1051-8223 (c) 2021 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information. This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI 10.1109/TASC.2021.3069732, IEEE Transactions on Applied Superconductivity ASC20-WK2EOR2B-08 1 Thermal Kinetic Inductance Detectors Camera: System Level Design, Strategy and Performance Forecast L. Minutolo, A. Wandui, J. Bock, C. Frez, B. Steinbach, A. Turner, and R. O’Brient Abstract—The next-generation instruments for millimeter- wave astronomy will require large arrays of detectors to sur- pass current performance [1], [2]. On account of integration challenges, focal planes relying on transition edge sensor (TES) bolometers will not scale gracefully. Thermal Kinetic Inductance Detectors (TKIDs) are highly scalable and have the design flexi- bility to solve this problem. We present an update on our detector development and systems engineering in preparation for a 2022 TKIDs camera deployment that includes noise performance, optical characterization, and a calibration strategy. Index Terms—KIDs, MKIDs, TKIDs, Observational Cosmol- ogy, Readout, RF I. I NTRODUCTION B ICEP array [3] consists of four interchangeable refractive telescopes operating on a common mount [4] situated at the Amundsen-Scott South Pole Station. The BICEP program’s goal is to constrain the amplitude of primordial gravitational waves by mapping the polarization of the cosmic microwave background (CMB) [5]. BICEP Array will map the same field on the sky at 30/40 [6], 95, 150, and 220/270 GHz to subtract the foreground signals from the B-mode polarization pattern of the CMB. We use TES bolometers to measure frequencies as high as 150GHz [7] because they are reliable and provide background limited performance. However, the TES bolometers require at least one set of wire-bonds between each detector. Its readout SQUID (superconducting quantum interference devices) is arduous and prone to handling failure at shorter wavelengths with larger numbers of detectors in the focal plane. Implementing and maintaining these connections is a significant point of failure in the telescope. To over- come this scalability challenge, we are retrofitting one of our telescopes to use a focal plane of thermal kinetic inductance detectors (TKIDs) in lieu of TES bolometers. These detectors are RF multiplexed by passive circuits lithographically defined and integrated onto the detector array, alleviating integration JPL RTD program (2016-19) Strategic support for TKIDs & readout; President Director Funds, 2018-2020; NASA SAT: started 2020 (JPL); Moore Foundation: started 2019 (Caltech) (Corresponding author: Lorenzo Minutolo) L. Minutolo, B. Steinbach, J. Bock and A. Wandui are with the Cali- fornia Institute of Technology, Pasadena, CA 91125, USA (email: minu- tolo@caltech.edu; awandui@caltech.edu; bsteinba@caltech.edu) A. Turner, C. Frez and R. O’Brient are with the Jet Propulsion Labo- ratory, Pasadena, CA 91109, USA (email: anthony.d.turner@jpl.nasa.gov; clifford.f.frez@jpl.nasa.gov; roger.o.brient@jpl.nasa.gov ) Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier will be inserted here upon acceptance. concerns. We will deploy a pathfinder TKIDs camera to the South Pole Station in the 2022 Austral Winter to observe at 150GHz, to facilitate comparisons with the prior generation TES bolometer cameras that our group has extensive experi- ence with and data from. II. TKIDS TECHNOLOGY Thermal kinetic inductance detectors [8], [9] differ from standard KIDs because the inductive element is not designed to absorb photons directly. Instead, an antenna couples the photons to a dissipative gold film on a suspended bolometer island (Fig. 1B), where the inductive element then detects thermalized phonons. The inductor is part of a high Q su- perconducting resonator, and when phonons on the island break Cooper pairs in the inductor, they shift the resonant frequency. A thermal link, formed by the supporting legs, restores the original island temperature. The legs’ thermal conductance (reported as G T in Fig. 2) determines the time constant to restore the detector’s idle condition. This detection enjoys multiple advantages over other technologies for our application. In contrast to KIDs, the inductor and absorber in TKIDs are distinct components that can be independently op- timized. Moreover, in contrast to TESes, the natural frequency multiplexing minimizes the number of wirebonds to connect to a single RF line per tile [10], [11]. Moreover, the cold electronics are passive - transmission lines and a low noise amplifier (LNA) [12], [13] with none of the active feedback required for readout through SQUIDs [14]. Our test camera will consist of four detector tiles, with 128 antenna-coupled detectors. The focal plane requires one LNA per tile- four total- which is substantially cheaper and easier to operate than the SQUID based multiplexing used for TES bolometer arrays. III. DEVICE CALIBRATION In addition to the gold antenna termination, we implemented another gold resistor on the island as a calibration heater or calibrator. We can directly bias this resistor with a room temperature current source to precisely mimic the antenna’s optical power level P opt ∼ 5pW . Fig. 3 well summarizes the calibration process. By knowing the power dissipated by the calibration heater and measuring the respective resonator properties, we can calibrate the non-linear resonant shift per pW of power in our detector. To limit the number of wires, we connect multiple detectors’ calibrators in series with a shared Authorized licensed use limited to: CALIFORNIA INSTITUTE OF TECHNOLOGY. Downloaded on April 19,2021 at 18:29:38 UTC from IEEE Xplore. Restrictions apply.