The fundamental unit of information in classical computing is a bit, which is represented by a 0 or a 1. The fundamental unit of information in quantum computing is the quantum bit, or qubit, which is often oversimplified as being in a state, a superposition, of 0 and 1 at the same time. Mathematically, however, a qubit can be represented as α∣0⟩+β∣1⟩, where α and β are complex numbers, and ∣0⟩ and ∣1⟩ are the basis states. The condition α^2+∣β^2=1 must be satisfied, and the squares of the absolute values of α and β represent the probabilities of measuring the qubit in the state ∣0⟩ or ∣1⟩, respectively. In other words, instead of being definitely 0 or definitely 1, like a bit, a qubit has some probability of being 0 and some probability of being 1. However, by using complex numbers, there is also information in the phase of the qubit, which has no analog in classical computing.
To visualize a bit, one can imagine a simple light switch with a 0 at the top and a 1 at the bottom. At any given time, the light switch is definitely in one of those two states. The “0 and 1 at the same time” analogy incorrectly implies that a qubit works like a slider, as if brightening or dimming a light between 0 and 1, only to snap to 0 or 1 upon observation. A qubit can indeed be used in such a manner, it can even be used like a bit, but doing so greatly underutilizes the qubit.
As the Quantum Inspire article “What is a qubit?” points out, the use of complex numbers allows qubits to be visualized as three-dimensional unit spheres, called Bloch spheres. An Azure Quantum article also titled “What is a qubit?” imagines a bit as a coin that is lying on a table as definitely heads or tails, while a qubit is a spinning coin that may land on either side. Extending the latter part of the analogy, consider how the face and tail of the spinning coin are pointing in horizontal directions. Therefore, there is some probability that the coin will fall face up and some probability it will fall face down, but imagine that there is also information in which horizontal direction the coin falls in.
The term “qubit” is not exclusive to quantum computing, and should not be mistaken for the “qubit” in the terms qubit fluorometer and qubit protocol.
While a bit is a transistor, a quantum computing qubit can currently take many forms. The most popular modality is superconducting qubits, which are electronic circuits that are cryogenically frozen to near Absolute Zero. Their popularity is due, at least in part, to the manufacture of superconducting qubits leveraging existing technologies, in particular semiconductor fabrication. Because of their behavior at such extremely low temperatures, they are sometimes characterized as “artificial atoms,” which is particularly ironic considering that another popular modality is actual atoms. In fact, neutral atoms and ionized atoms are both being researched and developed. Other well-known modalities include electron spins, photons, and nitrogen vacancies in diamonds.
It’s worth noting that all of these technologies are in early development stages. They each have different advantages and disadvantages in terms of error rates and scalability. They may also each have different applications that they are particularly well suited for, such as in the way the Rydberg states of neutral atoms are a natural fit for Maximum Independent Set (MIS) and related problems. It is too early to tell if the qubits of the future will be one of these existing modalities, something completely novel, or some combination of current and novel modalities.
Qubits are most often referenced in regard to quantum computation and quantum information, but they also play key roles in securing communication. Their four most commonly cited uses include:
Although quantum computing gained notoriety for threatening public cryptosystems, the bigger picture reveals that qubits can also provide security for communications. Although beyond the scope of this article, qubits are also used in quantum sensing, leveraging their sensitivity to make very small measurements.
The future involves using many physical qubits to provide error correction long enough to complete fault-tolerant computation. Many physical qubits are needed to correct what are called “bit-flip errors,” with which a measurement that should be 0 is incorrectly 1, and vice versa. Many more physical qubits are also needed to correct what are called “phase-flip errors,” which you can imagine as the face of a spinning coin pointing in the direction the tail should be pointing in. Together, these physical qubits form one “logical qubit,” and logical qubits are presumed to be necessary to achieve quantum utility. Some estimates are as high as 1,000 physical qubits will be needed for each logical qubit; for scale, there are zero quantum computers today with that many physical qubits. The closest is Aquila, which has 256 neutral atoms available via the cloud, but can fit 1,000 atoms in its vacuum chamber trap.
In addition to the modalities already mentioned, yet more are being researched. Azure Quantum’s “What is a qubit?” article, for example, mentions Microsoft’s research into topological qubits, which promise to be robust against noise and errors.
Some of the existing modalities are also making great progress. Thanks to a process called qubit shuttling, for example, neutral atoms can be moved around in a two-dimensional array, enabling something called “all-to-all connectivity.” The short of it is that any atom can be entangled with any other atom, groups of atoms can be set aside as some kind of memory, and other atoms can be set aside to be measured.
Neutral atoms have also demonstrated scalability, which is critical considering the sort of applications the world is waiting to run. Aquila’s 256 atoms dwarf all publicly-accessible ion trap, electron spin, nitrogen vacancy, and general-purpose photonic quantum computers.