Potential Fitting for Hindered Rotors in Cantherm #1007
Description
Currently hindered rotor potentials are fit to a Fourier series and a cosine/symmetry fit and the best is used. This has resulted in numerous problems related to using too few or too many Fourier coefficients and the cosine/symmetry fit being too simple to capture the potential.
Why don't we just use a simple set of cubic splines instead? The problem with a Fourier series is that it is a sum of oscillatory modes so adding an extra Fourier series term means adding an extra oscillatory mode to your potential, which can distort the potential shape. Cubic splines are local and thus won't have that problem. The only real issue with this is if quantum happens to be on, in which case we could switch back to the Fourier basis or use a different solution method.