Kuldeepsinh Raj, hunched over a glowing screen in a lab at Stony Brook University, isn’t running test tubes or adjusting dials—he’s training algorithms that could reshape how we fight climate change. Alongside his advisor, Professor Nav Nidhi Rajput, the Ph.D. researcher has helped identify six new solvents capable of turning carbon dioxide into valuable fuels like ethylene and ethanol—using only electricity and smart computational design. Their work, published in Cell Reports Physical Science in 2026, tackles one of clean energy’s most stubborn bottlenecks: finding the right liquid environment to make CO2 electroreduction efficient and scalable.
CO2 electroreduction holds immense promise. It could transform emissions from power plants and factories into usable chemicals, closing the carbon loop. But the reaction is finicky—its success hinges on the electrolyte, the liquid medium inside the reactor. With over a million possible molecular candidates, traditional lab testing would take lifetimes. That’s where Raj and Rajput’s breakthrough comes in. They built a computational framework blending physics-based simulations with machine learning, then used it to screen 1.3 million molecules in silico—a feat impossible through manual experimentation.
From that vast pool, six solvents emerged as standouts: five cyclic ethers and one nitrile, none of which had ever been tested for CO2 conversion before. These molecules don’t just dissolve CO2 effectively—they enhance its mobility and stabilize the reaction environment, two critical factors for industrial viability. Even more valuable than the solvents themselves are the design rules the team uncovered. By analyzing atomic structures, they revealed why certain molecules succeed, offering a blueprint for future innovation.
The real power of this work lies in its openness. All data, models, and findings are now freely accessible in COSMIC—the CO2 Solvent Materials Informatics Collection—a database built to accelerate global collaboration. Researchers from Tokyo to Toronto can now use this foundation to refine electrolytes, optimize devices, and speed up the timeline for commercial deployment. As Professor Dilip Gersappe, Department Chair at Stony Brook, put it, this work embodies the fusion of computation and chemistry to confront Earth’s most urgent challenges.
This isn’t just a step forward in materials science—it’s a leap toward a circular carbon economy. By turning waste CO2 into fuel with renewable electricity, technologies like this could one day decouple industrial growth from emissions. And with open tools like COSMIC, the pace of discovery won’t slow. The future of clean energy isn’t just being invented—it’s being shared.