Microprocessors in phones, computers, and data centers process information in a different way than our brains do. The manipulation of ion in liquid is what they rely on.
Researchers have been trying to develop ionics in a solution. Scientists think the diversity of ionic species could be harnessed for richer and more diverse information processing.
It's still in its infancy. No one has put many ionic devices together into a more complex circuit for computing until now.
A group of researchers at the Harvard John A. Paulson School of Engineering and Applied Sciences collaborated with a startup to develop an ionic circuit and perform neural net computing.
The research is in a journal.
A new type of ionic transistor was built by the researchers. The solution of quinone molecule is interfaced with two ring electrodes and a center disk electrode. In order to tune the local pH around the center disk, the two ring electrodes have to be lowered. An ion current is generated from the disk into the water when a voltage is applied to it. The local pH can be adjusted to increase or decrease the ionic current. The acid in the solution creates an ionic counterpart of the transistor.
The pH-gated ionic transistor was engineered in such a way that the disk current is a multiplication of the disk voltage and the weight of the local pH gating transistor. They put the transistors into a 16 16 array and used the array of local pH values to create a weight matrix for neural networks.
The most common calculation in neural networks for artificial intelligence is matrix multiplication, according to the first author of the paper. The matrix multiplication in water is performed by the ionic circuit in an analogue way.
The senior author of the paper was Donhee Ham, the Gordon McKay Professor of electrical engineering and applied physics at SEAS. The electrochemical matrix multiplication in water is charming in its own right, and has the potential to be energy efficient.
The team wants to make the system more complex.
Jung said that they have only used 3 to 4 ionic species to enable the gating and ionic transport in the transistor. It will be fascinating to see how we can use more diverse ionic species to make more money.
The research was written by a group of people.
More information: Woo‐Bin Jung et al, An Aqueous Analog MAC Machine, Advanced Materials (2022). DOI: 10.1002/adma.202205096 Journal information: Advanced Materials
All Rights Reserved © 2024
