I want to calculate potential energy curve and change in dipole moment for F- (fluorine anion) approach to the H2 molecule. Since, dipole moment of ionic molecule depends on the choice of reference, which reference is more convenient: center of mass of H2F- or center of charge or simply calculate at the center of mass of only H2 molecule?
The proper reference point is the center of charge. (Same formula as center of mass, but you need to replace the mass with the charge.) One can in principle then use the atomic partial charges derived by for example the Mulliken scheme. However, this will indeed be very sensitive to the basis set etc. and integrating over the charge distributions derived from the ab initio calculation should be better.
Thanks Saba. But it still did not answer my question, ie., where should I put the origin to calculate the dipole moment?
I think, in Dalton we just write a charge in the geometry of molecule.. After optimization, we use that geometry to calculate the dipole moment.... you can find the test examples in the folder, which already came with the program installation..
Thanks Alamgir. However, as charges obtained in ab-initio calculation are not observables and strongly dependent on the basis set. So a more accurate way of calculating is the application of dipole moment on the optimized wavefunction. To do that we need to define the dipole moment without any reference (for neutral molecules) or at some reference point for an anion.
The proper reference point is the center of charge. (Same formula as center of mass, but you need to replace the mass with the charge.) One can in principle then use the atomic partial charges derived by for example the Mulliken scheme. However, this will indeed be very sensitive to the basis set etc. and integrating over the charge distributions derived from the ab initio calculation should be better.
@Thomas. Thanks a lot. Could you please provide some references or web pages, (if you have any) where this procedure has been mentioned elaborately.
Check 'multipole expansions' on wikipedia and the book 'The Theory of Intermolecular Forces' by A.J. Stone. ISBN 019855883X, 9780198558835
@Somnath. If you wish to estimate the grade of deformation of fluoride anion, you could also consider AIM analysis due to R.Bader. In particular, atomic dipole (and, dependent on system geometry, quadrupole) moment of the anion will give you a useful hint. As atomic quantity, it always calculated with the nuclei position as the origin. For described task of the AIM analysis, free version of AIMAll will be enough.
@Oleg. Thanks a lot. I had a look at the at the AIMAII version. It is a great platform to perform to AIM analysis. However, I am not sure whether it can handle MRCI wavefunction, where most of our calculation are at this level.
@Somnath. Speaking practically, one may carry out AIM analysis for any wavefunction written by a QC package in .wfn file (at least until ignoring the energy decomposition results). Closer to theory, the AIM analysis requires only 3D map of electron density, which is represented in .wfn as natural expansion of the first-order density matrix, i.e. by a set of (orbital, occupation number) pairs. Therefore any wavefunction based on the exact Hamiltonian is completely OK. However, AIM energy decomposition task requires some caution. For non-equilibrium geometries, IQA energy decomposition is implemented in AIMAll; for DFT wavefucntions, the "decomposition" has nothing in common with the DFT energy.
For the aforementioned task of atom (ion) deformation, any wavefunction (even DFT) is in principle suitable. The physics produces other restrictions though. If DFT then it should be LC-corrected like spectroscopic functionals. Any CI wavefunction except FCI is not size consistent, so will not allow to correctly estimate the F-/H2 interaction energy by standard counterpoise scheme. I would use CCSD here.
Thomas:
I'm curious, what is the reason the center of charge is the proper origin.
Thanks!
@Carlos, Imagine that you have a system with just one point charge! If you now place the origin of your multipole expansion on the position of this charge (which is then also the centre of charge) you will get that the charge of the system equals to that charge and all higher multipoles will be zero. This is the most convenient situation. Of course you can use other points mathematically, but it does not give much sense to say that a system with a single point charge has higher multipole moments. In systems with real charge distributions you essentially remove the contribution to the multipole moments from the overall charge by using the centre of charge as origin. The overall charge then only contributes to the first (the charge) term. Reporting the origin you use is of course still a good idea as obviously not every know that the centre of mass is the best.
- Badges
- Science topic
- Similar topics
- Chemistry
- Theoretical Chemistry