At the KTH Royal Institute of Technology and Stanford University, scientists have fabricated a new material for computer components allowing the commercial practicality of computers that has the ability to mimic the human brain.
Electrochemical random access (ECRAM) memory components composed of 2D titanium carbide have demonstrated excellent potential for complementing classical transistor technology and contributing towards the commercialization of string computers being modeled after the brain’s neural network. These neuromorphic computers could be thousands of times more energy-efficient compared to the computers at present.
In computing, such progress is possible due to some basic variations from the classic computing architecture that is in use today. Also, the ECRAM, a component that serves as a kind of synaptic cell in an artificial neural network, states Max Hamedi, an associate professor from KTH.
Instead of transistors that are either on or off, and the need for information to be carried back and forth between the processor and memory—these new computers rely on components that can have multiple states, and perform in-memory computation.
Max Hamedi, Associate Professor, KTH Royal Institute of Technology
The researchers at KTH and Stanford have concentrated on testing better materials for building an ECRAM. This is a component in which switching takes place by embedding ions into an oxidation channel, in a way similar to the brain which also functions with ions. The materials that overcome the poor temperature stability of plastics and the slow kinetics of metal oxides are the ones required to make these chips commercially feasible.
The main material in the ECRAM units fabricated by the scientists is called a MXene. It is a two-dimensional (2D) compound, barely a few atoms thick, comprising titanium carbide (Ti3C2Tx). The MXene integrates the high speed of organic chemistry with the integration compatibility of inorganic materials in a separate device functioning at the nexus of electrochemistry and electronics, stated Hamedi.
The study’s co-author and Professor Alberto Salleo at Stanford University, states that MXene ECRAMs integrate the speed, write noise, linearity, switching energy and endurance metrics necessary for a parallel acceleration of artificial neural networks.
MXenes are an exciting materials family for this particular application as they combine the temperature stability needed for integration with conventional electronics with the availability of a vast composition space to optimize performance.
Alberto Salleo, Study Co-Author and Professor, Stanford University
As there are numerous other barriers to overcome before consumers could buy their own neuromorphic computers, Hamedi states the 2D ECRAMs represent a discovery at least in the area of neuromorphic materials.
This potentially results in artificial intelligence that could adapt to perplexing input and nuance, similar to the way the brain does with thousand times smaller energy consumption. Furthermore, this can enable portable devices that have the ability to execute much bulkier computing tasks without having to depend on the cloud.
Melianas, A., et al. (2021) High-Speed Ionic Synaptic Memory Based on 2D Titanium Carbide MXene. Advanced Functional Materials. doi.org/10.1002/adfm.202109970.