By Dr. Andre Saraiva, UNSW

In the war against quantum errors, it is important to pick your battles. Noise can cause all sorts of complicated multiqubit errors, but it is often the mundane phase flip of a single qubit that causes most trouble. This means that locally the qubits can flip erratically, but the collective properties of a large group of qubits remain intact. That is the idea behind quantum error correction – encoding logical qubits in many physical qubits, creating redundancy in the information.

The group of Professor Michel Devoret, from Yale University, took a completely different direction, experimentally showing a new class of error protected qubits that were theorised more than two decades ago (see here and here). These qubits consist not of a single elementary particle, but of a large group of particles collectively put in a superposition state. Hence the name “cat state” – these ensembles of particles can be so large that they are macroscopically distinguishable, but still follow quantum mechanical rules, in close analogy to Schrödinger’s proverbial cat.

Unlike quantum error correcting codes, this incarnation of error protection is implemented in a single device: a microwave antenna cavity of similar scale and fabrication demands as any superconducting qubit. In these systems, several quanta of microwave radiation, or photons, can be excited and collectively set to a superposition state. The challenge lies in getting the photons to interact with each other – we know that light beams cross each other without colliding, and the same occurs at the quantum level.

The group used a material that responds to the electromagnetic field in a way that effectively creates an interaction between photons, known among experts as a nonlinear Kerr medium. Once these Kerr-cat states are produced, they are not yet qubits. It is necessary to be able to create different superpositions of these collective photon states, called coherent states. This needs to be done fast enough to avoid the spontaneous decay of the qubit states.

The initial results described in the paper from Grimm et al are promising. While the single qubit gate fidelities are somewhat limited (a bit short of 86%, well below any fault-tolerance threshold), the qubits did show an improvement of a factor of 30 in phase flip error rates compared to their non-cat state counterparts (Fock qubits made out of a single photon). This improvement was limited by the fact that the researchers did not observe the expected improvement of the protection with larger number of photons – they stopped at an average of 2.6 photons. This limitation was associated with heating of the device at higher field intensities, which would compromise the stability of the two-level system, allowing it to escape into the large dimension of the Hilbert space of these quantum oscillators.

So, what’s next?

Before this technology can make a true impact on the competitive landscape of modern quantum computing, significant improvements will be required in the fidelity of the single Kerr-cat qubit manipulations. Being the same size as a superconducting device only translates into an advantage if it is shown that they outperform superconducting qubits.

For additional information, click on the link to view an abstract of the paper titled “Stablization and operation of a Kerr-cat qubit”.

Dr. Saraiva has worked for over a decade providing theoretical solutions to problems in silicon spin quantum computation, as well as other quantum technologies. He works on MOS Silicon Quantum Dot research and commercially-oriented projects at the University of New South Wales (UNSW).

September 6, 2020