Tiny robots and AI help to craft material solutions for cleaner environments

Mahshid Ahmadi, U. Tennessee

Human activities release pollutants into the air, water, and soil, threatening both ecosystems and human health. According to the World Health Organization, air pollution alone causes 4.2 million deaths annually. To address these issues, researchers are investigating solutions like photocatalysts, materials that can break down pollutants when triggered by light.

As a materials science and engineering researcher at the University of Tennessee, I am working with a team using robots and artificial intelligence (AI) to develop and test new photocatalysts that could help mitigate air pollution. Photocatalysts generate charged carriers when exposed to light, which in turn create reactive oxygen species that can bond with pollutants, breaking them down into harmless or even useful products.

However, existing photocatalysts have limitations. Many require high-energy light, like ultraviolet rays, to function, which makes them less effective in environments with lower-energy light such as visible or infrared light. Additionally, the charged particles often recombine too quickly, preventing full decomposition of pollutants. Surface changes during reactions can also reduce efficiency. Our goal is to overcome these challenges by developing new, more efficient materials that are also nontoxic, ensuring the cleaning process itself doesn’t create further pollution.

My team is focusing on materials called hybrid perovskites, which are nanocrystals that combine organic and inorganic components. These tiny crystals, about one-tenth the thickness of a human hair, have excellent light-absorbing properties due to their unique atomic structure. This allows them to efficiently generate the charged carriers needed for photocatalytic reactions. Hybrid perovskites are also used to improve the efficiency of solar panels and in LED lights, highlighting their versatility in energy applications.

There are thousands of potential types of hybrid perovskites, and testing them all manually would take an enormous amount of time. To speed up the process, we use automation and AI. Instead of handling samples by hand, our small robots can create and test up to 100 different materials in just an hour. These robots, which precisely handle tiny amounts of liquid, are controlled by computer systems that ensure accuracy and speed.

We also rely on machine learning to optimize this process. By analyzing the data from our experiments, machine learning algorithms can quickly identify patterns and insights, allowing us to adjust and improve our experiments for the next round of testing. This rapid learning and iteration process lets us test and refine materials far more efficiently than traditional methods.

Ultimately, the combination of automated experimentation and AI is helping us develop better photocatalytic materials. This approach allows us to quickly overcome complex challenges, advancing the field and bringing us closer to real-world solutions for cleaning up pollutants in our environment. Our work demonstrates how technology, when applied thoughtfully, can provide innovative tools to combat pressing environmental issues.

[Abridged]