When a material does something unexpected, we have to figure out why.
At Rice University, materials scientists have coaxed a long-studied compound into doing what it never could before: behaving as a powerful multiferroic at room temperature, without exotic conditions. By simultaneously tuning both the chemistry and the crystal strain of bismuth ferrite, Lane Martin's team achieved tenfold gains in magnetization and hundredfold gains in magnetoelectric coupling — properties that could allow a single material element to perform both memory and logic with a fraction of today's energy cost. The discovery arrives as computing's appetite for power approaches a civilizational threshold, and it suggests that the next chapter of electronics may be written not in silicon, but in the subtle interplay between electric and magnetic order.
- Computing's energy consumption is on a trajectory to claim a quarter to a third of all generated power within a decade — a crisis that conventional silicon architecture has no clear answer for.
- Multiferroics have promised a way out for twenty years, but the best candidates either worked only at extreme temperatures or delivered properties too weak to be practical.
- Rice researchers broke the impasse by turning two knobs at once — blending bismuth ferrite with barium titanate while engineering the crystal's strain — producing a material whose magnetization and magnetoelectric coupling shattered prior benchmarks.
- The result was so surprising that first author Tae Yeon Kim spent over six months repeating measurements and had a colleague independently reproduce the material before trusting what she was seeing.
- Synchrotron analysis across multiple national laboratories and universities confirmed the findings, transforming an anomaly into a reproducible, explainable phenomenon.
- The team now holds not just a new material but a design strategy — a recipe for combining chemistry and strain to unlock properties that conventional intuition would not predict.
Lane Martin's lab at Rice University has produced a modified bismuth ferrite that achieves something its predecessors could not: robust multiferroic performance at room temperature. Published in the Proceedings of the National Academy of Sciences, the work reports a tenfold increase in magnetization and a hundredfold increase in magnetoelectric coupling over standard bismuth ferrite — gains that place the material in a category of its own.
The breakthrough came from a deliberate double intervention. The team blended bismuth ferrite with barium titanate while simultaneously growing the compound as a thin film on a substrate engineered to distort its crystal structure. Martin described it as dialing two knobs — chemistry and strain — that no one had turned together before. The result was a new structure with a new combination of properties, including the counterintuitive finding that adding a nonmagnetic component made the overall material more magnetic.
The stakes behind the work are concrete. Martin has noted that computing could consume a quarter to a third of all generated power within the next decade. Multiferroics offer a potential escape: because an electric field can control their magnetism and vice versa, a single element could in principle handle both memory and logic, using orders of magnitude less energy than conventional electronics.
First author Tae Yeon Kim was not quick to celebrate. The magnetic measurements were so unexpected that she spent more than six months remaking and retesting samples, and recruited a colleague to independently reproduce the material from her recipe. The team extended its verification to synchrotron measurements at Lawrence Berkeley National Laboratory and drew on collaborators at MIT, UC Berkeley, Drexel, Northeastern, Bar-Ilan University, the University of Pennsylvania, and the U.S. Naval Research Laboratory.
What the effort yielded is more than a single promising compound. The combined chemistry-and-strain approach now stands as a broader design strategy — a way of engineering materials whose properties defy conventional expectation. The question Martin's team is turning toward is how this strategy scales, and whether it can help guide computing out of the energy trap it is steadily walking into.
Lane Martin's lab at Rice University has engineered a material that does something its predecessors could not: it works well at room temperature, in the real world, without requiring extreme cooling or specialized conditions. The achievement, published in the Proceedings of the National Academy of Sciences, describes a modified version of bismuth ferrite that shows a tenfold increase in magnetization and a hundredfold increase in magnetoelectric coupling compared to standard varieties of the same material.
The path to this discovery involved a deliberate act of combination. The researchers mixed bismuth ferrite with barium titanate while simultaneously growing the material as a thin film on a substrate engineered to distort its crystal structure. Martin, Rice's Robert A. Welch Professor of Materials Science and NanoEngineering, described the approach with precision: "Nobody had ever dialed both knobs — the strain and the chemistry — at once. We were able to combine two different material systems into a new material with a new structure and a new combination of properties." The strategy was not accidental. It was methodical, born from understanding that conventional approaches had left something on the table.
Why this matters becomes clear when you consider the trajectory of computing. Modern electronics move and store information by switching the flow of electrons on or off, a process silicon has mastered over decades. But that mastery has hit a wall. Martin put it plainly: "Electronics today have an energy problem. Within the next five to 10 years, computing could use up as much as a quarter to a third of all the power generated, which is unsustainable." The question driving materials scientists now is whether there are other properties of electrons—their spin, for instance, their magnetic character—that could form the basis for computation itself, using far less energy in the process.
Multiferroics have been the subject of serious research for two decades precisely because they possess multiple order parameters simultaneously. The ones Martin's team focuses on are both ferroelectric, meaning they hold a spontaneous electric polarization that can be switched with an electric field, and magnetic. The coupling between these properties—magnetoelectricity—is the real prize. An electric field can change the material's magnetism. A magnetic field can change its polarization. In theory, this switching could enable memory and logic operations in a single element, using orders of magnitude less power than conventional approaches.
Bismuth ferrite had long seemed like the natural candidate. It is ferroelectric and magnetic. But its magnetism is weak because its atomic moments cancel each other out. Adding barium titanate, a nonmagnetic component, in combination with carefully engineered strain, produced an unexpected result: the new material's overall magnetization increased while its electric properties remained strong. Tae Yeon Kim, a postdoctoral researcher in Martin's lab and the study's first author, was cautious about what she was seeing. "I did not expect such a large increase in magnetization," she said. "At first, I was excited by the new structure, but when I measured the magnetism, I became very anxious. We repeated the measurements many times to make sure it was real."
Kim's skepticism was justified. Measuring the magnetic properties of thin films is notoriously difficult. She spent more than six months making and testing samples, even recruiting another lab member to grow the same material independently using her recipe, to ensure the results could be reproduced. The work extended beyond Rice's walls. The team used synchrotron measurements at the Advanced Light Source at Lawrence Berkeley National Laboratory and collaborated with researchers at Bar-Ilan University, Drexel University, MIT, Northeastern University, UC Berkeley, the University of Pennsylvania, and the U.S. Naval Research Laboratory to understand the material's behavior from multiple angles.
What emerged from this effort is not just a promising new material but a broader strategy. By combining chemistry and strain engineering, researchers can create structures with properties that defy conventional expectation. The fact that adding nonmagnetic atoms made the material more magnetic is itself a guide for future design. Martin reflected on the moment of discovery: "This is the fun part of science. When a material does something unexpected, we have to then figure out why." The question now is what comes next—how this approach scales, how it translates from the lab to devices, and whether it can help computing escape the energy trap it is walking toward.
Notable Quotes
Nobody had ever dialed both knobs — the strain and the chemistry — at once. We were able to combine two different material systems into a new material with a new structure and a new combination of properties.— Lane Martin, Rice materials scientist
Electronics today have an energy problem. Within the next five to 10 years, computing could use up as much as a quarter to a third of all the power generated, which is unsustainable.— Lane Martin
The Hearth Conversation Another angle on the story
Why does room temperature matter so much here? Couldn't you just cool the material down?
You could, but then you've added a refrigeration system to every device. The whole point is to reduce energy consumption. A material that works at room temperature is a material that actually works in the world.
So the barium titanate is nonmagnetic, but adding it made the material more magnetic. How does that even happen?
That's the question they're still working to answer. The strain from combining the two materials seems to rearrange how the atomic moments interact. Instead of canceling each other out, they align. It's counterintuitive, which is why Kim had to repeat the measurements so many times.
What does magnetoelectric coupling actually do for a computer?
It means a single element can respond to both electric and magnetic signals. You could potentially store information magnetically and manipulate it electrically, or vice versa, all in one place. That's more efficient than having separate systems for memory and logic.
How close are we to seeing this in actual devices?
That's the honest question. They've proven the material works. Now comes the harder part—scaling it, integrating it into real circuits, making it manufacturable. That's years of work, probably.
Why did it take 20 years of research on multiferroics to get here?
Because the constraints were severe. You needed something that was both strongly ferroelectric and strongly magnetic at room temperature. Bismuth ferrite was close, but not close enough. Nobody had thought to combine it with barium titanate and engineer the strain simultaneously until Martin's team did.