PRL: New model proposes unexpected generalized quantum computing scheme

Quantum computers should be able to perform calculations that even the most powerful conventional supercomputers cannot. However, the technology is still in the early stages of development, and it is unclear which type of quantum bit is best. Currently, quantum bits based on superconducting circuits are the most advanced, but quantum bits based on arrays of cold ions have also been successful.

Recently, arrays of ultracold neutral atoms have also been investigated for use as quantum bits. Atoms are attractive because they are stable, scalable, identical in nature, and controllable due to advances in laser technology. Atoms can be excited to the Rydberg state, which allows the atoms to interact and create entanglement - a key process in quantum computing.

Laser manipulation of atomic states

In atomic arrays, lasers form regularly spaced optical tweezers that hold atoms in place. Other lasers are used to tune the quantum states of atoms by exciting them, nudging them to release energy and return to the ground state, or leaving them in a superposition of energy states.

Superpositions are useful for quantum computing.

Lasers that manipulate atomic states typically illuminate entire arrays, making it difficult to process quantum information in individual atoms. However, in 2022, a team of researchers in the U.S. and the U.K. demonstrated a method of targeting individual atoms with a laser beam; also that year, a team of researchers, including Pichler, used a different method of moving individual atoms in an array.

Pichler said, "I really like that approach." But, he added, there are benefits to methods that don't require too much control over individual atoms.

Chessa agrees, "It is true that the current results in local addressing are promising and very exciting, but this is still one of the most subtle aspects of computing with Rydberg atoms. Understandably, one prefers to use such subtle tools as little as possible and rely primarily on global control."

Globally driven quantum computing in the Rydberg Atom Array

In their new protocol, each quantum bit is a string of atoms called a "wire" (thread). Each wire can exist in one of two quantum states or in a superposition of two quantum states.

Chessa explains, "At each step of the computation, information is stored in a subset of atoms in each wire. This subset consists of 'interface atoms', which are located between two sections of wire consisting of atoms with different orderings of excited and ground states. In the standard configuration, the atoms on one side of the interface alternate between the ground and excited states, while the atoms on the other side are all in the ground state."

Within the wire, an atom cannot be excited when it is a certain distance (called the "Rydberg blockade radius") from another excited atom. This means that the incident pulse can only excite atoms on one side of the interface. Whether the first atom after the interface atom changes state depends on the state of the interface atom. In this way, when the system is pulsed, the interface and the information it encodes moves up along the wire, or backward along the wire if the pulse is reversed.

So far, the information moving up and down is constant. It changes when the interface atoms encounter "superatoms." Clusters of these atoms located at or between certain positions in the array of wires can change the state of the quantum bits. This effectively processes the quantum information in the array.

Pichler explains, "You can think of it as encoding algorithms in (configurations of) superatoms, or in sequences of pulses moving around the information. I think it links the natural dynamics of quantum many-body systems to quantum information processing in a very transparent and beautiful way."

Error correction is the next key task

Pichler noted that their protocol could complement techniques that target individual atoms "as additional know-how for designing quantum processors." Certain processes could use targeted approaches, while other subroutines could be efficiently implemented by globally addressing the entire array.

Chessa added, "By employing our ideas, it is possible to greatly reduce the number of invocations to individual atomic controls and decide wisely when to use it."

Mark Saffman of the University of Wisconsin, Madison, is an expert on targeting individual atoms. He describes the new protocol as "an unexpected solution for general-purpose quantum computing using globally controlled arrays of Rydberg interacting atoms."

The requirement to control the position and quantum state of individual atoms "places a heavy burden on the requirements of optical control systems," he said. Chessa and Pichler's global approach eliminates this requirement, which could shorten the path to scalability. However, he also notes that "the architecture does not yet incorporate error correction, which is undoubtedly necessary to realize quantum advantages in the most demanding applications."

Chessa and Pichler agree, and believe that error correction is the next key task. Pichler says, "This is a new kind of quantum processing, and it requires a new way of thinking about how to suppress errors." He notes that because each quantum bit uses a string of atoms, not just one, the process may be naively thought of as more error-prone; however, the impact of errors remains to be seen.

Chessa and Pichler have identified features that can be used to help correct errors, noting that most of the atoms in each stringy quantum bit have no information associated with them. "You don't need fully fledged quantum error correction to correct errors on such idle atoms."

Chessa and Pichler suggest that the protocol could also benefit other quantum computing platforms: such as those based on superconducting circuits.