Simulating morphologies of organic semiconductors by exploiting low-frequency vibrational modes

Generating morphologies of amorphous organic materials represents a significant computational challenge and severely limits the size of systems that can be studied. Furthermore, the dynamical evolution of a film at high density occurs on time scales impractical to simulate dynamically, limiting the number of independent states that can be generated. This is a problem in glassy systems as well as protein and polymeric systems. To overcome this problem, we identify rigid sections in molecules and construct an elastic network between them. Using normal mode analysis, we calculate the lowest frequency eigenmodes for the network and displace rigid sections along the low-frequency modes. The system undergoes fast structural relaxation, which allows us to generate many structurally independent approximations to a final atomistic morphology rapidly without force-field parameterization. Using these states as high-density starting configurations, we find equilibrium structures through short molecular dynamics simulations that show close agreement with other atomistic molecular dynamics studies. This method provides a convenient alternative for simulating morphologies of large molecular systems without access to high-performance computing facilities.