The opaque universe

The Epoch of Reionization (EoR) refers to the period in the history of the universe during which the predominantly neutral intergalactic medium was ionized by the emergence of the first luminous sources. These sources may have been stars, galaxies, quasars, or some combination of the above. By studying reionization, we can learn a great deal about the process of structure formation in the universe, and find the evolutionary links between the remarkably smooth matter distribution at early times revealed by CMB studies, and the highly structured universe of galaxies and clusters of galaxies at redshifts of 6 and below. The MWA is designed to provide detailed information on conditions in the intergalactic medium during and immediately preceding the EoR. In particular, Phase II with its two hexagonal sub-arrays with 72 tiles in a regular configuration, is designed to provide the precise calibration and high sensitivity required for this challenging experiment.

The diagram on the left provides a good graphical representation of the history of the universe, and where the Epoch of Reionization sits in the overall picture. After the Big Bang, the Universe was a hot, but quickly cooling soup of fundamental particles. After a few hundred thousand years, it cooled enough for protons and electrons to combine to form neutral hydrogen. This was a rather sudden event, and allowed the thermal glow of the fireball plasma, as it existed immediately before the hydrogen formation event, to radiate throughout the universe unimpeded by constant interactions with the charged particles of the now-absent plasma. This glow, redshifted by a factor of 1100 or so, is what we now observe as the Cosmic Microwave Background (CMB) in all directions. The CMB carries a frozen imprint of the density fluctuations in the early Universe, the study of which, by the observational cosmology community, is intense and sustained.

After the Universe became neutral, it became unobservable across much of the electromagnetic spectrum. Any short wavelength radiation that might have been emitted was quickly absorbed by the atomic gas, and a long interval known as the Dark Ages began. Slowly, gravitational collapse of overdense regions, the same regions we can see in the CMB imprint from earlier times, led to the formation of more and more pronounced structure in the neutral medium, and eventually the first stars, galaxies and quasars started to form. The exact mechanism and nature of this formation, poorly constrained by observation, is a topic of much research and great importance. We know what the Universe looked like at the time of the CMB, and we know what it looks like now, but how did it get from one to the other? In 2018, the EDGES experiment (Bowman et al., Nature) reported a detection of a large absorption trough in the observed temperature of this gas, when compared with the backdrop of the CMB photons, providing some early perplexing constraints on the rapidity of the growth and death of the first generation of stars. This preceded the Epoch of Reionization, in a period termed the Cosmic Dawn, but it helps set the conditions for how reionization proceeded.

Image: Graphical representation of the history of the universe, by Djorgovski et al. (Caltech).

As the collapse of structures proceeded, temperature variations developed. Gradually, energetic radiation emitted by the first sources caused local heating, and then ionization of the hydrogen in the Universe. It will have started with "bubbles" of ionized plasma surrounding the most energetic sources. As the bubbles grew and became more numerous, they started to overlap, and more and more of the neutral medium became exposed to the harsh ionizing radiation, which travels unimpeded through ionized regions.

The final phase of reionization of the Universe may have occurred swiftly. As soon as the bulk of the Universe was reionized, light at many wavelengths could escape from the early galaxies and quasars, revealing the distant Universe that we see today with optical and infrared telescopes.

Reionization was complete about 1 billion years after the Big Bang, corresponding to a redshift of about 6.5. Before that time, observations rapidly become more difficult. By and large, one must hope to find isolated, very luminous objects whose radiation in one form or another manages to reach us through the increasingly neutral medium. Perhaps the best hope for a more general and comprehensive probe of these early epochs is the 21cm hyperfine transition line of neutral hydrogen, redshifted to frequencies below 200 MHz. Sensitive observations of emission and absorption in this line can probe deeply into the reheating and reionization epochs, and give us a detailed view of the density, temperature and velocity field of the material. We would get a view, not just of isolated luminous objects and the material which happens to lie in front of them, but of large volumes of the Universe at the target redshifts. Such a view would yield a treasure trove of information from which to deduce the early history of structure formation, and the origin of the stars, galaxies, clusters and quasars that we see today. At this stage, the EDGES result is the only report of a detection of neutral hydrogen from this era.

The MWA is designed to deliver high-quality observations of the 21cm reheating and reionization signatures. The image to the right shows the sort of structure predicted to exist, and the MWA will look for such fluctuations using sophisticated statistical techniques, fueled by the precision treatment of an extraordinarily wide field of view that is the technological centerpiece of the MWA project.

Image: Simulation of redshifted 21cm emission/absorption at z~8.5, from Tozzi et al. (2000). The peak brightness is about 10 mK. The MWA EOR key project aims to characterize such structure, among other EOR diagnostics.

Detecting and Measuring the Statistical EOR Signal

The MWA does not have sufficient sensitivity to directly image individual features in the EoR HI signal. That capability must await a manyfold increase in physical collecting area. This lack of collecting area can, however, be effectively compensated for by increasing the effective field of view of the instrument. By measuring the intensity at many locations of the sky (and at many locations along the line of sight, by using different frequencies corresponding to different redshifts), we can obtain a statistical measure of the fluctuations even when the signal is too weak to see inany given pixel. This power spectrum measurement is similar to the type of measurement done by microwave background experiments, except that for the EoR we use three dimensions, not two. MWA has been designed to yield a high fidelity measurement of the EoR power spectrum shape using a few hundred hours of observing time.

Foreground Subtraction, Avoidance and Treatment

A major challenge for any EoR experiment is the fact that the signals being sought are far fainter than other types of emission from the sky. Filtering out or subtracting these troublesome foregrounds is a principal focus of instrument design and data analysis techniques. Fortunately, many of the strongest foregrounds have properties which are easily separable from the EoR signals, and one of the strongest discriminants is that most foregrounds are spectrally smooth. By contrast, the EoR signals are expected to show strong fluctuations over small ranges of frequency. One of the most difficult foregrounds to deal with is polarized emission from our own Galaxy. While it is intrinsically smooth, the clumpy interstellar medium imprints spatial and frequency structures on it which will be difficult to remove. A high precision polarimetric calibration is required, and has been an important design driver for the array.

MWA Results

The EoR collaboration of the MWA has been collecting data from 2013, in three primary observing fields across the sky. These fields are chosen to be relatively clean of foreground contamination, but also able to produce calibrated datasets. The collaboration has published several papers studying subsets of these data and producing statistical upper limits on the amount of signal from the EoR. With Phase II of the MWA being available from 2016, we have been able to use its increased capabilities to improve the calibration precision of the data, and produce high-quality science results. In tandem with the publication of limits on the conditions of the early Universe, the collaboration has undertaken a large and long-term effort to understand and treat the systematics in the data, which are strongly dominant over the signal, and are the ultimate limit to being able to make good progress in this challenging experiment. The results of these efforts are demonstrated in the improved understanding of the early Universe the MWA has delivered over the past five years.