Quantum computing based on trapped atomic ions has already proved itself to be a leading hardware platform for quantum information processing. Indeed, trapped ions have been used to realize quantum gates — the basic building blocks of a quantum computer — that have the smallest quantum-computation errors of any hardware platform1,2. The approach also stands out because it could allow practical machines to be built that do not require cooling to ultra-low (millikelvin) temperatures. However, there have been few comprehensive demonstrations of quantum-computing architectures capable of being scaled up to thousands of quantum bits (qubits). Writing in Nature, Pino et al.3 report the construction and operation of a prototype microchip-based, trapped-ion quantum computer that incorporates a promising architecture based on ion shuttling.
The concept of quantum computing relies on the strange phenomena of quantum physics, the counter-intuitive predictions of which Albert Einstein referred to as spooky. Quantum computers promise to perform calculations in hours or even minutes that might take millions of years to run on the fastest conventional supercomputer. Full-scale quantum computers containing millions of qubits would have transformative uses in nearly every industry, from simulating chemical reactions and helping to develop pharmaceuticals to disruptive applications in finance and machine learning. However, building such machines has been viewed as a holy grail of science because of the enormous difficulty involved.
The two most promising approaches to constructing practical quantum computers are based on superconducting circuits4 and trapped ions. A distinct advantage of superconducting qubits is that they are fundamentally chip-based. However, their working principle relies on cooling to millikelvin temperatures. The maximum cooling power that can be achieved at such low temperatures is very small. This makes it difficult to cool large devices, making it challenging to scale up machines to sizes large enough to hold millions of qubits.
By contrast, trapped ions have delivered world records in reducing errors that occur in the operation of quantum gates used to carry out calculations1,2, but many demonstrations have not been chip-based or have been accomplished with only 2–15 qubits. The next crucial step for trapped-ion quantum computing is the practical demonstration of architectures that enable scaling to thousands or even millions of qubits.
Pino and colleagues’ paper describes the realization of a prototype microchip-based, trapped-ion quantum computer. In the reported architecture, which is somewhat reminiscent of the Pac-Man video game, ions hover above the surface of the microchip and are moved along tracks by the adjustment of voltages applied to electrodes located on the chip surface (Fig. 1). Because of its similarity to microchips known as charge-coupled device (CCD) arrays, this design has been termed the quantum-CCD architecture5,6. The authors’ quantum computer is not the first realization of this design, but it is by far the most advanced implementation yet. A different trapped-ion architecture had already been demonstrated in an alternative microchip-based quantum computer, one in which laser beams manipulate the internal state of individual ions in a chain of trapped stationary ions7.
Quantum computers are often characterized by how many qubits they can host. Without the ability to correct for unavoidable errors, the number of usable qubits is limited both by the magnitude of individual errors and by the accumulation of all the errors in the system. A parameter known as the quantum volume provides a measure of how many usable qubits a machine contains, based on the overall system performance. Pino et al. determined that their device has a quantum volume of 64, which means that it can do generalized computations using up to 6 qubits. In principle, a single quantum-computing module built according to the quantum-CCD design could hold hundreds or even thousands of usable qubits. Therefore, it should be possible to scale up this architecture to a million-qubit machine using a modular approach8.
The authors also demonstrated a complete toolbox of features that make up the quantum-CCD architecture. These features include a full range of ion-transport operations, parallel operation zones and a mechanism called sympathetic cooling, whereby an ion that serves as a qubit is cooled by an ion of a different element to preserve the qubit’s internal state, which carries the quantum information.
Pino and colleagues’ work constitutes an impressive achievement, yet again illustrating the coming of age of trapped-ion quantum computing as a leading hardware platform. However, quantum computers that can tackle some of the most exciting real-world problems will need to host millions of qubits. Although trapped-ion technology is well suited for scaling up because it works at room temperature or requires only modest cooling, such scaling requires innovative ideas.
In the current paper, the execution of quantum gates is achieved using laser beams that need to be carefully aligned, posing a substantial engineering challenge when considering scaling to millions of qubits. Another highly promising approach is to use microwave technology that allows quantum gates to be implemented by applying voltages to the microchip9. Alternatively, the interaction of ions with the relevant laser beams could be facilitated using structures called waveguides in the microchip10.
Undoubtedly, the field of trapped-ion quantum computing is undergoing a transition from demonstrating the fundamental physics needed for quantum computations to addressing the difficult engineering problems involved in realizing a working computer. The solutions to these problems will enable the construction of quantum computers that will have transformative effects across many industry sectors. Indeed, this process has already started with the creation of a blueprint on how to build quantum computers hosting millions or even billions of qubits8. We are at the beginning of a fantastic journey towards constructing machines that were previously thought to belong to the realm of science fiction.
Competing Financial Interests
The author has shares in, and is partially seconded to, a company that is developing trapped-ion quantum computers, albeit using a different approach from that described by Pino and colleagues.