Phase Separation Key to Energy-Efficient AI

According to a University of Michigan-led study recently published in Matter, phase separation, when molecules separate like oil and water, works alongside oxygen diffusion to assist memristors—electrical components that preserve data using electrical resistance—retain information even when the power is turned off.

Until now, theories for how memristors preserve information without a power source, known as nonvolatile memory, have not been fully understood since models and examinations do not match.

While experiments have shown devices can retain information for over 10 years, the models used in the community show that information can only be retained for a few hours.

Jingxian Li, Study First Author and Doctoral Graduate Department of Materials Science and Engineering, University of Michigan

The researchers concentrated on a device known as resistive random access memory, or RRAM, an alternative to the volatile RAM used in classical computing and particularly promising for energy-efficient artificial intelligence applications, to gain an understanding of the underlying phenomenon driving nonvolatile memristor memory.

In the particular RRAM under study, a filament-type valence change memory (VCM), two platinum electrodes are sandwiched between an insulating layer of tantalum oxide. The cell enters a low resistance state, denoted by a “1” in binary code, when a certain voltage is supplied to the platinum electrodes, forming a tantalum ion bridge that passes through the insulator to the electrodes and allows electricity to flow.

A different voltage will cause the filament to dissolve because the tantalum ions and returning oxygen atoms will react, “rusting” the conductive bridge and causing it to revert to a state of high resistance, which corresponds to the binary number “0.”

It was once believed that RRAM stores information over time because oxygen takes too long to diffuse back. However, a series of studies indicated that prior models overlooked the importance of phase separation.

In these devices, oxygen ions prefer to be away from the filament and will never diffuse back, even after an indefinite period of time. This process is analogous to how a mixture of water and oil will not mix, no matter how much time we wait, because they have lower energy in a de-mixed state.

Yiyang Li, Study Senior Author and Assistant Professor, Department of Materials Science and Engineering, University of Michigan

To evaluate retention time, the researchers accelerated their experiments by raising the temperature. One hour at 250 °C is comparable to around 100 years at 85 °C, the average temperature of a computer chip.

The researchers used atomic force microscopy's exceptionally high-resolution imaging to observe filaments that were just approximately five nanometers or 20 atoms broad and formed within the one micron wide RRAM system.

We were surprised that we could find the filament in the device. It’s like finding a needle in a haystack,” added Li.

The study team discovered that varied filament sizes resulted in variable retention behavior. Filaments smaller than around 5 nanometers disintegrated, whereas filaments larger than 5 nanometers strengthened with time. Diffusion alone cannot explain the big discrepancy.

Experimental observations and simulations based on thermodynamic principles demonstrated that the production and stability of conductive filaments are dependent on phase separation.

The researchers used phase separation to increase memory retention from one day to well over ten years in a rad-hard memory chip, which is designed to survive radiation exposure for use in space exploration.

Other possibilities include in-memory computing for more energy-efficient AI applications and memory devices for electronic skin, which is a flexible electronic interface that mimics human skin's sensory characteristics. This material, also known as e-skin, has the potential to deliver sensory feedback to prosthetic limbs, produce new wearable fitness trackers, and assist robots in developing tactile sensitivity for delicate jobs.

Li further added, “We hope that our findings can inspire new ways to use phase separation to create information storage devices.”

This study was conducted by researchers from Ford Research in Dearborn, Oak Ridge National Laboratory, the University of Albany, NY CREATES, Sandia National Laboratories, and Arizona State University in Tempe.

The device was created in the Lurie Nanofabrication Facility and tested at the Michigan Center for Materials Characterization. The National Science Foundation (ECCS-2106225) provided the primary funding for the University of Michigan study.

Journal Reference:

Li, J., et al. (2024) Thermodynamic origin of non-volatility in resistive memory. Matter. doi.org/10.1016/j.matt.2024.07.018

Tell Us What You Think

Do you have a review, update or anything you would like to add to this news story?

Leave your feedback
Your comment type
Submit

While we only use edited and approved content for Azthena answers, it may on occasions provide incorrect responses. Please confirm any data provided with the related suppliers or authors. We do not provide medical advice, if you search for medical information you must always consult a medical professional before acting on any information provided.

Your questions, but not your email details will be shared with OpenAI and retained for 30 days in accordance with their privacy principles.

Please do not ask questions that use sensitive or confidential information.

Read the full Terms & Conditions.