<div dir="ltr">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 </div>