Northwestern researchers selected for Genesis Mission projects

Northwestern University researchers are participating in a group of projects selected for Phase I funding through the U.S. Department of Energy’s Genesis Mission, a national initiative designed to transform how complex scientific problems are investigated and solved.

The selected researchers gathered with fellow project teams July 22 in Washington, D.C., for the Genesis Mission Summit and the official launch of the initiative’s first cohort.

The Genesis Mission is building what the Department of Energy describes as the world’s most powerful integrated science discovery platform. By bringing together government, industry, universities and philanthropic organizations, the initiative will combine artificial intelligence, supercomputing, quantum systems and advanced scientific instruments to accelerate breakthroughs in energy, scientific discovery and national security.

Northwestern-affiliated projects include a University-led effort to discover new materials for lithium isotope separation; an Argonne National Laboratory-led project developing AI-enabled tools to co-design cryogenic CMOS circuits and superconducting electronics; another Argonne-led project using AI agents to design efficient quantum-computing workflows to simulate physical systems; and a project led by Ted Sargent, director of Northwestern’s Paula M. Trienens Institute for Sustainability and Energy, to find more energy-efficient ways to manufacture one of the world’s most widely produced plastics.

“The Genesis Mission is an important framework for advancing the future of AI-enabled energy research in the United States. We are proud that Northwestern researchers will help shape this historic initiative as members of its inaugural cohort,” said Sumit Dhar, Northwestern’s interim vice president for research. “Their selection reflects the high quality and depth of science being conducted at Northwestern as well as the strength of our critical partnership with Argonne National Laboratory.”

The Phase I awards are intended to identify promising pathways toward transformative scientific capabilities and establish a foundation for future investment and expansion. Project teams will design and demonstrate research workflows that integrate AI with scientific investigation, while evaluating whether these approaches can accelerate discovery, improve predictive capabilities, enhance experimentation or produce new scientific insights.

The Northwestern-affiliated projects are supported by the U.S. Department of Energy through the Genesis Mission.

Using AI to find materials essential to fusion energy

A Northwestern-led project headed by Randall Q. Snurr will develop an AI-enabled platform to discover materials capable of efficiently separating lithium isotopes, an unresolved challenge with important implications for the future of fusion energy.

Fusion energy offers the possibility of clean, reliable baseload power, but achieving it at scale will require a dependable domestic supply of lithium-6. This isotope is used in fusion “breeding blankets,” specialized layers surrounding a reactor that absorb neutrons and produce tritium, a key fuel for many proposed fusion systems. These blankets require lithium-6 at approximately 80% to 90% purity, whereas naturally occurring lithium contains only about 5% lithium-6.

The United States stopped domestic lithium-6 production in 1963 because the process then in use relied on liquid mercury, raising concerns about contamination and worker safety. The country has since relied on a diminishing Cold War-era stockpile, leaving the emerging domestic fusion sector without a viable long-term enrichment method.

Snurr’s team includes Northwestern co-investigators Omar K. Farha, Filip Formalik and Justin M. Notestein, along with N. Scott Bobbitt of Sandia National Laboratories. The researchers will use machine learning and physics-based simulations to search systematically for solid-state materials that can distinguish between lithium-6 and lithium-7. They will examine zeolites, metal oxides and two-dimensional chalcogenides, seeking materials that combine strong isotopic selectivity with the capacity required for practical enrichment.

The project will integrate surrogate machine-learning models, machine-learning interatomic potentials and AI-accelerated molecular dynamics simulations. Experimental measurements will then be used to validate the computational rankings, with Gaussian-process models and Bayesian optimization helping researchers select the most promising materials for testing.

The work will examine two complementary separation mechanisms: small differences in how the isotopes bind within a material and differences in how quickly they move through it.

Phase I is expected to produce an initial map of isotopic selectivity across chemically diverse materials and demonstrate whether the AI-enabled approach can outperform conventional computational modeling or trial-and-error experimentation. The resulting computational and experimental workflow also will help identify material characteristics that can be used to screen much larger collections of potential candidates during future phases.

The longer-term goal is to discover, synthesize and test materials that could support domestic lithium-6 enrichment at the scale required for fusion energy development.

“The scientific principles behind lithium isotope separation are understood, but the materials landscape remains largely unexplored,” Snurr said. “By combining physics-based simulation, machine learning and carefully targeted experiments, we hope to move beyond isolated discoveries and create a systematic way to identify materials with the performance needed for a viable domestic fusion fuel cycle.”

Using agentic AI to manufacture polyethylene from captured carbon

A Northwestern-led project headed by Ted Sargent, director of the Trienens Institute, will explore a more energy- and resource-efficient way to manufacture polyethylene, one of the world’s most widely produced plastics. The project, “Agentic AI-Driven Manufacturing of Polyethylene from Waste Carbon,” brings Northwestern together with Argonne National Laboratory and Lila Sciences, an AI science company.

Polyethylene is conventionally produced from ethylene generated through steam cracking of hydrocarbons, an industrial process that requires extremely high temperatures and substantial energy. The new project instead will investigate a tandem pathway that begins with captured waste carbon. In the first stage, an electrochemical process will convert carbon into ethylene. A second, thermochemical stage will then transform the ethylene into polyethylene.

The team will use Lila Sciences’ reasoning-and-planning AI to coordinate these two stages as a single interconnected system. Rather than optimizing each reaction independently, the agentic AI platform will reason across the complete manufacturing pathway, helping researchers identify combinations of catalysts and operating conditions that conventional human-guided workflows might not efficiently uncover.

By integrating AI with Northwestern’s longstanding strengths in catalysis and materials science and the advanced scientific capabilities of Argonne National Laboratory, the project aims to demonstrate a more atom-efficient and energy-efficient route to a high-volume industrial material. Phase I will establish and evaluate the AI-enabled research workflow while generating the experimental and computational foundation needed for potential expansion during a later phase of the Genesis Mission.

The longer-term objective is to show how captured carbon can become a useful manufacturing feedstock while reducing reliance on processes that employ hydrocarbons. The approach also could provide a model for using AI systems designed for scientific reasoning to coordinate complex, multistage chemical processes in which decisions made at one stage affect the performance of the entire production pathway.

“The Genesis Mission creates an opportunity to connect artificial intelligence directly with the physical process of scientific discovery,” Sargent said. “By bringing computation, experimentation and advanced scientific infrastructure into a shared workflow, we have the potential to investigate important problems with a speed and level of precision that previously were not possible.”

Using AI to design electronics for extreme cold

Northwestern researcher Jie Gu is participating in an Argonne National Laboratory-led project known as CMOS + X, a multi-institutional effort to develop AI-enabled tools for designing electronic systems that operate near absolute zero. The project also includes researchers from SLAC National Accelerator Laboratory and the University of Illinois Chicago.

Ultra-cold electronics — used in advanced particle detectors, X-ray sensors, and quantum devices — require experts in many fields to work together, but slow, siloed design processes limit progress. This project builds an AI-powered system with smart software "agents" that automatically design and optimize chip circuits, superconducting components, and thermal systems as one integrated problem. By using AI to predict performance and temperature effects, researchers can explore new device combinations faster, accelerating the development of next-generation sensors for scientific discovery.

Gu, a professor of electrical and computer engineering at Northwestern’s McCormick School of Engineering, brings expertise in complementary metal-oxide-semiconductor, or CMOS, circuit design — the foundational technology underlying modern computer chips. His group will investigate how superconducting elements incorporated into commercial chip-fabrication processes could improve the performance of cryogenic CMOS circuits. One promising approach involves using superconducting inductors, whose distinctive properties at extremely low temperatures may enable performance beyond what conventional CMOS circuits can achieve alone.

“CMOS is the workhorse of modern electronics, and understanding how to get the most out of it at cryogenic temperatures is essential,” said Professor Gu. “What excites me about this project is the opportunity to explore what becomes possible when you combine strong CMOS design with the unique capabilities that superconducting materials offer, all within a single commercial chip platform.”

Designing electronics for such extreme conditions requires researchers to account simultaneously for semiconductor behavior, superconductivity, circuit performance, and heat flow. Today, those areas often are addressed by separate specialists using different models and design tools. The CMOS + X project will use AI to connect these domains through a unified co-design framework that evaluates material, circuit and thermal considerations together rather than sequentially.

The project will develop AI agents that can draw on a continuously updated, multidisciplinary knowledge base, translate proposed designs into forms that can be modeled and tested, and identify relationships among design choices that influence performance. Fast AI surrogate models also will help predict thermal and superconducting behavior without requiring researchers to run a full, computationally intensive physics simulation for every potential design.

By replacing slower trial-and-error processes with more integrated and informed exploration, the project aims to advance electronics for quantum sensing, particle detection and other high-precision scientific instruments that must operate at exceptionally low temperatures.

The collaboration builds on connections developed through the Northwestern + Argonne Institute for Scientific and Engineering Excellence, or NAISE. Antonino Miceli, who leads the project at Argonne and is a NAISE Fellow at Northwestern, previously co-led the workshop “CMOS + X: Bridging CMOS Manufacturing with Quantum and Neuromorphic Technologies” with Northwestern’s Begum Gulsoy. The workshop was organized by NAISE in partnership with Northwestern’s Institute for Quantum Information Research and Engineering and the Argonne Microelectronics Institute.

AI-enabled optimization of quantum circuit design

Another Argonne-led project involving Northwestern researcher Ian Low will develop an AI-enabled framework to identify efficient quantum-computing workflows for simulating protons and neutrons, the particles that form atomic nuclei. Predicting how these strongly interacting particles behave remains a fundamental challenge in physics and is essential to understanding the structure of ordinary matter. The team also includes researchers from Dakota State University and Infleqtion, a quantum technology company that develops neutral-atom quantum computers and software.

Many important nuclear-physics calculations remain beyond the reach of existing classical methods. These include tracking strongly interacting particles in real time, deriving the behavior of protons and neutrons from the fundamental theory of the strong force, and making precise predictions for medium-mass and heavy nuclei.

Quantum computers may eventually make some of these simulations possible, but first researchers must determine how best to represent a physical system in qubits, which quantum algorithm to use and which hardware platform to target. Because these choices are interdependent, manually evaluating every combination is impractical and may miss more effective approaches.

Low, a professor of physics in Northwestern’s Weinberg College of Arts and Sciences who holds a joint appointment with Argonne, brings expertise in fundamental physics and quantum information science. The team will build AI agents that evaluate combinations of qubit encodings, algorithms and hardware platforms to identify efficient end-to-end workflows for nuclear-physics calculations.

The longer-term goal is to make quantum computing more accessible for fundamental research. By integrating the framework into the American Science Cloud, the team plans to make it available to researchers nationwide and accelerate the use of quantum computing to study matter and its interactions.

“Quantum computing and quantum simulation could open new ways to investigate the universe at its most fundamental level,” Low said. “Realizing that potential will require combining AI with scientific expertise to identify quantum-computing strategies that are both efficient and physically meaningful.”

Building a new model for scientific discovery

Northwestern’s participation in the first Genesis Mission cohort reflects the University’s strengths in AI-enabled materials discovery, energy research, microelectronics, quantum science and collaboration with the Department of Energy’s national laboratories.

Through Phase I, the selected teams will test not only individual scientific hypotheses but also a broader model of discovery in which AI systems work alongside researchers, advanced instruments and high-performance computing resources. Results from the projects will help the Department of Energy determine which approaches are most promising for future investment and expansion.

The July 22 Genesis Mission Summit at the Capital Hilton in Washington, D.C., brought the inaugural research cohort together with federal, national laboratory and industry leaders.

The daylong program included remarks from senior White House and Department of Energy officials, a technical keynote and AI-for-science presentations, a demonstration of the Genesis Mission platform, and discussions of the partnerships and shared infrastructure needed to advance the initiative. Northwestern-affiliated teams also participated in a science poster session and afternoon working groups focused on the resources, research priorities and collaborations needed to accelerate U.S. scientific discovery and innovation.