We investigate analytically whether in a close encounter with a star, a Dark Matter particle can be accelerated above the escape speed of a Globular Cluster and be ejected. We find that this mechanism is not sufficient to eject a massive, extended Dark Matter halo by the present time. Combined with observations of isolated Globular Clusters that may not have had their halos tidally stripped, these results cast doubt on the scenario in which Globular Clusters formed in Dark Matter halos.

White Dwarfs (WD) capture Dark Matter (DM) as they orbit within their host halos. These captured particles may subsequently annihilate, heating the stellar core and preventing the WD from cooling. The potential wells of WDs are considerably deeper and core temperatures significantly cooler than those of main sequence stars. Consequently, DM evaporation is less important in WDs and DM with masses $M_{\chi} \gtrsim 100\, \kev$ and annihilation cross-sections orders of magnitude below the canonical thermal cross-section ($\sigmav \gtrsim 10^{-46}\, \cm^3$/s) can significantly alter WD cooling in particular astrophysical environments. We consider WDs in globular clusters (GCs) and dwarf galaxies. If the parameters of the DM particle are known, then the temperature of the coolest WD in a GC can be used to constrain the DM density of the cluster's halo (potentially even ruling out the presence of a halo if the inferred density is of order the ambient Galactic density). Recently several direct detection experiments have seen signals whose origins might be due to low mass DM. In this paper, we show that if these claims from CRESST, DAMA, CDMS-Si, and CoGeNT could be interpreted as DM, then observations of NGC 6397 limit the fraction of DM in that cluster to be $f_{\mathrm{DM}} \lesssim 10^{-3}$. This would be an improvement over existing constraints of 3 orders of magnitude and clearly rule out a significant DM halo for this cluster...