Hi Cyrille,
I think that the two H2O molecules (i.e. the axial ligands) should stay oriented along the C axis only if the complex formed with the Co atom is octahedral. As far as I can see, the ligand field (i.e. the spatial distribution of coordinating ligands around the Co atom) in your system is not perfectly octahedral. This might drive your system away from an equilibrium geometry in which the two H2O molecules are axially oriented. Another possibility, is that the U that you are using is inappropriate and does not favor octahedral coordination of the ligands.
There could be two things that could be done to improve the situation:
1) I would first try to see if the U parameter that you are using is suitable to describe an octahedral complex. I would construct a Co(H2O)6 molecular complex, in which the 6 H2O molecules are coordinated in octahedral geometry (C-Co-C angles should be 120 degrees). If the U parameter that you are using is appropriate to describe an octahedral complex, then this octahedral complex should be stable upon geometry optimization. I would also try to distort this complex in the equatorial position, so that the angles resemble those that the Co atom has with the organic ligands in the 2D MOF, and relax it to see if it recovers an octahedral geometry.
2) I would try to start the optimization of the 2D MOF from a geometry where the coordination sphere around the ligand is perfectly octahedral. This means that the angles between four equatorial C-Co-C ligands should be perfectly 120 degrees.
Regards,
Daniele.
