[QE-users] vc-relax for getting polymer bond lengths gives error in overlap matrix

[MARIA CATERINA CEVALLOS BRENES]

Dear QE community: 

I am having several errors when trying to get the relaxed structure of
polyacetylene. I am supposed to get 1.36 and 1.43 as the dimer bond
lengths (for C=C and C-C respectively). 

I am using vc-relax method, just allowing z lattice parameter (along the
chain) to change. 

Here is my input file: 

&control 

calculation = 'vc-relax', 

prefix = 'pa', 

verbosity = 'low', 

pseudo_dir = '/home/ccevallos/pslibrary.1.0.0/pbe/PSEUDOPOTENTIALS', 

outdir = './tmp', 

wf_collect=.true. 

/ 

&SYSTEM 

  ibrav = 0, 

  nat=4, 

  ntyp=2, 

  ecutwfc=200.00, 

  occupations = 'fixed', 

  input_DFT='PBE', 

  nbnd= 12, 

/ 

&ELECTRONS 

/ 

&ions 

/ 

&cell 

cell_dofree='z' 

/ 

ATOMIC_SPECIES 

C 12.011 C.pbe-n-rrkjus_psl.1.0.0.UPF 

H 1.00784 H.pbe-rrkjus_psl.1.0.0.UPF 

CELL_PARAMETERS {angstrom} 

5.000000000000000   0.000000000000000   0.000000000000000 

0.000000000000000   5.000000000000000   0.000000000000000 

0.000000000000000   0.000000000000000   2.4 

ATOMIC_POSITIONS {angstrom} 

C 2.8 0 0 

C 2.1 0 1.16 

H 3.89 0 0 

H 3.19 0 1.16 

K_POINTS automatic 

2 2 25 1 1 1 

However I am getting the following error: 

     Error in routine cdiaghg (30): 

     S matrix not positive definite 

What am I doing wrong? I also tried using a larger supercell, with 4
Carbon atoms, and it gives the same error, but you can see the bond
length getting closer to 1.39 for al C-C bond lengths. In the latter
case, why am I not observing the Peierls distortion?  

Thank you very much in advance! 

Best regards, 

Caterina Cevallos. 

QCLab. University of Costa Rica

  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20190806/cd74ed6f/attachment.html>