Introduction to Cosmology Example Exam Question: Answer

Lecture 18: Cosmic Microwave Background: Analysis

You are an ESA astronomer who has been asked to design a next generation Cosmic Microwave Background experiment. Discuss the properties you would like this experiment to have, carefully explaining why they are important in the context of determining cosmological parameters. [5]
The wish-list you might reasonably come up with is:
  • whole-sky coverage (for sensitivity to small values of l);
  • good angular resolution (for sensitivity to large values of l);
  • location away from Earth – at L2, as for WMAP and Planck, or in a heliocentric orbit like Spitzer (to avoid the infra-red background produced by the Earth, and to be in a cool enough environment not to need active cooling);
  • multiple receivers covering a range of frequencies (so that foregrounds can be subtracted off based on their different spectra);
  • multiple (well, 2, anyway) receivers at each frequency (to reduce the contribution of instrumental noise by cross-correlating different receivers);
  • sensitivity to polarisation (to cross-check the temperature results, using the TE spectrum – ideally, to look for gravitational-wave contributions, using EE and EB spectra);
  • long lifetime, to be able to repeat the whole-sky scan at least once, in order to cross-check any anomalies.
This is more than enough points. There is a trade-off of breadth against depth here: many very briefly described points, as here, are equivalent to fewer points with more detailed discussion. The sort of details you could add would be, for example, more detail about wny it is desirable to measure high values of l (more well-defined peaks help you to reduces the ambiguities in the fit), what sort of wavelengths/frequencies you might want to be sensitive to (2.7 K implies a peak wavelength of just over 1 mm, so you want both far-infrared (tens or hundreds of microns) and microwave (few mm – few cm) sensitivity to ensure you cover both sides of the peak), the level of angular resolution you are looking for (if the first peak at l ∼ 200 corresponds to ∼1° then to get several peaks you probably want to get out to l of several thousand, which implies an angular resolution of ∼1 arc minute or so; Planck's best angular resolution is apparently 5 arc minutes, corresponding to l ∼ 2500 or thereabouts).

The mark scheme for this question is 0.5–1.5 marks per point, depending on level of detail.
[5]

(2005 Resit, Q4.)

Go back to the lectures page.