A groundbreaking collaboration between researchers and the Pacific Northwest National Laboratory (PNNL) is set to revolutionize the way scientists approach material science and chemistry. By harnessing the power of quantum-enhanced artificial intelligence (AI), these researchers aim to tackle some of the world’s most pressing challenges, including climate change and disease, years ahead of conventional timelines.
The project also delves deep into the consideration of over 32 million prospective battery materials. Its statutory mission is not ambiguous – it is to find safety saving, cost-saving, environment-enhancing alternatives. The full team leveraged AI that was trained on quantum-accurate data to dramatically reduce the pool of potential materials they were looking at. This would enable them to make quick predictions for analogous systems at a minuscule expense compared to conventional computing techniques.
Advancements in Material Evaluation
During 2023 and 2024, we worked alongside PNNL. As a team, we had to wrangle an enormous dataset—one with over 32 million candidates for new battery materials. This ambitious undertaking aimed to spur the development of transformational, breakthrough innovations that could make our energy storage systems more safe and sustainable.
Using these AI models, researchers were able to whittle this enormous list down to 500,000 promising, stable materials. This first filtering stage used machine learning processes to recognize candidates with the best potentially developable properties. The effort didn’t end there. That list was further narrowed down to only 800 of the most highly promising candidates.
These 800+ candidates then underwent further synthesis and strict functional testing at PNNL. It is only through this systematic approach that we will achieve the big leaps forward in battery technology that we need. It’s important for fulfilling increasing global energy demands and achieving climate ambitions.
The Role of Quantum Computing
Quantum computing is the key component that allows researchers to push the boundaries of AI’s capabilities in this work. That’s because conventional simulations typically require days — sometimes weeks — to return actionable results. Quantum-enhanced AI is rapidly increasing the depth and breadth of human scientific research. By applying this quantum data, scientists will be able to simulate important chemistry that was previously impossible with classical computation.
To get accurate results out of these kinds of simulations, you need hundreds to thousands of high-quality qubits, with the error rate below about 10^-15. To get to this point of precision, researchers are constantly working to be fault tolerant. They achieve this by redundantly encoding quantum information into logical qubits. Each new logical qubit will be composed of hundreds of physical qubits. Along with such a setup, we’ll eventually need at least one million physical qubits altogether.
This ambitious endeavor will streamline research processes. Finally, it will lead to a more fundamental understanding of correlation of electrons, the key ingredient in systems with strong electron correlations. Gaining a deeper appreciation for these interactions becomes all the more consequential when it comes to materials with nontraditional electronic properties, like high-temperature superconductors.
Future Implications for Science
Researchers at Stanford and Harvard University are working with PNNL to combine quantum computing with AI. This collaborative effort has the potential to be a major force in pushing material science and chemistry forward. AI offers researchers the opportunity to identify potential “first-time right” candidates. This helps them target their resources more effectively, working on just the most promising molecules for actual laboratory synthesis and testing.
This collaborative approach not only makes the entire research process more efficient, but offers the promise of millions in cost savings. With the help of AI, scientists can analyze and visualize massive datasets at lightning speed. This functionality allows them to focus on the most promising solutions rather than be held up by the unworkable alternatives.
While this research is in its early stages, it holds great potential to explore new sustainable materials. These materials have the potential to disrupt all sectors from energy storage to electronics.