Dec 2 2019
For the first time, researchers at the University of Oxford have developed an electro-optical device that bridges the fields of electronic and optical computing.
Image Credit: asharkyu/Shutterstock.com
This innovation offers a sophisticated solution to realize processors and memories that are not only faster but also more energy-efficient. The study was carried out in association with scientists at the University of Exeter and the University of Münster.
It has been an attractive but elusive prospect to compute at the speed of light. At present, this light-based computing is in tangible proximity, thanks to the latest development of the electro-optical device. When light is used to both encode and transfer data, such processes can take place at the extreme speed limit—that of light.
Experiments were recently conducted to demonstrate how light can be used to perform certain processes, but a compact device to interface with the electronic design of conventional computers is not available yet.
Light-based computing and electrical computing are not compatible, and this fact can be attributed to the different volumes of interactions operated by photons and electrons. Electrical chips can operate efficiently only if they are small, while optical chips have to be large because the wavelength of light is larger when compared to the wavelength of electrons.
In order to resolve this difficult issue, the researchers have developed a method in which light is confined into nanoscopic dimensions, as described in the scientists’ paper titled “Plasmonic nanogap enhanced phase-change devices with dual electrical-optical functionality.” The article has been published in the Science Advances journal on November 29th, 2019.
The researchers developed a design that they used to compress light into a nano-sized volume via the so-called surface plasmon polariton. The significant reduction in size along with the considerably increased energy density enabled the researchers to bridge the obvious incompatibility of electrons and photons for computation and data storage. In particular, it was demonstrated that when optical or electrical signals are transmitted, the state of an electro- and photo-sensitive material was changed between two varying states of molecular order.
The state of such a phase-transforming material was further read out by either electronics or light, thus making the device the first-ever electro-optical nanoscale memory cell that has non-volatile properties.
This is a very promising path forward in computation and especially in fields where high processing efficiency is needed.
Nikolaos Farmakidis, Study Co-First Author and Graduate Student, University of Oxford
According to Nathan Youngblood, the study co-author, “This naturally includes artificial intelligence applications where in many occasions the needs for high-performance, low-power computing far exceeds our current capabilities. It is believed that interfacing light-based photonic computing with its electrical counterpart is the key to the next chapter in CMOS technologies.”
The study was performed as part of the H2020 project Fun-COMP (#780848).