On Cross-Correlating Line Intensity Maps from SPHEREx during Reionization

Abstract

We have simulated Ly{\alpha}, H{\alpha}, H{\beta}, [OII], and [OIII] intensity maps which are observable by SPHEREx during cosmic reionization. We simulate these intensity maps including all significant sources of emission for each line, and include radiative transfer for the Ly{\alpha} intensity maps. We also include a simple model of dust extinction based on observations of galaxies at z<5. One of the main challenges of intensity mapping is interloping lines from galaxies at lower redshifts, which makes producing an auto-power spectrum challenging. We focus on cross-correlations between different lines, as this eliminates such foreground contamination of the signal. We have cross-correlated the simulated SPHEREx intensity maps to find the most observable cases. This includes modeling of interloping lines and masking bright interloping galaxies. Testing a range of cases motivated by observations, we find total signal-to-noise values up to 99 for the highest case of H{\alpha} cross-correlated with [OIII] at z=5 assuming no dust extinction. We also find cases which will not be detectable. We find that the dominant noise source in these intensity maps on most scales is from the instrument, except for Ly{\alpha} and [OII] and then only on the largest scales the interlopers are the dominant source. We find through intensity mapping we can probe galaxies with masses $M<4x10^{10}~M_{\odot}$ which are below the necessary luminosity for a 3{\sigma} signal-to-noise direct detection of galaxies by SPHEREx. However, the majority of our observable signal is dominated by large, directly detectable galaxies, rather than the smaller, fainter galaxies. We find marginal detections of the clustering portion of the power spectrum at z=5 for H{\alpha}x[OIII]. Detections of the clustering signal from other lines or at z>6 will require more sensitive instruments, such as the Cosmic Dawn Intensity Mapper.