A radical new chip design invented by engineers at the UNSW-based Centre of Excellence for Quantum Computation and Communication Technology (CQC2T) promises to make the manufacture of silicon-based quantum computers dramatically cheaper and easier than was previously thought possible.
The breakthrough allows quantum bits – the basic unit of information in a quantum computer – to be placed hundreds of nanometres apart and still remain coupled. Previously it was understood qubits would need to be spaced only 50 atoms apart, meaning the surrounding electronics would also need to be built to a nanometric scale.
“If we want to make an array of thousands or millions of qubits so close together, it means that all the control lines, the control electronics and the readout devices must also be fabricated at that nanometric scale, and with that pitch and that density of electrodes,” explains Andrea Morello, program manager at CQC2T.
“This new concept suggests another pathway. It’s a brilliant design, and like many such conceptual leaps, it’s amazing no-one had thought of it before.”
The design – dubbed the ‘flip flop qubit’ – was conceived by Guilherme Tosi, Morello, Fahd Mohiyaddin, Vivien Schmitt and Stefanie Tenberg from CQC2T, with collaborators Rajib Rahman and Gerhard Klimeck of Purdue University in the US.
The centre had already successfully built two functional qubits using the electron and nucleus of a phosphorus atom trapped in silicon. Those qubits, taken individually, have demonstrated world-record coherence times. But to perform calculations the qubits need to be placed only a few atoms apart so the electrons touch each other.
“We have now discovered that this is not necessary. We can make the phosphorus qubits talk to each other over much larger distances,” says Tosi.
(In quantum computing terms a large distance is “a good fraction of a micron” or 1000 nanometres; about one hundredth the width of a human hair.)
The flip flop qubit encodes quantum information in the combined state of the electron and the nucleus of the phosphorus atom.
“It is operated by pulling the electron away from the nucleus, and then oscillating the electron position around its equilibrium point. This means that we can now control a qubit in silicon using electric rather than magnetic signals. This makes it much easier to integrate with normal electronic circuits,” Morello adds. “Electric signals are significantly easier to distribute and localise within an electronic chip.”
The greater distances also allow the qubits to be more easily integrated with readily available circuitry.
“Once the negative of the charge of the electron is pulled away from the positive charge on the nucleus, the qubit creates an electric field that reaches over large distances. So we can now design a large scale quantum computer where there’s plenty of space to insert interconnects, control lines, a readout devices; without having to fabricate components at the scale of a few atoms,” Tosi says.
The breakthrough is a significant boost for the newly formed Silicon Quantum Computing, Australia’s first quantum computing hardware company which launched last month with funding from the federal and NSW governments, UNSW, the Commonwealth Bank of Australia and Telstra.
The company, based at UNSW’s Sydney campus has the goal of producing a 10 qubit integrated circuit prototype by 2022.
“This is a theory, a proposal – the qubit has yet to be built,” says Morello. “We have some preliminary experimental data that suggests it’s entirely feasible, so we’re working to fully demonstrate this.”
The silicon-based approach is just one of a number of potential paths to a fully functional quantum computer.
IBM and Google are pursuing the superconducting circuits approach. Their systems are larger and easier to fabricate and currently have the lead in the number of qubits that can be operated. However, their larger size means they may face challenges in the longer term when building arrays of millions of qubits that will be required by the most useful quantum algorithms.
Microsoft is taking a topological approach to forming qubits uses quasiparticles called non-abellian anyons. The quasiparticles are as yet unproven, but the company is confident, saying it had reached an ‘inflection point’ in its research.
The company cemented its long-standing quantum computing research relationship with the University of Sydney, with the signing of a multi-year investment deal understood to be in the multiple millions in July.
“Our new silicon-based approach sits right at the sweet spot,” adds Morello. “It’s easier to fabricate than atomic-scale devices, but still allows us to place a million qubits on a square millimetre.”