Cheng-Chien Chen, Ph.D., has already used machine learning and structure prediction to discover new superhard materials. With a new CAREER grant from the National Science Foundation, he's working on a super-hard problem: how to make superconductors that work at room temperature.Archaeologist Indiana Jones, Ph.D., went after hidden treasures with a bullwhip and a map. Physicist Cheng-Chien Chen, Ph.D., already has his bullwhip, powerful supercomputers — and he is building the map right now. With a new grant from the National Science Foundation and access to the fastest supercomputer in academia, Chen is blazing a trail toward his specialty’s own holy grail: the room-temperature superconductor.
Chen, an assistant professor in the Department of Physics, combines data-driven machine learning with first-principles evolutionary structure prediction to discover new materials. His research group has successfully used these approaches to discover several new superhard materials — which have numerous industrial applications. His new work could help move superconductors from a niche product to a mainstream breakthrough.
This January, Chen was awarded a five-year, $450,000 Faculty Early Career Development Program (CAREER) grant from the NSF that will fund his research into predicting, characterizing and discovering new superconducting materials. And because CAREER awards support early-career faculty with the potential to be academic role models in both research and education, UAB graduate and undergraduate students and even local high-schoolers will have an opportunity to take part in Chen’s investigations.
What are superconductors again?
Superconductors carry an electrical charge with no resistance — that is, no loss of energy through heat. In a normal electrical conductor, such as a copper wire, passing electrons cause the nuclei of copper atoms to vibrate, slowing the electrons and producing waste heat. But hook a loop of superconducting wire up to an electrical generator and the current will keep circling around essentially forever.
This seemingly magical property is the result of electrons’ pairing up at the quantum level. Chen uses the analogy of a three-legged race — with many, many more legs connected together. The pairing up of electrons creates a boson, which “can then condense into a quantum state with zero resistance — superconductivity — while a single electron cannot,” Chen said. “They move together, like a legged race, but with many people. Even if a single electron encounters an obstacle, the other electrons keep moving, and because the electrons are bound together, the electrical current will continue to flow.”
Closeup of a quantum computer being developed by IBM. These machines rely on the complex, counterintuitive nature of electrons on a quantum level, as do superconductors. Room-temperature superconductors, which will eliminate the need for large cooling apparatus, will "significantly impact computation and information technologies," Chen said. Image courtesy IBM.
Today, up to 10% of the electricity produced in power plants is lost as it travels to customers down transmission lines. Eliminating this waste could save American companies $20 billion per year, the Department of Energy estimates. Superconductors also repel magnetic fields, enabling frictionless travel in high-speed maglev trains. A superconducting magnet (an electromagnet made from coils of superconducting wire) is also the key component in giant particle accelerators such as the Large Hadron Collider and the more everyday magic of MRI machines.
Trains, MRI scanners and the Large Hadron Collider are all examples of superconductors in action right now. But they work only at extremely low temperatures. MRI machines require cooling equipment needed to keep their superconducting magnetic coils functioning just above absolute zero — zero Kelvin or –459.67 degrees Fahrenheit. The need for bulky, costly cooling also limits superconductors’ potential usefulness in new electronics. This is why new superconducting materials that function at higher temperatures are a critical need, Chen said. “A higher-temperature superconductor will also significantly impact computation and information technologies. One of the leading quantum computing platforms is based on superconducting quantum processing units, or qubits.”
Billions of dollars' worth of electricity is lost each year during transmission. Superconducting power lines could eliminate that waste and conserve energy.
Finding the right combination
Decades ago, researchers found that certain combinations of metals can reach superconductivity at warmer temperatures than any one element alone. In 1986, researchers found a copper-oxide material that made the transition into superconducting at about 30 Kelvin (–405 Fahrenheit). By the 1990s, superconductors were discovered that reached the transition at 133 K (–220 Fahrenheit). In 2020, researchers from the University of Rochester published the current record — a combination of carbon, sulfur and hydrogen that is a superconductor at about 58 degrees Fahrenheit (although at a pressure of 39 million pounds per square inch, or a million times larger than the air pressure in a car tire).
Register for CAREER Award training April 15
Cheng-Chien Chen, Ph.D., and Wenli Bi, Ph.D., both of the Department of Physics, and Gayan Wijeratne, Ph.D., Department of Chemistry, will lead a campus-wide training for junior faculty planning to apply for an NSF CAREER Award. Two sessions will be held online on April 15 and April 22; the second session builds off the first, so attendance at both sessions is recommended. Register by April 8.
Boosting pressure is one way to raise superconducting temperatures. So is the use of ultrafast lasers to push a material into a non-equilibrium state, which formed the basis for Chen’s high-profile paper in Physical Review X in 2021. “What a strong laser field can do is to suppress the tendency of correlated electrons to form insulating states and enhance the superconducting pairing tendency,” Chen said. “We have shown this mechanism in the Physical Review X paper. However, there may be other ways that lasers can help enhance superconductivity as well.”
But superconductors that work only at high pressures or for a limited time under laser bombardment have obvious limitations. What if you could make a material that kept on superconducting at room temperature without these supports? That is “the holy grail,” Chen said. “We would like to either induce a stable superconducting state after the laser is turned off or maintain the superconducting state by repeated pulse lasers. Unfortunately, both have not been realized in experiments yet.”
And so, we come to Chen’s NSF CAREER project: Correlated Superconductors under Extreme Conditions. It aims to add to the theoretical understanding of what happens to a superconducting material at a microscopic level when it is under high pressure or laser bombardment. Another goal is to predict new combinations of materials that could lead to a new generation of high-temperature superconductors. The various superconducting materials found so far have been largely the result of serendipity or trial and error. But as the field’s state-of-the-art advances into ternary or quaternary blends of materials (containing three or four different elements, respectively), the possible combinations are mind-bogglingly high.
The Large Hadron Collider, used to find the Higgs boson in one of this century's great physics achievements, is an example of superconducting in action. But the massive device requires massive cooling power as well.
Traffic jams and simulations
The computations involved in Chen’s research can be massive — think of an Excel spreadsheet with entries in 100 billion rows and 100 billion columns. Calculating formulas across all those entries, more than 1 trillion non-zero elements, “is possible only at national high-performance supercomputing facilities,” Chen said. The results “will help guide experiments in characterizing and discovering new superconducting materials.”
Chen’s research involves massive computer simulations of the quantum phenomena underlying superconductivity, and first-principles studies of materials themselves. His models include something most other simulations leave out. The interaction between electrons in a correlated superconductor is so strong that they cannot be treated as independent particles, as most models do, Chen says. He uses an automotive analogy to make the point: “If you are driving on an empty highway, you can go wherever you want,” he said. “But if you are in a traffic jam, the motion of one car will influence all the other cars. Electrons in highly correlated systems are in a massive traffic jam. This strong electron correlation could lead to new emergent quantum phenomena that are absent at the single electron level.”
Simulating all these electron connections takes serious computational horsepower. Chen’s research group performs calculations and machine-learning simulations on Cheaha, UAB’s research supercomputer. But even this is not enough power for his full-scale models. Some calculations in Chen’s CAREER proposal involve very large-scale matrices — think of an Excel spreadsheet with entries in 100 billion rows and 100 billion columns. Calculating formulas across all those entries, more than 1 trillion non-zero elements, “is possible only at national high-performance supercomputing facilities,” Chen said. Since 2019, Chen has been a Leadership Resource Allocation Awardee from the Texas Advanced Computing Center, home to Frontera, which is the most powerful supercomputer in an academic setting (and the 13th-fastest in the world).
Chen’s CAREER grant research “will predict the chemical compositions and corresponding crystal structures of new phonon (lattice)-mediated conventional superconductors,” Chen said. “In addition, my group will simulate unconventional superconductors driven out of equilibrium by intense ultrafast lasers.”
The results of these simulations “will help guide experiments in characterizing and discovering new superconducting materials,” Chen said. “For example, they could suggest new combinations of materials with the potential to be higher-temperature superconductors at reduced pressure, or details of the pumped-laser profiles for promising light-enhanced superconductivity experiments.”
An example calculation from Chen's group using the NSF-funded Frontera supercomputer in Texas. Chen's highly paralleled large-scale matrix diagonalization gets nearly maximum use of the supercomputer's power, achieving over 85% strong scaling performance with more than 100,000 CPUs. Image courtesy Cheng-Chien Chen, Ph.D.
Room on the edge
There is room for UAB students — graduate students and undergraduates — out on the cutting edge of computational physics, Chen says. He got his own introduction to computational condensed matter physics while an undergraduate student in his native Taiwan. “This experience and the mentorship of my undergraduate research adviser greatly influenced my later choice of research direction,” he said.
“This CAREER grant will support both graduate and undergraduate research assistants” with paid research opportunities, Chen explained. “The students will receive training in programming, machine learning and high-performance supercomputing for materials design and discovery.” (Chen’s CAREER award also will help fund science summer camps for Birmingham high school and community college students, particularly students underrepresented in STEM fields.)
Even if students don’t decide to pursue a science career, there are lots of related jobs in industry, such as software developers and/or data scientists, Chen adds. Two undergraduates who recently worked in his group are now software developers, and his first doctoral student, Wei-Chih Chen, “just received an offer to work for the Taiwan Semiconductor Manufacturing Company, which is the world’s largest chip manufacturer,” he said. “Wei-Chih will be working in an R&D department at TSMC, applying machine learning and supercomputing skills he developed in the UAB Department of Physics.”
