Section Fit

This section stores information about the fit functions, which are used for the Coulomb potential evaluation.

Unrestr.SumFrag

A logical that flags whether or not the fit coefficients have been set and stored for the sum-of-fragments, but adjusted for the unrestricted fragments option (see the keys UnrestrictedFragments, ModifyStartPotential).

coef_SumFrag

Fit coefficients pertaining to the sum-of-fragments charge density.

coef_SCF

SCF fit coefficients.

nfset

Total number of fit function sets (not counting the Cartesian sub functions, not counting the copies of the functions on the atoms of an atom type)

nfitpt

Index array: 1+the total number of fit function sets up to, but not including, the indicated atom type.

nqfit

Main quantum numbers of the fit sets

lqfit

Angular momentum quantum numbers of the fit sets

alffit

Exponential decay factors of the STO fit sets.

fitnmr

Normalization factors for the STO fit sets.

nfos

Total number of Cartesian fit functions, not counting copies on all atoms of an atom type, but including all (for instance, 6 for a d-set) Cartesian sub functions.

nfptr

Index array: 1+ total number of Cartesian (see variable nfos) fit functions, up to but not including the indicated atom type.

nprimf

Total number of Cartesian ('primitive') fit functions, counting also the copies on all atoms of each atom type.

nsfos

The total number of fully symmetric (A1 symmetry) fit function combinations that represent the true dimension (variational freedom) of the space of fit functions in the calculation.

na1ptr

Index array, like nfptr, but applying to the nsfos symmetric function combinations.

niskf

This refers to an atom-limited symmetry combination of primitive fit functions, in the code and some documentation indicated as a 'g'. A 'g' is the specific part of a molecule-wide A1 fit function combination (see nsfos) that consists of all the terms that are centered on one particular atom. The number niskf gives the total number of such 'g' function combinations.
To clarify this, consider an A1 fit function combination in the molecule. Assume, that it consists of a specific linear combination the following functions: a p-x function on atom A, its partner p-y function, and the corresponding p-x and p-y functions on atom B. (Atoms A and B must be symmetry equivalent). In this example we have one A1 function (in the list of nsfos such functions) and two 'g''s. Each 'g' consists of a p-x and a p-y function combination on a specific atom.

iskf

Compound index array. It runs over the niskf 'g' fit function combinations and has 4 entries for each function (1:4,1:niskf). The meaning of the entries is as follows. #1=number of the fit set (not counting the copies of fit functions on different atoms of an atom type, and not counting the Cartesian sub functions) this 'g' belongs to. #2=index where the combination coefficients for this 'g' start in the arrays cofcom and numcom (see next). #3=number of terms in the expansion of this 'g'. #4=number of the molecular fit A1 function combination this 'g' belongs to.

na1cof

Length of the arrays numcom and cofcom, see next

numcom

Numcom (and cofcom) consists of a sequence of smaller sub arrays. Each sub array gives the expansion of a 'g' function in terms of the Cartesian functions in the pertaining fit function set. The elements of numcom specify the particular Cartesian sub functions that participate in the expansion. Its values are therefore limited to lie between 1 and (L+1)(L+2)/2, where L is the maximum l-value occurring in the fit function sets.

cofcom

Compare numcom: cofcom gives the actual expansion coefficients for the expression of a 'g' function in primitive Cartesian fit functions.