Baryonic Signatures in Large-Scale Structure

TL;DR

The paper analyzes how a non-negligible baryon fraction alters structure formation in CDM cosmologies, predicting baryon-induced oscillations in the matter power spectrum that persist on large scales. Linear theory reveals a sequence of acoustic features with up to ~10% modulation and a characteristic spacing $\\Delta k \\sim 0.05\,h\,{ m Mpc}^{-1}$, while nonlinear evolution damps high-$k$ oscillations, leaving the second harmonic near $k\simeq 0.055\,h\,{ m Mpc}^{-1}$ as the most detectable signature. Through second-order perturbation theory and N-body simulations, the authors show that realistic nonlinear growth reduces the visibility of these calls, making detection challenging unless $\,\Omega_B h^2$ is sizeable and surveys are large and well-controlled. They argue that future 3D galaxy surveys could detect the baryonic features, whereas 2D methods face stronger limitations due to projection and mode counting. The work highlights the importance of precise normalization, bias handling, and redshift-space effects in interpreting large-scale clustering data for baryonic signatures.

Abstract

We investigate the consequences of a non-negligible baryon fraction for models of structure formation in Cold Dark Matter dominated cosmologies, emphasizing in particular the existence of oscillations in the present-day matter power spectrum. These oscillations are the remnants of acoustic oscillations in the photon-baryon fluid before last scattering. For acceptable values of the cosmological and baryon densities, the oscillations modulate the power by up to 10%, with a `period' in spatial wavenumber which is close to Delta k approximately 0.05/ Mpc. We study the effects of nonlinear evolution on these features, and show that they are erased for k > 0.2 h/ Mpc. At larger scales, the features evolve as expected from second-order perturbation theory: the visibility of the oscillations is affected only weakly by nonlinear evolution. No realistic CDM parameter combination is able to account for the claimed feature near k = 0.1 h/ Mpc in the APM power spectrum, or the excess power at 100 Mpc/h wavelengths quoted by several recent surveys. Thus baryonic oscillations are not predicted to dominate existing measurements of clustering. We examine several effects which may mask the features which are predicted, and conclude that future galaxy surveys may be able to detect the oscillatory features in the power spectrum provided baryons comprise more than 15% of the total density, but that it will be a technically challenging achievement.

Figures (13)

• Figure 1: A comparison of the power spectra for the CMB and LSS, for a model with $\Omega_0=0.4$, $\Lambda=0$, $h=0.65$ and $\Omega_{\rm B}=0.045$. The upper panel is contribution per $\ln(k)$ to the total variance in the CMB (scaled to unity at the smallest wavenumbers). The lower panel is the power per $\ln(k)$ in the matter, i.e. $\Delta^2(k)$, with the trend taken out using the BBKS fitting function (see text). The vertical dotted lines mark the positions of the peaks in the CMB power spectrum, which are $90^\circ$ out of phase with the corresponding oscillations in the LSS power spectrum.
• Figure 2: The position (in $k$ space in Mpc${}^{-1}$) of the second maximum in $T(k)$ compared to BBKS, as a function of $\Omega_0h^2$ and $\Omega_{\rm B} h^2$. Here $\Omega_0=\Omega_{\rm CDM}+\Omega_{\rm B}$. For measurements of distances in $\,h\,{\rm Mpc}^{-1}$ the values of the contours need to be divided by $h$. Apart from this scaling, the position of the peak is almost independent of cosmological parameters.
• Figure 3: The transfer function $T(k)$ for $\Omega_0=0.3$, 0.4 and $h=0.6$, 0.65 and 0.7. Note that the features in $T(k)$ are quite small on the scale of the variation of $T(k)$, but prominent when the trend is removed, as shown in Fig. \ref{['fig:del2']}.
• Figure 4: The linear theory matter power spectrum with the trend removed by dividing by the best fitting BBKS analytic fit to $T(k)$. Higher $\Omega_{\rm B}$ gives larger oscillations.
• Figure 5: The values of primordial spectral index required in order for CDM models to satisfy the constraints of correct normalization on both COBE and cluster scales, for the indicated values of $h$. This figure assumes no tensor contribution to the CMB anisotropy.