Abstract

Contributed Talk - Splinter LIM

Friday, 11 September 2026, 15:20   (MW-2050)

Considerations from line intensity mapping towards developing filterbank spectrometers for next generation mapper instruments

Himadri Saha (1), Jonathan Clarke (1), Yoko Okada (1), Christos Karoumpis (2), Netty Honingh (1), Dominik A. Riechers (1)
(1) I. Physikalisches Institut, Universität zu Köln, (2) Argelander-Institut für Astronomie, Universität Bonn

Line Intensity Mapping (LIM) in the sub-mm window aims to probe the large-scale aggregate emission from the star formation tracer [CII] line emitted by unresolved galaxy clusters, providing insights into star formation activity in early galaxies at high redshift (z ~ 3 – 8). A promising approach to broadband LIM surveys is to use Integrated Filterbank Spectrometers (IFBS) technology. An IFBS focal plane array can simultaneously measure all instrument voxels, and when complemented with wide field-of-view facilities such as CCAT/FYST, can enable rapid LIM surveys covering large spatial and redshift scales. In this work, we investigate how our IFBS development can be adapted for the LIM science. An IFBS is a compact planar device that uses a filterbank circuit to separate the sky signal into spectral channels, with Microwave Kinetic Inductance Detectors (MKIDs) integrated as bolometers for each channel. Our analysis focuses on the filterbank design, as it defines the spectral transmission characteristics, with the resolution and filter order kept as free design parameters. We evaluate different spectral profiles on two aspects in the context of LIM surveys with FYST: the effectiveness of [CII] signal recovery with techniques such as voxel masking, and the spatial-spectral coverage achievable in the mapping survey. A second-order filterbank is found to be more suitable for LIM than conventional first-order IFBSs, as its sharper spectral profile reduces crosstalk and thereby lowers the spread of interloping CO foregrounds into adjacent channels. This increases the [CII]-to-CO power spectrum ratio by a factor of two, making voxel masking more effective. However, increasing the spectral resolution provides only marginal improvement, indicating that a low-resolution spectrometer might be preferable as it keeps the device footprint small. Finally, we find that the expected atmospheric loading at the CCAT/FYST site on photon-noise limited MKID detectors is similar for different filter orders at the same spectral resolution, allowing detector optimisation to be decoupled from the filterbank design for our spectrometer development.