Correlated isocurvature perturbations from mixed inflaton-curvaton decay

TL;DR

This work generalizes the curvaton paradigm by treating curvaton decay as a field-driven process rather than dust and by including possible perturbations from the inflaton decay products. It derives a two-source primordial spectrum with correlated adiabatic and CDM isocurvature components, parameterized by two spectral indices and transfer efficiencies that map onto observable CMB spectra. Through an 11-parameter WMAP1 fit, the authors show that such correlated perturbations can suppress the low-l Sachs-Wolfe power and broaden the allowed cosmological-parameter space, though the overall fit is not decisively better than ΛCDM when penalized by degrees of freedom. The results highlight the sensitivity of cosmological parameter inference to the assumed primordial spectrum and demonstrate a mechanism for decoupling low- and high-l multipoles via anticorrelated isocurvature modes, with implications for how future data should constrain isocurvature contributions.

Abstract

We study cosmological perturbations in the case that present-day matter consists of a mixture of inflaton and curvaton decay products. We calculate how the curvaton perturbations are transferred to its decay products in the general case when it does not behave like dust. Taking into account that the decay products of the inflaton can also have perturbations results in an interesting mixture of correlated adiabatic and isocurvature perturbations. In particular, negative correlation can improve the fit to the CMB data by lowering the angular power in the Sachs-Wolfe plateau without changing the peak structure. We do an 11-parameter fit to the WMAP data. We find that the best-fit is not the 'concordance model', and that well-fitting models do not cluster around the best-fit, so that cosmological parameters cannot be reliably estimated. We also find that in our model the mean quadrupole (l=2) power is l(l+1) C_l/2pi = 1081 muK^2, much lower than in the pure adiabatic LCDM model, which gives 1262 muK^2.

Figures (6)

• Figure 1: Background energy densities (up) and $\xi_{comp}$ (down) for a large initial value of the field. On the left, a model where the field decays just as it is starting to oscillate and on the right, a model where the field oscillates before decaying. The dotted vertical lines mark the nominal start of oscillation and decay.
• Figure 2: Background energy densities (up) and $\xi_{comp}$ (down) for a small initial value of the field.
• Figure 3: The matter decay coefficient $\lambda_m$ on a logarithmic scale.
• Figure 4: The ratio of the decay coefficients for radiation and matter, $\lambda_r/\lambda_m$.
• Figure 5: The example model of the third column of Table \ref{['table:parameters']} with the WMAP data ($\bullet$) and the highest $l$ data points of other CMB experiments from Tegmark's compilation Tegmark:movie ($\blacklozenge$). The total angular power consists of the six components mentioned in the paragraph after equation (\ref{['ClTE']}): $C_l^{\rm tot} = C_l^{\rm adi1} + C_l^{\rm iso1} + C_l^{\rm cor1} + C_l^{\rm adi2} + C_l^{\rm iso2} + C_l^{\rm cor2}$. Upper panel: Temperature-temperature angular power $l(l+1)C_l^{TT}/2\pi$. Lower panel: Temperature-polarisation cross-correlation power $(l+1)C_l^{TE}/2\pi$.