Multiple inflation and the WMAP 'glitches' II. Data analysis and cosmological parameter extraction
Paul Hunt, Subir Sarkar
TL;DR
This paper investigates how features in the primordial curvature spectrum, motivated by the multiple inflation scenario with phase transitions during inflation, affect cosmological parameter estimation using WMAP3, SDSS, and LRG data. It implements a Markov Chain Monte Carlo analysis with two physically motivated spectra: a one-step 'Step' model and a two-step 'Bump' model, assessing their impact on standard ΛCDM parameters and exploring CHDM alternatives. The results show that a Step spectrum can modestly improve fits without dark energy, while a Bump spectrum can mimic an Einstein–de Sitter universe with very low $h$ and significant hot dark matter, though it struggles to fit BAO data. The work highlights strong degeneracies between primordial-spectrum features and background parameters, underscoring the need for broader, theory-driven analyses beyond a simple power-law spectrum and motivating Planck-era tests to discriminate these possibilities.
Abstract
Detailed analyses of the WMAP data indicate possible oscillatory features in the primordial curvature perturbation, which moreover appears to be suppressed beyond the present Hubble radius. Such deviations from the usual inflationary expectation of an approximately Harrison-Zeldovich spectrum are expected in the supergravity-based 'multiple inflation' model wherein phase transitions during inflation induce sudden changes in the mass of the inflaton, thus interrupting its slow-roll. In a previous paper we calculated the resulting curvature perturbation and showed how the oscillations arise. Here we perform a Markov Chain Monte Carlo fitting exercise using the 3-year WMAP data to determine how the fitted cosmological parameters vary when such a primordial spectrum is used as an input, rather than the usually assumed power-law spectrum. The 'concordance' LCDM model is still a good fit when there is just a 'step' in the spectrum. However if there is a 'bump' in the spectrum (due e.g. to two phase transitions in rapid succession), the precision CMB data can be well-fitted by a flat Einstein-de Sitter cosmology without dark energy. This however requires the Hubble constant to be h ~ 0.44 which is lower than the locally measured value. To fit the SDSS data on the power spectrum of galaxy clustering requires a ~10% component of hot dark matter, as would naturally be provided by 3 species of neutrinos of mass ~0.5 eV. This CHDM model cannot however fit the position of the baryon acoustic peak in the LRG redshift two-point correlation function. It may be possible to overcome these difficulties in an inhomogeneous Lemaitre-Tolman-Bondi cosmological model with a local void, which can potentially also account for the SN Ia Hubble diagram without invoking cosmic acceleration.