Although it is not well understood, artificial intelligence and machine learning apps are increasingly prevalent, performing a variety of jobs for us behind the scenes. AI and machine learning, for example, assist in the interpretation of voice instructions given to our phones and gadgets such as Alexa, recommends movies and music we would enjoy through services such as Netflix and Spotify, and is even driving autonomous vehicles. The reach of AI and machine learning applications is projected to expand even further in the near future, to more difficult jobs such as supporting space missions and defense operations, as well as generating new pharmaceuticals to treat sickness.
Researchers are working on new photonic materials that could one day aid in the development of low-power, ultra-fast light-based computers. The novel materials, known as topological insulators, resemble wires turned inside out, with the current running along the outside and the interior insulated. In their most recent work, the researchers demonstrated a revolutionary technique to creating materials that employs a linked honeycomb lattice pattern.
Researchers at the University of Central Florida are working on novel photonic materials that could one day aid in the development of low-power, ultra-fast light-based computers. Topological insulators are significant because they could be utilized in circuit designs that allow greater processing power to be jammed into a single space without creating heat, avoiding the overheating problem that today’s smaller and smaller circuits face.
Bimorphic topological insulators represent a new paradigm shift in the design of photonic circuitry by enabling secure transit of light packets with minimal losses. Our technology will increase speed, performance, and power consumption.
Georgios Pyrialakos
The researchers used a novel, linked, honeycomb lattice design to manufacture the materials in their newest work, which was published in the journal Nature Materials. The linked, honeycombed design was laser etched onto a sample of silica, a material frequently used to construct photonic circuits, by the researchers.
The design’s nodes allow the researchers to modulate the current without bending or stretching the photonic wires, which is required for directing the flow of light and consequently information in a circuit. The novel photonic material solves the shortcomings of current topological designs, which gave less features and control while allowing for far longer propagation distances for information packets by minimizing power losses.
The researchers believe that the new design approach presented by bimorphic topological insulators will lead to a divergence from existing modulation approaches, bringing light-based computer technology one step closer to reality.
Topological insulators could also one day lead to quantum computing as their features could be used to protect and harness fragile quantum information bits, thus allowing processing power hundreds of millions of times faster than today’s conventional computers.
The researchers used advanced imaging techniques and computer simulations to confirm their findings. At the Washington Nanofabrication Facility, the research team is attempting to combine the phase-change material with microelectronic circuitry. This will be accomplished by combining the phase-change material with a sophisticated silicon photonic technology manufactured in a commercial foundry. The technology enables the fabrication of thousands of photonic elements and millions of transistors in a cost-effective and scalable manner. In addition, the team will create computer models to replicate every component of the device.
“Bimorphic topological insulators represent a new paradigm shift in the design of photonic circuitry by enabling secure transit of light packets with minimal losses,” says Georgios Pyrialakos, a postdoctoral researcher with UCF’s College of Optics and Photonics and the study’s primary author.
“This is the first time we’ll be putting a non-traditional computer chip into the real world for practical applications, and I’m really enthusiastic about that,” Moazeni said. “It’s a manifestation of Moore’s law, which claimed that new materials will eventually need to be introduced into chip manufacturing in order to boost computer capacity and speed.”
“Our technology will increase speed, performance, and power consumption,” Li continued. “Perhaps most crucially, it will aid in the transition of AI computing to a more sustainable energy route.”
According to Demetrios Christodoulides, a professor in UCF’s College of Optics and Photonics and study co-author, the next steps for the research include the incorporation of nonlinear materials into the lattice, which could enable the active control of topological regions, thus creating custom pathways for light packets.
