Researchers at Northwestern University have shown that megalibraries,platforms that synthesize millions of material candidates simultaneously,can now design materials with tailored properties in hours rather than years, while generating the high-quality datasets AI systems need to accelerate future discoveries. In a study published in Science Advances, the team used the platform to identify and engineer a piezoelectric material that operates at a specific temperature, demonstrating a fundamentally new approach to materials science that moves beyond traditional trial-and-error experimentation.

What Are Megalibraries and How Do They Work?

Megalibraries are platforms that condense what traditionally takes years of searching into a single day by simultaneously synthesizing millions of tiny material candidates on a single chip. First introduced by Northwestern's Chad Mirkin in 2016, the technology allows scientists to explore chemical possibilities at a scale that would be impractical using conventional methods. The latest breakthrough uses a technique called second harmonic generation (SHG) microscopy, which allows researchers to review more than one million different material samples in less than 30 minutes.

What sets megalibraries apart from other emerging approaches is their massively parallel strategy. Rather than working step-by-step like "self-driving labs",automated systems that use robotics and artificial intelligence to propose, develop, and test materials iteratively,megalibraries generate and evaluate enormous numbers of candidates simultaneously. This difference in approach has significant implications for speed and scale.

How Can Scientists Design Materials With Specific Properties?

The Northwestern team demonstrated the platform's design capability by focusing on piezoelectric materials, which generate electricity when pressed, bent, or squeezed and are used in ultrasound imaging, sensors, motion detectors, and energy-harvesting devices. Using the megalibrary platform, the researchers identified a previously unknown, chemically complex material that would have been extraordinarily difficult to find through conventional experimentation. By analyzing how subtle changes in chemical composition affected performance, the team uncovered a useful relationship between material composition and operating temperature.

The researchers then engineered a piezoelectric material designed to maintain its function up to 80 degrees Celsius (176 degrees Fahrenheit). This ability to tune a material's performance means scientists can begin tailoring materials for specific technologies and operating conditions, including temperature-sensitive devices. The entire process, from identifying candidates to designing a functional material, took hours rather than the months or years traditional approaches would require.

"With the megalibrary format, we can synthesize materials faster than has ever been contemplated before," said Chad Mirkin, George B. Rathmann Professor of Chemistry at Northwestern University. "We have developed a screening capability based on a technique called second harmonic generation microscopy that allows researchers to review more than a million different material samples in less than 30 minutes."

Chad Mirkin, George B. Rathmann Professor of Chemistry at Northwestern University

Why Does This Matter for AI-Driven Materials Discovery?

Beyond materials discovery itself, the megalibrary platform addresses a critical bottleneck in artificial intelligence-driven science: the need for large, high-quality datasets built from real-world experiments. Artificial intelligence systems are only as powerful as the datasets used to train them. While scientists can increasingly automate materials synthesis, rapidly collecting meaningful information about how those materials behave has remained a major challenge.

By rapidly generating and screening vast numbers of materials, the megalibrary platform can produce massive datasets linking chemistry to performance. Machine-learning algorithms need this type of structured information to identify hidden patterns, predict promising candidates, and accelerate the future of discovery. In the recent study, the platform generated one million data points from one million different materials, providing researchers with the kind of comprehensive, real-world data that AI systems require to learn effectively.

"We've developed a screening capability that allows researchers to look at literally a million different materials, generating a million data points," explained Jun Li, co-first author of the study and now an assistant professor of mechanical engineering at the University of Colorado Boulder. "We can use that data to train algorithms."

Jun Li, Assistant Professor of Mechanical Engineering at the University of Colorado Boulder

Steps to Accelerate Materials Discovery With Megalibraries

• Parallel Synthesis: Simultaneously create millions of material candidates on a single chip rather than testing compounds one at a time, reducing discovery timelines from years to hours.
• High-Speed Screening: Use advanced microscopy techniques like second harmonic generation to evaluate more than one million samples in under 30 minutes, identifying promising candidates at scale.
• Dataset Generation: Collect structured data linking chemical composition to material performance, creating the high-quality datasets needed to train machine-learning algorithms for future discoveries.
• Targeted Design: Once promising candidates are identified, use insights about composition-performance relationships to deliberately engineer materials with specific properties for particular applications.
• Cross-Domain Application: Extend the megalibrary approach across multiple material types and properties, building data infrastructure for AI-assisted design in batteries, fusion energy, optics, and catalysis.

What's Next for Materials Science?

Mirkin envisions extending the megalibrary approach across many types of materials and properties, helping build the data infrastructure for the next era of AI-assisted materials design. The team has already found materials for piezoelectrics, catalysis, and photocatalysis, and plans to continue discovering materials for batteries, fusion energy, and optics. According to Mirkin, humanity has only explored a tiny fraction of materials possibilities so far, and the megalibrary platform represents a fundamental shift in how scientists approach discovery.

"We're going to repeat this process to find materials for batteries, for fusion, for optics. Our world depends on new materials, and we've only explored a tiny fraction of materials possibilities so far," Mirkin stated.

Chad Mirkin, George B. Rathmann Professor of Chemistry at Northwestern University

The implications extend beyond academic research. As megalibraries generate the massive, structured datasets that artificial intelligence systems require, the feedback loop between materials discovery and AI training accelerates. Better datasets lead to better-trained algorithms, which in turn help identify more promising candidates, which generate even richer datasets. This virtuous cycle could fundamentally reshape how quickly scientists can develop new materials for energy storage, semiconductors, medical devices, and countless other applications that depend on discovering materials with properties that don't yet exist.