A new collaboration between the Toyota Research Institute (TRI) and Northwestern University aims to develop the world’s first nanomaterial “data factory” using machine learning to accelerate the development of materials that will be critical in the clean energy transition.
Toyota Research Institute and Northwestern University have jointly developed a complex machine learning algorithm that can sift through and test nanomaterials at an unprecedented rate.
The new technology will accelerate the innovation of new materials that cannot be done through traditional research methods.
The nanomaterial data factory could be a tool to more efficiently discover new critical materials needed in a net zero world.
TRI and Northwestern will use a new AI-driven method to discover, design, and develop next generation materials that have a lower carbon footprint, use less critical minerals, and can increase efficiency of various applications.
By using machine learning, researchers at TRI and Northwestern University will be able to explore a vast array of parameter sets, collect data, and search the materials’ genome to find the best materials for a given application at a much quicker rate than this would take to do with traditional trial and error research methods.
“Meeting the growing demand for mobility without emitting carbon is a major challenge,” said TRI’s senior director of energy and materials Brian Storey. He explains that through this new partnership, they have “significantly reduced the time it takes to test and find new materials that can be used in batteries and fuel cells to decarbonise transport”.
What’s a data factory and why is it important for innovation?
TRI and Northwestern’s data factory brings together two main elements: an enormous amount of data and an AI algorithm capable of sifting through this data at record speeds.
The data used comes from Northwestern’s “Megalibraries”, which the university claims is a library containing more new inorganic materials than scientists have ever collected and categorised.
Using these high quality data sets to train complex algorithms will allow research teams to sift through, predict, and test the combination of different materials at an unprecedented rate, which can enable the rapid and efficient discovery of new materials.
“Before we were using machine learning, all of our materials design was trial and error. So we had to synthesise one material and then characterise it and then synthesise the next one to characterise that and hope that any one of those would give us properties of interest” explains Carolin Wahl, a graduate student at Mirkin Research Group.
The “data factory” takes away the guessing game and human error, thus increasing the possibility of discovering the innovative materials the clean energy transition will need to rely on to reduce environmental impact and accelerate speed of deployment.
“This ground-breaking research marks an inflection point in how we discover and develop critical materials” explained Chad Mirkin director of the International Institute for Nanotechnology and the George B. Rathmann professor of chemistry at Northwestern
“What’s exciting about this is that we don’t know what the limits are. We don’t know yet what the best performing devices and technologies that we can make can be” said Storey in a video explaining the new collaboration.
“Many of the active catalysts that go into fuel cells, that go into catalytic converters, that are used for catalysing all of the important processes out there, are nanomaterials”, explained Mirkin.
Nanomaterials can increase efficiency and reduce need for critical minerals
A nanomaterial is essentially what the name suggests: a very, very, very small particle. Most researchers define a nanomaterial as a material with at least one external dimension measuring 100 nanometers or less. One nanometer is one millionth of a millimetre – approximately 100,000 times smaller than the diameter of a human hair.
The National Nanotechnology initiative explains that when particles are created at the nanoscale, the properties of materials can change from those at large scales because quantum effects rule the behaviour and properties of particles. In simpler terms, working at the nanoscale allows researchers to “tune” a specific property of a material, which can open the door to the development of new and improved materials.
Nanomaterial research is now being explored in multiple sectors, such as health, information technology, agriculture, and transport. Using nanotechnology, materials can be fine-tuned to be made stronger, lighter, more durable, more reactive, better electrical conductors, among many other possible traits.
Nanotechnology can thus be key to unlocking the new materials needed for decarbonising transport by making batteries and fuel cells more efficient and affordable, while limiting environmental impact by reducing the need to mine for critical minerals.
“Nanomaterials make for good catalysts because they have a high surface area to volume ratio because they’re so small. So ideally when we make nanoparticles, we’re able to make particles that have a lot of available surface area for catalysing reactions and use less of a valuable material like platinum for example that goes into a fuel cell”, explained Joey Montoya, senior research scientist at Toyota Research Institute.
“The more active a catalyst is the more power it will produce for less fuel, so it basically means the fuel cell will be more efficient”, he added.
New materials are key for a net zero world
Mirkin explains that the deployment of the new platform is just the first step and that they need not only think about how they discover materials in a new way, “but ultimately how we deploy this platform in a company like Toyota to really make a difference in their business” said Mirkin.
The novel nanomaterial data factory could be an example for other sectors to accelerate the innovation of new materials that a net zero world will need, and ensure the businesses leading the way to net zero can access and apply the materials discovered through this research.
TRI and Northwestern believe that the research method they’ve developed could be used to support innovation in other areas in the future such as green hydrogen production, carbon removal, and high-efficiency solar cells.
For example, clean energy technologies such as solar and wind power currently require large amounts of critical minerals, and are the fastest-growing segment of critical mineral demand.
Investment in new materials fundamental for renewable energy sector
The intergovernmental body the International Energy Agency estimates that the world will require six times more mineral inputs in 2040 compared to today to stay aligned with their net zero by 2050 scenario. This dependency on critical minerals not only reduces the sustainability of these clean energy technologies, it also makes them more expensive and vulnerable to supply chain disruptions and price fluctuations.
While many companies in the renewable energy sector are heavily investing in the research and development of new materials to replace critical minerals and make their technologies more sustainable and efficient, the pace of these innovations has not been at the scale needed to deploy low carbon energy fast enough to meet climate targets.
Deploying a nanomaterial data factory model of research to other key sectors in the transition to net zero could accelerate the pace of innovation to the speed needed to avoid the worst impacts of climate change.
