Researchers Automated Discovery of Bio-based Monomers
A new machine learning framework identifies plant-derived replacements for industrial coatings and adhesives.
Updated on Sept. 29, 2026 in Chemistry
Live Poll
Do you believe AI-driven research will successfully make everyday consumer products more sustainable?
Researchers have developed a machine learning framework that predicts the chemical properties of bio-based monomers to streamline polymer development. The research-stage tool identifies viable, sustainable alternatives to conventional industrial formulations.
Why it matters
The framework addresses the inefficiency of trial-and-error experimental screening for new materials, which currently limits the adoption of sustainable chemicals. This approach accelerates the development of replacements for carbon-intensive petroleum-based polymers.
The framework uses neural networks and gradient boosting to predict properties including propagation rate constants and glass transition temperatures. It successfully identified bio-based alternatives for combinations using n-butyl acrylate and methyl methacrylate or styrene.
The players
Advanced Intelligent Discovery
A scientific journal focused on publishing research at the intersection of material science and artificial intelligence.
The details
The system acts as a unified selection tool that maps candidate monomer pairs to their likely physical characteristics. By predicting water solubility and reactivity ratios, the tool allows researchers to simulate potential performance before physical synthesis. This method was validated through two experimental case studies that compared the model outputs against established industrial counterparts.
Timeline
2026: The study was published in the journal Advanced Intelligent Discovery.
2035: Projected growth target for bio-based polymer production.
The Tech Race
This work directly addresses the resource bottleneck in sustainable chemistry that currently keeps bio-based materials at 1% of total global production. By automating the identification of viable monomers, this framework competes with traditional, slower experimental screening workflows.
This research provides a digital screening tool that will likely reduce the development cycle for manufacturers creating bio-based coatings and adhesives. While not yet a consumer product, the tool paves the way for sustainable alternatives to enter the industrial supply chain in the coming decade.
The takeaway
Automating property prediction replaces the time-intensive screening cycles that have historically slowed sustainable material science. Watch for further pilot-scale experimental validations of these model-identified monomers to confirm their viability in mass-market industrial applications.
Further reading
For broader context on material innovation, see our section on Chemistry.
More information
Read the academic study on machine learning models for the full technical analysis.
Live Poll
Do you believe AI-driven research will successfully make everyday consumer products more sustainable?
