New Delhi: Scientists at Brown University in the US have developed a way to make an extremely thin magnetic material without losing its essential properties, overcoming a major obstacle to developing energy-efficient computer chips.
The study was published in Nature Materials on 28 October, and demonstrated that a material called bismuth ferrite could retain its electrical and magnetic properties even when reduced to just a few atomic layers.
The researchers also found that their technique produced an unusual form of magnetism that could make the material useful for future computer memory and processing components.
“We’ve established that we can get rid of the dead layer,” Lucas Caretta, an assistant professor of engineering at Brown University who led the research, said in a university press release.
Overcoming the ‘dead layer’ problem
Bismuth ferrite belongs to a class of materials known as multiferroics, which possess linked electrical and magnetic properties.
These properties allow their electrical polarisation, or the arrangement of positive and negative charges, to be reversed using small voltages. This could help store and retrieve digital information while consuming relatively little electricity.
However, when multiferroics are made thin enough for use in computer chips, they tend to lose these properties, a phenomenon known as the ‘dead layer’ effect.
To overcome this, researchers deposited bismuth ferrite onto a layer of dysprosium scandate, creating compression that helped preserve its atomic structure.
They also added a metallic layer to counteract unwanted electric fields that could disrupt its electrical polarisation.
The resulting structure retained its electrical and magnetic properties at extremely small thicknesses.
Also read: How the Jensen Huang scandal unfolded—3 affairs, ‘spy’ concerns & $600 billion Nvidia loss
A different form of magnetism
The researchers also detected signatures of altermagnetism, an unusual magnetic state that combines characteristics associated with ferromagnets and antiferromagnets.
Ferromagnets produce strong magnetic signals that are relatively easy to detect, but their magnetic fields can interfere with neighbouring components.
Antiferromagnets, on the other hand, can operate faster without producing the same interfering fields, although the resulting signals are harder to detect.
“Altermagnets are kind of the best of both worlds,” Caretta said, explaining that they could enable faster and more densely packed electronic components.
Implications for AI computing
This combination could eventually help engineers develop computer memory and processing components that use less electricity. Something that could be particularly relevant as artificial intelligence drives increasingly high demand for computing power and electricity, straining natural resources.
However, the researchers have only demonstrated the properties of the engineered material, rather than build a functioning computer chip with it. Further development will be needed to establish whether it can be manufactured reliably and integrated into commercial electronics.
Caretta said the findings highlighted the potential of engineering materials at extremely small scales.
He hopes that the resulting material could sometimes perform better than its natural counterpart: “The trick is learning how to engineer them into a form that’s useful — and sometimes even better than what nature gives us.”
(Edited by Janaki Pande)