Optimizing a Transition Edge Sensor detector system for low flux infrared photon measurements at the ALPS II experiment

Abstract

This thesis investigates the optimization of a Transition-Edge Sensor (TES) for the Any Light Particle Search II (ALPS II) experiment. The ALPS II experiment searches for axions and axion-like particles, which are predicted by several beyond-standard model theories and are good candidates for dark matter. The TES is a superconducting detector sensitive to temperature-induced resistance changes. For ALPS II, the TES must be able to detect low-energy photons (1.165 eV) with high efficiency, good energy resolution and extremely low background rates in the order of one 1064 nm photon-like event every two days. Towards the goal of reducing the background rate, this thesis focuses on the understanding of the background and the optimization of the data analysis in an effort to reject them. A two-fold approach was developed to simulate the background observed when the TES is decoupled from the optical fiber used in the TES characterization. On the one hand, a Geant4-based framework was built to compute the energy deposition produced by muons from cosmic rays and radioactivity from the zirconia fiber sleeve used to couple the detector to the experimental setup. On the other hand, a COMSOL Multiphysics simulation was designed to model the TES physics. This was used to assess the effects of the energy depositions on the TES. The radioactivity in the zirconia was identified as the dominant intrinsic background source. When the optical fiber is coupled to the TES, an additional background component appears with respect to the aforementioned, predicted to be composed of Black Body Radiation (BBR). A framework was developed to simulate the BBR and compute the expected background rates in the TES. The uncertainties in the simulation results were mainly given by the bending of the optical fiber used to transmit the light to the TES. However, the importance of fiber bending for the rejection of low-energy BBR photons was demonstrated. Furthermore, this simulation demonstrated that BBR-induced background could be mitigated by enhancing the TES energy resolution. Consequently, another simulation tool was developed in order to understand the effect of the baseline noise on the TES energy resolution. The TES signal was modeled as the superposition of a non-distorted 1064 nm pulse and the baseline noise. Using this method, the obtained energy resolution was equal to the one computed from the measured data. This demonstrated that the energy resolution in the TES system is mainly governed by the electronic noise and not by variations in the photon absorption process in the TES. Finally, a frequency domain-based analysis was implemented to optimize the analysis and improve the energy resolution. Combining this analysis with the choice of the pulse height to describe the energy of the photon incident in the TES allowed an overall enhancement of the energy resolution by a factor of 2. Additionally, a pulse-finding algorithm based on signal deconvolution was implemented to determine the arrival times of pulses and the identification of pile-up. This algorithm improved fitting procedure stability and speed and enabled the rejection of false triggers and the identification of pulses with energies as low as 0.3 eV. The joint analysis optimization reduced the measured extrinsic background rates in the 1064 nm photon signal region by one order of magnitude.