Researchers at Brown University have solved a decades-old physics problem that has kept a promising class of magnetic materials from being used in modern computer chips. By engineering the atomic environment of these films, they have proven that the materials can function at thicknesses just a few angstroms, paving the way for significantly smaller and more energy-efficient electronics.
The Dead Layer Problem
For years, engineers have looked to multiferroics as a solution to the rising energy costs of computing. These materials possess interconnected electrical and magnetic properties, meaning a small voltage could theoretically flip a magnetic state to store data, rather than using energy-hungry electric currents. This dual-functionality could allow memory and logic devices to consume far less power than current silicon-based alternatives.
The catch was the “dead layer” effect. When multiferroic films are thinned out to the microscopic scales required for chips, their atomic structures and chemical environments change enough to disrupt their functionality. At the same time, internal electric fields destabilize the material’s polarization. Together, these factors eliminate the very ferroelectric and magnetic behaviors that make the material useful for electronics. The material effectively dies when it gets too thin.
Engineering the Solution
The Brown University team, led by assistant professor of engineering Lucas Caretta, overcame this by carefully engineering the layers surrounding the bismuth ferrite film. They grew the material on dysprosium scandate, a rare earth-based crystalline substrate with a smaller lattice spacing than the bismuth ferrite itself. This mismatch places the film under compressive strain, which helps preserve the specific atomic arrangement required for its electrical and magnetic properties to persist.
Derivative work: IamMM (CC BY-SA 4.0)
To address the destabilizing electric fields, the researchers incorporated a metallic layer that provides electrical shielding. This layer screens unwanted fields that would otherwise undermine the film’s electrical order. By controlling both the mechanical strain and the electrical boundary conditions, the team preserved the coupled behavior in ultrathin layers.
“We’ve established that we can get rid of the dead layer,” Caretta said. “We show that you can make this material basically as thin as you want — down to a few angstroms — and it’s still ferroelectric, and it’s still magnetic.”
An Unexpected Magnetic Bonus
The engineering didn’t just preserve the original properties; it unlocked a new one. The researchers found that the thin film exhibits a state called altermagnetism. This is unusual because conventional ferromagnets, which are useful for memory because they produce strong, readable signals, are relatively slow and create stray magnetic fields that interfere with neighboring components. Antiferromagnets operate faster and have no stray fields, but their signals are difficult to detect.
Altermagnets offer a compromise that could be ideal for dense chip integration. “Altermagnets are kind of the best of both worlds — they can be fast, densely integrated, and still provide a strong electrical signal for readout,” Caretta explained. This combination of speed, density, and signal strength could allow for faster, more compact electronic components that use less energy.
What This Means for Your Hardware
The practical upshot is a clearer path toward lower-energy AI and memory components. Current data centers and personal devices face a bottleneck where the energy required to move data between memory and processors often exceeds the energy used for the computation itself. Multiferroics could help bridge that gap by enabling logic and memory to operate with minimal power draw.
Think of it like moving from a heavy, inefficient incandescent bulb to a tiny, cold LED. The light is still there, but the heat and energy waste are gone. In this case, the “light” is the magnetic state that stores your data, and the “heat” is the stray electromagnetic interference and power consumption that currently limits how many components can fit on a single chip.
However, this is a materials science breakthrough, not a finished product. The study demonstrates that the material retains its properties at the necessary thicknesses, but it does not yet show a completed AI processor with measured energy savings. Engineers still need to translate these films into reliable working components and integrate them with other parts of a device. The amount of electricity an entire system saves will depend on the eventual device design and the specific work it performs.
The findings were published in Nature Materials, establishing a route to useful behavior in extreme thinness. While the full benefits for consumer hardware are still on the horizon, the removal of the dead layer barrier clears a significant hurdle for the next generation of semiconductors.
Sources
- Electrical engineering
- A Breakthrough in Magnetism Could Make Computer Chips Use Dramatically Less Energy
- A tiny laser beam could make computer memory 1,000 times faster
- Ultrathin magnetic material could help cut the energy used by AI chips
Photo: UCL Mathematical and Physical Sciences from London, UK (CC BY 2.0)
