Characterization of the ESPRESSO line-spread function and improvement of the wavelength calibration accuracy

• Published in: Monthly notices of the Royal Astronomical Society Vol. 530; no. 1; pp. 1252 - 1273
• Main Authors: Schmidt, Tobias M, Bouchy, François
• Format: Journal Article
• Language: English
• Published: Oxford University Press 05.04.2024
• Subjects: cosmology: observations; methods: data analysis; software: data analysis; instrumentation: spectrographs; line: profiles; techniques: spectroscopic
• ISSN: 0035-8711; 1365-2966
• DOI: 10.1093/mnras/stae920

ABSTRACT Achieving a truly accurate wavelength calibration of high-dispersion echelle spectrographs is a challenging task but crucially needed for certain science cases, e.g. to test for a possible variation of the fine-structure constant in quasar spectra. One of the spectrographs best suited for this mission is Very Large Telescope/Echelle SPectrograph for Rocky Exoplanets and Stable Spectroscopic Observation (ESPRESSO). Nevertheless, previous studies have identified significant discrepancies between the classical wavelength solutions and the one derived independently from the laser frequency comb. The dominant parts of these systematics were intra-order distortions, most-likely related to a deviation of the instrumental line-spread function from the assumed Gaussian shape. Here, we therefore present a study focused on a detailed modelling of the ESPRESSO instrumental line-spread function. We demonstrate that it is strongly asymmetric, non-Gaussian, different for the two slices and fibres, and varies significantly along the spectral orders. Incorporating the determined non-parametric model in the wavelength calibration process drastically improves the wavelength calibration accuracy, reducing the discrepancies between the two independent wavelength solutions from $50\,\rm{m\,s^{-1}}$ to about $10\, \rm{m\,s^{-1}}$. The most striking success is, however, that the different fibres and slices now provide fully consistent measurements with a scatter of just a couple m s−1. This demonstrates that the instrument-related systematics can be nearly eliminated over most of the spectral range by properly taking into account the complex shape of the instrumental line-spread function and paves the way for further optimizations of the wavelength calibration process.