Two Level System Defects Superconducting Qubits

Short answer: Two-level system (TLS) defects are atomic-scale imperfections in the materials around superconducting qubits that absorb energy and make qubit parameters fluctuate. Papers describe techniques to find and characterize individual defects, report that interface defects caused about 60 percent of the dielectric loss in one transmon sample, and propose a new TLS type formed by trapped quasiparticles. This is materials research, and it explains why fabrication matters for qubit coherence.

What TLS defects are

A 2015 Nature Communications paper says parasitic two-level tunnelling systems from structural material defects act as a source of noise in microfabricated devices, and superconducting qubits can be sensitive to even single defects residing in the tunnel barrier of their Josephson junctions. It says the solid-state approach carries the burden that the device material may host defects that give rise to two-level systems acting as a sparse decohering bath, even while superconducting qubits have shown excellent gate fidelities and long coherence times. A 2019 paper in npj Quantum Information likewise says superconducting qubits suffer strongly from energy dissipation and environmental fluctuations caused by atomic-scale defects in device materials, and that progress towards larger processors needs better fabrication guided by analysis of how defects form.

Studying single defects

The 2015 paper detects and fully characterizes a system of two strongly interacting defects using a high-resolution spectroscopy technique. It says mutual defect coupling had been conjectured to explain anomalies of glasses and suggested as the origin of low-frequency noise in superconducting devices, and that its study provides conclusive evidence of defect interactions. All its observations are consistent with defects generated by atomic tunnelling.

The 2019 paper presents a technique that tunes individual defects with applied electric fields, giving a spectroscopy method to extract each defect's energy distribution, electric dipole moments and coherence times. It also distinguishes defects in Josephson junction tunnel barriers from those at circuit interfaces. In the transmon sample investigated, defects at circuit interfaces were responsible for about 60 percent of the dielectric loss. About 40 percent of all detected defects were in the tunnel barriers of large-area parasitic Josephson junctions that arise collaterally in shadow evaporation, and only about 3 percent were strongly coupled defects presumably in the small-area qubit junctions. The authors say the technique can be applied to standard qubit samples. The percentages describe one sample.

Where the defects sit

A Science Advances paper says the most coherence-limiting TLS increasingly appear to reside outside the qubit junctions, on the surfaces of metals and dielectrics around them, and that their slowly fluctuating dynamics poses a serious challenge for quantum computation. It notes that charged surface TLS and paramagnetic impurities also create a stochastic backdrop for quasiparticles in the superconductor.

A newly proposed kind

The same paper reports a decoherence mechanism it calls previously unexplored: a new type of TLS originating from trapped quasiparticles, which can induce qubit relaxation. Using mapping of TLS-induced fluctuations in frequency-tunable resonators, the authors identify a highly coherent subset of the general TLS population with a low reconfiguration temperature of about 300 mK and a nonuniform density of states. They say these properties can be understood if the TLS are quasiparticles trapped in shallow subgap states formed by spatial fluctuations of the superconducting order parameter. They conclude that even very rare quasiparticle bursts would affect coherence over exponentially long time scales. That is the authors' interpretation of their data.

What this does not tell you

These papers study particular samples and structures, and the percentages are not universal. They do not offer a recipe that removes TLS losses, only methods to measure them and hypotheses about their origin.

FAQ

What is a TLS defect?

A material defect, often modelled as atomic tunnelling between two states, that acts as a noise source for superconducting qubits.

Where do the losses come from?

In one transmon sample, interface defects accounted for about 60 percent of dielectric loss, per the 2019 paper.

Can single defects be measured?

Yes. The 2015 and 2019 papers describe spectroscopy of individual defects, including tuning them with electric fields.

What is the new TLS type?

Quasiparticles trapped in shallow subgap states, which the Science Advances authors propose can cause relaxation.

Why does it matter?

The papers say defects limit coherence and that fabrication improvements need to be guided by understanding them.