SaddleSearchParameters

class SaddleSearchParameters(time_step=None, langevin_friction=None, minimization_period=None, escape_time_accuracy=None, max_time=None, optimizer_method=None, max_forces=None, max_steps=None, max_step_length=None, min_neb_images=None, max_neb_images=None, max_neb_steps=None, max_neb_forces=None, neb_image_distance=None, neb_spring_constant=None, neb_climbing_image=None, lanczos_relative_error=None, lanczos_maximum_iterations=None, finite_difference=None, finite_difference_method=None, structure_comparison_tolerance=None, structure_comparison_energy_tolerance=None, saddle_offset_length=None, max_energy_difference=None)

Class that stores all of the parameters associated with saddle searches.

Parameters:
  • time_step (PhysicalQuantity of type time) – The time-step interval used in the molecular dynamics simulation.
    Default: 1.0*fs
  • langevin_friction (PhysicalQuantity of type frequency) – The friction (coupling) with the reservoir.
    Default: 5.0*ps**-1
  • minimization_period (PhysicalQuantity of type time) – The time interval between minimizing the molecular dynamics trajectory to determine if the system has left the initial state.
    Default: 400*fs
  • escape_time_accuracy (PhysicalQuantity of type time) – The accuracy with which the escape time should be determined. This quantity must be less than or equal to the minimization period.
    Default: minimization_period/3
  • max_time (PhysicalQuantity of type time) – The maximum total length of time the MD simulation is allowed to run for.
    Default: 5000*fs
  • optimizer_method (LBFGS | FIRE) – The optimization algorithm to use for the minimizations, saddle optimization, and nudged elastic band calculations.
    Default: LBFGS
  • max_forces (PhysicalQuantity of type force) – Optimization is stopped when the maximum force on an atom is less than max_forces. There is a separate convergence criterion (max_neb_forces) for the NudgedElasticBand optimization.
    Default: 0.01*eV/Angstrom
  • max_steps (int) – The maximum number of steps during an optimization. There is a separate max steps (neb_max_steps) parameter for NudgedElasticBand optimization.
    Default: 400
  • max_step_length (PhysicalQuantity of type length) – The maximum step length the optimizer may take.
    Default: 0.2*Angstrom
  • min_neb_images (int) – The minimum number of NudgedElasticBand images in a band.
    Default: 3
  • max_neb_images (int) – The maximum number of NudgedElasticBand images in a band.
    Default: 7
  • max_neb_steps (int) – The maximum number of NudgedElasticBand optimization steps.
    Default: 100
  • max_neb_forces (PhysicalQuantity of type force) – The NudgedElasticBand optimization is stopped when the maximum force on an atom is less than max_neb_forces.
    Default: 0.1*eV/Angstrom
  • neb_image_distance (PhysicalQuantity of type distance) – The distance between the NudgedElasticBand images.
    Default: 0.5*Angstrom
  • neb_spring_constant (PhysicalQuantity of type force constant (e.g. eV/Angstrom**2)) – The spring constant used for the NudgedElasticBand relaxation.
    Default: 5.0*eV/Angstrom**2
  • neb_climbing_image (bool) – Flag indicating if the climbing image algorithm should be used to find a transition state.
    Default: False
  • lanczos_relative_error (float) – Stop the calculation when the relative change in the eigenvalue is less than this number.
    Default: 0.01
  • lanczos_maximum_iterations (int) – The maximum number of iterations in the Lanczos algorithm.
    Default: 20
  • finite_difference (PhysicalQuantity of type length) – The step size for the finite difference estimation of the second derivatives.
    Default: 0.01*Angstrom
  • finite_difference_method (Forward | Central) – The type of finite difference used to estimate the second derivatives.
    Default: Forward
  • structure_comparison_tolerance (PhysicalQuantity of type length) – The maximum distance mismatch between two atoms to determine if they are the same structure.
    Default: 0.1*Angstrom
  • structure_comparison_energy_tolerance (PhysicalQuantity of type energy) – This energy tolerance is used to determine if two structures are different based upon energy alone. Configurations that are closer than this tolerance a full geometry comparison.
    Default: 0.02*eV
  • saddle_offset_length (PhysicalQuantity of type length) – The distance along the negative mode from the saddle that a configuration is displaced before being minimized to determine if it is connected to the initial minimum.
    Default: 0.2*Angstrom
  • max_energy_difference (PhysicalQuantity of type energy) – If the final state is higher in energy than the initial state by this amount, then the search will be terminated. If the value is None then no energy comparison is performed.
    Default: None

Usage Examples

Setup a saddle search using a tight convergence criterion.

saddle_search_parameters = SaddleSearchParameters(
    max_forces=0.001*eV/Angstrom,
    )

parameters.py

Notes

These parameters are normally passed to the AdaptiveKineticMonteCarlo class.