Identification of low redshift groups and clusters of galaxies in the X-CLASS survey and the X-ray luminosity-temperature relation

Properties of the hot intracluster and intragroup medium are mostly set by the underlying gravitational potential well, although complex astrophysical processes at play during their buildup may leave a significant imprint. Observational constraints on the degree and scales of such non-gravitational processes require well-selected samples of objects and deep observations of their gas content. We aim to study the scaling relation between two global properties of the hot gas, namely its soft-band X-ray luminosity ($L_X$) and its temperature ($T$), by studying a sample of low-mass systems associated with precise redshifts, simultaneously accounting for sample selection biases and associated measurement uncertainties. This work takes as input a large catalogue of X-ray-selected galaxy clusters (X-CLASS). We perform a thorough revision of the redshifts of sources using deep photometric data from the Legacy Surveys and our own tailored spectroscopic follow-up of 52 low-redshift systems. We devise a spectroscopically complete sample of 155 low-redshift ($0.07<z<0.2$) systems, and we measure properties of their X-ray emitting gas, with median $\overline{T}=1.7$ keV and median $\overline{L_X}=10^{43}$ erg s$^{-1}$. We infer the relation between $L_X$ and $T$ in a Bayesian framework. Our sample of groups and clusters with median total mass $\sim 6 \times 10^{13}M_\odot$ reveals a relation $L_X-T$ steeper than predicted by the self-similar model, with a slope $B=3.2 \pm 0.1$. This result fits well within recent studies that together indicate a trend of increasing slope with decreasing median halo mass. This work supports a scenario of a stronger decrease in luminosity with decreasing mass in the group regime than for massive galaxy clusters. This effect is possibly due to strong and sustained feedback expelling gas efficiently from their relatively shallower potential wells.

💡 Research Summary

The paper presents a comprehensive study of the X‑ray luminosity–temperature (L_X–T) scaling relation for low‑mass galaxy groups and clusters using the X‑CLASS catalogue, an XMM‑Newton based survey of extended X‑ray sources. The authors first improve the redshift completeness of the X‑CLASS sample, which originally had spectroscopic redshifts for only ~60 % of its 1,646 candidates. They exploit the DESI Legacy Survey (DR9/DR10) photometry, combining optical (g, r, i, z) and infrared (W1, W2) bands, to compute photometric redshifts for individual galaxies via a Random Forest algorithm. Two cluster‑redshift estimators are then applied: a newly developed Bayesian code called photXclus, which incorporates the X‑ray extent as a prior, and the well‑known RedMaPPer algorithm, run in scan mode on the same data. By adopting photXclus for z < 0.15 (where RedMaPPer shows a higher outlier rate) and RedMaPPer otherwise, the authors achieve a combined photometric redshift performance of σ_NMAD = 0.0044 (1+z), negligible bias (−5.8 × 10⁻⁴) and an outlier fraction of 4.6 % when benchmarked against spectroscopic redshifts.

To obtain a fully spectroscopic low‑redshift subsample, the team carried out dedicated observations with the MISTRAL spectrograph, adding 52 new redshifts. Together with literature and NED entries, they assemble a spectroscopically complete sample of 155 systems in the range 0.07 < z < 0.2. The median temperature of these objects is T ≈ 1.7 keV and the median soft‑band (0.5–2 keV) luminosity is L_X ≈ 10⁴³ erg s⁻¹, corresponding to a typical total mass of ~6 × 10¹³ M_⊙.

X‑ray properties are measured using the X‑CLASS pipeline: surface‑brightness profiles are fitted with a β‑model, and temperatures are derived from APEC spectral fits within R₅₀₀. The authors then model the L_X–T relation in a Bayesian hierarchical framework that simultaneously accounts for measurement errors, intrinsic scatter, and the survey selection function. The selection function itself is built from extensive Monte‑Carlo simulations of mock XMM pointings that reproduce realistic background, point‑source contamination, and a variety of cluster morphologies.

The resulting scaling relation is expressed as log L_X = A + B log T, with a slope B = 3.2 ± 0.1 and an intercept A ≈ 42.5 (L_X in erg s⁻¹, T in keV). This slope is significantly steeper than the self‑similar prediction (B ≈ 1.5–2.0) and aligns with recent findings for low‑mass systems (e.g., Lovisari et al. 2015; Bahar et al. 2022). The authors interpret the steepening as evidence that non‑gravitational processes—particularly sustained AGN feedback—more efficiently expel or heat the intracluster medium in shallow potential wells, thereby suppressing X‑ray luminosity at a given temperature.

The paper also discusses the “up‑scattering” bias, where low‑luminosity clusters are preferentially scattered above the detection threshold, and shows how their careful modeling of the selection function mitigates this effect. All updated redshifts, gas property measurements, and the photXclus code are publicly released, providing valuable resources for future large‑scale optical–X‑ray cross‑correlation studies.

In summary, this work delivers a rigorously selected, spectroscopically complete low‑redshift sample, demonstrates a robust Bayesian methodology for scaling‑relation inference, and confirms that the L_X–T relation steepens toward lower halo masses, highlighting the dominant role of baryonic physics in shaping the observable properties of galaxy groups and clusters.