Underdog Technologies Gain Ground in Quantum-Computing Race

The race to build practical quantum computers might be entering a new phase. Some of the front-runner technologies are now facing size constraints, and others are rapidly coming up from behind.

For years, two leading approaches have enabled physicists to make progress partly by cramming devices with more and more qubits, the quantum equivalent of a computer’s memory bits. One of those methods encodes qubits as currents running on superconducting loops. The other uses excited states of individual ions trapped in a vacuum by electromagnetic fields.

But in the past two years, qubits that consist of single neutral atoms — as opposed to ions — and are held with ‘tweezers’ made of laser light have suddenly become competitive. And other techniques that are at an even earlier stage of development could yet catch up.

“Superconducting qubits and trapped-ion qubits have done the most-advanced experiments, with the most qubits under control,” says Barbara Terhal, a theoretical physicist at QuTech, a quantum-research institute at the Delft University of Technology in the Netherlands. “However, this is no guarantee that these platforms will stay in the lead.”

The quest for qubits

Quantum computers promise to solve problems that are out of reach for classical machines by harnessing phenomena such as quantum superposition, in which an object can exist in two simultaneous states — spinning both clockwise and anticlockwise, for example. Physicists call such states qubits to distinguish them from ordinary bits, which can be only ‘0’ or ‘1’.

Quantum states are notoriously fragile. In a quantum computer, the information they carry — which can extend across several qubits to form ‘entangled’ states — tends to degrade or get lost as a calculation progresses. To preserve the states for as long as possible, qubits must be kept isolated from the environment. But they cannot be too isolated from one another because they must interact to perform calculations.

This — among other factors — makes building a useful quantum computer is challenging. But the field has come further than QuTech director of research Lieven Vandersypen would have expected ten years ago. “The progress is actually impressive.”

Google made headlines in 2019 when it claimed that a machine made of 54 superconducting qubits had performed the first quantum computation that would have taken impossibly long on a classical computer, an achievement that researchers call quantum advantage. The technology company IBM, which has invested heavily in superconducting qubits, expects to reach a milestone in the next few months, when it will unveil a quantum chip named Condor, the first to breach the 1,000-qubit barrier.

Last November, the company announced its previous chip, the 433-qubit Osprey — a follow-up to the 127-qubit Eagle, which set a record in 2021. “We really wanted to lay a road map like you would expect from the semiconductor industry,” says Jerry Chow, who leads the quantum-computer programme at the IBM Thomas J. Watson Research Center in Yorktown Heights, New York.

Quality and quantity

Chow says that IBM’s aim is not only to scale up the number of qubits, but also to improve their quality. Some of the company’s superconducting elements can hold their quantum states for more than 300 microseconds, he says — a record for the technology. In another crucial measure, 99.9% of operations involving two qubits are now error-free.

Scaling up becomes impractical once the number of superconducting qubits on a chip goes much beyond 1,000, because each qubit needs to be individually wired to external circuits for control and readout. IBM will therefore take a modular approach. Starting in 2024, each further step on its road map will aim not to increase the number of qubits on a chip, but to link multiple chips into one machine — something that is not straightforward if the connection has to carry the quantum states unharmed or help to entangle qubits on separate chips. The chips are at the hearts of massive contraptions encased in cryogenic systems that keep the chips close to 0 kelvin.

Trapped-ion computers could have even more-stringent size constraints than superconducting ones, partly because they require a separate laser device to control each ion. Typically, that has meant limiting the traps to rows of around 32 ions per chip. But IonQ, a start-up company spun off from the University of Maryland in College Park, says its approach enables it to pack multiple rows of ions into a single chip, perhaps reaching as many as 1,024 qubits. To go beyond that, IonQ also plans to move to a modular approach, connecting multiple chips. In laboratory experiments, trapped ions have reached fidelities as high as 99.99%, according to a spokesperson for the company.

Tweezer tech

Another technique — which, until a few years ago, was barely on the radar — might soon break the 1,000-qubit barrier as well. It traps neutral atoms using tightly focused laser beams, called optical tweezers, and encodes qubits in the electronic states of the atoms or in the spins of atoms’ nuclei. The approach has been developing gradually for more than a decade, but now it’s “booming”, says Giulia Semeghini, a physicist at Harvard University in Cambridge, Massachusetts.

To assemble multiple qubits, physicists split a single laser beam into many, for example by passing it through a screen made of liquid crystals. This can create arrays of hundreds of tweezers, each trapping their own atom. The atoms are typically a few micrometres away from their neighbours, where they can persist in a quantum state for several seconds or more. To make the atoms interact, physicists point a separate laser at one of them to tickle it into an excited state, in which an outer electron orbits much farther away from the nucleus than normal. This boosts the atom’s electrostatic interactions with a neighbour.

Using tweezers, researchers have built arrays of more than 200 neutral atoms, and they are rapidly combining new and existing techniques to turn these into fully working quantum computers.

One major advantage of the technique is that physicists can combine multiple types of tweezers, some of which can move around quickly — with the atoms they carry. “Every time you want two of them to interact, you bring them together,” says Harvard physicist Dolev Bluvstein. This makes the technique more flexible than other platforms such as superconductors, in which each qubit can interact only with its direct neighbours on the chip. A team including Semeghini and Bluvstein demonstrated this flexibility in an April 2022 paper.

The tweezer-based qubits should soon be 99% error-free, although further improvements will take substantial work, Semeghini says.

The pace of improvement in neutral atoms has surprised the quantum-computing community. “The path to scale to thousands of atomic qubits is clear and will likely happen within two years,” says physicist Chao-Yang Lu at the University of Science and Technology of China (USTC) in Hefei.

Spin control

Other qubit technologies are still in their infancy, but advancing steadily. One method encodes information in the spin of individual electrons trapped by electric fields inside conventional semiconductors such as silicon. Last year, Vandersypen and his collaborators demonstrated a fully working six-qubit machine of this kind2. As in the case of optical tweezers, the electron spins can be shuttled around the device to bring them next to others on demand. But just like other types of qubit, a major difficulty is keeping the spins from influencing each other when they are not supposed to, in what physicists call crosstalk.

The benefit of semiconductor-based qubits would be the ability to make chips in the same type of factory where current computer chips are produced, although a team led by physicist Michelle Simmons at the University of New South Wales in Sydney, Australia, assembles the devices atom by atom using the tip of an automated scanning tunnelling microscope. “Everything is patterned with sub-nanometre precision,” she says.

Yet another approach is still at the conceptual stage, but it has received substantial investment, by Microsoft in particular. The technique aims to exploit ‘topological states’ to make qubits robust to degradation, just like a knotted string that can be twisted and pulled but not untied. In 2020, researchers observed the basic physical mechanism for one kind of topological protection, and they are now working on demonstrating the first topological qubits.

“Every platform that is pursued today has some promise, but developing it can require really novel ideas that you can’t predict,” says Vandersypen. Pan Jian-Wei, a physicist who works on multiple quantum-computing approaches at USTC, agrees. When it comes to the race to develop quantum computers, “it is still too early to say which candidate will win”.

This article is reproduced with permission and was first published on February 6 2023.

#Underdog #Technologies #Gain #Ground #QuantumComputing #Race

Leave a Comment