Proving the Power of Dwarf Galaxies in the Early Universe Using Individual Stars

Y. Choi (ychoi[at]stsci.edu), K. Gilbert (kgilbert[at]stsci.edu), and K. Gordon (kgordon[at]stsci.edu)

Shortly after the Big Bang, our expanding universe cooled enough for electrons to be combined with protons and became dark and neutral. The emergence of first stars and galaxies ended the dark ages by depositing ionizing photons into the intergalactic medium (IGM). These high-energy photons from galaxies gradually reionized the neutral IGM over the epoch of reionization, and observational evidence indicates that our universe was completely ionized by a redshift of 6 (~1 billion years after the Big Bang) and became transparent (e.g., Gunn & Peterson 1965; Fan et al. 2006, Stark et al. 2010; McGreer et al. 2015; Planck Collaboration et al. 2016).

It might sound like a happy ending for the early cosmic history, but our story starts from here with one of the key questions regarding the cosmic reionization process: what is the role of dwarf galaxies in cosmic reionization? Until very recently, there was no consensus on the main contributors to cosmic reionization: faint star-forming dwarf galaxies versus active galactic nuclei (e.g., Wise & Cen 2009; Madau & Haardt 2015). Pushing the rest-frame ultraviolet (UV) observational frontier up to redshift of 9–12 has led to a growing belief that ionizing photons escaping from numerous faint star-forming galaxies might be sufficient to reionize the early universe (e.g., Bouwens et al. 2012; Oesch et al. 2013; Finkelstein et al. 2015; Robertson et al. 2015). Confirming whether faint star-forming galaxies were indeed responsible for cosmic reionization requires constraining the fraction of ionizing photons that escape (the "escape fraction") into the IGM. Unfortunately, it is impossible to directly detect escaping ionizing photons from galaxies at the epoch of reionization due to the high neutral fraction of the IGM (the fraction of the IGM of all elements—mostly hydrogen—that are neutral, capturing all the radiation). Therefore, looking at lower-redshift analogs is currently the only way to constrain the physics that controls the escape fraction.

However, the actual escape fractions have been difficult to constrain, often because the neutral fraction is uncertain. To be consistent with reionizing the universe by a redshift of 6, the escape fraction must be at least 10%–30% (e.g., Finkelstein et al. 2012; Bouwens et al. 2015). Despite many attempts to detect escaping ionizing photons from galaxies, the escape fraction has been poorly constrained at any redshift, and only a small fraction of galaxies out of many candidates show any evidence for leakage, with most at a level of a few percent up to about 10%. Escape fractions at levels of >20% have been measured only very recently for a small number of galaxies (e.g., Vanzella et al. 2012; Shapley et al. 2016; Izotov et al. 2018a).

The explanation for low escape fraction measurements is inherent difficulties with the direct measurement of escaping ionizing photons. These include contamination from low-redshift intervening galaxies, uncertain IGM transmission, large uncertainty in UV background subtraction, and narrow opening angles of optically thin holes that are misaligned with our line of sight (i.e., lowering the chance to detect leaked ionizing photons). Of the possible limitations, the most fundamental issue may be the lack of sufficient spatial resolution to capture the local variation of the escape fraction within a galaxy. Constraining the escape fraction requires measuring (1) the intrinsic ionizing photon production rate, and (2) either the photon absorption rate by the interstellar medium (ISM) or the number of leaked ionizing photons. Because ionizing photons produced by clustered O/B stars must propagate through the complex, dusty ISM before eventually escaping to the IGM, all of the quantities needed to measure the escape fraction are sensitive to the distribution of hot stars and the ISM topology, which both vary significantly with position within a galaxy.

In Choi et al. (2020), we have developed a new method for measuring the escape fraction utilizing individual stars and the spatially resolved ISM up to a scale of individual star-forming regions to overcome the challenges listed above. We applied this new technique to the resolved photometry of NGC 4214, a metal-poor UV-faint starburst dwarf galaxy, which is an excellent laboratory for evaluating the plausibility of early star-forming galaxies as the source of cosmic reionization. High-resolution imaging with the Hubble Space Telescope (HST) resolves individual stars in NGC 4214. Furthermore, multiwavelength imaging with HST allowed us to correct for dust star-by-star by modeling their broad stellar spectral-energy distributions and thus infer the intrinsic ionizing flux precisely.

Showing absorbtion vs escape for ionized gas
Figure 1: Schematic of our method for deriving the escape fraction of ionizing photons by measuring the rate of (1) intrinsic production of ionizing photons from stellar SED fitting, (2) consuming ionizing photons by neutral hydrogen from the Hα luminosity, and (3) absorption of ionizing photons by dust. The effect of the relative geometry between stars and dust can be taken into account by introducing the covering factor, the fraction of H Ⅱ surface covered by the dust. We note that the gas/dust distribution presented here is for illustration purposes only. (Derived from Choi, Y., et al. 2020.)

Specifically, using the BEAST (Gordon et al. 2016), we forward modeled the UV through near-infrared spectral energy distributions of ~83,000 resolved stars to infer their intrinsic ionizing flux outputs (Fig. 2). We then constrain the local escape fraction by comparing the number of ionizing photons produced by stars to the number that are either absorbed by dust or consumed by ionizing the surrounding neutral hydrogen in individual star-forming regions. The total number of intrinsic ionizing photons was measured by combining the ionizing photons produced by all stars in a region of interest. The number of ionizing photons absorbed by dust before escaping the region was measured by employing a dust covering factor. Finally, the number of ionizing photons consumed by neutral hydrogen was measured based on the extinction-corrected Hα luminosity of the region.

Predicted maps vs observation for GALAX FUV
Figure 2: Left: Comparison of the reconstructed FUV map (grayscale) at the GALEX FUV angular resolution, derived from the BEAST SED fitting of individual stars, with actual GALEX FUV observations (contours). They show an excellent agreement in their local and global morphology. Right: Predicted map of the intrinsic ionizing photon production rate at the GALEX FUV angular resolution. Red crosses denote the 73 stars more massive than 100 solar masses and hotter than 25,000 K. These stars produce about a quarter of the total intrinsic ionizing photons per second. (Derived from Choi, Y., et al. 2020.)

We found substantial spatial variation in the local escape fraction (0%–40%) across the galaxy, depending on the ISM morphology and residing stellar populations (Fig. 3). Integrating over the entire galaxy yields a global escape fraction of 25% (+16%/–15%). This value is much higher than previous escape fractions of ~zero reported for this galaxy based on the direct detection method using spatially unresolved observations. We conclude that the main causes of the previous zero escape fraction measurements are the misaligned viewing angle of low-density holes, along which ionizing photons escape, and the lack of knowledge of the relative star/dust geometric effects due to lack of spatial resolution. Finally, if we assume that NGC 4214 has no internal dust, like many high-redshift galaxies with low metallicities, we found a global escape fraction of 59% (an upper limit for NGC 4214). This is the first nonzero escape fraction measurement for UV-faint (MUV=−15.9) galaxies at any redshift. Our results support the idea that starburst UV-faint dwarf galaxies can provide a sufficient number of ionizing photons to reionize the IGM at the epoch of reionization, and thus are likely the main drivers of cosmic reionization.

Map of LyC escape DEC (x) RA (y)
Figure 3: Map of the LyC escape fraction. The underlying grayscale shows the extinction-corrected Hα emission at the full HST/WFC3 resolution. Solid lines denote the individual SF regions and are color-coded by their measured escape fractions. The local escape fraction varies from 0% to 40%. (Derived from Choi, Y., et al. 2020.)

References

Bouwens, R. J., Illingworth, G. D., Oesch, P. A., et al. 2012, ApJL, 752, L5

Bouwens, R. J., Illingworth, G. D., Oesch, P. A., et al. 2015, ApJ, 811, 140

Choi, Y., Dalcanton, J. J., Williams, B. F., et al. 2020, ApJ, 902, 54

Fan, X., Strauss, M. A., Becker, R. H., et al. 2006, AJ, 132, 117

Finkelstein, S. L., Papovich, C., Ryan, R. E., et al. 2012, ApJ, 758, 93

Finkelstein, S. L., Ryan, R. E., Jr+, Papovich, C., et al. 2015, ApJ, 810, 71

Gordon, K. D., Fouesneau, M., Arab, H., et al. 2016, ApJ, 826, 104

Gunn, J. E., & Peterson, B. A. 1965, ApJ, 142, 1633

Izotov, Y. I., Schaerer, D., Worseck, G., et al. 2018, MNRAS, 474, 4514

Madau, P., & Haardt, F. 2015, ApJL, 813, 1

McGreer, I. D., Mesinger, A., & D’Odorico, V. 2015, MNRAS, 447, 499

Oesch, P. A., Bouwens, R. J., Illingworth, G. D., et al. 2013, ApJ, 773, 75

Planck Collaboration, Ade, P. A. R., Aghanim, N., et al. 2016, A&A, 594, A13

Robertson, B. E., Ellis, R. S., Furlanetto, S. R., & Dunlop, J. S. 2015, ApJL, 802, L19

Shapley, A. E., Steidel, C. C., Strom, A. L., et al. 2016, ApJL, 826, L24

Stark, D. P., Ellis, R. S., Chiu, K., Ouchi, M., & Bunker, A. 2010, MNRAS, 408, 1628

Vanzella, E., Guo, Y., Giavalisco, M., et al. 2012, ApJ, 751, 70

Wise, J. H., & Cen, R. 2009, ApJ, 693, 984