Results

After simulating 4000 stars, we find that ~17% experienced an MDE, ~2% were ejected from the cluster (pictured by the purple diamonds), ~1% experienced a TDE due to relaxation effects, and ~0.1% experienced a TDE due to a high-speed collision that altered its orbit. When looking at the stars that experienced an MDE (pictured below), we find that these collisions are typically due to high speed (average relative speed of ~6000 km/s) and directness (average of ~0.2 normalized periapsis separation). We also find that MDEs typically occur after a star has lost mass over a series of previous high-speed collisions (often 10 or more collisions). We can see that all of these collisions occur with a relative speed of at least 2000 km/s and at least a .26 normalized impact parameter. This value can be interpreted as at least 74% overlap between the two stars before their collision. These conditions are likely due to the high speeds and dense population near the SMBH. The stars with a final mass of 0 M_⊙ (MDE stars) all lie within 0.1 pc of the SMBH, with a majority lying within 0.01 pc for both α = 1.25 (bottom) and α = 1.75 (top).

In this small movie, we can see the evolution of the mass of the star cluster with α = 1.75. All 4000 stars are assumed to begin at 1 M_⊙ and a majority of stars don't experience a collision. As shown, stars that experience typical collisions have their mass slowly reduced over the course of multiple collisions. After a certain number of collisions, these stars experience an MDE which is pictured by the large bar on the far left (representing 0 M_⊙ stars). On the right of the 1 M_⊙ bar, we can see the merger cases which exhibit the same properties on a smaller scale: after merging, these stars will experience collisions that slowly reduce their mass over time. This population typically dies out faster due to their shorter lifetime. Below, we can see similar plots for α = 1.25.

Conclusion

Collisions between stars occur in the galactic cluster and are more common closer to the SMBH due to the high speeds and larger population density [7]. We can simulate how these collisions effect a stars evolution in these dense clusters and what conditions lead to MDEs, star ejections, and TDEs. We find that we are able to achieve qualitatively similar results to previous studies while using ML to predict the outcomes of collisions. Overall, we find that by using ML, we are able to efficiently and accurately classify the outcome of star collisions. By utilizing ML, we can accurately model the evolution of dense star clusters in the galactic center.

References