[Paper Review] CLOVER - A new instrument for measuring the B-mode polarization of the CMB

Angela C. Taylor, A. Challinor+12 more|arXiv (Cornell University)|Jul 7, 2004

Radio Astronomy Observations and Technology11 citations

TL;DR

CLOVER is a three-telescope mm-wave instrument designed to measure B-mode polarization of the cosmic microwave background at Dome C, Antarctica, using 128 background-limited TES bolometers per telescope at 90, 150, and 220 GHz. It achieves a lensing-confusion-limited sensitivity to gravitational wave signals, enabling a tensor-to-scalar ratio detection limit of r ≈ 0.005 over ℓ = 20–1000.

ABSTRACT

We describe the design and expected performance of Clover, a new instrument designed to measure the B-mode polarization of the cosmic microwave background. The proposed instrument will comprise three independent telescopes operating at 90, 150 and 220 GHz and is planned to be sited at Dome C, Antarctica. Each telescope will feed a focal plane array of 128 background-limited detectors and will measure polarized signals over angular multipoles 20 < l < 1000. The unique design of the telescope and careful control of systematics should enable the B-mode signature of gravitational waves to be measured to a lensing-confusion-limited tensor-to-scalar ratio r~0.005.

Motivation & Objective

• To detect primordial B-mode polarization from gravitational waves generated during cosmic inflation.
• To overcome extreme sensitivity and systematic challenges in measuring sub-muK-level B-mode amplitudes.
• To achieve lensing-confusion-limited sensitivity by minimizing systematics and enabling spectral subtraction of foregrounds.
• To deploy a multi-frequency, multi-telescope array at a high-altitude, dry site for optimal observing conditions.
• To enable a definitive test of inflationary models through detection of the tensor-to-scalar ratio r.

Proposed method

• CLOVER employs three independent telescopes operating at 90, 150, and 220 GHz to enable spectral subtraction of Galactic foregrounds.
• Each telescope uses four co-pointed optical assemblies, each with an 8×8 feed-horn array, producing 256 total horns but only 64 simultaneously observed pixels.
• Polarization is measured via pseudo-correlation receivers using phase modulation to extract Stokes parameters Q and U without moving parts.
• The phase modulation scheme allows instantaneous measurement of Q and U, while intensity I is measured via telescope scanning.
• Signals from common pixels across four optical assemblies are incoherently summed and detected by voltage-biased transition edge sensors (TES) at 300 mK.
• TES detectors are read out via frequency multiplexing, with eight detectors per SQUID, enabling high dynamic range and low noise operation.

Experimental results

Research questions

• RQ1 Can a multi-frequency, multi-telescope CMB polarimeter achieve the sensitivity and systematics control required to detect primordial B-mode polarization?
• RQ2 What is the minimum detectable tensor-to-scalar ratio r achievable with a ground-based instrument at Dome C?
• RQ3 How effectively can foreground contamination from Galactic emissions be subtracted using multi-frequency observations?
• RQ4 To what extent does weak gravitational lensing of E-modes limit the detection of primordial B-mode signals?
• RQ5 Can a compact, rotating telescope design with co-pointed optical assemblies improve systematics calibration and sky coverage?

Key findings

• CLOVER is projected to achieve a one-sigma detection limit for the tensor-to-scalar ratio of r ≈ 0.0037, limited by sample variance from lensing.
• The instrument’s effective sensitivity after foreground subtraction is approximately 0.24 μK per 15-arcmin resolution element in Stokes parameters Q and U.
• The thermal noise contribution after foreground removal is expected to be 0.24 μK per resolution element, with 0.1 μK arising from lensing B-modes and 0.3√r μK from primordial gravitational waves.
• For a two-year observing campaign at Dome C, Clover can measure the B-mode power spectrum over the multipole range ℓ = 20–1000 with a lensing-confusion-limited sensitivity.
• The design enables cross-checks on systematics through rotation of the optical structure, improving calibration and sky coverage density.
• The instrument’s performance is limited not by instrumental noise but by the intrinsic lensing confusion of E-mode power, setting the ultimate sensitivity floor for r detection.