Spectral Line De-confusion in an Intensity Mapping Survey

Spectral line intensity mapping has been proposed as a promising tool to efficiently probe the cosmic reionization and the large-scale structure. Without detecting individual sources, line intensity mapping makes use of all available photons and measures the integrated light in the source confusion limit, to efficiently map the three-dimensional matter distribution on large scales as traced by a given emission line. One particular challenge is the separation of desired signals from astrophysical continuum foregrounds and line interlopers. Here we present a technique to extract large-scale structure information traced by emission lines from different redshifts, embedded in a three-dimensional intensity mapping data cube. The line redshifts are distinguished by the anisotropic shape of the power spectra when projected onto a common coordinate frame. We consider the case where high-redshift [CII] lines are confused with multiple low-redshift CO rotational lines. We present a semi-analytic model for [CII] and CO line estimates based on the cosmic infrared background measurements, and show that with a modest instrumental noise level and survey geometry, the large-scale [CII] and CO power spectrum amplitudes can be successfully extracted from a confusion-limited data set, without external information. We discuss the implications and limits of this technique for possible line intensity mapping experiments.