Quantum computing engineers are setting a new standard for silicon chip performance

Two milliseconds – or 2 thousandths of a second – is an extraordinarily long time in the world of quantum computing.

On these time scales, a blink of an eye – to a 10th of a second – is like an eternity.

Now a team of researchers from the University of New South Wales (UNSW) in Sydney have broken new ground by proving that ‘spin qubits’ – properties of electrons representing the basic units of information in quantum computers – can contain information up to two milliseconds. Known as “coherence time,” the amount of time qubits can be manipulated into increasingly complicated calculations, the achievement is 100 times longer than previous benchmarks in the same quantum processor.

“Longer coherence time means you have more time for your quantum information to be stored, which is exactly what you need when performing quantum operations,” says PhD student Amanda Seedhouse, whose work in computer science quantum theory have contributed to the realization.

“Coherence time basically tells you how long you can perform all the operations in the algorithm or sequence you want to perform before you lose all of your qubit information.”

In quantum computing, the longer you can keep spins moving, the better the chance that information can be retained during calculations. When the spin qubits stop spinning, the computation collapses and the values ​​represented by each qubit are lost. The concept of coherence extension has already been confirmed experimentally by quantum engineers from UNSW in 2016.

Making the task even more difficult is the fact that working quantum computers of the future will need to track the values ​​of millions of qubits if they are to solve some of humanity’s greatest challenges, such as finding effective vaccines, modeling systems meteorological conditions, and forecast the impacts of climate change.

Late last year, the same team at UNSW Sydney solved a technical problem that had stuck engineers for decades on how to manipulate millions of qubits without generating more heat and interference. Rather than adding thousands of tiny antennas to control millions of electrons with magnetic waves, the research team found a way to use a single antenna to control all of the chip’s qubits by introducing a crystal called dielectric resonator. These results have been published in Scientists progress.

This solved the problem of space, heat and noise which would inevitably increase as more and more qubits are brought online to perform the mind-blowing calculations which are possible when the qubits represent not just one or zero like conventional binary computers, but both, using a phenomenon known as quantum superposition.

Global control vs individual control

However, this proof-of-concept achievement still left some challenges to be resolved. Lead researcher Ingvild Hansen has joined Seedhouse to address these questions in a series of journal articles Physical examination B, Physical examination A and Applied physics exams– the latest article published just this week.

Being able to control millions of qubits with a single antenna was a big step forward. But while controlling millions of qubits at once is a great feat, working quantum computers will also need them to be manipulated individually. If all spin qubits spin at roughly the same frequency, they will have the same values. How can we control them individually so that they can represent different values ​​in a calculation?

“We first showed theoretically that we can improve coherence time by continuously spinning qubits,” says Hansen.

“If you imagine a circus performer spinning plates, while they are still spinning, the show can go on. Similarly, if we continuously drive qubits, they can hold information longer. We have shown that such “dressed” qubits have coherence times of more than 230 microseconds [230 millionths of a second].”

After the team showed that coherence times could be extended with so-called dressed qubits, the next challenge was to make the protocol more robust and show that globally controlled electrons can also be individually controlled so that they can contain different values ​​needed for complex processes. calculations.

This was achieved by creating what the team dubbed the “SMART” qubit protocol – Sinusoidally Modulated, Always Rotating and Tailored.

Rather than spinning the qubits in circles, they manipulated them to rock back and forth like a metronome. Then, if an electric field is individually applied to any qubit – bringing it out of resonance – it can be put into a different tempo from its neighbors, but still moving at the same rate.

“Think of it like two kids on a seesaw moving forward and backward pretty much in sync,” says Seedhouse. “If we give one of them a push, we can get them to reach the end of their arc at opposite ends, so one can be a zero when the other is now a one.”

The result is that not only can a qubit be controlled individually (electronically) under the influence of a global control (magnetically), but the coherence time is, as mentioned before, significantly longer and suitable for quantum calculations.

“We have shown a simple and elegant way to control all qubits at once, which also comes with better performance,” says Dr. Henry Yang, one of the team’s lead researchers.

“The SMART protocol will be a potential route for large-scale quantum computers.”

The research team is led by Professor Andrew Dzurak, CEO and founder of Diraq, a UNSW spin-off company that develops quantum computing processors that can be fabricated using standard silicon chip fabrication.

Next steps

“Our next goal is to show it works with two-qubit computations after showing our proof of concept in our experimental paper with one qubit,” says Hansen.

“After that, we want to show that we can also do this for a handful of qubits, to show that the theory is proven in practice.”

– This press release originally appeared on the University of New South Wales website

Sherry J. Basler