[ase-users] Vibrational calculation for Cu-Cl on Cu fails

[Andrew Peterson]

Hi Dominik,

s042547 at student.dtu.dk wrote:
> Dear Ase users,
>
>  I would like to perform a vibrational calculation for Cu-Cl adsorbed on a
> Cu(2x2x3) slab. The calculator I am using is dacapo, and the script is as
> follows:
>
> ..
> calc = Jacapo(nc=name_system+'.nc',
>               kpts=(4,4,1),
>               pw = 26*Ry,
>               dw = 35*Ry,
>               nbands = nelecbands,
>               atoms = slab,
>               xc = 'RPBE',
>               dipole = True)
>
> vib = Vibrations(slab, indices = [12,13])
> vib.run()
> vib.summary(T=523)
> ..
>
>  As you see only Cu and Cl are displaced. When I run the script, it
> generates 12 pckl files (as expected, all have the same size, so nothing
> seems to be wrong there) and then I get the error:
> ..
> File "/home/dobr/APPS/ase-3.4.1.1765/ase/vibrations.py", line 210, in read
>     omega2, modes = np.linalg.eigh(self.im[:, None] * H * self.im)
>   File
> "/home/dobr/APPS/Numpy-1.3.0//lib64/python2.4/site-packages/numpy/linalg/linalg.py",
> line 924, in eigh
>     raise LinAlgError, 'Eigenvalues did not converge'
> numpy.linalg.linalg.LinAlgError: Eigenvalues did not converge
> ..
>
>  If I only displace the chlorine atom, then everything works fine, and I
> get 3 frequencies.
>
>  Do you have any ideas why numpy fails when calculating the
> eigenfrequencies and how to correct this problem?
>   

The issue is in your input trajectory file: your Cu atoms have no masses 
set, although your Cl atom does. (The atomic masses are necessary for 
the calculation of vibrational frequencies, but not for the electronic 
structure.) If you create your atoms objects in ASE, they should 
automatically have the masses attached. If you create your atoms in 
another fashion, you can attach the masses using the set_masses() method 
of the Atoms object. I've attached a modified version of your script 
that has this added, and it runs fine.

Also, I just updated the vibrations module of ASE to give a more 
sensible error message when this occurs.

Best,
Andy


> (I attach the full script and the error message)
>
> Best regards
> Dominik Bjørn Rasmussen
> M.Sc. Student, DTU
>   
> ------------------------------------------------------------------------
>
> _______________________________________________
> ase-users mailing list
> ase-users at listserv.fysik.dtu.dk
> https://listserv.fysik.dtu.dk/mailman/listinfo/ase-users

-------------- next part --------------
A non-text attachment was scrubbed...
Name: start.py
Type: text/x-python
Size: 2087 bytes
Desc: not available
URL: <http://listserv.fysik.dtu.dk/pipermail/ase-users/attachments/20101026/c88b8b39/attachment.py>