[DFTB-Plus-User] Fwd: Question about barostat in DFTB+

[Longtao Han]

Dear DFTB+ community,

I am a user of DFTB+1.2, and currently I have encountered some problems
with the barostat in DFTB+ when I try to simulate a gas phase system.

I am trying to simulate the association and crystallization behavior of
atomic borons and nitrogens at a given temperature. So I constructed a
cubic supercell system with 343 atoms, with about equal numbers of B and N
atoms. Initial distances between atoms are 4nm, therefore they don't
interact. The system is subject to thermostat and barostat. Below is my
input file for the test.
*******************
Geometry = GenFormat {
 <<< "bnnt.gen"
 }
 Driver = VelocityVerlet{
  MovedAtoms = 1:-1
  TimeStep [Femtosecond] = 1.0
  Thermostat = NoseHoover {
    Temperature [K] = TemperatureProfile {
        constant      1 8000
        linear    40000 2000
        constant  10000000 2000
        }
    CouplingStrength [cm^-1] = 3000
  }
  Barostat = {
    Pressure [Pa]=500000.0
    Timescale [ps]=0.01
    }
  MDRestartFrequency = 500
 OutputPrefix = "geo_stp"
 }

 Hamiltonian = DFTB{
    Charge = 0
    Filling = Fermi{
        Temperature [K] = 3000
    }
    KPointsAndWeights = SupercellFolding{
        1       0       0
        0       1       0
        0       0       1
        0.0     0.0     0.0
    }
    MaxAngularMomentum = {
       B = "p"
       N = "p"
  }
  SlaterKosterFiles = Type2FileNames {
    Prefix = "/gpfs/home/lonhan/software/executable/SK-files/matsci-0-3/"
    Separator = "-"
    Suffix = ".skf"
    LowerCaseTypeName = No
  }
 }
Options={
  ShowFoldedCoords=Yes
}
****************

The pressure reported in the MD output is orders of magnitude higher than
it should be. And after some thousands of steps, the run will fail with the
following error:

*ERROR!*
*-> Failure in diagonalisation routine dsygvd, non-positive definite
overlap! Minor 485107 responsible.*

I tried to increase the coupling strength up to 10000 (cm^-1), but I still
get this error.
The input geometry and output files are attached.
Could anyone​​​​​​ help me identify if this is a problem with my simulation
approach or with the barostat?


Best regards,

Longtao Han
PhD student
IACS, Stony Brook University
Email: longtao.han at stonybrook.edu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.zfn.uni-bremen.de/pipermail/dftb-plus-user/attachments/20170112/d1ef7fe0/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: bnnt.gen
Type: application/x-genesis-rom
Size: 18588 bytes
Desc: not available
URL: <http://mailman.zfn.uni-bremen.de/pipermail/dftb-plus-user/attachments/20170112/d1ef7fe0/attachment.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: md.out
Type: application/octet-stream
Size: 48900 bytes
Desc: not available
URL: <http://mailman.zfn.uni-bremen.de/pipermail/dftb-plus-user/attachments/20170112/d1ef7fe0/attachment.obj>