The future of plastics

New software could help transform the budding bioplastics industry.

By Brittany St. Jacques | July 25, 2022

Abstracts Turtle Opt
Dreamstime

More than 380 million tons of plastic waste is generated worldwide each year. Approximately half of this waste is single-use plastics, which are typically made with chemicals sourced from fossil fuels, such as petroleum. Single-use plastics are generally used once and then discarded. Few are recycled, an often-tedious process that uses tremendous resources. Most end up in landfills and are discarded into the environment, where they are a significant source of pollution, taking decades to degrade.

“As a society, we generate a staggering amount of plastic waste, with severe consequences on the environment and human health,” says Joseph H. Dumont, an engineer at Los Alamos National Laboratory. “We could mitigate the overwhelming problem of petroleum-derived plastic pollution with the development of bioplastics that safely, completely biodegrade.”

Bioplastics are made from renewable resources and can more easily be broken down to recycle or degrade. However, currently available biopolymers—essentially the building blocks of bioplastics—need significant research and development (R&D) to have the functionality to compete with and replace petroleum-derived plastics. R&D is an exhaustive, expensive process that cycles continuously as researchers work toward polymer optimization, without any guarantee of success.

“We are still decades away from developing large-scale, competitive bioplastic options,” Dumont says.

That’s why Dumont and a team of researchers at Los Alamos have created BioManIAC (Bioplastics Manufacturing with Intelligent Adaptive Control), a software platform that integrates machine-learning and big-data approaches to rapidly discover new biopolymers and enable plastic producers to more quickly transition to sustainable bioplastics.

Carl Iverson@2x
Carol Iverson, of the Lab's Inorganic Isotope and Actinide Chemistry group, analyzes a polymer solution using a special type of spectrometer that uses magnetic fields to identify molecular structures.

 

BioManIAC fills the gaps in polymer data to gain insight into polymer properties and predict new polymer structures. BioManIAC guides researchers to the most promising polymers and limits real-world experiments to those few necessary to validate the machine-learning platform’s predictions. In this way, BioManIAC reduces development costs up to 75 percent and accelerates the discovery of new bioplastics, at a rate about ten times faster than traditional R&D.

“BioManIAC accelerates the development of new bioplastics, using a machine learning–based approach to predict biopolymer structures with the desired features necessary to replace conventional plastics,” Dumont explains. “These new bioplastics will help us reduce our impact on the environment now and for generations to come.”