The formation of structure in the Universe is one of the most outstanding problems of Astronomy. The goal is to understand how small inhomogeneities generated in a primordial phase of the Universe give rise to the complex pattern of structure we observed today.
At IAS we have been using hydrodynamical N-body simulation methods and developing new numerical techniques to study a number of aspects of the formation of cosmological structure. These include performing large scale structure (LSS) simulations, and developing a number analysis tools to extract cosmological information from simulations.

Hydrodynamic N-body simulation methods constitute the most powerful tool to study the formation and evolution of Cosmological structure. These methods track the growth of structure by following the trajectories of a multitude of baryonic (gas) and dark matter particles evolving under the action of gravity, gas dynamics and other relevant processes, such as radiative gas physics and adiabatic cosmological expansion.
At IAS, we have been performing LSS simulations to study a number of aspects of the galaxy cluster physics and cosmology. These simulations have been carried out using modified versions of the publically avilable software codes GADGET and Hydra, and run at the national parallel computing centres of CINES/IDRIS. Small/medium sized simulations and post-simulation process tasks are performed locally using the IAS computing facilities.

 

Cosmological simulation  from the CLEF-SSH collaboration (IAS, LATT, Oxford, Sussex)

Simulation of systematic effects on CMB polarization measurements

Given the low amplitude of the expected CMB polarization signal, great care must be taken in the control of systematic effects. This is one of the major drivers in the design of the future CMB polarization dedicated missions. The difficulty in the simulation of these effects is that they are difficult to separate from one another and that there is in general no straightforward mathematical formalism to express them. One therefore often has to choose between important computational time or approximations. Although sophisticated methods to estimate these effects on real data have been developped e.g. in the Archeops collaboration, there still was a need for an intermediate formalism that enables the prediction of their amplitude as a function of intuitive instrumental parameters, e.g. beam ellipticity, beam differential size, gain offsets, pointing uncertainties...

We have developped a method to compute the systematic effects induced on temperature and polarization by an instrument as a function of such parameters. This formalism also accounts for the exact coupling between the beam and the scanning strategy of the considered experiment. This allows us to optimize the designs of the projects in which we are involved : BICEP, EBEX, Pilot, EPIC, SAMPAN.

Simulation of the polarization from diffuse dust thermal radiation
                                                                               
Large dust grains in the interstellar medium are aspherical and align mainly orthogonally with the local magnetic field. Both this asymetry and alignment polarized the thermal radiation of dust. This polarization was proven to be significant compared to the CMB by Archeops. However, few experimental data exist about dust polarization on large angular scales at high latitudes, those regions that are precisely the most relevant for CMB studies.
                                                                               
Based on our past experience on Archeops and expertise on the physics of dust and the interstellar medium, we are developping simulations of this polarized foregrounds, to help build intuition on this component and to developp techniques to remove it from CMB data (Planck-HFI in particular).

IAS members involved in this theme: F. Boulanger, M.-A. Milville-Deschenes, N. Ponthieu

Full sky simulated maps in I, Q, U

Simulation of IR galaxies

Using empirical models accounting fot the evolution of IR galaxies, we simulate maps of those sources including their correlation properties.

The simulations use the number counts and spectral energy distribution of the IAS model for galaxy evolution that are consistent with the galaxy model (see Lagache et al. (2003)). They include as well the correlation of the IR galaxies with the dark matter with a bias of 3. The dark matter power spectrum is calculated using the method presented in Knox et al (2001)

These are the maps for Spitzer (160 microns), 500 Herschel, and 550 Planck. The maps for Spitzer and Herschel simulate the same region of the sky of size 1024x1024 pixels (2 arcsec²/pixel). The Planck simulation does not
cover the same region of the sky as the former two, its size is 1024x1024 pixels of (25 arcsec²/pixel)