[DFTB-Plus-User] Problem to MD run with Supercell of CuSi system with 1372 atoms

[Faruq, Muhammad]

Dear All

I have read a post regarding this problem. The guy suggests that he has to change the Supercell to Cluster. But in my case I have to keep the supercel . Please help me to sort this out. Here is the details of the simulation input.

The problem is  ----------

********************************************************************************
** Geometry step: 0
********************************************************************************

WARNING!
-> Atoms     1 and     5 too close to each other! (dist= 0.000000E+00)

WARNING!
-> Atoms     2 and     6 too close to each other! (dist= 0.000000E+00)


Input HSD format:

Geometry = GenFormat {
  <<< CuSiS_MD.gen
}
Driver = VelocityVerlet {
  MovedAtoms =  "1:-1"
  Steps = 10000
  TimeStep = 2.0e-15                #Time interval between two MD steps
  KeepStationary = No               #Remove translational motion from the system
  MDRestartFrequency=10
  Velocities = {}
  ConvergentForcesOnly = No
  Barostat = {
    Pressure = 0
   Timescale =10
  }
  Thermostat = NoseHoover{
    Temperature[Kelvin] = 2500
    CouplingStrength =100
  }
  OutputPrefix = "MDCuSi.out"
}

Hamiltonian = DFTB {
  SCC = Yes
  SCCTolerance = 1.0e-5
  EwaldParameter = 0.1
  MaxSCCIterations = 100
  Mixer = Broyden {
    MixingParameter = 0.2
}
  MaxAngularMomentum = {
    Cu = "d"
    Si = "p"
  }
  Charge =0.0
  Filling = Fermi {
    Temperature [Kelvin] = 300
  }
  SlaterKosterFiles = Type2FileNames {
    Prefix = "/home/maf2mpo/dftb_tutorial/sk/matsci-0-3/"
    Separator = "-"
    Suffix = ".skf"
  }
   KPointsAndWeights = SupercellFolding {
     2 0 0
     0 2 0
     0 0 3
     0.5 0.5 0.0
  }
  ReadInitialCharges = No
}



Thanks
Muhammad Faruq

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.zfn.uni-bremen.de/pipermail/dftb-plus-user/attachments/20150806/111ac77b/attachment-0001.html>