摘要:
   Open clusters are formed in the Galactic disk and will contribute to field stars after the disruption. They are good probes for the formation and evolution of the disk. The disruption of open clusters is a consequence of internal and external dynamical processes, which will lead to a dramatic change in the 3D morphology of open clusters. We study a dozen open clusters, plus a cluster pair and a stellar complex within 500 pc in the solar neighborhood based on the data from Gaia EDR3, we determine members of open clusters in 5D parameter space (3D spatial positions + 2D proper motions) via the self-adapted unsupervised learning algorithms based on an artificial neural network, which is advantageous at identifying low density and extended tidal structures. The hierarchical structures of the stellar complex are also disentangled via the same algorithm. Through the spatial positions of member stars, we will establish the 3D morphology of open clusters, and do statistical quantification and analysis of the morphology by parametrization of an ellipsoid model. The dynamical state of open clusters is investigated with 3D velocities and velocity dispersions. Most clusters in our samples are supervirial and expanding after gas expulsion (including the cluster pair and the hierarchical stellar complex), except for one cluster NGC 2232 that is probably undergoing revirialization. N-body models of rapid gas expulsion with an SFE of ≈ 1/3 are consistent with clusters more massive than 250 M⊙ , while clusters less massive than 250 M⊙ tend to agree with adiabatic gas expulsion models. The expansion rate in the youngest stellar complex is the most significant. We propose that supernova plays a role of quenching star formation in its hierarchical structures, which will keep expanding for the future 100 Myr and become unbound according to our simulation results.