Forecastfiso planck

From CMB-S4 wiki
Revision as of 20:45, 2 June 2016 by Wlwu (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigationJump to search


We can improve constraints on isocurvature perturbations from Planck with observations from S4.

Here we consider the curvaton model: there is one isocurvature mode: the CDM isocurvature, and the spectral index of the CDM mode is the same as that of the adiabatic scalar perturbations, n_s. In this scenario, the CDM mode is either totally correlated or anti-correlated with the adiabatic perturbations.

The parameter we are constraining is f_iso, the primordial fraction of CDM isocurvature amplitude to the scalar perturbation amplitude. Specifically, the final spectra is

Cl^{tot} = Cl^{ad} + f_iso^2 Cl^{cdm iso} + 2* f_iso * Cl^{correlated}

where Cl^{cdm iso} is generated such that it has the same primordial power A_s as the scalar perturbations, and similarly for Cl^{correlated}. And the anti-correlated CDM modes contributes to the total spectrum in an analogous way.


Cosmology used in the forecast LCDM + f_iso. Fiducial value of f_iso=0. Use lensed TT/EE/TE spectra.

Correlated CDM isocurvature

  • vary beam
Isocorr varybeam.png
  • vary noise
Isocorr varynoise.png
  • fixed effort
Isocorr fixedeffort.png

Anti-correlated CDM isocurvature

  • vary beam
Isoanti varybeam.png
  • vary noise
Isoanti varynoise.png
  • fixed effort
Isoanti fixedeffort.png

Planck test

To check that we are getting reasonable results, we run the Fisher matrix with LCDM+Mnu+f_iso with Planck noise levels, beams, and fsky as specified in previous forecast checks.

We tested for fiducial cosmology f_iso=0 with 100% correlated CDM isocurvature modes; The following constraints use TT/EE/TE lensed spectra.

The following are the constraints we get:

och2 0.001523
obh2 0.000163
onuh2 0.005396
10^9 As 0.2589
ns 0.00486
tau 0.0597
hubble 4.84
f_iso 0.00391

We are doing a literature search to find previous forecasts that uses a similar parameterization (as opposed to the alpha or beta parameterization) to compare our results.