Do galaxies that leak ionizing photons have extreme outflows?

The early universe made a few phase transitions. Initially, the universe was hot and highly ionized, but it cooled as it expanded such that hydrogen recombined to form the Cosmic Microwave Background Radiation. 

If this was the end of the story, the present-day universe would look very different. Neutral hydrogen absorbs photons blueward of Lyman alpha, and if the present universe was neutral we would not be able to observe galaxies in the far ultra-violet. Fortunately, we can observe galaxies bluward of Lyman alpha because the universe was reionized at some point early in its history (see the picture to the right for our best understanding of this timeline). 

An ionized universe is surprising because there must be an ionizing source (i.e. something must ionize the universe). Theoretically, this is challenging because even small amounts of neutral hydrogen within galaxies absorbs nearly all of the ionizing photons. Some suggest that star formation can inject energy and momentum into the neutral gas within the galaxy, and eject the gas out as a large scale galactic outflow. Removing this gas creates low density paths through which ionizing photons "leak" out of galaxies. The goal of this work is to determine whether galaxies that are known to leak ionizing photons have extreme galactic outflows as compared to "normal" local galaxies.  If you'd prefer to just skip right to the paper, it has been accepted for publication in Astronomy & Astrophysics, just click the button below.
I have also worked with a graduate student (Simon Gazagnes) to characterize the escape fraction from the Lyman Series (UV H I absorption lines). 
Picture
Schema of the evolution of the universe. The early universe transitioned from being neutral to being nearly fully ionized. How ionizing photons escape galaxies is the subject of this work.
Picture
The comparison of the Si II and Si III equivalent width (EW). The equivalent width is a proxy for the strength of the transition. Galaxies that are known to leak ionizing photons are in red, and the control sample is in blue. Leakers have small equivalent widths.

Leakers have small silicon equivalent widths

To determine the outflow properties of galaxies, we composed a sample of all of the nine known local galaxies that leak ionizing photons and compared their Si II and Si III absorption line properties to galaxies that are unlikely to leak ionizing photons.

Si II is an important tracer of partially neutral gas, and the equivalent width acts as a proxy of the strength of the transition. Unfortunately, the equivalent width is hard to interpret because a small equivalent width can be produced with a small amount of gas or a very clumpy gas distribution.

To the left I am plotting the Si II and Si III equivalent width of galaxies that leak ionizing photons (red points) and galaxies that do not leak ionizing photons (blue points). Galaxies that leak ionizing photons have small silicon equivalent widths. This observational fact can act as a diagnostic of leaking galaxies, but it's interpretation is challenging. ​

Leakers' equivalent widths are largely set by their metallicities

While the interpretation of equivalent widths is challenging, host galaxy properties give clues to what causes the small silicon equivalent widths. In previous work, I showed that the outflow equivalent width strongly scaled with star formation rate (SFR) and stellar mass (M*). Since this new leaking sample spans a different range of SFRs and M*, we are able to further test these relations.

The figure to the right shows that the equivalent width does not scale strongly with SFR, but it does scale strongly with the metallicity. Not only does this have important consequences for leaking photons, it provides tighter constraints for galactic outflows. 

The leakers' small silicon equivalent widths are likely because they have small metallicities. This has important consequences for the early universe when the average metallicities were also much lower. ​
Picture
The Si II equivalent width versus the star formation rate (SFR; upper panel) and metallicity (lower panel) of its host galaxy. The metallicity correlates strongly with the Si II equivalent width, but not with the SFR. The red points are galaxies that leak ionizing photons.
Picture
The maximum outflow velocity versus stellar mass. The leakers, shown in red, follow similar velocity trends as the non-leaking galaxies. Leakers do not have extreme outflow velocities.

Leakers do not have extreme outflows

An important characteristic of galactic outflows is how fast the outflow is moving. To characterize the maximum measurable velocity, we measure the velocity at 90% of the continuum. This velocity correlates strongly with the stellar mass (see plot to the left), and the leakers do not vary statistically from the control sample trends. 

This demonstrates that the confirmed leakers do not have extreme outflows, rather they have similar outflow properties to galaxies with similar stellar masses, star formation rates, and metallicities. 

The silicon equivalent width scales with Lyman alpha properties

The key question in this work is what the outflow properties, as measured by the silicon absorption lines, tell us about ionizing photons escaping galaxies. It has been shown that the escape of Lyman Alpha photons and ionizing photons are related, so we studied how the outflow properties scale with Lyman alpha properties.

The velocity separation of double peaked emitters correlates with silicon equivalent widths (see the plot to the right). Theory suggests that the peak separation is related to the neutral hydrogen column density, and this strong correlation could indicate that leakers have small Si II equivalent widths because they have low H I column densities.

This suggests that galaxies leak ionizing photons because they have less neutral hydrogen gas. 

​
Picture
The correlation between the velocity separation of double peaked Lyman Alpha emission lines and the Si II equivalent width. The velocity separation is thought to be related to the H I column density, implying that galaxies leak ionizing photons because they have low H I column densities.

Summary

Here I have summarized a paper that compared the silicon outflow properties between galaxies that leak ionizing photons (red points in all the graphs above) and those that don't (blue points).

The main points of the study are:
  1. Leakers do not have higher outflow velocities than non-leaking galaxies
  2. Leakers have smaller silicon equivalent widths than non-leaking galaxies
  3. The silicon equivalent width is strongly correlated with metallicity
  4. The silicon equivalent width is strongly correlated with Lyman Alpha velocity separation, which is suggested to correlate with neutral hydrogen column density
​ 
These observations have strong implications for how ionizing photons escape galaxies to reionize the universe. These observations suggest that ionizing photons escape through something called "density-bounded nebula" instead of through holes created by outflows (see the diagram to the right). This implies that photons escape because there is not enough gas to absorb the ionizing radiation, not because there are holes in the gas.

These observations have important implications for which galaxies should emit ionizing photons: low-metallicity, compact, galaxies that are highly ionized. These observations will guide future surveys trying to characterize the escape of ionizing photons. 

This situation is actually more complicated than these initial observations suggested.
Picture
Image from Zackrisson et al. 2013 depicting the two scenarios for the escape of ionizing photons: through holes created by outflows (left picture) or through a region that does not have enough gas to absorb all of the photons (the right picture). Our observations favor the low column density region scenario.