[Ape-users] Pseudopotential for uranium

[Micael Oliveira]

Dear Vasily,

  When using the Troullier-Martins scheme, only one state can be used 
per angular momentum channel to generate the pseudopotentials. If you 
want to include semi-core states, you can still do it by setting the 
PPSemiCoreComponents block, although in that case the valence states 
will not be used during the pseudopotential generation. As an 
alternative, you can also use the MRPP scheme, which takes both valence 
and semi-core states into account when generating the pseudopotentials.

  Best regards,

Micael


On 12/07/2011 08:35 AM, Василий Писарев wrote:
> Dear APE users,
>
> I am trying to generate a PP for Uranium. AE calculation goes OK, but the
> program crashes on generating PP.
>
> APE 1.1.1, libxc 1.1.0, tried with GSL 1.13 and 1.14
> Input for PP:
> -----------------------------------------
> CalculationMode = pp
> Verbose = 40
>
> PPScheme = rtm
>
> %PPComponents
>   6 | 0 | 1.20
>   6 | 1 | 1.50
>   5 | 3 | 1.00
>   6 | 2 | 1.80
>   7 | 0 | 2.40
> %
>
> ODEIntTolerance = 1e-07
> PPCalculationTolerance = 1e-07
> #################################
>
> Program output:
>
>                   APE - Atomic Pseudopotentials Engine
>
>                Program started on 2011/12/07 at 12:22:01
>
> Compilation Info
>    Version: 1.1.1
>    Revision: 5986 [fc_command_line.m4 2009-10-26]
>    Build time: Mon Dec  5 14:18:28 MSK 2011
>    C compiler: gcc
>    C compiler flags: -g -O2 -I/usr/include
>    Fortran compiler: gfortran
>    Fortran compiler flags: -g -O2
>
> Calculation Type:
>    PseudoPotential Generation
>
> Setting units
>    Input units system:  Atomic Units
>    Output units system: Atomic Units
>
> Initializing Wave-Equations Integrator
>    Stepping function: Embedded 8th order Runge-Kutta Prince-Dormand
>                       method with 9th order error estimate
>    ODE Integrator tolerance:  1.000E-07
>    ODE Integrator maximum number of steps:  50000
>    ODE Integrator minimum step size:  1.000E-15
>
> Eigensolver Info
>    Method:    Brent's method
>    Tolerance:  1.000E-05
>
>
>                     -- Pseudopotential Generation --
>
> Pseudo atom information:
>    PP generation scheme : Troullier-Martins Relativistic extension
>    Wavefunction info:
>      State   Occupation    Node radius   Peak radius   Default core radius
>      6s0.5      2.00          0.802         1.342             0.000
>      6p0.5      2.00          0.824         1.496             0.000
>      6p1.5      4.00          0.919         1.667             0.000
>      5f2.5      1.29          0.442         1.052             0.000
>      5f3.5      1.71          0.442         1.081             0.000
>      6d1.5      0.40          1.081         2.371             0.000
>      6d2.5      0.60          1.110         2.503             0.000
>      7s0.5      2.00          1.760         3.464             0.000
>
> Pseudopotential Generation:
>    State: 6s0.5
>      Core radius:   1.200
>        Iteration number:   1
>        Coefficients:
>          c0  =  3.999999474E-01   c2  = -3.509959070E+00
>          c4  =  4.848535087E-01   c6  =  9.128076237E+00
>          c8  = -1.117226099E+01   c10 =  5.123093791E+00
>          c12 = -8.444621103E-01
>        Residues:
>          c2**2 + c4(2l+5)            =  7.715764818E-02
>          ln(norm)[AE] - ln(norm)[PP] =  1.474408022E+01
> ...
> Pseudopotentials Self-Consistency:
>    State  Eigenvalue [H ]    Norm Test   Slope Test
> Segmentation fault
> #############################
>
> Input for AE:
> -----------------------------------------------------------
> Title = "Uranium"
> CalculationMode = ae
>
> Verbose = 40
>
> WaveEquation = dirac
>
> XCFunctional = lda_x + lda_c_pw
> XCCorrections = rel_x
>
> NuclearCharge = 92
> %Orbitals
>   "Rn"
>   5 | 3 | 3
>   6 | 2 | 1
>   7 | 0 | 2
> %
>
> ConvAbsDens = 5.0e-6
>
> EigenSolverTolerance = 1e-07
> ODEIntTolerance = 1e-07
>
> MeshStartingPoint = 1.0e-7
> MeshOutmostPoint = 250.0
> MeshNumberOfPoints = 800
>
> TypeOfMixing = broyden
> Mixing = 0.5
> MixNumberSteps = 5
> ######################################
>
> I would appreciate any help.
> Best regards,
> Vasily
> _______________________________________________
> Ape-users mailing list
> Ape-users at tddft.org
> http://www.tddft.org/mailman/listinfo/ape-users
>