[Pw_forum] "charge is wrong" error in semiconductor slab calculation

[J. J. Ramsey]

I'm doing a slab calculation somewhat similar to that in Phys. Rev. B 77, 165323 (2008), except that (1) the geometry is simplified down to 10 atoms and (2) the goal is to relax the positions of the pseudohydrogens used to passivate the bottom of the slab while keeping all other atoms fixed. It's partly a dummy problem to shake out the kinks in doing a simulation in Quantum Espresso. Here is the input file:

&CONTROL
    title = "Finding positions of pseudohydrogens",
    calculation= "relax",
    prefix = findPseudoInitPos_PWscf,
    pseudo_dir = "Pseudopotentials/"
/
&SYSTEM
    ibrav = 8,
    a = 3.92232131524178,
    b = 3.92232131524178,
    c = 22.207375,
    nat = 10,
    ntyp = 4,
    ecutwfc = 40
/
&ELECTRONS
    conv_thr = 1D-6
/
&IONS
    bfgs_ndim         = 3,
    pot_extrapolation = "second_order",
    wfc_extrapolation = "second_order",
/
ATOMIC_SPECIES
Ga 69.723 Ga.pz-bhs.UPF
In 114.818 In.pz-n-bhs.UPF
As 74.92160 As.pz-bhs.UPF 
H* 1.25 01_25-H-tm.fhi.UPF
ATOMIC_POSITIONS angstrom
H* 0 0.980580328810445 0 1 1 1
H* 0 2.94174098643133 0 1 1 1
Ga 0 0 0.693375 0 0 0
As 1.96116065762089 0 2.080125 0 0 0
Ga 1.96116065762089 1.96116065762089 3.466875 0 0 0
As 0 1.96116065762089 4.853625 0 0 0
In 0 0 6.240375 0 0 0
As 1.96116065762089 0 7.732125 0 0 0
In 1.96116065762089 1.96116065762089 9.223875 0 0 0
As 0 1.96116065762089 10.715625 0 0 0
K_POINTS automatic
4 4 1 1 1 1

The error that I get is
WARNING: integrated charge=    34.15000000, expected=    34.50000000

 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
     from electrons : error #         1
     charge is wrong
 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

     stopping ...

This happened at the first iteration.

The pseudopotentials for Ga, In, and As are from the pseudopotential tables, and I tested them by using them to find the lattice constants of GaAs and InAs, which were found to be about 5.547 and 5.967 angstroms, respectively, which appears reasonable. The pseudopotential for pseudohydrogen was generated from fhi98pp from this input file,

 1.25  0  1  3  0.00 : z  nc  nv iexc rnlc 
    1  0   1.25      : n  l   f 
0  t                 : lmax  s_pp_def 
0  0.00  0.00  t     : lt  rct  et  s_pp_type

and converted to UPF with the fhi2upf utility.

Input, output, and pseudopotential for pseudohydrogen are attached in the job-stuff.tgz tarball.

Any ideas as to what's wrong?


      
-------------- next part --------------
A non-text attachment was scrubbed...
Name: job_stuff.tgz
Type: application/x-compressed-tar
Size: 19533 bytes
Desc: not available
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20081202/2ac7a7d8/attachment.bin>