Search for features in the spectrum of primordial perturbations using Planck and other datasets

TL;DR

This paper reconstructs the primordial power spectrum using Planck and other datasets, revealing features and an infrared cutoff, with implications for early universe models.

Contribution

It introduces a non-parametric reconstruction method combining multiple datasets to analyze primordial perturbations, identifying features and cutoff in the spectrum.

Findings

Reconstructed spectrum shows an infrared cutoff at k < 5 x 10^-4 Mpc^-1.

Detected features at specific wavenumbers consistent with previous data.

Planck data suggests a higher significance feature at k ~ 0.12 Mpc^-1, but with reduced global significance.

Abstract

We reconstruct the power spectrum of primordial curvature perturbations by applying a well-validated non-parametric technique employing Tikhonov regularisation to the first data release from the Planck satellite. To improve the reconstruction on small spatial scales we include data from the ground-based ACT and SPT experiments, the WiggleZ galaxy redshift survey, the CFHTLenS tomographic weak lensing survey, and spectral analysis of the Lyman-alpha forest. The reconstructed scalar spectrum (assuming the standard LCDM cosmology) is not scale-free but has an infrared cutoff at k < 5 x 10^-4 Mpc^-1 and several ~2-3 sigma features, of which two at wavenumber k/Mpc^-1 ~ 0.0018 and 0.057 had been seen already in WMAP data. A higher significance feature at k ~ 0.12 Mpc^-1 is indicated by Planck data, but may be sensitive to the systematic uncertainty around multipole l ~ 1800 in the 217x217 GHz cross-spectrum. In any case accounting for the 'look elsewhere' effect drops its global significance to ~2 sigma.