Roughly one-third of stars in the Milky Way are in binary or multiple systems, but only ~1–10% of stars in globular clusters have a companion. Did globular clusters start with fewer binaries than the solar neighborhood, or have they since been wiped out?
Survival of the Tightest
Map of the semi-major axis–primary mass parameter space for binary systems. Solid lines show the boundary between hard (below the line) and soft (above the line) binaries for a variety of models. Models A–K span a variety of cluster masses and virial radii (a proxy for cluster density). [O’Connor et al. 2026]
It turns out that the fate of a binary system depends on whether it is dynamically “hard” or “soft.” Hard binaries are generally tightly bound, and they tend to get “harder” from stellar flybys, which are common in a crowded cluster environment. Soft binaries are the opposite — they have much shorter lifetimes since encounters with other stars tend to loosen them. The team found that, assuming GCs have the same initial distribution of binaries as the solar neighborhood, the dissolution of soft binaries is sufficient to explain the present-day scarcity of binaries in GCs.
Burning Away the Binaries
What actually tears the soft binaries apart? The answer is a process called “black hole burning.” As some of the most massive objects in a globular cluster, black holes in binary systems can transfer energy when they interact with passing stars. This influx of energy can excite soft binaries to the point of dissolution, the study shows.
Black hole burning can also prevent core collapse in GCs. Left to itself, a cluster’s core gradually loses energy to its outskirts, causing the core to grow hotter and denser with stars. This leads to more dynamical interactions between stars that can then get flung from the core or enter new tight binaries. Black hole burning pumps energy back into the core and prevents this from happening.
Putting the Theory to the Test
The total number of black holes in a cluster (top) and the binary fraction (bottom) as a function of time for three different cluster models. In the densest cluster (blue), binary dissolution starts early on, mainly driven by stellar collisions. For the other two models, binary dissolution mainly occurs after the number of black holes reaches its peak and black hole burning commences. [Adapted from O’Connor et al. 2026]
In general, the final binary fractions of the simulated clusters agree with those predicted by the team’s analytical formalism. This further supports the theory that there is a universal initial binary distribution between galactic GCs and our solar neighborhood. Constraining this distribution is pivotal for refining dynamical models of GCs.
In short, GCs and the solar neighborhood likely started with the same amount of binary systems, and interstellar dynamics handled the rest.
Citation
“A Common Origin for Binary Fractions in the Field and in Dense Clusters,” Christopher E. O’Connor et al 2026 ApJ 1009 L80. doi:10.3847/1538-4357/aea078