Mahlet Shiferaw

I conducted galaxy formation research as part of the 2018 Summer Research Experiences for Undergraduates (REU) program at Northwestern University's Center for the Interdisciplinary Exploration of Astrophysics (CIERA). I worked with Claude-Andre Faucher-Giguere, Aaron Geller, and Alex Gurvich on developing Firefly, an interactive web-based visualization application for particle-based simulation data.

Firefly allows the user to quite literally explore the data by zooming in, rotating around, re-sizing, coloring, and even filtering out different particle types. This level of sheer interactivity allows us to gain an intuitive understanding of the data in a way that traditional plots and figures simply do not allow. During the REU program, I worked on implementing new colormapping functionality into Firefly's GUI. This feature now allows the user to apply a unique colormap to each particle type based on a certain attribute.

Try Firefly

View my Poster

Find us on Github

Firefly has enabled us to examine the chemical processes involved in galaxy formation, and in particular, how regions of ionized hydrogen spatially relate to young stars. Using cosmological zoom-in galaxy formation simulation data from the Feedback in Realistic Environments (FIRE) Project, we studied a Milky Way-mass isolated galaxy with stars and gas only, at a snapshot of 500 Myr.

We used CHIMES, a time-dependent chemistry module developed by Alex Richings which predicts the chemical abundances of such galaxy simulations. Using this data, we hoped to see how well these HII regions matched the Stromgren Radius, an idealized model which predicts perfectly spherical bubbles of HII that form around young stars due to their strong ultraviolet radiation.

$$R_S \simeq \left({3N \over 4{\pi}\alpha}\right)^{1/3} n_H^{-2/3}$$

To explore these HII regions in Firefly, we filter out stars younger than 10 Myr, apply a blue-white-red divergent colormap, and zoom in to find dense bubbles of highly ionized gas that have formed around these young stars. We aso notice a sharp cutoff between neutral and ionized hydrogen, made apparent by a dramatic change in color, just as the Stromgren Radius predicts.

Performing a more quantitative analysis, however, reveals that the Strömgren Radius is consistently smaller than the radius observed in the simulation data, which we refer to as the “Ionizing Radius”. This is because the Strömgren model does not take into account the effects of stellar clustering: it assumes isolated stars, while in the simulation, as well as in real observations of star-forming regions, we find that young stars tend to cluster together.

This stellar clustering greatly increases the ionizing luminosity of an HII region, and as a result, the Ionizing Radius as well; the canonical Strömgren Radius therefore predicts too small a radius. In the future, we hope to also examine other chemical abundances from our simulation, such as H2 and CO. We also aim to further improve Firefly, and visualize how exactly these molecular clouds relate to HII regions around young stars.