If the atomic force does not converge, maybe decreasing 'MD.MaxCGDispl' can help. Sometimes it works for me.
2011-02-28 tauhall 发件人： loc duong ding 发送时间： 2011-02-28 10:22:53 收件人： siesta-l 抄送： 主题： Re: [SIESTA-L] fluctuation of energy in optimization process Dear TaooHee, In my understanding, DM.MixingWeight and DM.NumberPulay option is used for SCF convergence. In my case, there is no problem with SCF convergence. Maybe you misunderstood my problem. The problem is the total energy of optimization steps is not convergence. I tried to increase the mesh cut-off to 450 but still not work. Do you have another instructions for solving this? Sincerely, ----------------------------------------------- Loc Duong Dinh Ms-Ph.D Student Sungkyunkwan Advanced Institute of Nanotechnology, Sungkyunkwan University, Suwon, 440-746, Korea Email: [email protected], [email protected] > >From: taoohee <[email protected]> >To: siesta-l <[email protected]> >Sent: Mon, February 28, 2011 10:13:04 AM >Subject: Re: [SIESTA-L] fluctuation of energy in optimization process > > >Have you tried tuning 'DM.MixingWeight' and 'DM.NumberPulay' ? >Best, > >2011-02-28 ________________________________ >tauhall ________________________________ >发件人： loc duong ding >发送时间： 2011-02-26 07:32:24 >收件人： siesta-l >抄送： >主题： [SIESTA-L] fluctuation of energy in optimization process >Dear all, >I am a new user of siesta. I try to use siesta to optimized molecule on the >top >of graphene. However, I have a problem with configuration optimization. >The energy is fluctuation, that is not smooth reduce. I attach the Energy of >optimized steps. > siesta: E_KS(eV) = -33776.8773 >siesta: E_KS(eV) = -33776.9352 >siesta: E_KS(eV) = -33776.9810 >siesta: E_KS(eV) = -33777.0382 >siesta: E_KS(eV) = -33776.9230 >siesta: E_KS(eV) = -33777.0397 >siesta: E_KS(eV) = -33777.0442 >siesta: E_KS(eV) = -33777.0603 >siesta: E_KS(eV) = -33777.0774 >siesta: E_KS(eV) = -33777.0543 >siesta: E_KS(eV) = -33777.0788 >siesta: E_KS(eV) = -33777.1002 >siesta: E_KS(eV) = -33777.0986 >siesta: E_KS(eV) = -33777.1054 >siesta: E_KS(eV) = -33777.1351 >siesta: E_KS(eV) = -33777.1034 >This is my input file: >SystemName bilayer-graphene-72-tfsi >SystemLabel G72T >#Species and Atom >NumberOfSpecies 5 >NumberOfAtoms 159 >#ChemicalSpeciesLabel >%block ChemicalSpeciesLabel > 1 6 C > 2 7 N > 3 8 O > 4 16 S > 5 9 F >%endblock ChemicalSpeciesLabel ># Basis >PAO.EnergyShift 200 meV >PAO.BasisSize DZP >#Lattice Information >LatticeConstant 1.0 Ang >#LatticeVectors >%block LatticeVectors > 12.778200000 -7.3775000000 0.00000000 ># Atomic coordinates >AtomicCoordinatesFormat ScaledCartesian >%block AtomicCoordinatesAndAtomicSpecies >23.45320923 13.58855965 16.37730642 1 >25.58232873 14.81751862 16.37072604 1 >27.71084039 16.04652041 16.38185762 1 >29.83988172 17.27720442 16.37977254 1 >19.19150216 11.13037036 16.37722270 1 >21.32244584 12.36011162 16.38424376 1 >24.87273016 13.58709504 16.36155225 1 >25.58372289 12.35970141 16.33168688 1 >27.00152025 12.35827450 16.29914305 1 >.......... >32.18813116 9.68042097 20.75094507 5 >30.90013751 8.91478097 19.13337767 5 >%endblock AtomicCoordinatesAndAtomicSpecies >%block kgrid_Monkhorst_Pack > 4 0 0 0.000 > 0 4 0 0.000 > 0 0 1 0.000 >%endblock kgrid_Monkhorst_Pack >xc.functional VDW # Exchange-correlation functional >xc.authors DRSLL # Exchange-correlation version >MeshCutoff 180. Ry # Mesh cutoff. real space mesh ># SCF options >MaxSCFIterations 500 # Maximum number of SCF iter >DM.MixingWeight 0.1 # New DM amount for next SCF cycle >DM.Tolerance 1.d-4 eV # Tolerance in maximum difference >DM.NumberPulay 3 >Diag.DivideAndConquer .false. >SolutionMethod diagon # OrderN or Diagon >ElectronicTemperature 0.1 meV # Temp. for Fermi smearing ># MD options >MD.TypeOfRun cg # Type of dynamics: >MD.NumCGsteps 500 # Number of CG steps for >MD.MaxCGDispl 0.3 Ang # Maximum atomic displacement >MD.MaxForceTol 0.04 eV/Ang # Tolerance in the maximum >#MD.TargetPressure 2.0 GPa >#%block MD.TargetStress ># 0.0 0.0 -1.0 0.0 0.0 0.0 >#%endblock MD.TargetStress >#OutPutOption >#WriteMDhistory .true. >#WriteMDXmol .true. >Could you give some instruction to show this problem? >I appreciate to all your help. > >Sincerely, >----------------------------------------------- >Loc Duong Dinh >Ms-Ph.D Student >Sungkyunkwan Advanced Institute of Nanotechnology, >Sungkyunkwan University, >Suwon, 440-746, Korea >Email: [email protected], [email protected] >
