AI-assisted design of chemically recyclable polymers for food packaging

TL;DR

Packaging sustainability challenges arise from durable, multi-layer plastics that hinder recycling. The authors implement a polymer informatics workflow that combines curated data, Gaussian process regression for enthalpy and deep learning for other properties, and a Virtual Forward Synthesis to predict eight properties for around $7.4$ million ROP polymers, prioritizing an enthalpy window of $-10$ to $-20$ kJ/mol. This approach identifies thousands of viable candidates, and poly($p$-dioxanone) (poly-PDO) is experimentally validated as a single-layer recyclable polymer with strong water-barrier performance and exceptional chemical recyclability ($>95 ext{%}$ monomer recovery within six hours). The work demonstrates a scalable, data-driven route to accelerate sustainable polymer discovery across existing and novel chemistries, enabling more circular packaging options.

Abstract

Polymer packaging plays a crucial role in food preservation but poses major challenges in recycling and environmental persistence. To address the need for sustainable, high-performance alternatives, we employed a polymer informatics workflow to identify single- and multi-layer drop-in replacements for polymer-based packaging materials. Machine learning (ML) models, trained on carefully curated polymer datasets, predicted eight key properties across a library of approximately 7.4 million ring-opening polymerization (ROP) polymers generated by virtual forward synthesis (VFS). Candidates were prioritized by the enthalpy of polymerization, a critical metric for chemical recyclability. This screening yielded thousands of promising candidates, demonstrating the feasibility of replacing diverse packaging architectures. We then experimentally validated poly(p-dioxanone) (poly-PDO), an existing ROP polymer whose barrier performance had not been previously reported. Validation showed that poly-PDO exhibits strong water barrier performance, mechanical and thermal properties consistent with predictions, and excellent chemical recyclability (95% monomer recovery), thereby meeting the design targets and underscoring its potential for sustainable packaging. These findings highlight the power of informatics-driven approaches to accelerate the discovery of sustainable polymers by uncovering opportunities in both existing and novel chemistries.

Figures (4)

• Figure 1: Sustainable food packaging design strategy and performance targets. (a) Conventional food packaging with complex multi-layer designs where each layer performs a specific function, complicating chemical recycling, can be replaced with two simplified, chemically recyclable alternatives, including a single-layer multi-purpose polymer and a multi-layer structure where each layer is a single-function, chemically recyclable polymer. (b) The design criteria are compared against the property profiles of common polymers (PP, EVOH, PE), highlighting the performance gaps that the newly designed polymer candidates must fill.
• Figure 2: AI-assisted design workflow for recyclable polymer discovery. A schematic overview of the data-driven methodology, illustrating the key stages: goal setting, defining target properties, developing predictive ML models, candidate design using virtual forward synthesis (VFS), predicting/screening by applying models to filter candidates, and experimental verification of promising cases.
• Figure 3: Schematics for the multi-stage screening process to achieve sustainable polymer candidates in single- and multi-layer packaging architectures.
• Figure 4: Structure of monomers for polymerization, and predicted property values of five (of 1,548) candidates found as promising food packaging replacement polymers.