Would you like qubits with that new laptop? What challenges remain before we all have our own quantum PCs?
Quantum computing is one of the most exciting (and hyped) fields of research right now, but if you ask a scientist how far away quantum computers are from ‘making it’, you’ll get the classic scientist “five to 10 years” response.
So, if a new quantum reality is still potentially a decade away, how far away are we from personal quantum technology in our computer or laptop?
Although there are still some very big barriers to overcome in quantum computing, some research groups are already connecting very small quantum processors to traditional computers, and the future (as many years away as it might be) seems bright.
“With quantum computing, you’re harnessing an otherwise untapped, physical phenomenon,” says Dr Andrew Horsley, CEO of Quantum Brilliance, an Australian team working on diamond-based quantum technology.
“The last time we did that was electricity – that’s a pretty big step change in the technology underpinnings of society.”
One of the reasons quantum computing has been so hard to get going is that the architecture supporting it needed to be built entirely from scratch.
Quantum computers do calculations by using “qubits”, the quantum equivalent of computing “bits”. But instead of being either on or off (0 or 1 in binary) the qubit can be in a mixture of 0 and 1. Imagine a qubit like a globe or spinning atom where the qubit state is a point on the globe creating a mixture between 0 and 1 in both longitude and latitude.
Unfortunately, the price of adding quantum to these bits is much more room for noise – random variations that can change the value of the result. Without specialised error correction this can mean the qubit provides the wrong value. Even the 2019 quantum supremacy demonstration by Google was 99% noise and only 1% signal.
Despite our systems getting slowly better over time, we’re still in what’s known as the noisy intermediate-scale quantum era. This means that although we have the technology to build systems of up to a couple of hundred qubits, the systems are still incredibly sensitive to their environment and can lose “coherence” after just a few seconds of work.
The actual physical materials we’re using to make quantum computers are not helping the matter at all.
“In a classical system, hard disks are made out of magnetic memory, and magnetic material has the property that it retains a memory of its state for a long time… nature actually does a lot of error correction for us for free,” says Tom Stace, the Deputy Director of the ARC Centre of Excellence in Engineered Quantum Systems in Queensland.
“For quantum computers, we don’t have an analogous substance. We don’t have something that preserves quantum states indefinitely. So, we really have to engineer something that’s like a quantum magnet, because a thing that’s able to store quantum information indefinitely doesn’t exist in nature in any form.”
There have been multiple approaches to this problem.
Groups like IBM and Google are using “superconducting transmon” qubits made of materials like niobium and aluminium on silicon. Some teams, like the quantum group at the University of New South Wales are using silicon and phosphorus atoms to make “semiconductor” quantum computing, while Quantum Brilliance is using nitrogen atoms inside a diamond lattice to create its diamond-based qubits.
Each of these technologies has its pros and cons.
Superconducting qubits are better developed and more accurate, but still have high noise-to-signal ratios compared to traditional computers. These qubits are also more easily able to “interact” with each other, which is an important way for this technology to scale. However, this makes them incredibly complex machines. As Stace describes it, it’s qubits “all the way down”, needing physical qubits in one section to do error correction and logical qubits in another area to do the actual calculation.
Then there’s the issue of temperature – being a superconductor means that the machine needs to be almost at absolute zero temperatures. Even if you could put that in a laptop, the energy cost to run it mean you probably wouldn’t want to.
Silicon and diamond aren’t as advanced as their supercomputing counterparts when it comes to noise and error correction, but they do have the advantage of not having to be at super-cold temperatures.
Silicon qubits recently reached over 99% accuracy, which is an exciting milestone. It means researchers can start implementing error correction in the same way that the superconducting qubits do. But this technologystill requires colder temperatures and has only two qubits in a system, so there’s still a long way to go before it’s likely to be ready for manufacture at scale.
Then there’s diamond-based qubits. The tech itself has similar issues to silicon – although diamond qubits can be run at room temperature – but Quantum Brilliance is seemingly much further ahead, with the team soon to provide a quantum chip to the Pawsey Supercomputing Centre in Perth.
“Diamond is one of the most widely used quantum technologies, but mostly it’s just used for sensing,” says Horsley. “It’s room temperature, it’s a very simple system, it’s very high performing.
“The challenge has been scaling it beyond a handful of qubits.”
The reason why it’s hard to scale up is because of the particular way they’re made. In a process known as shotgun implantation, nitrogen atoms or electrons are fired at a piece of synthetic diamond to create something called a “nitrogen vacancy centre”. The problem is, despite the many atoms being fired, the researchers might only get one or two of these sitting at the right level to be used as a qubit.
Instead, the Diamond Brilliance team are working on a system where they implant the nitrogen vacancy, and then grow more diamond, and then implant another nitrogen vacancy, and so on.
They have grand plans for a 50-qubit system built in this way, which would make the technology useful for implementation with a classical computer to speed up time-sensitive, processor-hungry requests like speech-to-text conversion, particularly where it wouldn’t be easy to access the cloud and additional processing power.
This goal is still a fair way off though – despite exciting new software to connect the quantum system to the classical one, the box going to the Pawsey centre has just two lonely qubits.
“The really exciting thing about that is less its computation power, [and] more that we’ve been able to then take a really complex set of tabletop systems, put that in a box and ship that 3,500 kilometres away and run it in a supercomputing centre there,” says Horsley.
“I’m very curious – what are all the weird things that [that will stem from] simply having a box in their facilities?”
Quantum computing – don’t give up
Although there are some very large technological hurdles to overcome before quantum computing is likely to be in our lives, both Stace and Horsley suggest to not put away dreams of owning a personal quantum laptop.
“If you put yourself back in the 1940s, and people were inventing the first serious digital computers, you couldn’t even ask the question, ‘will we have a laptop?’ because nobody could even conceive that as a thing to have,” says Stace.
“The challenges you would had to have anticipated solving back then would be sort of inconceivable, but nonetheless, we have solved it 80 years later. I think anything’s possible.”
Jacinta Bowler is a science journalist at Cosmos. They have a undergraduate degree in genetics and journalism from the University of Queensland and have been published in the Best Australian Science Writing 2022.