Abstract:Context. Galaxy clusters are the most massive bound objects in the recent history of the universe; the number density of galaxy clusters as a function of mass and redshift is a sensitive function of the cosmological parameters. To use clusters for cosmological parameter studies, it is necessary to determine their masses as accurately as possible, which is typically done via scaling relations between mass and observables. Aims. X-ray observables can be biased by a number of effects, including multiphase gas and projection effects, especially in the case where cluster temperatures and luminosities are estimated from single-model fits to all of the emission with an overdensity radius such as r 500c . Using simulated galaxy clusters from a realistic cosmological simulation, our aim is to determine the importance of these biases in the context of Spectrum-Roentgen-Gamma/eROSITA observations of clusters. Methods. We extracted clusters from the Box2_hr simulation from the Magneticum suite, and simulated synthetic eROSITA observations of these clusters using PHOX to generate the photons and the end-to-end simulator SIXTE to trace them through the optics and simulate the detection process. We fitted the spectra from these observations and compared the fitted temperatures and luminosities to the quantities derived from the simulations. We fitted an intrinsically scattered L X − T scaling relation to these measurements following a Bayesian approach with which we fully took into account the selection effects and the mass function. Results. The largest biases on the estimated temperature and luminosities of the clusters come from the inadequacy of single-temperature model fits to represent emission from multiphase gas, and from a bias arising from cluster emission within the projected r 500c along the line of sight but outside of the spherical r 500c . We find that the biases on temperature and luminosity due to the projection of emission from other clusters within r 500c is comparatively small. We find eROSITA-like measurements of Magneticum clusters following a L X − T scaling relation that has a broadly consistent but slightly shallower slope compared to the literature values. We also find that the intrinsic scatter of L X at given T is lower compared to the recent observational results where the selection effects are fully considered.

The eROSITA Final Equatorial-Depth Survey (eFEDS): A Machine Learning Approach to Infer Galaxy Cluster Masses from eROSITA X-ray Images

Using X-Ray Morphological Parameters to Strengthen Galaxy Cluster Mass Estimates via Machine Learning

The cosmological analysis of X-ray cluster surveys: VI. Inference based on analytically simulated observable diagrams

Benchmarks and Explanations for Deep Learning Estimates of X-ray Galaxy Cluster Masses

Estimating cluster masses from SDSS multiband images with transfer learning

Mass Estimation of Planck Galaxy Clusters using Deep Learning

Measuring the X-ray Luminosities of SDSS DR7 Clusters from ROSAT All Sky Survey

CircleZ : Reliable photometric redshifts for active galactic nuclei computed solely using photometry from Legacy Survey Imaging for DESI

Effects of multiphase gas and projection on X-ray observables in simulated galaxy clusters as seen by eROSITA

A deep learning approach to infer galaxy cluster masses from Planck Compton-y parameter maps

A Multi-Wavelength Technique for Estimating Galaxy Cluster Mass Accretion Rates

CircleZ: Reliable Photometric redshifts for AGN computed using only photometry from Legacy Survey Imaging for DESI

The SRG/eROSITA All-Sky Survey. Dark Energy Survey year 3 weak gravitational lensing by eRASS1 selected galaxy clusters

An Observationally Driven Multifield Approach for Probing the Circum-Galactic Medium with Convolutional Neural Networks

X-ray luminosity-star formation rate scaling relation: Constraints from the eROSITA Final Equatorial Depth Survey (eFEDS)

Probing dark energy with the next generation X-ray surveys of galaxy clusters

Photometric redshift estimation with a convolutional neural network: NetZ

Estimating the mass of galactic components using machine learning algorithms

The SRG/eROSITA All-Sky Survey: Weak-Lensing of eRASS1 Galaxy Clusters in KiDS-1000 and Consistency Checks with DES Y3 & HSC-Y3

Total and dark mass from observations of galaxy centers with Machine Learning

Deep Learning generated observations of galaxy clusters from dark-matter-only simulations