Skip to main content

Theory for superconducting few-photon detectors

Time: Fri 2022-06-10 14.00

Location: FB42, Roslagstullsbacken 21, Stockholm

Video link:

Subject area: Physics

Doctoral student: Mattias Jönsson , Kondenserade materiens teori

Opponent: Karl Berggren, Massachusetts Institute of Technology

Supervisor: Jack Lidmar, Kondenserade materiens teori; Mats Wallin, Kondenserade materiens teori; Marcin Swillo, Kvant- och biofotonik


High-performance photon detectors are essential for fiber communication, which is the foundation of the modern internet. Emerging quantum information technologies, such as quantum key distribution, impose new requirements on the photon detectors used. Superconducting single-photon detectors (SSPDs) exhibit high efficiencies, low dark count rates and fast recovery times, which makes them commonly used for quantum applications.

The ability to resolve photon numbers in a wave packet is useful in applications like imaging, characterization of light sources, and in optical quantum computation. Ordinary single-photon detectors like SSPDs are not photon-number resolving (PNR), and are only capable of determining if light is present or not. However, photon-number resolution may be achieved by combining multiple single-photon detectors in an array and split the input over them.

In this thesis, we introduce and model PNR detectors based on multiplexing single-photon detectors. Using these models, we investigate the requirements on the single-photon detectors when they are used in a multiplexed scheme and we investigate how a PNR detector may be used in imaging applications. We experimentally realize a temporally multiplexed PNR detector based on SSPDs and show that it is capable of accurately determining the mean photon number for a series of wave packets.

We also model a SSPD using the generalized time-dependent Ginzburg-Landau model to investigate how the geometry of the SSPD affects the performance of the detector. We show that the geometric reduction of the critical current in turnarounds is less pronounced than previously reported, which relaxes design restrictions.