Quantum computing race Neutral atoms rise to the top

The race to build practical quantum computers may be entering a new phase. Some leading-edge technologies are now facing size constraints, while others are rapidly catching up.

For years, two leading approaches have enabled physicists to make progress, in part by stuffing devices with more and more quantum bits (the quantum equivalent of a computer's memory bits). One approach encodes quantum bits as currents running on superconducting loops. Another method uses the excited state of a single ion trapped in a vacuum by an electromagnetic field.

IBM's quantum computing system has a cryostat at its center to cool the quantum chips.Credit: IBM

Barbara Terhal, a theoretical physicist at QuTech, a quantum research institute at Delft University of Technology in the Netherlands, said, "Superconducting quantum bits and trapped ion quantum bits have done state-of-the-art experiments with the largest number of quantum bits being controlled. However, this does not guarantee that these platforms will remain on the leading edge."

Requirements for quantum bits

Quantum computers promise to solve problems that classical machines cannot by exploiting phenomena such as quantum superposition, in which an object can exist in two states at once: rotating both clockwise and counterclockwise. Physicists refer to such states as quantum bits to distinguish them from ordinary bits, which can only be 0 or 1.

Quantum states are notoriously fragile. In quantum computers, the information they carry can be extended to several quantum bits to form "entangled" states: often degraded or lost during computation. To preserve these states for as long as possible, quantum bits must be isolated from their environment. But they cannot be too isolated from each other, because they must interact with each other in order to perform computations.

This is something that makes it challenging to build a useful quantum computer. But progress in the field is further along than QuTech's research director Lieven Vandersypen expected a decade ago: "The progress is actually impressive."

Google made headlines in 2019 when it claimed that a machine made up of 54 superconducting quantum bits had performed the first quantum calculations that would take an impossible amount of time on a classical computer, an achievement that researchers are calling the quantum advantage. IBM, a technology company that has invested heavily in superconducting quantum bits, expects to reach a milestone in the coming months when it will launch a quantum chip called Condor, the first to break the 1,000-quantum-bit barrier.

Last November, the company announced its last chip: the 433-bit Osprey, a successor to the 127-bit Eagle, which set a record in 2021. "We really want to set a roadmap, just like we would expect from the semiconductor industry." said Jerry Chow, who leads the quantum computer program at the IBMThomas J. Watson Research Center in Yorktown Heights, New York.

Quality, quantity of quantum bits

Chow said IBM's goal is not only to expand the number of quantum bits, but also to improve their quality. Some of the company's superconducting elements can maintain their quantum state for more than 300 microseconds, a record for the technology, he said. In another key measure, 99.9 percent of operations involving two quantum bits are now error-free.

Once the number of superconducting quantum bits on a chip exceeds well over 1,000, scaling up becomes impractical because each quantum bit needs to be individually connected to external circuitry for control and readout. Therefore, IBM will take a modular approach. Each step on its roadmap, starting in 2024, is not about increasing the number of quantum bits on a chip, but about connecting multiple chips to a single machine; not a simple task if the connections must carry quantum states unharmed or help entangle the quantum bits on different chips. These chips are at the heart of a huge device wrapped in a cryogenic system that brings the chips close to 0 Kelvin.

Trapped-ion computers may have tighter size constraints than superconducting computers, in part because they require a separate laser device to control each ion. Typically, this means limiting the traps to about 32 ions per chip. But IonQ - a startup spun out of the University of Maryland, College Park - says its approach makes it possible to pack multiple rows of ions into a single chip, perhaps up to 1,024 quantum bits. To surpass that goal, IonQ also plans to use a modular approach, connecting multiple chips.

A company spokesman said that in laboratory experiments, the fidelity of captured ions has been as high as 99.99 percent.

Optical tweezer technique captures neutral atoms

Another technique may soon break the 1,000-quantum-bit barrier as well: It uses tightly focused laser beams called "optical tweezers" to capture neutral atoms and encode quantum bits in the electronic states of the atoms, or in the spin of the nucleus. Harvard physicist Giulia Semeghini says the method has been evolving for more than a decade, but now it's "booming.

To assemble multiple quantum bits, physicists divide a laser beam into many beams, for example, and pass it through a screen made of liquid crystal. This creates an array of hundreds of optical tweezers, each capturing its own atoms. These atoms are usually only a few micrometers away from their neighbors, where they can remain in a quantum state for a few seconds or more. To get the atoms to interact, physicists point a separate laser at one of the atoms, sending it into an excited state in which the outer electrons orbit farther from the nucleus than normal.

This promotes electrostatic interactions between that atom and neighboring atoms.

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

A major advantage of the technique is that physicists can combine multiple types of tweezers, some of which can move quickly, with the atoms they carry. Every time you want two of them to interact, you put them together," says Harvard physicist Dolev Bluvstein. This makes the technique more flexible than other platforms, such as superconductors, in which each quantum bit can only interact with its direct neighbor on the chip. A team, including Semeghini and Bluvstein, demonstrated this flexibility in an April 2022 paper [1]." Semeghini said that optical tweezers-based quantum bits should soon achieve a 99 percent fault tolerance threshold, although further improvements will require a lot of work.

The speed of improvement in neutral atoms has surprised the quantum computing community. "The path to scaling to thousands of atomic quantum bits is clear and could be achieved within two years," said Lu Chaoyang, a physicist at the University of Science and Technology of China.

Spin quantum bits

Other quantum bit technologies are still in their infancy, but are making steady progress.

One approach is to encode information in the spin of a single electron captured by an electric field within a conventional semiconductor, such as silicon. Last year, Vandersypen and his collaborators demonstrated such a fully working six-quantum-bit machine [2]. As in the case of optical tweezers, the spins of the electrons can be shuttled through the device in order to bring them next to other electrons as needed. However, as in the case of other types of quantum bits, a major difficulty is to prevent the electron spins from influencing each other when they should not, what physicists call crosstalk.

The benefit of semiconductor-based quantum bits is the ability to build chips in the same type of factories that currently produce computer chips, and a team led by physicist Michelle Simmons of the University of New South Wales in Sydney, Australia, is assembling the devices one by one using the tip of an automated scanning tunneling microscope. "Everything is designed with sub-nanometer precision," she says.

There is another approach that is still in the conceptual stage, but it has already received significant investment, notably from Microsoft. The technique aims to make quantum bits robust to degradation using "topological states," like a knotted string that can be twisted and pulled, but not untangled. 2020, researchers observed a fundamental physical mechanism for topological protection, and they are now working to demonstrate the first topological quantum bits.

"Every platform pursued today holds some promise, but developing it may require truly novel ideas that we can't predict." Vandersypen said. Pan Jianwei, a physicist working on multiple quantum computing methods at the University of Science and Technology of China, agrees and says that when it comes to the race to develop quantum computers, "it's too early to say which candidate route will win."

Reference links:

[1]https://www.nature.com/articles/s41586-022-04592-6

[2]https://www.nature.com/articles/s41586-022-05117-x