Scaling Kinetic Inductance Detectors (KIDs), 15-R8311
Peter W. A. Roming
Inclusive Dates: 05/07/13 – 09/07/13
Background — For approximately the last three decades, charged coupled devices (CCDs) and hybrid complementary metal-oxide-semiconductor (CMOS) detectors have dominated the field of optical and IR imaging. Despite their dominance, they lack simultaneous timing and spectral information, and only have good quantum efficiencies over a narrow wavelength range. The development of superconducting tunnel junctions (STJs) and transition edge sensors (TESs) has alleviated these shortcomings. However, STJs and TESs suffer from the major challenge of constructing large format arrays, which are required for most imaging applications. Kinetic inductance detectors (KIDs) are a relatively new alternative superconducting technology that have many of the same desirable characteristics of STJs and TESs, but have great promise for creating large-scale formats like current CCDs and CMOS detectors.
The largest current working KID-based instrument has an ~2,000 pixel array, a far cry from the needed 64 kpixel to 4 Mpixel KID-based arrays. Multiplexing such large arrays has been challenging. Because of SwRI's experience with several detector technologies, extensive electronics background, and radio frequency expertise, SwRI is positioned for making large format KID arrays a viable alternative to CCDs and CMOS detectors in such areas as precision scientific, military, and medical imaging.
The current approach for multiplexing arrays relies on detecting changes in resonance by injecting a signal with multiple frequency components as the source signal. The frequency components within the source signal are designed to match the resonant frequency of each sensor element. If the array has N elements, then N frequencies are required to stimulate all of the elements in the array. The output of the array is a signal with multiple frequency components, ideally one for each sensor element. The spectral response of the array is measured with no photons present. When photons are absorbed by an element, its resonant frequency will change, and detecting the change in frequency indicates detection of photon absorption.
While this method has been proven to enable detection of photons, it has two key drawbacks that limit its utility in remote and in size-constrained applications. The first is in device characterization. Each element is designed with a different resonant frequency, but because of manufacturing tolerances and variation due to temperature, the exact frequency is unknown. Each element of the array needs to be individually characterized to identify its resonant frequency under precisely controlled conditions. Then, combining all resonant frequencies generates a single source signal. A second, even greater challenge is that the resonant frequency of each element changes with micro-Kelvin changes in temperature. This necessitates in-system re-calibration of the resonant frequency of each element with and without photon absorption. After re-calibration, the electronics that generate the source signal need to be modified with the new set of frequencies, as do the electronics and signal processing for sensing the changes in resonant frequency due to photon absorption. The large number of elements in arrays required for many sensing applications exacerbates the problem.
Approach — The multi-frequency source signal is replaced with white noise, a signal that contains equal power within every frequency across the band of interest. The constant power spectral density enables every element in the array to resonate, as long as its resonant frequency is within the band of interest. Measurement of the array output is very similar to that of the multi-frequency source. The white noise causes each sensor element to resonate, and the spectral response of the array will contain a notch at the resonant frequency of each sensor element. Instead of needing to calibrate each element to determine its resonant frequency, a measurement of the response is taken during dark conditions, and only changes in resonances need to be detected for detection of photon absorption.
Accomplishments — The basic design of the output electronics for detecting frequency shifts has been completed.