Survey Strategy

Spectral Scan Strategy

The purpose of the REBELS program is to characterize the most massive galaxies at z>6.5 through the detection of the dust continuum and line emission from bright sources. In theory, one would accomplish this by using ALMA to target a redshift-complete, mass-complete selection of z>6.5 galaxies.

The difficulty with this approach is that most of the brightest (and perhaps most massive) sources at z>6.5 do not have spectroscopic redshifts, either because of the intrinsic faintest of Lyman-alpha in such systems or because Lyman-alpha is absorbed by an increasingly neutral IGM at z>6.5. Additionally, when a spectroscopic redshift becomes available for a source or two at z>6.5, those sources are immediately targeted for follow-up with ALMA, leaving few if any sources for a large survey program to study in the epoch of reionization.

Given the lack of mass-complete selections of z>6.5 galaxies with spectroscopic redshift information, we are forced to rely on high-fidelity bright selections of z>6.5 galaxies constructed using photometry. How does this work, in practice, for the REBELS program? For bright galaxies with redshifts in the range z=6.5-7.2, the depth of existing observations and sensitivity of the z-band, y-band, and especially Spitzer/IRAC fluxes to the redshift of sources (as well as multiplicity of bands) made it possible to constrain source redshifts with a typical uncertainty of sigma(z) <~ 0.10. For bright galaxies with redshifts in excess of z=7.2, source redshifts were slightly more uncertain (due to the gap between the VISTA Y/J filters), but thanks to the partial detection of most sources in the Y, J, or a space-based Y-band filter (where the Lyman-break is partially through a bandpass), the uncertainties on source redshifts were typically only a factor of 2 poorer than at z=6.5-7.2.

Spectral scans for bright ISM cooling lines have been employed to determine the redshift for both galaxies at z~7 and z~8-10, but at z~7 the line used was the 157.74 micron [CII] line (Smit+2018; Schouws+2019, in prep) while at z~8-10 the line used has been the 88.36 micron [OIII] line (Tamura+2018; Hashimoto+2018; Tamura+2019, in prep). To maximize the efficiency of our proposed program, we calculated the approximate break-even luminosity L[OIII]/L[CII] ratio for [OIII] vs. [CII] spectral scans as a function of redshift. For the typical luminosity ratios observed for z~7 galaxies, we determined that [CII] was the most efficient line to use for our searches from z~6.5 to z~8.5, while at z>8.5, we determined that [OIII] would be the most efficient line to use.

Our sensitivity requirements for REBELS relies on our experience with searches in our pilot programs. There the detected [CII] lines had peak fluxes of 1.5-4.0 mJy (both in the Smit+2018 and Schouws+2019 programs). To guarantee the selection of similar sources at z∼7 with REBELS, the peak flux sensitivity must be ∼340 μJy in a 66 km/s channel such that sources with a peak flux of 1 mJy can be detected at 5σ when combining multiple channels. This is equivalent to a limiting luminosity (5σ) of ∼2×10⁸ L_⊙ at z∼7 and requires ~20 minutes of integration time per tuning. Reaching the same limiting luminosity at z∼8 requires a 300μJy sensitivity and ∼30-40 minutes of integration time in band 5 or 6 (for z > 8 and z < 8 [CII] searches, respectively). In scanning for the bright [OIII]_88.36 lines in our z ∼ 9 targets, the equivalent sensitivity requirement is 450μJy.

For the purposes of illustration, the limiting [CII] and [OIII] luminosities probed as a function of source redshift for the spectral scans planned for sources in REBELS is provided in the figure below. As should be clear from the figure, REBELS will detect line emission from sources if the [CII] luminosities exceed ∼2×10⁸ L_⊙ and, for z>8.5 targets, if the [OIII] luminosities exceed ∼7×10⁸ L_⊙.

Based on the typical [CII]/far-IR luminosity ratios, detecting sources in the dust continuum generally requires significantly longer integration times than detecting ISM-cooling lines alone. As such, if the goal is to integrate long enough to detect the dust continuum in individual sources, the use of multiple scan windows to search for [CII] carries no significant additional observational cost.

63% of the z~7 galaxies targeted with observations from our pilot programs have yielded lines with luminosities in excess of this limit, resulting in redshift measurements and line luminosities/widths. Given that sources with line detections are also detected in the dust continuum in essentially every case (and those lacking line detections also lack detections in the dust continuum), it seems likely that the observed incompleteness in our [CII] searches is simply the result of the [CII] lines being fainter than the ∼2×10⁸ L⊙ detection limit of those searches (and perhaps much fainter!).

While we could, in principle, integrate longer on individual sources to improve the completeness of our line searches, sources which remain undetected in these searches are almost certainly less dusty and have lower masses than the sources we can efficiently identify with ALMA using our scan strategy. The purpose of REBELS is to identify a large statistical sample of massive/dusty galaxies in the z>6.5 universe to study the build-up of dust and stellar mass in galaxies (while probing the dynamical state of these massive galaxies), and it is for these galaxies (at least within the UV-bright population of galaxies) that we expect to be complete.

Given the similar selection criteria used for the REBELS large program to that used in our pilot programs, we would expect a similar detection rate for the full program, as in our pilot, i.e., 63%, suggesting 25 line detections out of 40 targeted sources. Combining the expected new results with earlier results, future samples of bright [CII]/[OIII] emitting galaxies at z>6.5 should include >~35 sources in total, sufficient for a detailed physical characterization of massive galaxies at z>~6.5 and a study of their evolution.