For our second seminar of the new year, Miguel Zumalacárregui from the University of Barcelona talked to us about his recent work on constraining void models, based on his paper 1201.2790 with Juan Garcia-Bellido and Pilar Ruiz-Lapuente.
Miguel began by reviewing the conventional Lambda-CDM model of the universe which tells us that roughly 5% of the universe is ordinary matter, 25% dark matter and 70% vacuum energy. The second and third of these are somewhat more mysterious than the first, and provide the necessary grist for the mill of theoretical speculation.
One such speculation is that the apparent (and Nobel prize winning) acceleration of the universe might be caused by a failure of the Copernican principle. Miguel pointed out that there are several assumptions that go into the Lambda-CDM model, and therefore to test the model we should probe all of the said assumptions, including that of statistical homogeneity.
Just as a large over-density would slow down the apparent expansion rate for an observer sitting in the center of the over-dense region, a void will increase the rate for a similarly central observer. Since the isotropic CMB tells us we must be with 10 Mpc of the center of such a void, such an inhomogeneous universe is well described by the spherically symmetric Lemaitre-Tolman-Bondi metric.
After discussing the form of the LTB metric, Miguel explained the assumptions he and his collaborators made in analysing the sub-class of models described by the LTB metric and the void profile of Garcia-Bellido and Haugboelle. As well as the using the GBH parametrization, the analysis was further simplified by working with the constrained GBH profile, where the big bang is homogeneous and the perturbation is adiabatic.
Miguel then told us how cosmological data constrained this model of the void. First he argued that supernovae and BAO data could both be used to analyze void models, if suitable care were taken to account for subtleties that arose from working with a LTB rather than FRW metric.
The results from comparing MCMC analyses of the BAO data to supernovae data were that SNe and BAO disagreed at the level of 3 sigma when fitting the local matter fraction and Hubble parameter. Consequently, the particular constrained void model considered was ruled out by the evidence.
Miguel finished his talk by summarizing his conclusions that BAO and SNe strongly rule out constrained GBH void models and then by discussing the various ways that the result could be evaded. These included changing the profile, having an inhomogeneous big bang and large scale isocurvature. Finally, he concluded by reiterating that future BAO measurements would be a powerful probe of void models.