One step closer to large-scale quantum chips

Quantum computing is the type of technology that is difficult to oversell, with the potential to perform calculations in a single step that would take hundreds of thousands of years on a conventional computer.

The issue is that the same physics loophole that gives quantum computing its incredible power also makes it nearly impossible to control reliably – but researchers believe that may be about to change, and large-scale quantum chips capable of finally delivering on the promise of quantum computing may be here much sooner than we anticipated.

According to new research published in Science Advances on August 13, a new technique could enable quantum computing engineers to reliably control millions of qubits, removing one of the major roadblocks to commercialising quantum computing.

Quantum engineers at UNSW Sydney have overcome a major impediment to the realisation of quantum computers: they discovered a new technique for controlling millions of spin qubits – the fundamental units of information in a silicon quantum processor.

Until now, quantum computer engineers and scientists have used a proof-of-concept model of quantum processors to demonstrate control over a small number of qubits.

The issue with qubits is that they rely on a quantum mechanics phenomenon called superposition, which enables a subatomic particle to possess two mutually exclusive properties (for example, the spin of an electron) simultaneously.

Quantum computing engineers use this superposition to represent the ones and zeroes that comprise the fundamental unit of digital technology – the bit – but a qubit can be both one and zero at the same time due to superposition (thus, making it a quantum bit, or qubit for short).

This enables a quantum computer to perform unfathomably complex computations that would take a billion years for an Intel Rocket Lake processor to complete in one go by computing all possible outcomes simultaneously.

The issue is that when you “look” at a qubit, its superposition collapses into a defined state and it reverts to being a plain old bit, effectively eradicating qubits’ incredible computing power.

This makes effective control over them to perform calculations extremely difficult, requiring a variety of equipment to filter out external interference and keep the qubits as close to absolute zero as possible to ensure they remain mostly still and do not collide, all of which counts as “looking” in quantum mechanics.

This has hampered engineers’ ability to reliably control dozens, hundreds, or even a few thousand qubits, but now researchers at the University of New South Wales (UNSW) claim to have solved the qubit control problem, potentially unlocking the power of quantum computing for some of the most pressing real-world problems, including medical research, climate forecasting, and a whole lot more.

“Until now, controlling electron spin qubits required us to deliver microwave magnetic fields via a wire adjacent to the qubit,” said Dr. Jarryd Pla, a faculty member at UNSW’s School of Electrical Engineering and Telecommunications. “This presents some significant challenges if we are to scale up to the millions of qubits required by a quantum computer to solve globally significant problems, such as vaccine design.”

The issue is that adding more qubits requires adding more wires to generate the magnetic field required to control them. However, wires generate heat, and excessive heat can cause qubits to collapse into bits, so adding additional wires to a quantum processor will simply not work.

The researcher solved this problem by eliminating the wires entirely and applying magnetic control fields from above the quantum chip via a crystal prism called a dielectric resonator. This allows for simultaneous control of all of the qubits.

“We removed the wire connecting the qubits and then devised a novel method for delivering microwave-frequency magnetic control fields throughout the system,” Dr. Pla explained. “Therefore, we could theoretically deliver control fields to up to four million qubits.”

Making quantum computing on a large scale a reality
“I was completely taken aback when Dr. Pla approached me with his new idea,” said Prof. Andrew Dzurak, a UNSW engineering colleague who had spent years working on quantum logic implementation on silicon chips. “We immediately set to work figuring out how to integrate it with the qubit chips developed by my team.”

“We were ecstatic when the experiment was successful,” he continued. “This problem of controlling millions of qubits had been a source of concern for me for a long time, as it was a significant impediment to building a full-scale quantum computer.”

While this research may prove to be a necessary first step toward widespread, large-scale quantum computing, much more work remains to be done. One of the difficulties is that, while a quantum computer can calculate as many results as the number of qubits allows, reading the desired answer from those same qubits results in the same quantum decoherence as heat or other interference. Thus, even if a quantum computer calculates all possible outcomes, you can only ever access one of them.

“The trick is to cleverly design your algorithm in such a way that the correct answer reveals itself at the conclusion of the calculation while still utilising parallelism,” Dr. Pla explained via email. “That is why a quantum computer can perform only a few tasks faster than a classical computer (such as factoring large composite prime numbers, searching unsorted databases, and so on), because it is difficult to design such clever algorithms – though people are improving at it and more useful examples are emerging almost daily.”

Other engineering challenges remain, such as refining error correction to reduce the number of qubits required to construct quantum circuits.

“It is critical to distinguish between a ‘physical qubit’ (in our case, a single electron spin) and a ‘logical qubit,'” Dr. Pla explained. “If all of your physical qubits could be controlled and measured with infinite precision (no errors), you’d have a 4-million-qubit quantum computer capable of solving virtually any problem we can think of right now.

“However, qubits contain errors, which accumulate rapidly in a quantum circuit. As a result, you must implement some form of error correction in the case of qubits encoded in groups of qubits (this is called quantum error correction). Logical qubits are qubit groups that are error-protected. The number of qubits required in the groups is highly system-dependent, i.e. how well connected the qubits are and the actual error rates.

“Thus, we may require somewhere in the neighbourhood of 1000 physical qubits to generate a useful logical qubit capable of being used in computations. This reduces the four million count to four thousand – which is still quite useful. At that level, you can decrypt 2048-bit number encryptions and simulate complex chemical processes, as well as deduce the structures of proteins.”

While this is a start, and we would not have the modern information age without the room-sized ENIAC, we should not have to wait long to see the promise of quantum computing realised.

Reference
“Single-electron spin resonance in a nanoelectronic device using a global field” 13 August 2021, Science Advances. DOI: 10.1126/sciadv.abg9158