The epoch of reionization (EoR) is associated with the formation of first stars and galaxies. The UV light from the first stars ionizes the intergalactic medium and the resulting free electrons leave signatures on a variety of probes such as the quasar spectra, cosmic microwave background (CMB) anisotropies and neutral hydrogen maps from 21 cm experiments. Studies of reionization have not only implications for our astrophysical understanding of stellar and galaxy formation, but also impact our interpretation of important cosmological parameters such as neutrino mass (for determining neutrino hierarchy) and cosmic acceleration.
Chen Heinrich’s work on reionization has focused on signatures of reionization in the CMB power spectra, by means of a principal component (PC) analysis. The PC analysis has advantages over conventional approaches by completely extracting all reionization signatures in the polarization data.
In Heinrich et al. 2016, Chen, together with Prof. Wayne Hu and Dr. Vinicius Miranda, derived PC constraints on the optical depth evolution during reionization with the Planck 2015 data, and found that the data allows for a high-redshift optical depth for z ~ 15 at two sigma. These results suggest that a PC analysis is necessary for obtaining unbiased results, especially when applied to the Planck 2017 data which could be the best large-angle CMB polarization maps we have for many years.
Subsequently, Chen worked with Dr. Miranda, Prof. Adam Lidz and Prof. Hu to derive implications of their results for early star-formation and pointed that, some models (e.g. Pop-III stars) that are excluded by previous approaches (which assume step-like or power-law forms for the ionization history), are actually allowed by the data in a PC analysis that is complete.
In particular, models including metal-free Pop-III stars that best-fit the data required a self-regularization mechanism to the Pop-III stars formation (e.g. UV dissociation of molecular hydrogen), leading to plateau like ionization fraction at high-redshift.
ReLike – A Reionization Likelihood for Fast and Effective Model Testing
For the first time, Heinrich et al. developed an effective likelihood code to provide fast and effective testing of ionization models using reionization PC amplitudes from the CMB. The first version ReLike v1.0 is available on this website by request.
The PC approach was first developed by Prof. Hu and Prof. Gil Holder in 2003. It is meant to capture all information of reionization present in the large-angle CMB polarization data with minimum number of parameters. Constraints on these parameters (PC amplitudes) can be turned into constraints on any model of ionization history.
The PC technique will come to its full fruition when applied to the final Planck 2017 data (better precision expected!). Stay tuned for ReLike v2.0!