Scaling of polarization noise with number of simultaneous detector pairsin BICEP/Keck data

The BICEP/Keck collaboration has set constraints on the tensor-to-scalar ratio

r using a series of CMB polarization telescopes at the South Pole. BICEP

Array, the next stage of the experiment, will operate with an order of

magnitude more detectors than the previous stage, and is projected to reach a

sensitivity of σ(r) ≤ 0.003 using data upto 2027. This will be possible only if the

polarized noise scales down as expected with the increase in number of

simultaneously observing detector pairs. As the BICEP/Keck telescopes are

ground-based, the atmosphere acts as an unavoidable source of noise.

Large-scale atmospheric noise is correlated across detectors and, if it has a

polarized component, might cause a deviation from simple noise scaling,

impeding the progress of the experiment. We check for this “failure to scale”

in our polarization datasets. We observe that polarized noise in the high-l

regime, corresponding to small scale structures, scales well but that there is

a failure to scale in the low-l regime. We show that this issue can be

potentially addressed by down-weighting periods of elevated 1/f noise in the

pair difference timestreams, which results in a reduction in the low-l noise

excess in the final maps.