Australian researchers believe they’ve solved a decades-old problem that could ultimately lead to the development of larger and more useful quantum computers.
Quantum computing has the potential to solve certain problems millions of times faster than conventional computers.
But despite billions of dollars of investment, it’s remained tantalisingly out of reach.
In research published today in Science Advances, the team of researchers from the University of New South Wales has proposed a new way of controlling millions of qubits, which are the building blocks of quantum computers.
The current largest quantum computers have only a few dozen qubits, but it will take millions of qubits to develop new pharmaceutical drugs or construct more accurate climate models.
Standing in the way historically has been the problem of how to control millions of qubits at once without making the thumbnail-size quantum chip too crowded or too hot.
Known as the problem of achieving “global control”, it was identified in the late 1990s and sat in the “unsolved” basket for two decades.
But one day in mid-2019, Jarryd Pla, a young researcher, walked into the office of Professor Andrew Dzurak, one of the leading quantum computing researchers in the country with an idea about how to use magnetic fields to control qubits.
“I remember when Jarryd came into the office and talked to me about the idea,” Professor Dzurak said.
“I was like, ‘Oh my God, if that works it’s going to solve all our problems’.
“And it has.”
A very cramped, very cold problem
Before we get to Dr Pla’s idea, we need to understand why heat is the great enemy of quantum computers.
Qubits are very small and strange – traditional computers store information as either zeros or ones, but in a quantum computer, qubits can be both numbers at the same time.
This ability, known as superposition, means quantum computers have the ability to perform multiple calculations at once.
But, there’s a catch. To retain their quantum abilities, qubits need to be kept very cold.
“It’s unbelievably cold,” said Andrew Doherty, a quantum physicist at University of Sydney, who was not involved in the research.
“It’s like a 20th of the temperature of outer space,” Professor Doherty said.
Many of the quantum computing technologies operate at around 0.1 Kelvin, or just above absolute zero – which is -273.15 degrees Celsius.
Quantum chips have to operate within vast cryogenic dilution fridges, in which isotopes of liquid helium are pumped through a system of tubes that look a bit like a chandelier or an upside-down bird’s nest.
And that’s the easy bit.
To be useful for computation, qubits also need to be controlled.
This is often done by putting a current through a wire close to the qubit to create a magnetic field that controls its spin.
The direction of spin is equivalent to a zero or one in a binary code, explained Professor Dzurak.
“You can think of the spin down as a zero and the spin up as one,” he said.
These control-wires transfer heat. As more qubits are installed, more wires are needed to control them, and the task of keeping the qubits cool becomes harder.
“You can’t control millions of qubits, which you’re going to need to do to solve these big important problems,” Professor Dzurak said.
Then there’s the problem of real estate: as well as being kept cold, qubits need to be packed close together to retain their quantum properties.
In some models, they’re separated by 100 nanometres, which is a tenth of a micron. A human hair is about 70 microns thick.
The control-wires take up valuable space on a thumbnail-sized quantum chip that also needs to contain millions of qubits.
Controlling four million qubits at once
To get around these problems, Dr Pla came up with the idea to get rid of the wires altogether and replace them with a magnetic field from above the chip that can manipulate all the qubits simultaneously.
Enter something called a “dielectric resonator”.
Though it may sound like the heart of a time machine, it’s actually a fairly standard piece of equipment – as an antennae for microwave frequencies, it’s used in mobile phones to send high-frequency signals.
The one developed by the researchers is a crystal prism made out of potassium tantalate, which allows it to work at very low temperatures.
“It’s basically a small transparent crystal,” Dr Pla said.
The crystal is zapped with microwaves, which get trapped and bounce around.
In doing so, they generate a magnetic field that emanates from the bottom of the crystal in a flat and uniform plane.
The researchers found the field generated by the resonator could control an area that could potentially fit four million qubits; enough for global control.
In addition, they found it requires relatively little power to create the magnetic field, which means, crucially, not too much heat.
“The elegance of Jarryd’s design is basically that he’s taken all of that microwave wiring off the chip, and it just sits it above it,” Professor Dzurak said.
A quantum processor chip Photo: Wikimedia Commons
Promising but just the ‘beginning of the story’
David Reilly, an experimental physicist and director of the University of Sydney’s quantum computing lab was not involved in the research, but said the results were “very interesting”.
“It’s an important development that solves a niggling problem that many researchers were concerned about,” he said.
Professor Doherty said the solution was a “promising approach to an important problem”.
“It’s great science with excellent results,” he said.
Though promising, Professor Doherty said results were preliminary, and there was still a long way to go before definitely proving global control.
“They roughly speaking have one qubit sitting under a magnetic field,” he said.
“The thing they’re not showing is they don’t have a million qubits.”
Another challenge would be manufacturing qubits in large quantities.
“In order to get a million on a chip, qubits need to be not a handmade thing but a large-scale fabrication,” Professor Doherty said.
“That’s going to be very challenging.”
How far are we away from seeing a quantum computer?
While the reality of quantum computing is still some time away, Professor Doherty, who has recently returned from a two-year stint with Silicon Valley quantum computing company PsiQuantum, is confident we’ll get there.
“Twenty-five years ago, I used to say that it’ll be 50 years before we have a quantum computer. Now I think we might not need those 50 years,” he said.
He said he was encouraged by the amount of money that’s being spent on quantum computing research.
In 2020, investors poured US$557.5 million into 28 venture deals for quantum-computing companies based in the US and Canada; that’s more than three times the amount spent in 2019.
In July this year, PsiQuantum, co-founded by two Australians, raised close to US$450 million in funding to build what it claims will be the world’s first commercially viable quantum computer.
“The UNSW team’s solution to global control “is the beginning of a story rather than the end of one,” Professor Doherty said.
Originally published at Rnz