MILCA, a modified internal linear combination algorithm to extract astrophysical emissions from multifrequency sky maps

This analysis of current cosmic microwave background (CMB) experiments is based on the interpretation of multifrequency sky maps in terms of different astrophysical components and it requires specifically tailored, component separation algorithms. In this context, internal linear combination (ILC) methods have been extensively used to extract the CMB emission from the WMAP multifrequency data. We present here a modified internal linear component algorithm (MILCA) that generalizes the ILC approach to the case of multiple astrophysical components for which the electromagnetic spectrum is known. In addition, MILCA corrects for the intrinsic noise bias in the standard ILC approach and extends it to a hybrid space-frequency representation of the data. It also allows us to use external templates to minimize the contribution of extra components but still using only a linear combination of the input data. We applied MILCA to simulations of the Planck satellite data at the frequency bands from 100 GHz to 857 GHz. We explore the possibility of reconstructing the Galactic molecular CO emission and the thermal Sunyaev-Zeldovich effect from the Planck maps. We conclude that MILCA is able to accurately estimate those emissions, and it has been successfully used for this purpose within the Planck collaboration.