[CP2K:5496] Huge number of imaginary frequencies

[hut... at chem.uzh.ch]

Hi

no real help from my side, but
- have you checked that you really used the converged geometry
- that all finite difference SCF calculations did converge (also
  converge to the same state)

regards

Juerg 
--------------------------------------------------------------
Juerg Hutter                         Phone : ++41 44 635 4491
Institut für Chemie C                FAX   : ++41 44 635 6838
Universität Zürich                   E-mail: hut... at chem.uzh.ch
Winterthurerstrasse 190
CH-8057 Zürich, Switzerland
---------------------------------------------------------------

-----cp... at googlegroups.com wrote: -----
To: cp... at googlegroups.com
From: Tobias Kraemer <161brun... at gmail.com>
Sent by: cp... at googlegroups.com
Date: 07/03/2014 03:08PM
Subject: [CP2K:5496] Huge number of imaginary frequencies

Hello,


I have completed my first vibrational frequency calculation, following a geometry optimization of an organometallic molecular
crystal. However, the output of the freq job reveals 400+ imaginary modes, ranging from -6516 cm-1 to -40 cm-1. To me this 
looks suspiciously like something went pretty wrong. As far as cutoffs and other parameters are concerned, I have done testing
to (hopefully) ensure good values are used. So I ended up using a cutoff of 1000 Ry and a rel. cutoff of 60 Ry. For consistency I have used
DZVP-MOLOPT-SR-GTH as basis on all atoms (Rh, P, C, H, B, F) in combination with the GTH pseudos. The structure seems to have 
converged to a reasonable "minimum", so from this point of view I didn't see anything odd going on. The system contains
a large number of fluorine atoms due to the presence of BArF4 (=[B[3,5-(CF3)2C6H3]4]−) counterions. I understood from previous forum posts
that fluorine can be difficult at times. I have also done an optimization using a higher cutoff (1200 Ry) and tighter convergence criteria,
but other than taking about twice as many steps as before, the bond metric parameters remained essentially unchanged.   

I used coordinates from the optimized geometry for the subsequent vibrational analysis, basically leaving all parameters untouched,
just changing corresponding keywords in the VIBRATIONAL_ANALYSIS block. All input files are attached to the post (output files were
too large to be uploaded I am afraid). Like I've said above, the geometry 'looks' perfectly fine to me, and I could have lived with perhaps 2-3 imaginary frequencies. But 450 is a bit awkward, especially with such large energy (>6000 cm-1).  

 VIB|                        NORMAL MODES - CARTESIAN DISPLACEMENTS
 VIB|
 VIB|                         1                    2                    3
 VIB|Frequency (cm^-1) -6416.230457         -6359.574009         -5920.931547
 VIB|Intensities           1.151212             1.257309             0.570843
 VIB|Red.Masses (a.u.)     1.251703             1.632844             1.281401
 VIB|Frc consts (a.u.)    -5.431171            -6.837998            -4.031972
  ATOM  EL             X     Y     Z        X     Y     Z        X     Y     Z
     1  F             0.00  0.00  0.00    -0.00 -0.00  0.00     0.00  0.00  0.00
     2  F            -0.00  0.00  0.00    -0.00  0.00 -0.00    -0.00 -0.00  0.00
     3  F             0.00 -0.00 -0.00    -0.00 -0.00 -0.00     0.00 -0.00  0.00
     4  F             0.00 -0.00  0.00     0.00  0.00 -0.00     0.00 -0.00  0.00
     5  F             0.00 -0.00 -0.00     0.00  0.00  0.00     0.00 -0.00 -0.00
     6  F             0.00  0.00 -0.00     0.00 -0.00  0.00    -0.00  0.00 -0.00
     7  F            -0.00  0.00 -0.00     0.00  0.00  0.00    -0.00 -0.00 -0.00
     8  F             0.00 -0.00  0.00     0.00 -0.00  0.00    -0.00  0.00  0.00
     9  F             0.00 -0.00  0.00     0.00 -0.00  0.00    -0.00 -0.00  0.00
    10  F             0.00 -0.00  0.00     0.00  0.00  0.00     0.00 -0.00 -0.00
    11  F             0.00 -0.00  0.00    -0.00  0.00  0.00    -0.00 -0.00 -0.00
    12  F            -0.00  0.00 -0.00     0.00 -0.00  0.00    -0.00  0.00 -0.00
    13  F            -0.00 -0.00  0.00    -0.00  0.00 -0.00    -0.00 -0.00 -0.00
    14  F             0.00  0.00  0.00    -0.00 -0.00 -0.00     0.00 -0.00  0.00

....

I wonder what I am doing wrong, since I am not sure if I trust this result. According to the vibrational analysis, this geometry is not a minimum at all,
but I don't know how reliable this is. Thought that spring to mind: should I rather perform a full cell relaxation without fixing the unit cell? Can the fluorine
atoms cause this problem, and if yes, what would be a remedy for this? Should I decrease the DX value?  

Note that the geometry optimisation was done on a local cluster, using version 2.5.1, while the frequency job was
performed externally with 2.6. I suspect that this should not make a difference though. 


If anyone has some spare time to look at my inputs and the structure, I would very much appreciate this. I am a newbie to cp2k, and following trial-and-error, have to ask a lot of questions here. 


Best


Tobias






  
  -- 
 You received this message because you are subscribed to the Google Groups "cp2k" group.
 To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns... at googlegroups.com.
 To post to this group, send email to cp... at googlegroups.com.
 Visit this group at http://groups.google.com/group/cp2k.
 For more options, visit https://groups.google.com/d/optout.
 

[attachment "opt.xyz" removed by Jürg Hutter/at/UZH]
[attachment "rhdcpenba_vib.inp" removed by Jürg Hutter/at/UZH]
[attachment "rhdcpenba.inp" removed by Jürg Hutter/at/UZH]
[attachment "rhdcpenba.xyz" removed by Jürg Hutter/at/UZH]