A measurement by BOOMERANG of multiple peaks in the angular power spectrum of the cosmic microwave background

TL;DR

BOOMERANG measures the CMB angular power spectrum from $l=75$ to $1025$ using four 150 GHz channels, revealing a harmonic series of acoustic peaks predicted by inflationary, adiabatic models. A MASTER-based pipeline with improved beam models and dipole-calibration yields calibrated $C_\ell$ with quantified uncertainties, demonstrating consistency across independent analyses. Within weak priors, the data constrain $\Omega_{tot}$ to unity and $\Omega_b h^2 \approx 0.022$, with $n_s \approx 0.96$, and, when combined with LSS and SN1a priors, infer $\Omega_\Lambda \approx 0.65$–0.70 and $\Omega_c h^2 \approx 0.12$, supporting a flat, $\Lambda$CDM universe aged $\sim 13$–$15$ Gyr. These results solidify the role of CMB peak structure in constraining cosmology and highlight the presence of dark energy and dark matter, while aligning with inflationary predictions and improving parameter precision.

Abstract

This paper presents a measurement of the angular power spectrum of the Cosmic Microwave Background from l=75 to l=1025 (~10' to 5 degrees) from a combined analysis of four 150 GHz channels in the BOOMERANG experiment. The spectrum contains multiple peaks and minima, as predicted by standard adiabatic-inflationary models in which the primordial plasma undergoes acoustic oscillations. These results significantly constrain the values of Omega_tot, Omega_b h^2, Omega_c h^2 and n_s.

Figures (4)

• Figure 1: Sky coverage. The upper panel shows the BOOMERANG 150GHz map. The locations of the three bright quasars are circled. The sky subset used in B00 (rectangle) and this paper (ellipse) are shown. The bottom panel shows the integration time/pixel. The ellipse for this analysis was chosen to include the well sampled sky, and to avoid the poorly sampled sky
• Figure 2: The angular power spectrum of the CMB, as measured at 150 GHz by BOOMERANG. The vertical error bars show the statistical + sample variance errors on each point. There is a common 10% calibration uncertainty in temperature, which becomes a 20% uncertainty in the units of this plot. The points are also subject to an uncertainty in the effective beam width of $\pm1.4'$ ($1 \sigma$). The effect of a $1\sigma$ error in the beam width would be to move the red points (all up together if the beam width has been underestimated, or all down together if the beam width has been overestimated) to the positions shown by the black triangles. The blue points would move in a similar fashion. The blue and red points show the results of two independent analyses using top-hat binnings offset and overlapping by 50%. This shows the basic result is not dependent on binning. While each of the independent spectra (red circles or blue squares) are internally nearly uncorrelated, each red point is highly correlated with its blue neighbors, and vice versa. These data are listed in Table \ref{['table:spec75']}
• Figure 3: Selected best fit models normalized to the best compromise amplitude between COBE-DMR and BOOMERANG are shown overlayed on the BOOMERANG spectrum. The upper panel shows the points plotted as listed in table \ref{['table:spec']}. The best-fit models for the "weak" and "no priors" cases coincide (blue, solid curve) with $\Omega_{tot} = 1.05, H_0 = 50, \Omega_{\Lambda} = 0.5, \omega_b = 0.020, \omega_c = 0.120, \tau_c= 0, n_s=0.925, t_0 = 15.8 Gyrs$. Strong Hubble prior gives the best fit model with parameters $\Omega_{tot} = 1., H_0 = 68, \Omega_{\Lambda} = 0.7, \omega_b = 0.020, \omega_c = 0.120, \tau_c= 0, n_s=0.925, t_0 = 13.8 Gyrs$. The full analysis takes into account the calibration and beam uncertainties which best fit models take advantage of. This explains the apparent mismatch between some of the models in the upper panel and the plotted central values of Boomerang band powers. The green (dash-dot) curve is the best fit model ($\Omega_{tot} = 1.15, H_0 = 42, \Omega_{\Lambda} = 0.7, \omega_b = 0.020, \omega_c = 0.060, \tau_c= 0.2, n_s=0.925, t_0 = 20 Gyrs$) when both beam and calibration uncertainties are switched off. The model fits closely the central values as expected. To demonstrate the effect of beam and calibration uncertanties, in the lower panel the data points have been replotted with a $4\%$ decrease in calibration ($0.4 \sigma$) and a 0.5 arcminute change in beam size ($0.4 \sigma$). The plot makes it clear that the best-fit conventional CDM models are indeed good fits to the data, once these uncertainties are correctly accounted for.
• Figure 4: Likelihood functions for a subset of the priors used in Table \ref{['table:parameters']}. $\Omega_bh^2$, and $n_s$ are well constrained, even under the "whole database" case, and are insensitive to additional priors. $\Omega_{\rm tot}$ is poorly constrained over the whole database, but when the weak priors are applied, it becomes stably consistent with the flat case. With the weak priors, $\Omega_\Lambda$ and $\Omega_c h^2$ are poorly constrained, but become significant detections with the addition of the other priors considered. For all cases, $\tau_c$ is poorly constrained, but does prefer low values.