AMS Driver
Complex MD, MC, and PES tasks with any engine
AMS driver is a powerful central tool in the Amsterdam Modeling Suite for complex potential energy tasks, such as molecular dynamics, Monte Carlo, PES scans, and finding transition states with our atomistic modules ADF, BAND, DFTB, MOPAC, ReaxFF, Machine Learning Potentials, Force Field, as well as hybrid (multi-layer) simulations. External programs can also be used. The AMS driver provides efficient and reliable implementations for common tasks like geometry optimization or transition state searches as well as more advanced modeling and simulation options like Grand Canonical Monte Carlo, force bias Monte Carlo, and the molecule gun.
Complex PES exploration tasks are made easy by AMS, minimizing the effort for the user, improving the computational efficiency, and enabling high-throughput screening scripting workflows with PLAMS. You can easily switch between computational engines, enabling workflows or high-throughput screening to increase accuracy by switching from force fields to tight-binding to density functional theory codes.
Automated reaction networks with AMS
The potential energy surface (PES) exploration tasks in the central AMS driver enable researchers to automatically discover transition states and local minima with any of our or external engines:
- Process search: find minima and the transition states connecting them
- Saddle search to find nearby transition states
- Basin hopping to find local minima
- Refinement of previously located TSs and minima with different level of theory
- Determine and visualize binding sites on a cluster or surface
Selected functionality of the AMS driver
- Advanced PES scans: scan along many degrees of freedom, any periodicity
- PES exploration: automatic reaction networks
- More robust optimization, FIRE algorithm for fast codes
- Quick approximate Hessians for Transition States searches
- Advanced thermo- and barostats to drive Molecular Dynamics
also supporting AIMD with ADF, BAND, MOPAC & DFTB - Acceleration methods: CVHD, fbMC, parallel replica dynamics
- Grand Canonical Monte Carlo to find thermodynamic minima
- Molecule gun: simulate sputtering and deposition (ALD, CVD) processes
- Properties: frequencies, phonons, stress and elastic tensors
- Easily script your workflows across codes with PLAMS and AMS
- Double parallelization for numerical (second) derivatives
AMS example applications
- Accelerate reactive MD with CVHD
- Run advanced molecular dynamics with MOPAC
- Study chemical vapor deposition with DFTB and the molecule gun
- Scan H2 dissociation and H2-surface coordinates for dissociation on a surface (2D)
- Find a TS with geometry constraints with DFTB, then refine the result with ADF
- Screen UV/VIS properties with TDDFTB, refine with sTDA for the promising candidates