Nokia and Toshiba to build photon chips based on Bristol breakthrough
Scientists from Bristol will this week show a quantum computing chip, which partners such as Toshiba and Nokia hope to build into commercial phones and other systems within the next few years, securing communications and attacking previously insolvable problems.
Quantum computing has the potential to massively speed up calculations and provide more secure communications. The Bristol Centre for Quantum Photonics hopes to establish and patent key technologies which can be licensed to manufacturers, becoming a central player in the technology which could push electronics beyond the limitations of Moore’s Law.
Quantum computing at commercial scale
“We are envisaging a model like ARM or CSR,” said Dr Mark Thompson, the Bristol physicist leading the work, referring to two British chip designers whose low-power processors are manufactured in the Far East, but dominate the world’s smartphones and tablets in the case of ARM, and are important for radio communications in the case of CSR. “Already, the South West [of England] has the biggest cluster of silicon designers outside of Silicon Valley.”
The proposal is more radical than ARM or CSR, however, as Thompson’s team are building quantum computers with optical waveguides on silicon, instead of simply creating a better design for a conventional chip.
Quantum computing potentially allows a system to process trillions of inputs simultaneously, because of “superposition”. Multiple quantum states exist at the same time, like the cat in Schrodinger’s famous thought-experiment, which is both alive and dead, until you open the box which contains it. Multiple states of a quantum computer can each process a different input to a problem.
The trouble is “decoherence”: when a quantum system interacts with the world, it decays to one state, and the superposition is over. Previous quantum computing approaches have required very low temperatures and systems which are isolated from the outside world, to maintain their systems in a state of quantum coherence long enough to do work.
The Bristol-based team, who work with Toshiba, Nokia and UK-based Oclaro, as well as Heriot-Watt University in Scotland and Delft University in the Netherlands, have developed an approach using pairs of photons (light particles) to produce useful quantum computing systems at room temperatures, in chips which can be mass produced for real devices.
“Photons don’t see temperature,” said Thompson. “They act like they are at zero Kelvin. Every other quantum system will decohere, but photons don’t.”
Quantum photonics systems have been built on vibration-free optical benches in the laboratory, but the last few years has seen a “Renaissance” in opto-electronics, which the quantum photonics group has been able to capitalise on, explained Thompson.
Waveguides have been built to get information on and off chips, said Thompson, because companies like IBM, Intel and Toshiba suddenly realised they needed them to keep up with Moore’s law, which states that the number of transistors that can fit on a chip doubles roughly every two years. Such manufacturers also wanted to shrink chips without them overheating. “What generates heat is getting information in and out of the chips. They are looking at silicon waveguides to move it optically,” the researcher said.
Thompson’s team has been using these waveguides to build optical computing components at the chip level. “It was the missing building block,” he told TechWeekeurope.”People can generate and detect photons in silicon, but how do you control and manipulate those photons?”
Real quantum computing within ten years?
Thompson expects commercial chips to be used for quantum cryptography will be available within three years, as that task is comparatively simple given that it only needs one pair of entangled photons. It is currently very expensive, but that cost should come down thanks to Thompson’s team’s findings. At this stage, the system would use 20 photons to produce 10 quantum bits or “qubits” to perform some specific calculations faster than conventional computers.
Within ten years, Thompson expects to have systems that can produce hundreds of qubits and perform useful calculations such as determine the actual shape of a molecule in what’s known as protein folding, which would be useful for designing drugs.
The chip to be shown at this week’s British Science Festival in Aberdeen shows a single light splitter, and the team will also demonstrate other elements of eventual quantum photonic computers.
For those who want more technical details, here are recent papers from the group:
- Photon Pair Generation in Silicon Micro-Ring Resonators with Reverse Bias Enhancement
- Quantum interference and manipulation of entanglement in silicon wire waveguide quantum circuits (New Journal of Physics)
How are you on tech patents? Try our quiz!