Torch

Torch is an open-source code to simulate coupled gas and N-body dynamics in astrophysical systems, and it is primarily used for simulating star cluster formation. Torch combines the FLASH code for gas dynamics and the ph4 code for direct N-body evolution via the AMUSE framework. An introduction to Torch is given by Wall et al. 2019, 2020, and Joshua Wall's Ph.D thesis (pdf mirror). The Torch project is supported by NASA grant no. NNX14AP27G and NSF grant no. AST18-15461 and AST23-07950.

Torch Theses

Joshua E. Wall (PhD, 2019): The Formation and Early Evolution of Stellar Clusters
Claude Cournoyer-Cloutier (MSc, 2021): Dynamical Modification of a Primordial Population of Binaries in Simulations of Star Cluster Formation
Sean C. Lewis (PhD, 2023): New Perspectives on the Formation and Early Evolution of Stellar Clusters
Sabrina M. Appel (PhD, 2024): Exploring the effect of protostellar feedback on star formation and molecular cloud dynamics
Brooke Polak (PhD, 2024): The Secret Lives of Young Massive Star Clusters
Claude Cournoyer-Cloutier (PhD, 2025): Dynamics and Feedback of Massive Binaries in Young Massive Star Clusters

Research Highlights

Evolution of primordial binary populations during cluster formation (Cournoyer-Cloutier et al. 2024)

Evolution of a star cluster forming from a 2x10⁴ M_⊙ cloud. Primordial binary formation is included. The colours of the stars change with age, going from white to golden yellow.

Runaway stars as fossils of subcluster mergers (Polak et al. 2024b)

This is the subcluster ejection scenario (SCES), where a subset of stars from an infalling subcluster is ejected out of the cluster via a tidal interaction with the contracting gravitational potential of the assembling cluster. Runaways formed in the same SCES have similar ages, velocities, and ejection directions. This movie shows formation and ejection of an SCES runaway star group. Left: Stars as they are forming in the cluster. Right: Gravitational potential of both the stars and gas. The trajectories of the runaway stars are shown as lines. Once each runaway star forms, it appears with a star marker.

Centrally concentrated gas collapses further during star formation (Assilkhan et al. in prep)

Stellar density profiles and gas slices compared to analytic prediction (blue) of Parmentier & Pfalzner (2013) that assumes in situ collapse of gas into stars.

Abundance distribution inherited by stars from gas (Andersson et al. in prep)

Turbulent, spherical clouds of 10⁴ M_⊙ with different initial tracer distributions shown on the right (Kolmogorov spectra with ℓ=L/f_max). Larger-scale variation in gas widens stellar distribution.

High star formation efficiency in young massive clusters (Polak et al. 2024a)

Models of 10⁴, 10⁵, and 10⁶ M_⊙ star-forming turbulent clouds (M4, M5, and M6). Stars below 4 M_⊙ are agglomerated into dynamical "super"-star particles.

Massive interacting binaries enhance feedback in star-forming regions (Cournoyer-Cloutier et al. subm.)

Modelling binary stellar evolution in Torch has a big impact on the strength of stellar feedback and the size of the feedback bubble. More ionizing radiation escapes due to binary interactions stripping the star and exposing its hot core. More kinetic energy is injected into the bubble due to mass transfer events.