First Steps with SPFIT/SPCAT
SPFIT and SPCAT are two programs written by Herbert Pickett. SPFIT fits the effective Hamiltonian to assignments which connect the experimental lines with the correct quantum numbers. Then, SPCAT creates predictions from the effective Hamiltonian. In this course, we will use LLWP to create the assignments between the experimental lines and the quantum numbers.
SPFIT takes as input a *.lin file with the assignments and a *.par file defining the starting values for the effective Hamiltonian. It creates a backup file (*.bak extension) of the *.par file and updates the *.par file with the new best parameters. The summary of the fit is reported in the *.fit file and the parameters with their uncertainties are written to the *.var file. The *.var and the *.int file (holding the dipole moments and other intensity related parameters) are then used by SPCAT to create the predictions in *.cat format. Additionally, a summary of the predictions is provided in the *.out file and if requested the energy levels are reported in the *.egy file and the transition dipole moments in the *.str file.
For the formats of the different file types, see the official documentation and its annotated version on the PROPSE page.
To summarize the different file types:
| File | Content |
|---|---|
| *.lin | Assignments which are the center positions of the transitions and their respective quantum numbers |
| *.par | The parameters of the effective Hamiltonian (rotational constants, centrifugal distortion constants, …) and any settings that are tied to the symmetry |
| *.fit | Summary of the SPFIT run |
| *.bak | Backup of the *.par file from before running SPFIT; Revert to this file if your model breaks |
| *.var | Similar to the *.par file but also gives the uncertainty for each parameter |
| *.int | Dipole moments, partition function, temperature, … |
| *.cat | Predicted transitions |
| *.egy | Energy levels |
| *.str | Transition dipole moments |
| *.out | Summary of the SPCAT run |
Secondary literature on these programs exists (e.g. by Brian Drouin) but often worked examples of similar molecules are the most helpful resources.
How To Start
For most use cases it makes sense to first set up an initial model. Starting values can either come from simple assumptions (as seen in the previous section) or from quantum chemical calculations. Running SPCAT on this initial model will yield predictions that match the experimental spectrum qualitatively but not quantitatively. Then we, the spectroscopists, have to correctly assign the predictions to the lines in the spectrum. With the assignments we can refine the model by running SPFIT and subsequently creating new predictions by running SPCAT. This starts a bootstrapping process of assigning more transitions and further refining the model.
