically tailored for undergraduate students to acquire the skills necessary to employ ML models in the classification of infrared spectral data from different types of polyvinyl butyrals (PVBs). The course encompasses fundamental knowledge of PVB structure and synthesis, a comprehensive spectral analysis workflow for constructing classification models, specific data processing techniques, and practical implementation of a student-synthesized PVB product in a laboratory demonstration. Assessment of students' knowledge acquisition is conducted through assignments, and student attitudes toward this course via submitted self-reflection surveys are discussed. This study underscores the efficacy of classroom examples in developing students' abilities and fostering their interest in amalgamating chemistry and artificial intelligence. The knowledge and techniques acquired in this course hold practical implications for quality control, process monitoring, and material identification in industry.