Quantum computing has gained significant traction in recent years and number of companies now provide access to their state of the art quantum computing platforms through the cloud.
The most popular hardware implementations have, so far, been based on superconducting qubits or trapped-ion qubits, with photonic based qubits following closely behind. Up to now, silicon based qubits have lagged behind in popularity.
Recently, a number of groups have demonstrated that silicon qubits can perform logic operations with >99% accuracy (or fidelity). In this blog, we briefly discuss why reaching this milestone matters.
Noise is present in all computing systems, whether classical or quantum. In classical computing, a bit may have a value of, say, “0” but as the bit is processed in the system, it may become corrupted by noise and instead appear as “1”. A solution with classical computing is to use repetition or multiple copies of the bit to reduce the incidence of errors. However, such solutions are not applicable to qubits because duplicating quantum states is not possible (this is known as the no cloning theorem). Other ways of correcting errors in qubits do exist (these will not be discussed here) but generally, achieving a gate fidelity (or accuracy) of 99% is regarded as the minimum performance required to implement such quantum error correction codes in practice.
The demonstration of >99% accuracy in silicon qubits matters exactly for this reason; the >99% accuracy makes silicon qubits a viable candidate for implementing a practical fault-tolerant quantum computer.
Silicon based qubits are also likely to be scalable (and scaling the number of qubits is key to unlocking useful applications for quantum computing). This is because a silicon based architecture will benefit from the engineering knowledge developed over several decades and at huge costs by the semiconductor industry. Already, transistors a few atoms long can be made precisely, donor atoms can be implanted in silicon with nanometre scale precision (which is how the some silicon qubits are implemented), while wiring and readout electronics may also be positioned precisely in relation to the qubit. While further improvements to the manufacturing process will be needed to make silicon qubits practical, one would bet that the semiconductor industry, with its adherence to Moore’s law, will deliver.
At that point, silicon qubits should become the new qubit on the block!
