Researchers at Oregon State University are using artificial intelligence to dramatically speed up the creation and testing of nanomaterials, cutting production time from roughly 24 hours to about 15 minutes. This breakthrough, demonstrated during the university's 2026 AI Week, shows how machine learning and automation are transforming materials science by making the discovery process faster and more reproducible.

Why Is Nanomaterial Discovery So Difficult Today?

Creating nanomaterials, which are particles so small they're measured in billionths of a meter, has long been a bottleneck in biomedical and environmental research. The core problem is reproducibility. When scientists use traditional batch synthesis methods, even with tightly controlled conditions like temperature, pH, and stirring rates, they often end up with heterogeneous mixtures of nanoparticles that vary in size, shape, and surface chemistry. This variability makes it nearly impossible to reliably study the materials or scale them up for real-world applications like cancer therapies or stem cell imaging.

Marilyn Rampersad Mackiewicz, an associate professor of chemistry at Oregon State, recognized this challenge and began building a solution that combines artificial intelligence with continuous flow chemistry. Her lab studies how engineered materials interact with living systems and how those interactions affect health and environmental outcomes. The key insight was that controlling nanoparticle behavior in biological environments is critical, since properties like size, shape, and surface chemistry can determine everything from toxicity to therapeutic effectiveness.

How Does the AI-Powered System Work?

Mackiewicz's lab has developed two custom-built platforms that integrate artificial intelligence, continuous-flow chemistry, and real-time spectroscopy into a single workflow. The first platform can generate liposome-coated nanoparticles in about 15 minutes, compared to roughly 24 hours using conventional methods. This speed enables rapid production and screening of hundreds of formulations.

The second system connects pumps, mixers, and inline ultraviolet-visible spectroscopy to monitor reactions in real time as materials form. This setup produces continuous data streams that feed into analysis tools and machine learning models, allowing the system to learn from each reaction and help guide subsequent experiments. Rather than replacing scientists, the platform keeps humans in control of the process while automating the repetitive and time-consuming parts.

Steps to Accelerate Materials Discovery with AI

• Real-Time Monitoring: Use inline spectroscopy and sensors to collect continuous data streams during chemical reactions, enabling immediate feedback and adjustment rather than waiting for batch results.
• Machine Learning Integration: Feed experimental data into machine learning models that learn from each reaction and predict optimal conditions for subsequent experiments, reducing trial-and-error cycles.
• Automation of Synthesis: Implement continuous-flow systems with automated pumps and mixers that can rapidly produce hundreds of material formulations without manual intervention between batches.

The practical impact is already visible. Mackiewicz's system is being applied to a range of projects, including materials for cancer therapies, stem cell imaging, glaucoma treatment, and dental applications. The broader aim is to accelerate the path from materials design to clinical and practical use.

"By combining automation with real-time analysis, the platform enables faster iteration across large libraries of nanomaterials," explained Marilyn Rampersad Mackiewicz, associate professor of chemistry at Oregon State University.

Marilyn Rampersad Mackiewicz, Associate Professor of Chemistry, Oregon State University

What Does This Mean for Future Materials Science?

The shift from batch synthesis to AI-enabled continuous flow represents a fundamental change in how materials scientists approach discovery. Instead of waiting days or weeks for results from a single experiment, researchers can now run dozens of experiments in parallel, collect real-time data, and use machine learning to identify the most promising candidates for further study. This acceleration is particularly important for biomedical applications, where the ability to rapidly test different formulations can shorten the timeline from laboratory discovery to clinical trials.

Mackiewicz emphasized that while AI supports experimental design and data analysis, scientists remain in control of the process. This human-in-the-loop approach reflects a broader trend in AI-assisted research, where the technology augments human expertise rather than replacing it. As more labs adopt similar systems, the ability to rapidly discover and optimize new materials could unlock breakthroughs in drug delivery, diagnostics, and environmental remediation that would otherwise take years to achieve.