mm Universe @ NIKA2
configuration [2, 3]. The case of ground-based experiments has an additional requirement: the atmospheric fluctuations have to be addressed. For this purpose, FTSs have to be coupled with fast detectors, and kinetic inductance detectors (KIDs) are the fastest available in large format arrays at mm-wavelength [4, 5].
In this scientific framework, we have developed the KIDs Interferometer Spectrum Survey (KISS), which uses two arrays of KIDs coupled to an MPI. The reference source of such a differential FTS configuration can be set between the image of the de-focused sky and a cryogenics stage, see [3] for more details. KISS allows us to exploit a wide instantaneous field of view (1 deg) and a spectral resolution as fine as 1.45 GHz in the 120–180 GHz frequency band. The instrument is installed on the 2.25-meter Q-U-I JOint TEnerife (QUIJOTE) telescope at the Teide Observatory, in Tenerife and has been operational since February 2019. KISS is also the pathfinder of the new CarbON CII line in the post-rEionization and ReionizaTiOn epoch project (CONCERTO) [6].
In this paper, we give an overall description of the spectral mapping paradigm and we present recent results from the last year of observations. In Section 2, we describe the necessity for fast scanning strategy and we explain the spectral mapping technique. Section 3 presents the preliminary on-sky results with point sources in spectral mapping reduction and we address the data reduction of the KISS project.
2 Spectral mapping with FTS fast scanning
Ground-based experiments have to deal with a particular additional issue with respect to space-borne telescopes. Despite the ease to debug and modify the hardware of telescopes on the ground, even after the deployment, an additional source of contamination is present: the atmosphere, both in terms of absorption and emission. This limits the frequency range over which they can be used. The emission has two main effects: 1) it increases the background (photon) noise with respect to what is expected for satellite experiments, and 2) leads to low-frequency noise induced by low-frequency drifts in the atmospheric emission. For the first, it requires an increase of the necessary integration time for a given signal-to-noise ratio. However, this requirement is compensated by the lower cost of a ground-based facility. For reducing the latter, there are specific observation strategies for different instruments that prevent the introduction of systematic errors in the signal. This problem has to be carefully addressed to avoid affecting the accuracy of the calibration. In the case of FTS, the solution is to acquire the full interference pattern, the so-called “interferogram”, on times scales shorter than those of the atmospheric fluctuations.
State-of-the-art detectors in the mm domain reach a white noise level dominated component is the photon noise, which is the intrinsic noise derived by the random nature of the incoming photons. The best noise performance is matched if this condition is verified and the detectors are named photon-noise limited. The second higher noise component is produced predominantly by the electronic noise. In addition, there is the noise coming from the state fluctuation of the superconducting particles, called generation-recombination noise, in the KID case [7]. The slow fluctuations of the atmosphere are translated to a 1/f noise in the frequency domain, whose intersection with the white noise forms a knee shape. The frequency of the intersection, the knee frequency, is particularly important for ground-based experiments. The recording of every single interferogram at stable atmospheric absorption (in other words, faster than the 1/f knee) is a prior requirement, in the specific KISS case. We use the so-called “fast scanning” technique for interferometric pattern sampling: we record the interferogram on the fly while continuously oscillating the roof mirror, which introduces the optical path difference (OPD) while pointing the telescope on-sky (see [8] for a description
