First Estimations of Cosmological Parameters From BOOMERANG

Abstract

The anisotropy of the cosmic microwave background radiation contains information about the contents and history of the universe. We report new limits on cosmological parameters derived from the angular power spectrum measured in the first Antarctic flight of the BOOMERANG experiment. Within the framework of inflation-motivated adiabatic cold dark matter models, and using only weakly restrictive prior probabilites on the age of the universe and the Hubble expansion parameter $h$, we find that the curvature is consistent with flat and that the primordial fluctuation spectrum is consistent with scale invariant, in agreement with the basic inflation paradigm. We find that the data prefer a baryon density $Ω_b h^2$ above, though similar to, the estimates from light element abundances and big bang nucleosynthesis. When combined with large scale structure observations, the BOOMERANG data provide clear detections of both dark matter and dark energy contributions to the total energy density $Ω_{\rm {tot}}$, independent of data from high redshift supernovae.

Figures (3)

• Figure 1: CMB angular power spectra, ${\cal C}_\ell \equiv \ell (\ell+1)\langle \vert T_{\ell m}\vert^2 \rangle/(2\pi)$, where the $T_{\ell m}$ are the multipole moments of the CMB temperature. The closed (green) circles show the B98 data. The magenta crosses are a radical compression of all the data prior to B98 into optimal bandpowers BJK98toco98mauskopf99, showing the qualitative improvement provided by B98 except in the $\ell \mathrel{\hbox{$\mathchar"218$} \hbox{$\mathchar"13C$}} 20$ DMR regime, where the COBE data are represented as a single bandpower (open black circle). (Note that the B98 and prior CMB points at $\ell =150$ lie on top of each other.) The smooth curves depict power spectra for several maximum likelihood models with different priors chosen from Table \ref{['parameters']}, with ($\Omega_{\rm {tot}},\omega_b,\omega_c,\Omega_\Lambda,n_s,\tau_c$) as follows: P1, short dashed line,(1.3,0.10,0.80,0.6,0.80,0.025); P4, dot-dashed line,(1.15,0.03,0.17,0.4,0.925,0); P8, short-long dashed line,(1.05,0.02,0.06,0.90,0.825,0); P11, solid line,(1.0,0.03,0.17,0.70,0.95,0.025). These curves are all reasonable fits to the B98+COBE data. For comparison, we plot a $H_0=68$, $\Omega_\Lambda=0.7$ "concordance model" which does not fit (dotted line labelled C), with parameters (1.0,0.02,0.12,0.70,1.0,0).
• Figure 2: Likelihood functions for a subset of the priors used in Table \ref{['parameters']}. Panel 1 (top left) shows the likelihood for $\Omega_k \equiv 1-\Omega_{\rm {tot}}$; the full-database (P1, dotted line) prefers closed models, but reasonable priors (P2, dashed blue line; P4, solid blue line; P0, dot-dash red line; note that P2 and P4 lie on top of one another in every panel in this plot but are distinct in Figure \ref{['6-panel-prior']}) progressively move toward $\Omega_k =0$. We caution the reader against agressively interpreting any 2$\sigma$ effects. Likelihood curves for $\Omega_{\Lambda}$ are shown in panel 2 (top center). In panels 2 and 4-6, the cases and line types are as in panel 1, except that dot-dashed now denotes the weak+LSS prior, P5. With weak priors applied, there is no significant detection of $\Omega_{\Lambda}$ (P2 and P4, overlapping as solid blue line in all remaining $\mathcal{L}(x)$ panels). Only by adding the LSS prior is $\Omega_{\Lambda}$ localized away from zero (P5, red dot-dash in all remaining $\mathcal{L}(x)$ panels). Panel 3 (top right) shows the contour plot of $\Omega_k$ and $\Omega_{\Lambda}$, for which the first two panels are projections to one axis. The bold diagonal black lines mark $\Omega_m$=1 and $\Omega_m$=0. The blue contours are those found with the weak prior (P4), plotted at 1, 2, and 3$\sigma$contourerrors. Red contours are similarly plotted for the weak+LSS prior (P5). SN1a constraints are plotted as the lighter (black) smooth contours, and are consistent with the CMB contours at the $1\sigma$ level. Panel 4 (bottom left) shows the contours for $\omega_b$; the full database analysis results in a bimodal distribution with the higher peak concentrated at very high values. These high $\omega_b$ models are eliminated by the application of a weak $h$ prior or weak BBN prior (P2 and P4, overlapping as blue here). Panel 5 (bottom center) shows a localization of $\omega_c$ for the weak $h$ and BBN prior cases, but this is partially due to the effect of the database structure coupling to the $h$ and age priors. Only the LSS prior (P5, red dot-dash) allows the CMB to significantly constrain $\omega_c$. Panel 6 (bottom right) shows good localization and consistency in the $n_s$ determination once any priors are applied. The inflation-motivated $\Omega_{\rm {tot}}$=1 priors (P10, P11) give very similar curves localized around unity. See Figure \ref{['6-panel-prior']} to see the effects of the database and priors on these curves.
• Figure 3: Likelihood functions similar to those in Figure \ref{['6-panel']}, but computed without using the B98 data. These curves show the effect of the database constraints and applied priors alone. The identification of the curves is the same as in Figure \ref{['6-panel']}, with the addition of the dotted magenta curve in panels 2-6, which shows the likelihood given weak priors and the COBE DMR data. In panel 3, only the 1$\sigma$ (red) contour is shown for the prior only and prior+DMR cases, while 1, 2 and 3$\sigma$ (light black) contours are shown for SN1a. The curves for P2 (solid blue) and P4 (dashed blue) are slightly separated in this figure, in contrast to Figure \ref{['6-panel']}, where they overlapped. Of particular interest here are the slope induced across $\Omega_k$, the slight localization of $\Omega_c h^2$ with the weak priors, and the significant localization of $\Omega_c h^2$ and $n_s$ with just weak+DMR+LSS (dotted magenta).