Addressing the research question: Is the outlying charge error significant in the gradient calculations? Is the outlying charge correction enough to remedy the issue?
TODO: Develop a set of molecules to test the error (Jenny and Matthew)
- Made a directory called benchmark with a file for unoptimized anions containing both small(less than 50 atoms) and large(bigger than 50 atoms) anions
TODO: Decide on the COSMO parameters to test (Matthew)
- Dft
- grid 5
- func pbe0
- weight derivatives
- stepsize .001
- scfconv 6
- denconv 1d-7
- rij
- def2-SVP
cosmo parameters:
- nppa 1082
- nspa 92
- dielectric constant 80 (water)
Based on prelimiary results, these parameters are chosen to maintain consistancy with the ground state calculations.
TODO: Compute the outlying charge error (Jenny and Matthew)
- Determine the converged single point energy calculations for each molecule with the given parameters
- Read, understand, and, if necessary, modify the NumForce script to compute the gradients (dE/dxyz) with COSMO
- Determine the significance of the outlying charge error