Matthew Cornachione Tuesday, May 1, 2018 at 10:00 AM (110 INSCC)
Title: Quantifying the Dark Matter Substructure Mass Function Using Strongly Lensed Lyman-Alpha Emitter Galaxies
One of the greatest mysteries in modern day astrophysics is that of the nature and composition of dark matter. Numerous lines of research have sought to detect and characterize this unknown material. Cosmological simulations in particular have proven very successful at re-creating observed large scale structure given a few simple assumptions about dark matter.
These simulations, however, appear to fail on smaller scales. Tension between simulations and observations still exists in the predicted abundance of dark matter subhalos, small, gravitationally-bound dark matter halos. These subhalos are believed to host dwarf galaxies in the Milky Way. The observed abundance of dwarf galaxies, however, is too low by an order of magnitude.
Strong gravitational lensing allows us to probe subhalo abundance beyond the Milky Way. This technique has detected dark matter substructure in several distant galaxies, allowing us to empirically tested cosmological simulations. This thesis extends of current results using a particular class our source galaxies known as Lyman-alpha Emitters (LAEs) to achieve lower mass detection thresholds, down to 1e7 solar masses. Our sample of high-resolution HST images promises the greatest substructure detection power of any sample to date.
I develop the statistical framework to measure the mass fraction and slope of the subhalo mass function (SHMF). My results give a substructure mass fraction f=0.0021 (+0.0030, -0.0014) which is lower than previous studies, but consistent with both observational results and theoretical predictions. For the slope I find alpha = 0.971 (+0.506, -0.545) which is significantly lower than cosmological predictions. Given this tension, I suspect that we may not yet be resolving the lowest mass substructure. I propose an alternative analysis that may definitively determine whether this apparent difference is real or an artifact introduced in data reduction.