neb_simulation
Differences
This shows you the differences between two versions of the page.
|
neb_simulation [2009/12/04 10:53] prokop |
neb_simulation [2011/02/18 13:13] |
||
|---|---|---|---|
| Line 1: | Line 1: | ||
| - | == Input == | ||
| - | Here we provide brief description how to run Nudged Elastic Band ([http://theory.cm.utexas.edu/henkelman/research/saddle/ NEB]) simulations to explore reaction pathways between two known atomic configurations. | ||
| - | More details about the method including list of references can be found [http://theory.cm.utexas.edu/henkelman/research/saddle/ here]. | ||
| - | To start, we need an initial and final atomic configuration. | ||
| - | |||
| - | The initial configuration is defined in input ''*.bas'' file defined in ''' fireball.in''' input file. Final configuration | ||
| - | is defined in other ''*.bas'' file, which name is defined in ''neb.optional'' file (see bellow). | ||
| - | |||
| - | Next step, we need to activate ''ineb'' flag in the section OPTION of the '''fireball.in''' input file: <br \> | ||
| - | |||
| - | <code> | ||
| - | &OPTION | ||
| - | ineb = 1 | ||
| - | &END | ||
| - | </code> | ||
| - | |||
| - | In addition, we create an extra input file ''neb.optional'' including parameters to control NEB searching process. This file has to be presented in the working directory (This file will probably converted into section format of ''fireball.in'' soon, for details please checkout your current version). | ||
| - | |||
| - | The format of '''neb.optional''' file is following: | ||
| - | |||
| - | <code> | ||
| - | 1.0 ! spring strength (k_neb) | ||
| - | 0.03 ! tolerance of displacement (tol_displ_neb) | ||
| - | 0.05 ! tolerance of force (tol_ftot_neb) | ||
| - | 0.003 ! tolerance of energy (tol_etot_neb) | ||
| - | 16 ! neb images (including the initial and final image) (nimg_neb) | ||
| - | 1000 ! max. number of neb iterations (niter_neb_max) | ||
| - | 2.0 ! time step for NEB optimization verlet method (dt_neb) | ||
| - | 0 ! restart (irestart) | ||
| - | final.bas ! file name including the final configuration (file_neb)) | ||
| - | </code> | ||
| - | |||
| - | == Output & Convergence== | ||
| - | Information about the convergence status/progress can be checked out searching lines with a string ''NEB'' in output file; e.g. typing from command line: | ||
| - | <code> | ||
| - | grep "NEB" $output_name | ||
| - | </code> | ||
| - | |||
| - | you'll get a summary in the following format: | ||
| - | <code> | ||
| - | ... | ||
| - | NEB_RES no. iter = 52 | ||
| - | NEB_RES displ : RES = 0.00977406 TOL = 0.03000000 | ||
| - | NEB_RES etot : RES = 0.05848069 TOL = 0.00300000 | ||
| - | NEB_RES ftot : RES = 0.64556157 TOL = 0.05000000 | ||
| - | NEB_RES no. iter = 53 | ||
| - | NEB_RES displ : RES = 0.00968433 TOL = 0.03000000 | ||
| - | NEB_RES etot : RES = 0.05707478 TOL = 0.00300000 | ||
| - | NEB_RES ftot : RES = 0.62786898 TOL = 0.05000000 | ||
| - | NEB_RES no. iter = 54 | ||
| - | NEB_RES displ : RES = 0.00960022 TOL = 0.03000000 | ||
| - | NEB_RES etot : RES = 0.05630221 TOL = 0.00300000 | ||
| - | NEB_RES ftot : RES = 0.62082779 TOL = 0.05000000 | ||
| - | NEB_RES no. iter = 55 | ||
| - | ... | ||
| - | </code> | ||
| - | |||
| - | In addition actual optimized position of each NEB image is stored in '''answer_neb_''XX''.bas''' file; where ''XX'' means image no.. File '''answer_neb_''XX''.xyz''' contains whole optimization history of an individual image. | ||
| - | Files '''answer_image.bas''' and '''answer_image.xyz''' contain all, last optimized, NEB images including the initial and final ones.<br/> | ||
| - | Note: The files '''answer_image.bas''' can be directly used for a restart of NEB-optimization process. | ||
| - | |||
| - | -- | ||
| - | Main sources of convergence troubles: | ||
| - | # '''large time step'''<br/>initial forces can be pretty large, therefore a small time step is advisable to avoid any numerical instabilities of [http://en.wikipedia.org/wiki/Verlet_integration Verlet] alghorithm used to integrate the equations of motion. | ||
| - | # '''wrong initial guess of NEB images''' <br/>Linear extrapolation is used to build up set of NEB images at begin. In special cases, where the reaction trajectories are complicated, can lead to unrealistic configurations with atoms close to each other. Such configurations will suffer large repulsive forces which consequently may lead to numerical instabilities during optimization. | ||
| - | # '''weak spring constant'''<br/>If the potential energy surface is complicated, the imaginary elastic band can get kinked. On other hand, large spring constant will lead to less precise estimation of the saddle point. | ||
| - | |||
| - | == Restart == | ||
| - | There is possibility to make a restart of previous run on penultimate line in the file '''neb.optional''': | ||
| - | |||
| - | <code> | ||
| - | 1.0 ! spring strength (k_neb) | ||
| - | 0.03 ! tolerance of displacement (tol_displ_neb) | ||
| - | 0.05 ! tolerance of force (tol_ftot_neb) | ||
| - | 0.003 ! tolerance of energy (tol_etot_neb) | ||
| - | 16 ! neb images (including the initial and final image) (nimg_neb) | ||
| - | 1000 ! max. number of neb iterations (niter_neb_max) | ||
| - | 0.5 ! time step for NEB optimization verlet method (dt_neb) | ||
| - | 1 ! restart (irestart) | ||
| - | final.bas ! file name including the final configuration (file_neb)) | ||
| - | </code> | ||
| - | |||
| - | In this case, an initial NEB images will be uploaded from a file declared on the next line of '''neb.optional''' file. | ||
| - | Format of this file has to be indentical to the '''answer_image.bas''' file, where the first line containing total number of NEB images. | ||
neb_simulation.txt ยท Last modified: 2011/02/18 13:13 (external edit)
Page Tools
Except where otherwise noted, content on this wiki is licensed under the following license: CC Attribution-Noncommercial-Share Alike 4.0 International
