Hi Ben
You may be referring to something else,
sorry if I misunderstand you. The virial
relation you refer to is the one holding
for particles interacting via 1/r potentials,
which is not the case here: cores interact
with valence electrons via pseudopotentials.
In an all-electron calculation you should
recover the virial result, since there you
do have electrons and nuclei, all seeing
each other by 1/r.
(A minor point there: even for all electron,
if you work with a finite basis like our LCAO,
the virial result is only recovered in the
converged basis limit, unless you scale your
basis functions and minimise the energy w.r.t
the scaling factors).
Emilio
On Dec 8 2009, Ben Martin wrote:
I'm having trouble resolving the energy output in SIESTA, specifically
that it seems to violate the virial theorem. The virial theorem requires
that the EXACT solution must satisfy kinetic energy = - total energy, but
I'm not getting anywhere close to that. Below is the energy output of an
optimization I ran:
Final energy (eV):
Kinetic = 5497.928558
Hartree = 2350.057707
Ext. field = 0.000000
Exch.-corr. = -1786.932583
Ion-electron = -9917.735789
Ion-ion = -3791.753153
Ekinion = 0.000000
Total = -7648.435261
The relaxation runs fine and the relaxed coordinates are good, but as you
can see there is a discrepency of ~2000 eV (or about 30%) between Kinetic
and -Total.
Am I missing something?
Thanks,
Ben
*************************************
#General system specifications
SystemName Olivine
SystemLabel Ol
NumberOfAtoms 28
NumberOfSpecies 3
%block ChemicalSpeciesLabel
1 12 Mg # Species index, atomic number, species label
2 14 Si # Species index, atomic number, species label
3 8 O # Species index, atomic number, species label
%endblock ChemicalSpeciesLabel
AtomicCoordinatesFormat Ang
%block LatticeVectors
4.6300 0.00000 0.00000
0.0000 9.96000 0.00000
0.0000 0.00000 5.98000
%endblock LatticeVectors
%block AtomicCoordinatesAndAtomicSpecies
28 atom coordinates, etc.
%endblock AtomicCoordinatesAndAtomicSpecies
%block GeometryConstraints
position 1
%endblock GeometryConstraints
%block PAO.Basis
bases from the SIESTA database
%endblock PAO.Basis
XC.Functional LDA
XC.Authors CA
MeshCutoff 300.0 Ry
MaxSCFIterations 600
DM.MixingWeight 0.01
MD.TypeOfRun CG
MD.NumCGSteps 600
MD.VariableCell
SolutionMethod diagon
AtomCoorFormatOut Ang
MD.MaxForceTol 0.004 ev/Ang
DM.Tolerance 1.d-4
DM.EnergyTolerance 1.d-4 ev
MD.MaxCGDispl 0.000635 Ang
WriteForces .true.
--
Emilio Artacho
Department of Earth Sciences, University of Cambridge
Downing Street, Cambridge CB2 3EQ, UK
Tel. (+44/0) 1223 333480, Fax (+44/0) 1223 333450
[email protected], http://www.esc.cam.ac.uk/~emilio
