The WMAP space mission, launched by NASA in june 2001, has made available to the scientific community high signal-to-noise, high resolution observations of the sky in 5 frequency bands ranging from 23 to 94 GHz.
I have worked on WMAP data to separate CMB from galactic and extragalactic foregrounds, to measure the temperature and polarisation power spectra of the CMB emission (see below), and to measure scaling laws that connect the SZ flux Y to the X-ray luminosity and mass of clusters (see my SZ cluster page).
Component separation consists in identifying, in the observed frequency maps, what part of the radiation is due to which astrophysical process of emission. This is illustrated on a small patch of sky in the following figure, where we show how the WMAP K-band map at 23 GHz (top left) can be decomposed into a CMB part (top right), and a noisy foreground map (bottom left). The latter can be further processed to reduce contamination by instrumental noise, and obtain a clean foreground map (bottom right), here by Wiener-filtering the noisy foreground map with a latitude-dependent filter.
The extraction of the CMB map by needlet ILC is described below, and in this paper. The extraction of the foreground map has been done with Tuhin Ghosh while he spent a month at APC as a visiting student, and is described here.
Together with a postdoctoral fellow, Soumen Basak, we have applied the needlet ILC method to measure the temperature and polarisation power spectra of the CMB.
The analysis is done by first applying a needlet ILC to find the best set of linear combinations of needlet coefficients that reject foreground contaminations in maps of CMB intensity and of E and B-mode polarisation. These filters are then applied to each individual year of observation from the WMAP 7-year data set. Spectra are estimated as noise-weighted averages of (local, in pixel space) band-averaged covariances.
Results are published in the following papers: