Microsoft’s artificial intelligence searches for the perfect battery material


Artificial intelligence is already revolutionizing areas such as research, drastically reduces otherwise very long and complex procedures. The latest demonstration comes from a consolidated collaboration between Microsoft and the US Department of Energy’s Pacific Northwest National Laboratory (PNNL) to address the need minimize the amount of lithium that is normally used inside batteries. To achieve this, the pool of potential alternative materials was something like 32 million candidates, but AI condensed the work that would take a large team of human researchers for decades in just a few weeks. And at the beginning of the year, the first functional prototype was created.

In the last minutes, Microsoft announced the expansion of cooperation with Pnnl, making available such an impressive tool as the analysis of millions of data using AI models and high performance computing (HPC) technologies in the cloud using the platform. Azure Quantum Elements, designed to accelerate scientific research. The comparison with traditional methods is significant, as digital solutions based on artificial intelligence worked in times estimated to be weeks, thus avoiding exhaustive work that could be spread over many years. We went step-by-step using simulations, quantum computing and cloud computing to select a new electrolyte for sustainable energy storage using as little lithium as possible. From the original 32 million candidates, half a million stable options were selected, so we were down to 800, then 18, and therefore only one was selected, a real needle in a haystack. Everything happened at a speed that was unthinkable until recently, as Pnnl synthesized and validated the new material in a short period of time, proving its potential and unique properties in the field. The result was the creation of a real functional prototype, capable of powering a light bulb and operating even at high temperatures.

Solid batteries are safer and more efficient those that use lithium in a liquid or gelatinous state. Lithium is relatively rare in nature and requires mining, which is environmentally problematic and wasteful of resources. Being able to reduce its use by 70% would be an epochal turning point, especially considering the millions of battery-powered devices that are manufactured every month around the world. This project is a demonstration that artificial intelligence is not just a tool for entertainment or to increase productivity, but can bring benefits to society as a whole, starting from crucial areas such as scientific research.


Source link

Leave a Comment