Bolometric interferometry: the concept

In bolometric interferometry, bolometers do not see the sky directly but they see the interference pattern produced by the sky through a bunch of horns. The signal collected by each horn is separated into its two orthogonal polarization components, one of the polarization is rotated by 90 degrees so that they can interfere, and each channel is appropriately phase shifted. The signals from all different channels are combined in a beam combiner to produce the interference patterns. The combined signals are sent on a bolometer array. Bolometric interferometry is an additive interferometry: this means that the visibilities, the Fourier transform modes of the sky signal, must be separated from the total bolometer output. This can be done by modulating the phase shifts. The difficulty is to separate the 2N (N − 1) complex visibilities corresponding to all incoming couples of N primary horns and 2 polarizations (a couple of horns is called a baseline in interferometry).

The Quasi-Optical Combiner

The goal of the quasi-optical combiner is to add together a fraction of the signal coming from each of the 2N secondary back horns. It can be seen as a telescope located inside the cryostat, on the focal plane of which are the bolometer array: all parallel rays re-emitted by the back horns in the same direction reach the bolometer array on the same pixel. The additional geometrical phase shifts create the interference patterns.

Reconstructing the visibilities

The visibilities are reconstructed (independently for each bolometer) by solving a linear problem whose coefficients depend on phase shifts modulation. We have shown in [Charlassier et al., 2008] that these sequences must respect a particular phase shifting scheme we call "coherent summation of equivalent baselines". This scheme has been further optimized in [Hyland et al., 2008].


We have shown in [Hamilton et al., 2008] that a detector with the QUBIC specifications will be competitive with imaging experiments and heterodyne interferometers for the measurement of CMB B-modes. A bolometric interferometer appears to be roughly two times less sensitive than an imager with the same number of horns, but this seems to be the price to pay for having reduced systematics...

1 horn

Experimental validation

The work principle of bolometric interferometry has been demonstrated by the DIBO experiment [Ghribi et al., 2009]. The MBI-4 prototype [Tucker et al., 2008], whose design is very closed to the one of the QUBIC first module, has validated the work principle of bolometric interferometry with quasi-optical combining.

Instrument characteristics

The final QUBIC configuration can still evolve but a typical one would be as follows: 6 modules, each equipped with 144 input horns, (about 10000 baselines) and 288 back horns in a compact square array and a focal plane of about 900 Transition Edge Sensors. The primary beams would cover approximately 14 degrees. To check for foreground contamination we would have 3 frequency channels : 90, 150 and 220 GHz, with 25% bandwidth. The cryogenics would involve a 4K pulse tube for each module and a 100-300 mK dilution unit for the focal plane. The multipole range would be approximately 25 < l < 200. Such an instrument would allow to reach a tensor to scalar ratio of 0.01 in one year of continuous operation at Dome C in Antarctica.

1 baseline

1 baseline

1 baseline

total signal

Related links

See our publications about bolometric interferometry Publications.htmlshapeimage_6_link_0