The interplay of CMB temperature lensing power reconstruction with primordial non-Gaussianity of local type

TL;DR

This paper investigates how primordial non-Gaussianity of local type affects the bias in CMB lensing potential power spectrum reconstruction, finding that the bias is larger than expected but negligible for current experiments like Planck.

Contribution

It provides a detailed simulation-based analysis of the bias introduced by primordial non-Gaussianity in CMB lensing reconstruction, extending previous analytical estimates.

Findings

Bias due to non-Gaussianity is larger than previous estimates.

For current f_NL values, the bias is negligible except at large scales.

Simulations confirm the bias's dependence on primordial non-Gaussianity.

Abstract

In the current era of high-precision CMB experiments, the imprint of gravitational lensing on the CMB temperature is exploited as a source of valuable information. Especially the reconstruction of the lensing potential power spectrum is of great interest. The reconstruction from the optimal quadratic estimator of the lensing potential, though, is biased. As long as the intrinsic CMB fluctuations are Gaussian this bias is well understood and controlled. In the presence of non-Gaussian primordial curvature perturbations, however, the CMB also acquires a non-Gaussian structure mimicking the lensing signal. Concentrating on primordial non-Gaussianity of local type, we address the resulting bias by extracting the lensing potential power spectrum from large samples of simulated lensed CMB temperature maps comprising different values of f_NL. We find that the bias is considerably larger than previous analytical calculations suggested. For current values of f_NL and a sensitivity like that of the Planck mission, however, the bias is completely negligible on all but the largest angular scales.