Foreground-immune CMB lensing with shear-only reconstruction

Abstract

CMB lensing from current and upcoming wide-field CMB experiments such as AdvACT, SPT-3G and Simons Observatory relies heavily on temperature (vs. polarization). In this regime, foreground contamination to the temperature map produces significant lensing biases, which cannot be fully controlled by multi-frequency component separation, masking or bias hardening. In this letter, we split the standard CMB lensing quadratic estimator into a new set of optimal "multipole" estimators. On large scales, these multipole estimators reduce to the known magnification and shear estimators, and a new shear B-mode estimator. We leverage the different symmetries of the lensed CMB and extragalactic foregrounds to argue that the shear-only estimator should be approximately immune to extragalactic foregrounds. We build a new method to compute separately and without noise the primary, secondary and trispectrum biases to CMB lensing from foreground simulations. Using this method, we demonstrate that the shear estimator is indeed insensitive to extragalactic foregrounds, even when applied to a single-frequency temperature map contaminated with CIB, tSZ, kSZ and radio point sources. This dramatic reduction in foreground biases allows us to include higher temperature multipoles than with the standard quadratic estimator, thus increasing the total lensing signal-to-noise beyond the quadratic estimator. In addition, magnification-only and shear B-mode estimators provide useful diagnostics for potential residuals. Our Python code LensQuEst to forecast the signal-to-noise of the various estimators, generate mock maps, lense them, and apply the various lensing estimators to them is publicly available at https://github.com/EmmanuelSchaan/LensQuEst .