Detecting Primordial Gravitational Waves: a forecast study on optimizing frequency distribution of next generation ground-based CMB telescope

Abstract

Abstract Probing primordial gravitational waves is one of the core scientific objectives of the next generation CMB polarization experiment. Integrating more detector modules on the focal plane and performing high accurate observations are the main directions of the next generation CMB polarization telescope, like CMB S4. Also, multi-band observation is required by foreground analysis and reduction, as it is understood that foregrounds have become the main obstacles of CMB polarization measurements. However, ground observation is limited by the atmospheric window and can be usually carried out in one or two bands, like what BICEP or Keck array have done in the south pole. In this paper, we forecast the sensitivity of tensor-to-scalar ratio r that may be achieved by a multi-frequency CMB polarization experiment, basing on which to provide guidance for further expanding frequency bands and optimize the focal plane of a telescope. At the same time, the realization of having two frequency bands in one atmospheric window is discussed. With fixed number of detectors, the simulation results show that, in order to get a good limit, more frequency bands are needed. Better constraints can be obtained when it includes at least three bands, i.e., one CMB channel (95 GHz) + one dust channel (high frequency) and one synchrotron channel (low frequency). For example, 41 + 95 + 220 GHz, which is better than only focusing around the CMB band, like 85 + 105 + 150 GHz, and 95 + 135 + 155 GHz, and this frequency combination is even better than the combination of 41 + 95 + 150 + 220 GHz. As CMB S4 plans to consider two frequency bands in each atmospheric window, and along this way, we find that one CMB band and more bands in synchrotron and dust channels are helpful, for example, 2 bands in lower frequency, 30 + 41 GHz, 2 bands in higher frequency, 220 + 270 GHz, i.e. 30 + 41 + 95 + 220 + 270 GHz, can get better constraints, and in this case, more detectors are asked to be assigned in the CMB channel.