A Study Of Foreground Modeling Of Radio Halos And EoR Signal Separation Method For The SKA EoR Experiment

The epoch of reionization(EoR)is an important stage in the early Universe that is expected to last from about 300 Myr to 1 Gyr after the Big Bang,corresponding to a redshift range of about 6-15.During the EoR,the first stars and galaxies have just formed and start emitting ultraviolet and soft X-ray photons,which gradually reionize the surrounding neutral baryonic matter.Studying the EoR is invaluable in understanding the properties of the first stars and galaxies and the structure for-mation of the early Universe.Among the methods to probe the EoR,detecting the 21 cm line of neutral hydrogen originating from the EoR in the low-frequency ra-dio band(-50-200 MHz)is regarded as the most promising and effective method.However,the EoR signal is extremely faint and is buried in the overwhelming fore-ground contamination of about 4-5 orders of magnitude stronger.Therefore,it is indispensable to comprehend every foreground component and to develop specific foreground removal and EoR signal separation methods.Among various foreground components,the Galactic synchrotron and free-free radiations and the extragalactic point sources are the major components and have been rather extensively investigated in the literature.Their low-frequency ra-dio properties and contamination on the EoR detection are basically understood.On the other hand,radio halos in galaxy clusters,which are common extragalactic extended sources,are also expected to have an impact on the EoR signal detection.However,only several works that investigate the EoR foreground have preliminarily explored radio halos.Those works made oversimplifications in simulating the im-ages and spectra of radio halos and did not further analyze the impacts of radio halo emission on the EoR detection.In this work,we carry out two studies concerning the EoR foreground contamination and signal separation problems.First,we con-struct a more complete and physical model for simulating the radio halo emission and build a more realistic foreground model.By taking into account the SKA1-Low instrument effects,we evaluate the contamination of radio halo emission on the EoR signal detection.Secondly,based on the improved foreground model,we de-velop a new EoR signal separation method by utilizing the deep learning algorithms,in order to achieve accurate separation of the EoR signal.By employing the Press-Schechter formalism and the turbulent re-acceleration theory,we model the formation and evolution of radio halos and simulate their sky maps in the low-frequency band.Then,we adopt the latest SKA1-Low layout config-uration and simulate the SKA1-Low images of radio halos with instrument effects incorporated.By comparing the one-dimensional power spectra of radio halos and the EoR signal in the 120-128,154-162,and 192-200 MHz frequency bands,we find that radio halos are generally about 10 000,1000,and 300 times more luminous than the EoR signal on scales of 0.1 Mpc-1<k<2 Mpc-1(corresponding to scales of about 1.2'<s<24')in the three bands,respectively.After examining the two-dimensional power spectra inside the appropriately defined EoR windows,we find that the power leaked by radio halos can still be significant,as the power ratios of radio halos to the EoR signal on scales of 0.5 Mpc-1(?)k(?)1Mpc-1(correspond-ing to scales of about 2.4'(?)s(?)4.8')can be up to about 230-800%,18-95%,and 7-40%in the three bands,when the 68%uncertainties caused by the variation of the number density of bright radio halos are considered.Furthermore,we find that frequency artifacts resulted from instrument response can remarkably increase the power leakage of radio halos in the EoR window,which becomes more severe due to the radio halos located inside the side-lobes.These results show that radio halos are severe foreground sources and need serious treatments in future EoR experiments.Based on the improved foreground model and simulated SKA1-Low images ob-tained above,we further investigate the beam effects of interferometers and their impacts on the spectral smoothness of the foreground emission.The frequency-dependent beam effects will cause rapid fluctuations along the frequency dimen-sion and damage the spectral smoothness of the foreground emission,which makes traditional foreground removal methods inapplicable.To address this issue,we propose a deep-learning-based method that employs a 9-layer convolutional denois-ing autoencoder(CDAE)to separate the EoR signal.After being trained and tested on the simulated SKA1-Low images,the CDAE achieves excellent performance as the mean correlation coefficient between the reconstructed and input EoR signals reaches ?cdae=0.929±0.045.In comparison,the two representative traditional methods,namely the polynomial fitting method and the continuous wavelet trans-form method,both have difficulties in modeling and removing the foreground emis-sion that is complicated with the beam effects,yielding only ?poly=0.296±0.121 and ?cwt=0.198±0.160,respectively.In consequence,the CDAE can effectively deal with the spectral smoothness damage caused by the beam effects and thus ac-curately separate the EoR signal.This result also exhibits the great potential of deep-learning-based methods in future EoR experiments.