Non-degradable plastic waste is a major environmental problem. Bio-synthesized and biodegradable alternatives, such as polyhydroxyalkanoates (PHAs), offer a potential solution. This study develops multitask deep neural network property predictors using experimental data for nearly 23,000 PHA homo- and copolymer chemistries. These predictors identified 14 PHA-based bioplastics from a search space of almost 1.4 million candidates as potential replacements for seven petroleum-based commodity plastics (accounting for 75% of global plastic production). Potential synthesis routes for these promising materials are also discussed.

Christopher Kuenneth, Jessica Lalonde, Babetta L. Marrone, Carl N. Iverson, Rampi Ramprasad, Ghanshyam Pilania