Search

Measurement of the largest angular scale ( ℓ < 30) features of the cosmic microwave background (CMB) polarization is a powerful way to constrain the optical depth to reionization and search for the signature of inflation through the detection of primordial B -modes. We present an analysis of maps covering 73.6% of the sky made from the 40 GHz channel of the Cosmology Large Angular Scale Surveyor (CLASS) from 2016 August to 2022 May. Taking advantage of the measurement stability enabled by front-end polarization modulation and excellent conditions from the Atacama Desert, we show this channel achieves higher sensitivity than the analogous frequencies from satellite measurements in the range 10 < ℓ < 100. Simulations show the CLASS linear (circular) polarization maps have a white noise level of $125(130)\,\mu {\rm{K}}\,\mathrm{arcmin}$ . We measure the Galaxy-masked EE and BB spectra of diffuse synchrotron radiation and compare to space-based measurements at similar frequencies. In combination with external data, we expand measurements of the spatial variations of the synchrotron spectral energy density (SED) to include new sky regions and measure the diffuse SED in the harmonic domain. We place a new upper limit on a background of circular polarization in the range 5 < ℓ < 125 with the first bin showing D _ℓ < 0.023 $\mu {{\rm{K}}}_{\mathrm{CMB}}^{2}$ at 95% confidence. These results establish a new standard for recovery of the largest-scale CMB polarization from the ground and signal exciting possibilities when the higher sensitivity and higher-frequency CLASS channels are included in the analysis.

The dynamic atmosphere imposes challenges to ground-based cosmic microwave background observation, especially for measurements on large angular scales. The hydrometeors in the atmosphere, mostly in the form of clouds, scatter the ambient thermal radiation and are known to be the main linearly polarized source in the atmosphere. This scattering-induced polarization is significantly enhanced for ice clouds due to the alignment of ice crystals under gravity, which are also the most common clouds seen at the millimeter-astronomy sites at high altitudes. This work presents a multifrequency study of cloud polarization observed by the Cosmology Large Angular Scale Surveyor experiment on Cerro Toco in the Atacama Desert of northern Chile, from 2016–2022, at the frequency bands centered around 40, 90, 150, and 220 GHz. Using a machine-learning-assisted cloud classifier, we made connections between the transient polarized emission found in all four frequencies with the clouds imaged by monitoring cameras at the observing site. The polarization angles of the cloud events are found to be mostly 90° from the local meridian, which is consistent with the presence of horizontally aligned ice crystals. The 90 and 150 GHz polarization data are consistent with a power law with a spectral index of 3.90 ± 0.06, while an excess/deficit of polarization amplitude is found at 40/220 GHz compared with a Rayleigh scattering spectrum. These results are consistent with Rayleigh-scattering-dominated cloud polarization, with possible effects from supercooled water absorption and/or Mie scattering from a population of large cloud particles that contribute to the 220 GHz polarization.

The Cosmology Large Angular Scale Surveyor (CLASS) is a telescope array that observes the cosmic microwave background over 75% of the sky from the Atacama Desert, Chile, at frequency bands centered near 40, 90, 150, and 220 GHz. This paper describes the CLASS data pipeline and maps for 40 GHz observations conducted from 2016 August to 2022 May. We demonstrate how well the CLASS survey strategy, with rapid (∼10 Hz) front-end modulation, recovers the large-scale Galactic polarization signal from the ground: the mapping transfer function recovers ∼67% (85%) of EE and BB ( VV ) power at ℓ = 20 and ∼35% (47%) at ℓ = 10. We present linear and circular polarization maps over 75% of the sky. Simulations based on the data imply the maps have a white noise level of $110\,\mu {\rm{K}}\,\mathrm{arcmin}$ and correlated noise component rising at low- ℓ as ℓ ^−2.4 . The transfer-function-corrected low- ℓ component is comparable to the white noise at the angular knee frequencies of ℓ ≈ 18 (linear polarization) and ℓ ≈ 12 (circular polarization). Finally, we present simulations of the level at which expected sources of systematic error bias the measurements, finding subpercent bias for the Λ cold dark matter EE power spectra. Bias from E -to- B leakage due to the data reduction pipeline and polarization angle uncertainty approaches the expected level for an r = 0.01 BB power spectrum. Improvements to the instrument calibration and the data pipeline will decrease this bias.