Surface Spins and 1/f Flux Noise in SQUIDs and Qubits

Short answer: Magnetic flux noise, with a power spectrum that varies roughly as 1/f, limits superconducting quantum interference devices (SQUIDs) and flux-sensitive superconducting qubits. A 2016 paper reported that adsorbed molecular oxygen is the dominant source of magnetism in the superconducting films it studied, and that surface treatment cut the noise power by up to a factor of 5. A thesis abstract describes measurements and models of the same noise.

What is flux noise?

The 2016 Physical Review Applied abstract says magnetic flux noise is a dominant source of dephasing and energy relaxation in superconducting qubits. The noise power spectral density varies with frequency as 1/f to the power alpha, with alpha at most about 1, and spans 13 orders of magnitude. The authors say recent work indicates the noise comes from unpaired magnetic defects on the surfaces of superconducting devices.

A doctoral thesis abstract hosted on an eScholarship repository gives the longer history: for over three decades flux noise with a spectrum scaling roughly as 1/f^alpha has been known to limit the low-frequency performance of dc SQUIDs, and experiments indicate the same noise persists up to frequencies of 1 GHz and is a dominant source of dephasing in flux-sensitive qubits. The thesis calls its reduction a major hurdle for scalable computers based on flux qubits.

What did the 2016 paper find?

The authors demonstrate that adsorbed molecular oxygen, O2, is the dominant contributor to magnetism in superconducting thin films. They show this magnetism can be reduced by suitable surface treatment or by improving the vacuum environment of the sample. They observe a suppression of static spin susceptibility by more than an order of magnitude and a suppression of 1/f magnetic flux noise power spectral density by up to a factor of 5. The authors say these advances open the door to superconducting qubits with improved coherence, which is their stated prospect and not a demonstrated qubit result.

What does the thesis work cover?

The thesis abstract describes experimental measurements, theoretical analyses and numerical simulations meant to give a more detailed picture of microscopic and macroscopic properties of flux noise. It reports flux noise measurements on ten SQUIDs with systematically varied geometries, in the temperature range from 0.1 K to 4 K, and states that alpha increases as temperature is lowered, with each spectrum pivoting about a nearly constant frequency. The mean square flux noise grows rapidly with temperature and, at a given temperature, is roughly independent of the outer dimension of the SQUID washer. The authors say these results are incompatible with a model based on random reversal of independent spins. The rest of the abstract was cut off in the page we read, so this post does not describe it. A second listed source, a University of Wisconsin dissertation from 2013 on 1/f flux noise, showed only its title page.

How to read this

Flux noise has several proposed microscopic origins, and these sources show evidence about particular materials and devices, not a single settled explanation. The factor-of-5 figure applies to the films and treatments in that paper. This post is based on abstracts and visible summary text and has not checked full texts. It makes no claim about products.

FAQ

What does 1/f mean here?

The noise power falls with frequency roughly as one over frequency to the power alpha, with alpha at most about 1, according to the 2016 abstract.

What caused the magnetism in the 2016 films?

The authors report adsorbed molecular oxygen as the dominant contributor.

How much did flux noise drop?

Up to a factor of 5 in power spectral density, with a susceptibility drop of more than an order of magnitude.

Is the origin of flux noise settled?

These abstracts do not say so. The 2016 paper points to surface magnetic defects, and the thesis tests models against SQUID data.