Huge number of imaginary frequencies

[Tobias Kraemer]

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






-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20140703/1fdf4f41/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: opt.xyz
Type: chemical/x-xyz
Size: 27765 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20140703/1fdf4f41/attachment.xyz>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: rhdcpenba_vib.inp
Type: chemical/x-gamess-input
Size: 2171 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20140703/1fdf4f41/attachment.inp>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: rhdcpenba.inp
Type: chemical/x-gamess-input
Size: 1784 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20140703/1fdf4f41/attachment-0001.inp>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: rhdcpenba.xyz
Type: chemical/x-xyz
Size: 36305 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20140703/1fdf4f41/attachment-0001.xyz>