Ursa Major II – reproducing the observed properties through tidal disruption

Abstract

Recent deep photometry of the dwarf spheroidal Ursa Major II's morphology, and spectroscopy of individual stars, have provided a number of new constraints on its properties. With a velocity dispersion ∼6 km s −1 , and under the assumption that the galaxy is virialised, the mass-to-light ratio is found to be approaching ∼2000 -apparently heavily dark matter dominated. Using N-Body simulations, we demonstrate that the observed luminosity, ellipticity, irregular morphology, velocity gradient, and the velocity dispersion can be well reproduced through processes associated with tidal mass loss, and in the absence of dark matter. These results highlight the considerable uncertainty that exists in measurements of the dark matter content of Ursa Major II. The dynamics of the inner tidal tails, and tidal stream, causes the observed velocity dispersion of stars to be boosted to values of >5 km s −1 . These dispersion boosts occur at each apocentre, and last throughout the time the galaxy is close to apocentre. The model need not be close to destruction to have a boosted velocity dispersion. We additionally note that the velocity dispersion at apocentre is periodically enhanced substantially (e.g >20 km s −1 ). This occurs most strongly when the model's trajectory is close to perpendicular with the Galaxy's disk at pericentre. This effect is responsible for raising the velocity dispersion of our model to (and beyond) the observed values in UMaII. We test an iterative rejection technique for removing unbound stars from samples of UMaII stars whose positions on the sky, and line-of-sight velocities, are provided. We find this technique is very effective at providing an accurate bound mass from this information, and only fails when the galaxy has a bound mass less than 10% of its initial mass. However when <2% mass remains bound, mass overestimation by >3 orders of magnitude are seen. Additionally we find that mass measurements are sensitive to measurement uncertainty in line-of-sight velocities. Measurement uncertainties of 1-4 km s −1 result in mass overestimates by a factor of ∼1.3-5.7.

Could Segue 1 be a destroyed star cluster? – a dynamical perspective

“…Even though those three methods do not agree very well with each other, they all show a boosted dispersion by an order of magnitude compared to the bound particles alone. As shown in e.g., Smith et al (2013) the boosting is higher the closer the object is to destruction and the closer we see the object to its apo-galacticon, when the tails get compressed. Both effects are at work at our solution for Segue 1.…”

“…• If sufficient mass is lost (usually more than 90%; see, e.g., Smith et al 2013) the object is at the brink of destruction and a final peri-centre passage will completely dissolve the object, turning it into a pure stellar stream. Now, for this study, we are more interested in the general trends, how the initial parameters influence the results of the simulations:…”

“…We repeat the analysis on our best fitting model (see next section) with different observational methods as described in Smith et al (2013) and find all methods to lead to extremely boosted velocity dispersions.…”

“…If Segue 1 is a globular cluster that is undergoing tidal disruption, ⋆ rdominguez@astro-udec.cl then extra-tidal stars may not be so easy to distinguish from gravitationally bound members. Studies of Smith et al (2013); Blaña et al (2015, e.g.,) have shown that objects on the brink of destruction and/or close to apo-galacticon are surrounded by sufficient extratidal stars to boost the measured velocity dispersion by an order of magnitude or even more. This boosted velocity dispersion measurement, which does not represent the 'real' velocity dispersion of a bound object in equilibrium, will lead to a significant overestimation of the dynamical mass and therefore to the postulation of a heavily DM dominated object.…”

The sensitivity of harassment to orbit: mass loss from early-type dwarfs in galaxy clusters

“…In prac- tice the total bound mass is found within 5-10 iterations. We find the snowballing method of measuring bound masses to be robust and trustworthy, and we have applied this technique successfully in several previous studies (Smith et al 2013a;Smith et al 2013b;Smith et al 2013c). …”