Prospects for Delensing the Cosmic Microwave Background for Studying Inflation

TL;DR

This paper assesses how effectively gravitational lensing B modes in the CMB can be removed using CIB or CMB data, to improve measurements of inflationary parameters like the tensor tilt.

Contribution

It provides forecasts for delensing performance using CIB and CMB maps, highlighting their effectiveness at different noise levels for studying inflation.

Findings

CIB maps are as effective as CMB maps for delensing at current noise levels.

CMB maps outperform CIB maps at polarization noise below 1 μK-arcmin.

CMB delensing can measure tensor tilt with an accuracy of 0.02 or better.

Abstract

A detection of excess cosmic microwave background (CMB) B-mode polarization on large scales allows the possibility of measuring not only the amplitude of these fluctuations but also their scale dependence, which can be parametrized as the tensor tilt $n_T$. Measurements of this scale dependence will be hindered by the secondary B-mode polarization anisotropy induced by gravitational lensing. Fortunately, these contaminating B modes can be estimated and removed with a sufficiently good estimate of the intervening gravitational potential and a good map of CMB E-mode polarization. We present forecasts for how well these gravitational lensing B modes can be removed, assuming that the lensing potential can be estimated either internally from CMB data or using maps of the cosmic infrared background (CIB) as a tracer. We find that CIB maps are as effective as CMB maps for delensing at the noise levels of the current generation of CMB experiments, while the CMB maps themselves will ultimately be best for delensing at polarization noise below $\Delta_P$=1 $\mu$K-arcmin. At this sensitivity level, CMB delensing will be able to measure $n_T$ to an accuracy of 0.02 or better, which corresponds to the tensor tilt predicted by the consistency relation for single-field slow-roll models of inflation with $r=0.2$. However, CIB-based delensing will not be sufficient for constraining $n_T$ in simple inflationary models.