Developing a Fourier-transform Infrared (FT-IR) Spectroscopy Classification Tool to Identify Common Beach Plastics on Hawaiʻi Island
Date
2021-12
Authors
Contributor
Advisor
Instructor
Depositor
Speaker
Researcher
Consultant
Interviewer
Narrator
Transcriber
Annotator
Journal Title
Journal ISSN
Volume Title
Publisher
Volume
Number/Issue
Starting Page
Ending Page
Alternative Title
Abstract
Many studies rely on Fourier-transform infrared spectroscopy (FT-IR) to identify the plastic polymer types of micro to macro plastics collected from marine and coastal environments. FT-IR research predominantly uses pre-installed reference library software to identify unknown plastic polymer types. However, plastic recovered from the outdoor environment is often altered by heat and ultraviolet light, which changes its chemical composition and consequently its FT-IR spectrum. Reference library limitations make proper identification of weathered plastic spectra challenging. In the following study, various separation and machine learning techniques were developed and evaluated to create a FT-IR beach plastic classification tool. The multivariate classification algorithm of Principal Component Analysis (PCA) served as the model framework for the tool, and two PCA models were developed and tested. The first PCA model served as a baseline and was built based on the spectra of known unweathered plastic polymers, referred to as standards. After five preprocessing techniques were applied to the spectra, the first PCA model separated the 96 standards into eight distinct groups within the modeled 99% confidence limit. Initially, the second PCA model, composed of the same standards with the addition of unknown beach plastics calibrated into the model, was unable to separate plastic polymer types into eight groups. The second PCA model was then modified to reduce the number of plastic polymer standards and beach plastics to the three most common beach plastic types sampled. Polyethylene (PE), polypropylene (PP) and polystyrene (PS) served as the framework for the new model, which classified 77% of beach plastics sampled, compared to the first PCA model, which classified 42% of all beach plastic samples. As a result of numerous metrics and preprocessing applications, developing a robust identification tool for common beach plastics remains a challenge. The results indicate the new PCA model provided an improvement in plastic polymer classification and holds promise to serve as a tool to identify weathered plastics found not only on Hawai'i Island beaches but worldwide.
Description
Keywords
Chemistry, Environmental science, Computer science, Infrared Spectroscopy, Machine Learning, Marine Debris, Plastic Polymers
Citation
Extent
76 pages
Format
Geographic Location
Time Period
Related To
Related To (URI)
Table of Contents
Rights
All UHH dissertations and theses are protected by copyright. They may be viewed from this source for any purpose, but reproduction or distribution in any format is prohibited without written permission from the copyright owner.
Rights Holder
Local Contexts
Email libraryada-l@lists.hawaii.edu if you need this content in ADA-compliant format.