Artificial Intelligence (AI) tools like machine learning and large language models have fundamentally shifted the way humanity solves problems. Today, everything from writing an email to discovering new proteins can be done by our silicon companions, perhaps better than humans can. One of the latest in this list of AI-assisted technologies is in the discovery of new electronic polymers, a class of materials that exhibit both electrical conductivity and the mechanical properties of polymers. These polymer films are the building blocks for many of the flexible and wearable electronic devices we see today.
In a new study, an international team of researchers, including those from Argonne National Laboratory, USA, the Indian Institute of Technology (IIT) Madras, Chennai, and the University of Chicago, USA, have developed an autonomous laboratory named Polybot to explore the creation of electronic polymer films.
Did You Know? AI algorithms can sift through massive datasets of material properties and structures to identify promising candidates for new materials, enabling researchers to design materials with specific properties like strength, conductivity, or optical characteristics. |
Electronic polymers are materials that can conduct electricity like metals but are also flexible and can be processed from a liquid solution. This makes them ideal for creating things like flexible displays, sensors for our bodies, and even efficient solar cells. However, making these polymer films with just the right properties, like high conductivity and very few imperfections, has been a challenge. The process involves many different steps, and each step has many variables that can affect the final outcome. This is where Polybot comes in.
The researchers designed Polybot as a self-driving laboratory. It's a robotic system equipped with everything needed to make these electronic films: stations for mixing solutions, coating the liquid onto a surface, and even testing the final product. What makes Polybot different from a normal assembly robot, however, is its AI brain. It uses a technique called importance-guided Bayesian optimisation, which allows it to learns from its experiments.
Instead of trying every single possible combination of settings, Polybot uses AI to predict which combinations are most likely to lead to the desired outcome – in this case, highly conductive and low-defect films. It learns from each experiment, figuring out which factors are most important and using that knowledge to guide its next steps. This allows it to efficiently explore a vast processing space of seven different variables, including things like the type and amount of additives in the polymer solution, the speed and temperature of the coating process, and even the type of solvent used for post-processing.
The AI-guided system was able to discover recipes for creating transparent conductive thin films that achieved an average conductivity of over 4500 S/cm. This is a significant achievement, placing these films among the best-performing ones currently available. More importantly, Polybot also helped the researchers understand why certain combinations worked better than others. By analysing the data, they identified key factors, like the concentration of a solvent called DMSO, that significantly influence the quality and conductivity of the final film. This deeper understanding is crucial for developing new materials and manufacturing processes.
While other studies have used AI to optimise a few parameters for a single material property, Polybot tackles multiple objectives (conductivity and low defects) across a complex, multi-dimensional processing space. It also emphasises the quality and repeatability of the data, which is essential for reliable AI predictions. The platform can complete an entire experimental cycle, from formulation to testing, in about 15 minutes, allowing it to test around 100 samples per day. This is a massive leap in efficiency compared to traditional methods that rely on human intuition and trial-and-error, which can take years.
By enabling faster and more efficient development of electronic polymers, Polybot and similar AI-driven platforms can accelerate the creation of next-generation electronic devices. The ability to precisely control material properties means we can design devices that are not only stronger, more durable and flexible, but also sustainable and environmentally friendly. The work represents a significant step towards transforming how we manufacture advanced materials, making the development of future technologies more accessible and efficient.
This article was written with the help of generative AI and edited by an editor at Research Matters.