Computers are believed to work more efficiently than humans. They can instantly solve a complex math problem. However, the human brain is capable of quickly processing large amounts of information, such as recognizing a face after seeing it only once in a lifetime.

Such simple human tasks require a lot of energy and time to process data from computers, and results are produced with varying degrees of accuracy.

Brain-like computers with low energy consumption could revolutionize many aspects of modern life. At the heart of these studies are quantum materials for energy efficient neuromorphic computing (Q-MEEN-C).

Alex Franjo, Assistant Professor of Physics and Mathematics at the University of California at San Diego and director of Q-MEEN-C, is leading the center’s work in stages. In the first step, he and his colleagues found a way to create simulations of the properties of individual neurons in quantum materials.

In the second phase of the Q-MEEN-C study, it was demonstrated that electrical stimuli can be transmitted between adjacent electrodes and between electrodes that are far apart.

The discovery is an important step towards creating new types of devices that mimic brain functions known as neuromorphic computing.

As the researchers developed this idea, the simulation turned into a real device.

To do this, the scientists took the thinnest film of nickelate, a ceramic “quantum material” with unique electron beam properties, added hydrogen ions to it, and placed a metal conductor on top of it. Through this, an electrical impulse was applied to the nickel-plated surface, which caused the gel-like hydrogen atoms to slide into a particular structure. This configuration remained stable when the signal was interrupted.

“It’s basically like a memory. The device remembers that you influenced the material,” Franyo said.

It is now possible to fine-tune the movement of ions to create conductive paths for electric current.

Typically, building networks that transmit enough power to power devices like laptops requires complex circuits with unbreakable ports. It is ineffective and unprofitable. The design principle of the Q-MEEN-C is much simpler because not all the wires in the chain are necessarily connected together.

This is similar to how the brain learns – not sequentially, but in complex layers. Each of the learning parts creates connections in some parts of the brain; this not only enables one to distinguish dogs from trees, poodles from shepherds, or oaks from palm trees.

Scientists are looking for revolutionary changes in hardware. The next step will be the creation of more voluminous arrays with an increased number of electrodes and a more complex configuration.