Polymer Biodegradation in Aquatic Environments: A Machine Learning Model Informed by Meta-Analysis of...

Area of Science:

• Polymer Science
• Environmental Science
• Computational Chemistry

Background:

• Polymers pose environmental challenges due to low recycling rates and persistence.
• Biodegradable polymers offer a solution but are underutilized.
• Existing research on polymer biodegradability is limited to specific polymer types.

Purpose of the Study:

• To create a comprehensive dataset of polymer biodegradation.
• To analyze factors influencing polymer biodegradation.
• To develop a predictive model for polymer biodegradation in aquatic environments.

Main Methods:

• Curated an extensive aerobic biodegradation dataset (74 polymers, 1779 data points).
• Conducted a meta-analysis on experimental conditions and polymer structure-biodegradability relationships.
• Developed and validated a machine learning model using polymer features and experimental data.

Main Results:

• The machine learning model achieved an R _test ² of 0.66.
• Key predictors included polymer substructures (e.g., R-O-R, -OC(═O)-), molecular weight, thermal decomposition temperature, and side chains.
• Feature importance analysis identified specific structural elements that enhance or inhibit biodegradation.

Conclusions:

• The study provides a robust machine learning tool for predicting polymer biodegradation.
• Findings advance the understanding of structure-biodegradability relationships across diverse polymers.
• This research facilitates the design of novel, environmentally friendly biodegradable polymers.