Problem with QMMM simulation: MM H2O too close to QM NH3.

[stclair]

Dear CP2K experts:

      Recently I am trying to do QMMM simulation for water system. First I 
fixed the QM region and try to relax the MM H2O first. But after certain 
number (~3ps) of MD runs, it shows error info: Geometry is wrong or 
EMAS_SPLINE is too small. From the trajectory file, one of MM H2O moved too 
close to QM NH3 (ON distance < 1 angstrom). 

      In my simulation, H2O-NH3 and H2O-N2O4 force fields are both not 
provided. But I define them as QM region including 4 H2O molecules between 
them. I used GAUSS coupling method to compute QM-MM interaction, so I think 
QM-MM electrostatic interaction suppose to prevent H2O and NH3 from moving 
too close (both O and N are negatively charged). Input files are attached.

      Any idea will be welcome.
      Thank you in advance!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20180403/11c2ec18/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: qmmm.inp
Type: chemical/x-gamess-input
Size: 8224 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20180403/11c2ec18/attachment.inp>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: water.psf
Type: application/octet-stream
Size: 49087 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20180403/11c2ec18/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: geometry.xyz
Type: chemical/x-xyz
Size: 38570 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20180403/11c2ec18/attachment.xyz>