Cerjan-Miller Saddle Point Search Algorithm

Python implementation of Cerjan-Miller method for saddle point search using GAUSSIAN as calculation engine.

Usage:

Create a CM_TSS object and pass the path to the settings.json file. The settings.json file is formated as (default values in parantheses):

{
    "N": {number of atoms} (required),
    "Dim": {dimesions of the system} (3),
    "working-dir": {path to directory with all the files} (required),
    "submit-f-dir": {path to Gaussian submission file} (required),
    "basis-f-name": {filename of the basis file} ("" skipped),
    "init-f-name": {filename of the initial structure} (required),
    "gaussian-f-name": {name for gaussian input file} ("in"), 
    "charge": {charge of the system} (0),
    "spin": {spin of the system} (1),
    "N-procs": {number of processors to use} (8),
    "max-iter": {max number of iterations} (10),
    "energy-header-calc": {Gaussian header for energy calculations} ("#P wB97XD/6-31G** nosymm force"),
    "hess-header-calc": {Gaussian header for Hessian calculations } ("#P b3lyp/6-31G** nosymm freq"),
    "conv-radius":{atom position convergence radius} (1e-6),
    "conv-grad":{gradient convergence radius} (1e-6),
    "R-trust": {Trust radius} (0.2)
}

A GAUSSIAN submission script is necessary. This is a system depandant file.

basis-f-name is the name of the file containing basis specifications if necessary and will be added to the bottom of the GAUSSIAN input file.

Input coordinates file should have the format:

{atomic number} {-1 for frozen 0 otherwise} {x} {y} {z}

There's an example calculation for HCN available under the HCN directory. The CM_TSS.py should either be added to PATH or copied to the same directory as the test script.