Date of Award

May 2023

Degree Type

Thesis

Degree Name

Master of Science

Department

Biological Sciences

First Advisor

John A Berges

Committee Members

John A Berges, Marcia Silva, Rebecca Klaper, Dong Fang Deng

Keywords

machine learning, Microplastics, microscopy, Nile Red

Abstract

Identifying and quantifying microplastics (i.e., plastic particle that is smaller than 5 mm) in the environment is critical to our understanding of them because of potential harm to humans and wildlife. Traditional identification tools such as observing and sorting under microscopes are tedious and error-prone, especially in sediments. The lipid stain Nile Red can also stain plastics and has been used previously with some field samples. In the present study I applied it to some of the smaller size categories of particles than have previously been considered, in freshwater beach sediments. My objectives were: 1) to develop a Nile Red staining method and 2) to demonstrate its feasibility by sampling beach sediments on part of the coast of Lake Michigan. Beach sediment samples were collected from multiple sites along the Lake Michigan coast in March, June, and August, year. Plastics were separated out using floatation (with 10 M ZnCl2) and digestion (with 30% H2O2 and 7.2 mM Fe(II)SO4) protocols, followed by Nile Red (1 µg/mL in 55% DMSO) staining. Measuring the fluorescence of stained particles (epifluorescence microscopy at 40 x magnification, with excitation at 480 nm and emission at 535 nm) allowed for several plastic polymers to be easily and quickly identified, and quantification was automated using image processing software (Fiji package of ImageJ) and a machine learning algorithm. Particle numbers per gram of sediment ranged from 450 to 1211 and particle size from 5.1 µm (the resolution limit of reliable detection of particles at this magnification) to 458 µm. It was estimated that plastics averaged approximately 2.85 µg per kg of sediment on these beaches. With one exception (Lakeshore Park on June 2nd, 2021) there were no significant (p > 0.05) changes in particle density throughout the seasons sampled, nor among sample sites. The present study detected far more microplastic particles (a total of 50,276 in present study), and far smaller particles (15.1 – 30.0 μm mean diameter) than previous studies. I conclude that the coupling of Nile Red epifluorescence staining with automated analysis may be invaluable in correctly accounting for the microplastics currently in waterways, sediments, and wildlife.

Available for download on Sunday, January 26, 2025

Share

COinS