Author Topic: A question about ' common neighbor analysis'  (Read 188 times)

keepoptimistic

  • Newbie
  • *
  • Posts: 3
A question about ' common neighbor analysis'
« on: June 05, 2018, 11:37:29 AM »
why the results of 'common neighbor analysis' deviate from the fact? For example,my input file is a perfect model of 4000 FCC atoms set by lammps,but my analysis result of ovito show there are only 3000 FCC atoms,please see 4000FCC.png and  my input file Ag.0.cus. I have set a periodic boundary conditions for x y z directions. The results by ovito are also different from other visualization software.For another input file(not attached) ,Please see other.png obtained by another software,showing there are 3999 FCC atoms,and it meets the fact,but the ovito.png show there are no FCC atom.I feel confused and I wish to recieve your answers ,thank you!

Alexander Stukowski

  • Administrator
  • Sr. Member
  • *****
  • Posts: 475
Re: A question about ' common neighbor analysis'
« Reply #1 on: June 05, 2018, 01:29:57 PM »
Your input file seems to be wrong. The simulation cell, which is visible in the last screenshot you posted, is much too small.

Normally, all particles should be positioned inside the periodic simulation cell, but in your case most of the particles are located outside. Note that the CNA calculation needs the simulation cell information in order to correctly implement periodic boundary conditions. Without the correct size information of the periodic domain, the calculation will yield wrong results (typically, the lattice structure cannot be identified correctly, because the periodic images of the crystal do not fit together).

keepoptimistic

  • Newbie
  • *
  • Posts: 3
Re: A question about ' common neighbor analysis'
« Reply #2 on: June 06, 2018, 09:38:12 AM »
 ;)thank you very much,you're right that my input file was wrong,and I have solved my problem.