Wednesday, September 3, 2025
12:00 - 14:00
Abstract: Traditionally we design materials with exactly the properties we want and try to make them stable for decades – we intentionally avoid mutations. This means that we avoid degradation processes like rusting, cracking and warping, but we also exclude the possibility that materials improve over time or adapt to their environment. The idea of a bridge becoming more stable or a computer becoming faster with use may sound absurd, but such performance enhancements over time are a hallmark of biological evolution. We are not surprised now that AI models become better over time and even design them to evolve and improve, so why don’t we take such an approach with materials and devices? This lecture outlines the requirements for such evolvable materials and proposes spatiotemporal patterning of light as a tool to direct the evolution. It looks at two easily mutable systems – metal nanoparticles and halide perovskite semiconductors – as platforms to study material evolution. I will highlight the ways that light can both control and measure the properties of these materials in space and time. I will then show several examples of adaptable, self-optimizing and (re)programmable functions in photovoltaics, optics and catalysis and our first results on materials that display memory and elements of learning. I will end with my vision for the material evolution revolution and the exciting possibilities it presents.
VU
Vrije Universiteit, VO Research Building
Spectrum 5
Colloquium
complexity
Erik C. Garnett