Abstract
We report measurements of the CMB polarization power spectra from the 2003 January Antarctic flight of BOOMERANG. The primary results come from 6 days of observation of a patch covering 0.22% of the sky centered near R.A.=82.5°, decl.=-45°. The observations were made using four pairs of polarization-sensitive bolometers operating in bands centered at 145 GHz. Using two independent analysis pipelines, we measure a nonzero 〈EE〉 signal in the range 201 < l < 1000 with a significance of 4.8σ, a 2σ upper limit of 8.6 μK² for any 〈BB〉 contribution, and a 2σ upper limit of 7.0 μK² for the 〈EB〉 spectrum. Estimates of foreground intensity fluctuations and the nondetection of 〈BB〉 and 〈EB〉 signals rule out any significant contribution from Galactic foregrounds. The results are consistent with a ΛCDM cosmology seeded by adiabatic perturbations. This is the first detection of CMB polarization with bolometric detectors.
Introduction
While the temperature fluctuations of the Cosmic Microwave Background (CMB) have been studied extensively to map the early universe, its polarization holds a wealth of additional information. Polarization is generated by Thomson scattering during the epochs of recombination and reionization. Specifically, this study sought to measure the $E$-mode (curl-free) polarization spectrum. In a universe dominated by adiabatic perturbations, the peaks in the $E$-mode spectrum are predicted to be roughly 180 degrees out of phase with the peaks in the temperature spectrum. Measuring this signal confirms our understanding of the physics of the photon-baryon fluid just 380,000 years after the Big Bang and helps constrain parameters like the timing of reionization.
Methods
The team utilized the BOOMERANG telescope, a balloon-borne instrument launched from McMurdo Station, Antarctica. Key technical aspects included:
- Polarization Sensitive Bolometers (PSBs): BOOMERANG used pairs of bolometers at 145 GHz designed to measure orthogonal linear polarization states directly, proving that bolometric detectors—which are more scalable than interferometers—could reach the required sensitivity.
- Dual Analysis Pipelines: Data was processed through two independent pipelines: the "NA" (North American) pipeline using the Xfaster quadratic estimator, and the "IT" (Italian) pipeline based on the MASTER method.
- Consistency Checks: Rigorous "jackknife" tests were performed by splitting data in time and by detector position to ensure the signal was not an artifact of noise or systematic drift. 500 Monte Carlo simulations characterized the signal transfer function and noise bias.
Observations and Data
The primary dataset consists of 205 hours of CMB observations from the 2003 (B03) flight, divided into two main regions:
- Deep Region: A 0.22% patch of the sky where integration time per pixel was maximized to boost the signal-to-noise ratio for polarization.
- Shallow Region: A larger 1.8% patch used primarily for temperature anisotropy and cross-correlation analysis.
Multi-frequency checks at 245 and 345 GHz were used to monitor and rule out contamination from Galactic dust.
Results
The primary result is the detection of a nonzero $\langle EE \rangle$ signal in the multipole range $201 < l < 1000$ with a statistical significance of 4.8$\sigma$. The measured $E$-mode peaks align with the troughs of the temperature spectrum, exactly as predicted for a universe seeded by adiabatic fluctuations.
The study also placed 2$\sigma$ upper limits of 8.6 $\mu$K² on the $\langle BB \rangle$ spectrum and 7.0 $\mu$K² on the $\langle EB \rangle$ spectrum. The results are highly consistent with a $\Lambda$CDM cosmology, with a "Probability to Exceed" (PTE) of 0.63 for the NA pipeline. The null hypothesis (no polarization) was ruled out with a PTE of $1.3 \times 10^{-3}$.
Discussion and Conclusion
This measurement was a milestone, proving that bolometric missions (like the Planck satellite) could successfully map the polarized sky. However, limitations remain. The "deep" patch covered only 0.22% of the sky, limiting constraints on large-scale features. Furthermore, the instrument was not sensitive enough to detect $B$-modes, which are much fainter and could provide evidence for primordial gravitational waves from inflation. Future experiments will require even tighter control over "T-to-P leakage" to detect these weaker signals.
Summary updated