New materials for ICs Efficiency in data processing with materials that behave like axons

Related Vendor

FAULHABER Drive Systems

The imitation of natural processes (biomimetic materials) has already led to groundbreaking developments, from superconductivity to self-healing materials. Researchers have discovered materials that behave like axons and amplify electrical impulses.

Biomimetic materials, which mimic processes from nature, have already led humanity to some technological developments that previously seemed difficult, if not impossible, to achieve. A team of researchers from Texas A&M University, Sandia National Lab, and Stanford University has drawn inspiration from the brain in the search for materials for more efficient data processing.

In doing so, they came across a material that mimics the behavior of an axon. This occurs by the material class spontaneously transmitting an electrical signal as it moves along a transmission line. Axons, to explain, are the long extensions of nerve cells that transmit electrical signals over long distances in the body. They enable communication between nerve cells and are thus crucial for motor and sensory functions.

If these developed materials are able to spontaneously amplify electrical impulses in a way that is efficient, scalable, and cost-effective, it could represent a real technological advancement. This could open up a whole new paradigm particularly in areas such as quantum computing, high-speed communication, or even neural networks. However, if the discovery is still associated with practical challenges, such as stability under various conditions, it will still be exciting to follow the developments further.

Farewell, signal loss?

Every electrical signal that propagates through a metallic conductor loses strength due to the metal's natural resistance. Modern computer processors and graphics processors can contain about 29.8 miles of fine copper wires that transmit electrical signals within the chip. These losses accumulate quickly and require amplifiers to maintain pulse integrity. These design constraints impact the performance of current chips with high interconnect density.

To counter this limitation, the research team drew inspiration from axons. "Often we want to transmit a data signal from one place to another, more distant location," says lead author Dr. Tim Brown, postdoctoral researcher at Sandia National Lab and former PhD student in materials science and engineering at Texas A&M. The study on the material was published in Nature.

"We need to transmit an electrical pulse from the edge of a CPU chip to the transistors near its center, among other things. Even with the most conductive metals, resistance at room temperature causes the transmitted signals to scatter and be lost, so we usually have to tap into the transmission line and amplify the signal, which costs energy, time, and space. Biology does it differently: Some signals in the brain are transmitted over short distances through axons made of resistant organic matter. And without ever interrupting and amplifying the signals."

According to Dr. Patrick Shamberger, a professor in the Department of Materials Science and Engineering at Texas A&M, axons are communication highways. They transmit signals from one neuron to a neighboring neuron. While neurons are responsible for processing signals, axons are like fiber optic cables that transmit signals.

Spontaneous amplification

Like the axons in the brain, the discovered materials are also in a prepared state that allows them to spontaneously amplify a voltage pulse as it travels down the axon. The researchers harnessed an electronic phase transition in lanthanum cobalt oxide, which causes the material to become significantly more electrically conductive when heated. This property interacts with the small amounts of heat generated as a signal passes through the material, resulting in a positive feedback loop.

The result is a range of exotic behaviors not observed in ordinary passive electrical components—resistors, capacitors, inductors. These include the amplification of small disturbances, negative electrical resistance, and unusually large phase shifts in alternating current signals. According to Shamberger, these materials are unique because they are in a half-stable "Goldilocks" state; that is, just right, but susceptible to disturbances. In this case, electrical impulses neither decay nor lead to thermal runaway and collapse.

Instead, the material naturally oscillates when held under constant current conditions. The researchers found that they could exploit this behavior to generate a spiking behavior and amplify a signal transmitted over a transmission line. "We are essentially leveraging internal instabilities in the material to further amplify an electronic pulse as it travels through the transmission line. Although this behavior was theoretically predicted by our co-author Dr. Stan Williams, this is the first confirmation of its existence."

Insights with future potential

These insights could be crucial for the energy-intensive future of data processing. It is expected that data centers will consume eight percent of electricity in the United States by 2030. And artificial intelligence could dramatically increase this demand.

In the long run, the discovery of the new materials is a step towards understanding dynamic materials and harnessing biological inspirations to promote more efficient data processing. "I had the idea for an active transmission line on the edge of chaos twelve years ago," said Dr. Stan Williams, co-author of the study, director of REMIND (Reconfigurable Electronic Materials Inspired by Nonlinear Neuron Dynamics), and professor in the Department of Electrical and Computer Engineering at Texas A&M. "To bring it into experimental reality required the resources, expertise, and teamwork of REMIND."

Founded in 2022, the REMIND EFRC (Energy Frontier Research Center) aims to develop fundamental scientific insights that underpin the operation of massively reconfigurable computer architectures. Real-time learning and embedded intelligence in these architectures are intended to emulate the specific neuronal and synaptic functions of the human brain.

Other members of the EFRC include Jenny Chong, a master's student in the Department of Materials Science and Engineering at Texas A&M, who contributed to developing a simulation to understand the phenomenon and designed transmission lines for better signal amplification. The team also includes researchers Dr. Suhas Kumar, Dr. Elliot Fuller, and Dr. Alec Talin from Sandia National Lab. The research work was conducted in collaboration with Dr. Eric Pop's research group at Stanford University. (sb)

Link: Axon-mimicking Materials for Computing