A Clearer View of the Epoch of Reionisation

17 July 2017

The Epoch of Reionisation (EoR) is the transition from the ‘dark ages’, billions of years ago, before galaxies existed, to the modern era of rich galaxy populations. It is one of the last unexplored periods of cosmic history.

The best method we have for exploring the EoR is to measure the emission of un-ionised (neutral) hydrogen (HI), which was abundant in the early universe. HI gas produces a unique spectral line detectable at radio wavelengths. As early galaxies formed and heated this gas, it became ionised (the atoms lost their sole electron) and no longer produced the HI spectral line. 

When all of the neutral hydrogen was ionised, the signal vanished, marking the end of the EoR. Unfortunately, as well as being very distant, the HI signal is also very weak. Thus, we need massively powerful and carefully calibrated radio telescopes to detect the HI spectral line.

An animation of HI detection over cosmic time in a square patch of sky. The closer we look to earth (the smaller the "z" value), the less neutral hyrdrogen we detect (yellow), as compared to the Epoch of Reionisation.

The EoR also hides behind a gamut of obscuring ‘foregrounds’. The atmosphere, our Galaxy, and every other galaxy in the Universe together drown the signal at least a thousandfold. Thus, a major part of the EoR hunt is to develop sophisticated statistical methods to differentiate the signal from these foregrounds.

Because describing the foreground ‘noise’ is complex, only simplified models of the foregrounds have thus far been investigated: in particular, mathematical models have included the assumption that galaxies are uniformly distributed across the sky, whereas observations indicate that they tend to 'cluster' in a 'cosmic web'. In their new paper, Steven Murray (Curtin University) and his collaborators extend the mathematical description of the foreground ‘noise’ to include the clustering of foreground galaxies. 

A "top-down" view of tens of thousands of "clustered" galaxies as seen from the Milky Way. Imaged as part of the 2dF Galaxy Redshift Survey (credit to the 2dFGRS Team)

This analysis has produced a couple of important new results. First, including spatial clustering adds an extra term to the previous calculations, which tends to boost the predicted noise on large on-sky scales. The magnitude of this effect can be anywhere from insignificant to overwhelming, depending on the precise layout of the galaxies and their brightness distribution: the more faint sources there are in a given area of the sky, the more important it becomes to take clustering into account. Factoring this in will be crucial for future observations, such as those done with Square Kilometre Array.

Second, for certain choices of the population parameters, a failure to account for spatial clustering would lead to a false detection. Because of the mismatch between expected and actual noise, certain measurements could appear to be detections of the EoR signal when in fact they were the result of un-suppressed foregrounds. This finding is extremely important for upcoming analyses of observations designed to detect the EoR.


Publication details:

Steven Murray, Cath Trott, Christopher Jordan in The Astrophysical Journal (2017): "An Improved Statistical Point-Source Foreground Model for the Epoch of Reionization"