Programação

  • Big Bang Nucleosynthesis & the Cosmic Microwave Background


    The second topic of our course will be the thermal history of our Universe, in particular Big Bang Nucleosynthesis and the  Cosmic Microwave Background (CMB). 

    After a revision of BBN we will see how the CMB can give us a glimpse on the early Universe, and how we can extract information from the observations.

    You should start by downloading the Einstein-Boltzmann codes that compute the CMB spectrum (as well as transfer functions that will be useful later, when we study structure formation). There are two main codes nowadays which are widely used:

    - Camb

    - CLASS

    As I wrote in an e-mail, a good guide to actually using, e.g., CAMB  in a working code was produced with the help of Hugo Camacho (thanks a lot, Hugo!) . Here is that guide, again:

    1) Basic documentation for CAMB:
    http://camb.readthedocs.io/en/latest/

    2) The last versions already come with a Python wrapper:
    https://github.com/cmbant/CAMB/tree/master/pycamb

    3) Check out the Python notebook with a demo:
    http://camb.readthedocs.org/en/latest/CAMBdemo.html

    You could also start looking at the CMB data. A good start would be to explore the ESA Planck website, which has the latest and most accurate measurements of the CMB to-date. Take a look at the Cosmology section of that website (http://pla.esac.esa.int/pla/#cosmology), and get a feeling about how the data is organized. I will provide you with the CMB data from Planck in a convenient format, but you could try yourself to download the FITS files and use them, if you are able to figure it out.


    Task for next classes

    1. Compute the chi^2 for the best-fit cosmological model of Planck (see the Planck documentation or the Planck papers), using the theory calculated using CAMB or CLASS, and comparing it to the Planck data (the links for the dataset appears below). Do this first for a simple model, not including running of the scalar index (i.e., take d n_s/ d lnk = 0).

    2. Now, compute the log-likelihood again, but searching for a smaller chi^2 if you vary the parameter d n_s/ d lnk . Use some criterium to decide whether or not it was worth the effort.

    3. Start thinking about how you could explore the space of parameters, in order to calculate a likelihood given the Planck data and some model. How would you do it? If you know, start doing it!


◄1: Supernovas
►Inflação