I am trying to optimize the geometry of a reactant complex obtained from the IRC. The geometry of the product complex is getting optimized, but in the reactant complex, the maximum force is not converged.
Item Value Threshold Converged?
Maximum Force 0.001652 0.000450 NO
RMS Force 0.000225 0.000300 YES
Maximum Displacement 0.000169 0.001800 YES
RMS Displacement 0.000041 0.001200 YES
tried using keywords such as opt=calcfc with the initial guess of the geometry, opt=caclall with the geometry where force is not converged, and slightly modifying the geometry obtained from the IRC.
Any other suggestions would be highly appreciated.
Perhaps the reason is that the reactant complex does not exist as a whole, but parts fall apart.
Related to the answer above from Alexey - assuming that your reaction involves two molecules coming together, their interaction may be extremely small such that the potential energy surface has no discernible minimum in this region. The idea of a "reactant complex" on the PES may not be valid.
I'd probably be optimizing the two species separately in this case. But if you really want to optimize the "reactant complex" you can play around with the dynamic trust radius (i.e. the optimization step size) with the maxstep and notrustupdate options for the opt keyword. Be aware this may take a while...!
Thank you, Alexey N. Masliy and Robert Paton, for your answers. When I checked for Energy v/s steps during the optimization. It was seen that the energy was minimized when the three criteria converged. Therefore, I changed the convergence criteria of Gaussian by using iop(1/7=x) and used the geometry where the three criteria converged. The energy values are comparable to similar structures optimized without this issue.
- Badges
- Science method