Unraveling Multiphase Growth Dynamics of Phosphorus on Ag(111): Machine Learning‐Driven Design Principles for Controlled Synthesis

Abstract Elemental phosphorus (P) nanomaterials exhibit extraordinary polymorphism that dictates their electronic and quantum properties, yet controlling phase evolution during growth remains a fundamental challenge. By integrating atomic‐scale scanning tunneling microscopy (STM) with a high‐precision machine learning force field (MLFF), the multiphase competition governing P growth on Ag(111) is decoded. Through cumulative molecular dynamics simulations exceeding 600 nanoseconds, three distinct regimes are revealed: 1) low‐coverage chain dominance via rapid P 4 decomposition (0.25 atomic layers, AL), 2) collision‐driven pentamer formation at intermediate coverage (0.71 AL), and 3) high‐temperature random nucleation of blue phosphorene (1.5 AL). Crucially, pentamer‐mediated kinetic pathways are identified as the origin of polycrystalline blue phosphorene boundaries, explaining experimental observations of limited domain sizes. This STM‐MLFF synergistic framework establishes design principles for polymorph control, demonstrated by achieving room‐temperature pentamer arrays through coverage engineering. The approach provides valuable insights into the synthesis of low‐dimensional phosphorus structures on metal substrate surfaces.