Are you an EPFL student looking for a semester project?

Work with us on data science and visualisation projects, and deploy your project as an app on top of GraphSearch.

Learn more about Graph Apps.

Person

Maurizio Martinelli

Copyright © 2024 EPFL, all rights reserved

Euclid preparation XXI. Intermediate-redshift contaminants in the search for z > 6 galaxies within the Euclid Deep Survey

Hakim Atek, Gianluca Castignani, Frédéric Courbin, Marcello Farina, Maurizio Martinelli, Richard Massey, Georges Meylan, Austin Chandler Peel, Jean-Luc Starck, Yi Wang

Context. The Euclid mission is expected to discover thousands of z > 6 galaxies in three deep fields, which together will cover a similar to 50 deg(2) area. However, the limited number of Euclid bands (four) and the low availability of ancillary data could make the identification of z > 6 galaxies challenging. Aims. In this work we assess the degree of contamination by intermediate-redshift galaxies (z = 1-5.8) expected for z > 6 galaxies within the Euclid Deep Survey. Methods. This study is based on similar to 176 000 real galaxies at z = 1-8 in a similar to 0.7 deg(2) area selected from the UltraVISTA ultra-deep survey and similar to 96 000 mock galaxies with 25.3 6 recovery of 91% (88%) for bright (faint) galaxies. For the UltraVISTA-like bright sample, the percentage of z = 1-5.8 contaminants amongst apparent z > 6 galaxies as observed with Euclid alone is 18%, which is reduced to 4% (13%) by including ultra-deep Rubin (Spitzer) photometry. Conversely, for the faint mock sample, the contamination fraction with Euclid alone is considerably higher at 39%, and minimised to 7% when including ultra-deep Rubin data. For UltraVISTA-like bright galaxies, we find that Euclid (I-E - Y-E) > 2:8 and (Y-E - J(E)) < 1.4 colour criteria can separate contaminants from true z > 6 galaxies, although these are applicable to only 54% of the contaminants as many have unconstrained (I-E - Y-E) colours. In the best scenario, these cuts reduce the contamination fraction to 1% whilst preserving 81% of the fiducial z > 6 sample. For the faint mock sample, colour cuts are infeasible; we find instead that a 5 sigma detection threshold requirement in at least one of the Euclid near-infrared bands reduces the contamination fraction to 25%.

EDP SCIENCES S A2022