Recovering unbiased CMB polarization maps using modern ground-based experiments with minimal assumptions about atmospheric emission

We present a study of unbiased reconstruction of cosmic microwave background (CMB) polarization maps from data collected by modern ground-based observatories. Atmospheric emission is a major source of correlated noise in such experiments, complicating the recovery of faint cosmological signals. We consider estimators that require only minimal assumptions about the properties of the unpolarized atmospheric emission, instead exploiting hardware solutions commonly implemented in modern instruments, such as pairs of orthogonal antennas in each focal plane pixel, interleaving of the antenna orientation from pixel to pixel, and polarization signal modulation via a continuously rotating half-wave plate (HWP). We focus on two techniques: (i) statistical downweighting of low-frequency atmospheric signals and (ii) pair differencing (PD), which involves differencing the signals collected by two detectors in the same focal plane pixel. We compare their performance and contrast them with the idealized case where the atmospheric signal is perfectly known and can be cleanly subtracted. We show that PD can be derived from the maximum likelihood principle under very general assumptions about the atmospheric signal and is therefore optimal in terms of map sensitivity for essentially any such signal. Remarkably, in the absence of instrumental systematics, but with reasonable variations of detector noise properties, PD yields polarized sky maps with noise levels only slightly worse (typically a few percent) than the ideal reference case. While the downweighting method could in principle match or surpass this performance, it requires highly accurate models of the atmospheric signal, which are not readily available. The performance of PD will be affected by instrumental systematics and those leaking the atmospheric signal to the difference time stream may have particularly adverse impact. However, we expect that some of these effects, as shown in the case of gain mismatch, are efficiently mitigated by a continuously rotating HWP and, therefore, that in the context of modern CMB experimental setups PD could provide a competitive, numerically robust, and efficient practical solution for CMB polarization mapmaking without the need for atmospheric modeling.