English

Constraints on primordial non-Gaussianity from Quaia

Cosmology and Nongalactic Astrophysics 2026-05-19 v2 General Relativity and Quantum Cosmology

Abstract

We analyse the large-scale angular clustering of quasars in the \gaia-\unwise quasar catalog, \quaia, and their cross-correlation with maps of the lensing convergence of the Cosmic Microwave Background (CMB), to constrain the level of primordial non-Gaussianity (PNG). Specifically, we target the scale-dependent bias that would be induced by PNG on biased tracers of the matter inhomogeneities on large scales. The \quaia sample is particularly well suited for this analysis, given the large effective volume covered, and our ability to map out the main potential sources of systematic contamination and mitigate their impact. Using the universality relation to characterise the response of the quasar overdensity to PNG (pϕ=1p_\phi=1), we report constraints on the local-type PNG parameter fNLf_{\rm NL} of fNL=20.518.1+19.0f_{\rm NL} =-20.5^{+19.0}_{-18.1} (68\% C.L.) by combining the quasar auto-correlation and its cross-correlation with CMB lensing in two tomographic redshift bins (or fNL=28.724.6+26.1f_{\rm NL} =-28.7^{+26.1}_{-24.6} if assuming a lower response for quasars, pϕ=1.6p_\phi=1.6). The error on fNLf_{\rm NL} can be further improved if the cross-correlation between the tomographic redshift bins is included. Using the CMB lensing cross-correlations alone, we find fNL=13.825.0+26.7f_{\rm NL} =-13.8^{+26.7}_{-25.0} and fNL=15.634.8+42.3f_{\rm NL} = -15.6^{+42.3}_{-34.8} for pϕ=1p_\phi=1 and pϕ=1.6p_\phi=1.6 respectively. These are the tightest constraints on fNLf_{\rm NL} to date from angular clustering statistics and cross-correlations with CMB lensing.

Keywords

Cite

@article{arxiv.2504.20992,
  title  = {Constraints on primordial non-Gaussianity from Quaia},
  author = {Giulio Fabbian and David Alonso and Kate Storey-Fisher and Thomas Cornish},
  journal= {arXiv preprint arXiv:2504.20992},
  year   = {2026}
}

Comments

Matches version published in JCAP. Material relevant for this analysis is available at https://github.com/gfabbian/quaia-fnl