Abstract
In this thesis we are looking at the foreground cleaned maps of Smica, Sevem, NILC, and CR and trying to see if there are any residuals left after the cleaning. This is done by looking at the simulations for these methods and comparing them to the real Planck data. The simulations are used to make mean and standard deviation maps for each method, and also for combined maps where we take the difference of two of the maps so that we ideally only have the noise and foregrounds left. After that this is displayed on the form x-( x )/ σ x. When we look it on this form we can see that there are some discrepancies near the galaxy, and we can explore this closer by looking at correlation coefficient between the difference maps and its components to find which method(s) contains the residuals. Looking at the correlation coefficient we see that the NILC and Smica methods have residuals of the same size, and that the Sevem method contains slightly less foreground than the previously mentioned methods but the difference is too small to be significant. The combinations with CR are to spread out to be certain, but for j = 2 CR has the least amount of residuals. The residuals are in general too big to say anything specific.
Veileder: Professor Frode Kristian Hansen, Institutt for teoretisk astrofysikk, UiO
Intern sensor:Førsteamanuensis Boris Vilhelm Gudiksen, Institutt for teoretisk astrofysikk, UiO
Ekstern sensor: Forsker Elina Kaihänen, University of Helsinki, Department of Physics