A qubit is the quantum analogue of the bit, and is the fundamental unit of quantum information. The physical representation of the bit is typically a transistor, although some variety exists in the implementations. The physical representation of a qubit is much more varied than for a bit, with several major modalities currently in development, more modalities being researched, and most of the modalities having variations.
One of the major modalities is the silicon spin qubit. While implementations vary, the fundamental concept is the encoding of quantum information into the intrinsic spins of electrons. Qubits are two-level systems, which we represent as a 0 or a 1, or as a superposition of both. Electron spins are spin up or spin down, which makes them natural two-level systems to serve as qubits.
There are qubits that encode quantum information within nuclear spins, which are the intrinsic spins of atomic nuclei. Those are different. For more information on electron spin qubits, be sure to check out the following:
- “Quantum Computing Breakthrough: Silicon Encoded Spin Qubits Achieve Universality” by SciTechDaily includes a 3-minute video explainer.
- “Quantum Computing Milestone: Researchers Compute With ‘Hot’ Silicon Qubits” by IEEE Spectrum discusses achievements using relatively high temperatures.
- “Semiconductor Spin Qubits” by a team from the University of Konstanz, HRL Laboratories LLC, the University of Rochester, and Princeton University is a 63-page paper.
What are Silicon Spin Qubits?
Despite the name “silicon spin,” spin qubits have been implemented in other semiconductors, specifically gallium arsenide and germanium. A fourth substrate they’ve been implemented in is graphene. Again, implementations vary, but typically these substrates have microscopic wells known as quantum dots, and then above each well is a transistor’s gate electrode. The spins of one or more electrons in the well are manipulated by applying voltages to the gate of the transistor.
Silicon Spin Qubit Systems
Coincidental to there being four major substrates, there are also four classifications of spin qubits. These classifications are:
- Single spin qubits encode quantum information within the intrinsic spins of electrons, and perform computation by manipulating those spins.
- Instead of quantum dots, donor spin qubits use the valence electrons of atoms in the substrate; the atomic nuclei can be thought of as the well.
- Singlet-triplet spin qubits use two electrons, and quantum information is encoded in whether their spins are opposite or aligned.
- Exchange-only spin qubits use three electrons and execute gate operations with exchange interactions, potentially making them less prone to errors.
Research continues into this modality. One novel approach is to hybridize a spin qubit quantum computer, which uses semiconductor technology, with a superconducting quantum computer.
Advantages of Silicon Spin Qubits
Silicon spin qubits offer several advantages over some of the other qubit modalities:
- There has been 26 years of research and development since they were first proposed.
- Coherence times are relatively long.
- The use of semiconductor fabrication technology makes them relatively inexpensive.
- Their size might make it possible to have huge numbers of qubits on a single chip.
- They operate at relatively high temperatures, which lowers operating costs.
- The substrate provides some protection from environmental noise.
- Fast gate speeds allow more gates to be executed within coherence times.
Spin qubits also boast greater control than some other qubits. This is due to how the individual electrons can be isolated from each other and shielded by the substrate from the environment.
Challenges in Silicon Spin Qubit Development
Despite the numerous aforementioned advantages, silicon spin qubits are not without their development challenges. Some of the most pressing challenges include:
- Charge noise, which is caused by electrostatic fluctuations
- Valley splitting, when energy levels in the quantum dot are too close together
- Spatial variations in electron g-factor, which reduce control over the electrons
- Fabrication, which can result in inconsistencies, aka defects
- Instability across ranges of temperatures
- Spatial noise correlations, when qubits are affected by the noise affecting other qubits
- Integration with classical control electronics is challenging
Other challenges are comparable to the general challenges facing most qubit modalities on the path to achieving large-scale fault-tolerant quantum computing (FTQC). Perhaps the greatest challenge, however, is that there are currently no publicly available spin qubit quantum computers. Public usage has accelerated the development of neutral atom quantum computers, as well as other modalities.