Microwave Kinetic Inductance Detectors (MKIDs) have great potential for large very sensitive detector arrays for use in, for example, sub-mm imaging. Being intrinsically readout in the frequency domain, they are particularly suited for frequency domain multiplexing allowing $\sim$1000s of devices to be readout with one pair of coaxial cables. However, this moves the complexity of the detector from the cryogenics to the warm electronics. We present here the concept and experimental demonstration of the use of Fast Fourier Transform Spectrometer (FFTS) readout, showing no deterioration of the noise performance compared to low noise analog mixing while allowing high multiplexing ratios.
Deep Dive into FFTS readout for large arrays of Microwave Kinetic Inductance Detectors.
Microwave Kinetic Inductance Detectors (MKIDs) have great potential for large very sensitive detector arrays for use in, for example, sub-mm imaging. Being intrinsically readout in the frequency domain, they are particularly suited for frequency domain multiplexing allowing $\sim$1000s of devices to be readout with one pair of coaxial cables. However, this moves the complexity of the detector from the cryogenics to the warm electronics. We present here the concept and experimental demonstration of the use of Fast Fourier Transform Spectrometer (FFTS) readout, showing no deterioration of the noise performance compared to low noise analog mixing while allowing high multiplexing ratios.
The Far-Infrared (FIR) and sub-mm wavelength bands (10 -1000 µm) contain a plethora of information on the evolution and formation of galaxies, stars and planetary systems. Future FIR and sub-mm space missions, such as SPICA [1] and Millimetron [2] require large arrays of detectors with an unprecedented high sensitivity, expressed in a Noise Equivalent Power (NEP) < 2×10 -19 W/Hz 1/2 . Also, many ground based telescopes like IRAM [3], APEX [4] and future observatories like CCAT [5] require very large arrays for faster observation speeds and larger instantaneous fields-of-view.
Microwave Kinetic Inductance Detectors MKIDs [6] are a rapidly developing new detector technology [7,8,9] that has the potential to be used to fabricate very large imaging arrays for the FIR, sub-mm [10], optical and x-ray [11]. The main advantages of MKIDs over other detector technologies are: i) ease of fabrication, ii) the intrinsic adaption for frequency domain multiplexing (FD-MUX) using microwave readout signal frequencies and iii) operation with a wide dynamic range. Currently MKIDs have shown sufficient sensitivities in a lab environment for both ground-based and space-based instruments [12] and there has been an initial demonstration at the CSO telescope on Mauna Kea, Hawaii [10].
However, FDMUX of MKIDs has yet to be satisfactorily demonstrated, particularly anywhere near the (power and weight) requirements needed for space. Previous measurements have been non-scalable using either analog mixing to readout one MKID, which has proven useful in single detector characterization in a laboratory environment, or digital mixing using specialist chips [13]. Using the latter technique a readout of 16 MKIDs has been demonstrated using one commercial demultiplexer card. However, the technique is hard to scale to very large pixel numbers [13]. In this Letter we present a new MKID readout scheme based upon frequency division multiplex-ing and a digital Fast Fourier Transform Spectrometer (FFTS) [14,15]. Digital FFTS systems are presently being used as back-end electronics for heterodyne mixers on several ground based telescopes [15]. They can process a real time data stream with a bandwidth of up to 1.8 GHz into a frequency spectrum of ∼ 8000 points. We propose here to use the FFTS to read out a set of single frequency probe signals that have passed through a chip containing one MKID detector per probe signal. The proposed readout is able to read out ∼ 1000 pixels using 1 probe signal generator and 1 FFTS board. To demonstrate the principle, we describe an experiment where we have read out 8 MKIDs simultaneously and we have measured both the dark detector NEP and the response to an optical signal for 8 MKIDS. These results are compared to a conventional analog single MKID readout scheme and show that the FFTS based multiplexed readout does not deteriorate the intrinsic system noise of our setup, which is limited by the first stage cryogenic amplifier.
MKIDs are superconducting pair breaking detectors that sense the change in the complex surface impedance of a thin superconducting film due to radiation absorption with a (sky) frequency F rad > 2∆/h, which is ∼ 80 GHz for aluminum. An MKID consists of a thin superconducting film that is incorporated in a resonance circuit which is either capacitively [6] or inductively [17] coupled to a through line. Changes in the surface impedance of the film are converted to changes in resonator quality factor and resonance frequency. These changes can be read out by measuring the phase and amplitude modulation of a probe tone at a probe frequency equal or close to the resonator resonance frequency F 0 , which is typically a few GHz.
MKIDs take advantage of the fact that a superconductor at Tc/10 has negligible losses at the probe frequency f 0 . This enables very low 3 dB bandwidth δf resonators with high Q factors f 0 /δf ∼ 10 6 . Hence, close packing of the resonators in (readout) frequency space is only limited by manufacturing tolerances (∼1 MHz at f 0 ∼ 3 GHz). For example, using conventional one oc- tave bandwidth amplifier (4-8 GHz) one can read out ∼ 4000 MKIDs using just one pair of coaxial cables.
To demonstrate the FFTS based readout of MKIDs we use the readout scheme shown in Fig. 1 to readout 8 MKIDs. A set of 8 Intermediate Frequency (IF) probe tones in a band from 0.1 -300 MHz are generated by a 1 GSPS (1 × 10 9 samples per second) commercial arbitrary waveform generator (AWG) with a 12-bit Digitalto-Analog Converter (DAC). Subsequentially they are mixed to the required readout at Radio Frequency (RF) of 4 -4.3 GHz by a commercial single sideband upconverter using an Local Oscillator (LO) frequency of 4 GHz from SG1 (example tones shown in the inset Fig. 1). The resulting RF probe signal consists of a set of RF frequency probe tones, where each tone corresponds to one MKID while additional blind tones can be added to measure the system noise. The LO of the mixer is a low noise synthe
…(Full text truncated)…
This content is AI-processed based on ArXiv data.