Impact of Residual Gain Errors on the Detectability of Ionized Bubbles during Reionization with SKA 21-cm observations

Abstract

Detecting ionized bubbles and characterizing their properties and the surrounding intergalactic medium (IGM) using the HI 21-cm signal provides direct probes of cosmic reionization. Here, we focus on the detection of individual ionized bubbles around luminous sources using a matched-filter technique. We investigate the impact of time- and frequency-correlated residual antenna gain errors on the detectability and characterization of individual ionized regions with SKA-Low. We simulate mock SKA-Low AA* observations that include the HI 21-cm signal, astrophysical foregrounds, residual gain errors, and system noise corresponding to 100 h of observing time. We subtract the foreground components from the gain-corrupted data and subsequently apply the matched-filter technique to the residual visibilities. We find that the matched-filter technique is able to detect the ionized bubble and constrain its properties in the presence of residual gain errors with the rms amplitudes up to $σ_A\sim1\%$ and rms phases up to $σ_ϕ\sim1^\circ$. Residual gain errors at similar levels prevent the reliable identification of ionized regions in reconstructed HI 21-cm images and bias the estimated HI 21-cm power spectrum. Furthermore, we find that increasing the temporal correlation length increases the level of gain-induced artifacts in the reconstructed image. In contrast, increasing the frequency correlation length reduces these artifacts, as some fraction of the spectrally smooth gain-induced residuals is removed during foreground subtraction. We further find that the inferred bubble and IGM parameters remain largely unaffected by residual gain errors, although their uncertainties can increase. These results demonstrate the potential and robustness of the matched-filter technique for detecting and characterizing individual ionized bubbles in the presence of substantial residual calibration errors.