DParFit 3.3

DParFit performs least-squares fits of a diatomic molecule spectroscopic data set consisting of any combination of microwave, infrared or electronic vibrational bands, fluorescence series, photo-association spectroscopy binding energies and/or input Bv values, involving one or more singlet electronic states and one or more isotopomers, to parameters defining the energy levels of each state. The levels for each electronic state may be described by one of:
  1. band constants Gv, Bv, Dv, ... etc. for each vibrational level of each isotopomer,
  2. generalized Dunham expansions,
  3. near-dissociation expansions (NDE), or
  4. mixed NDE/Dunham functions (MXS), or
  5. independent term values for each vibration-rotation level of each isotopomer.
The effect of lambda doubling or doublet-sigma spin splitting is represented either by band-type constants for each vibrational level of each isotopomer, or by Dunham-type expansions in (v+1/2). Dunham, pure NDE or MXS fits automatically incorporate normal first-order semiclassical multi-isotopomer mass scaling, and allow for inclusion of atomic-mass-dependent Born-Oppenheimer and first-order JWKB breakdown corrections (collectively called BOB corrections) for each electronic state.

The source code for DParFit 3.3 and the associated user manual University of Waterloo Chemical Physics Research Report CP-660 (2005) are distributed free of charge. However, I would appreciate it if use of this program were acknowledged by citation of the above report. The manual describes the various fitting and level-energy representation options and provides a detailed description of the nature of the input data files and the options invoked by various input parameter choices. It is available as a PDF file which may be downloaded or read and printed from the above www link. The FORTRAN source code and the sample data file used to generate the sample outputs shown in the manual are automatically distributed (by e-mail) to anyone who fills in the form: Download DParFit .

DParFit 3.3 has a number of additional capabilities beyond those of early versions of this code. Noteworthy items new to version 3.3 include 1.(b), 1.(c) and 9.

  1. Lambda-Doubling Changes, Corrections and Extensions
    (a) The isotopomer mass-scaling of Lambda doubling expansion in Eq.(17) in the manual and in the code has been now utilizes the proper inverse-square reduced-mass factor associated with each power of [J(J+1)].
    (b) The treatment of Lambda doubling has been generalized to treat cases with Lambda > 1 properly, so the leading term now has the proper J and reduced mass dependence {[J(J+1)]/mu2}Lambda
    (c) The input of the parameters controlling the treatment of Lambda doubling have been separated from those associated with mechanical rotation. This has significant implications w.r.t. the structure of the input data file. In particular: the old reads #7 & 19 are gone, and reads #9 & 21 are changed (to omit parameter NQC).
  2. Electronic states with doublet-sigma symmetry may now be treated, and the associated spin-rotation parameters may be represented in either band-constant or (mass-scaled for different isotopomers) Dunham-type expansion format.
  3. Absolute binding energies, such as those obtained from photo-association spectroscopy, may now be introduced as a distinct data type.
  4. For any given electronic state, the fit may may be used to determine an independent parameter (the "term value") for each distinct vibration-rotation-parity level of each isotopomer.
  5. Vibrational energies and inertial rotational constants Bv may be represented by the MXS "mixed" Dunham/NDE functions proposed by Tellinghuisen
  6. Pure NDE (or MXS) expressions for Bv may now use the full range of types of NDE functions, rather than only the exponential-NDE form.
  7. (a) Band-constant fits to CDC's may be combined with analytic (Dunham, etc.) treatments of Gv & Bv, and
    (b) Band-constant fits to all mechanical rotational contants (Bv's and CDC's) may be combined with analytic (Dunham, etc.) treatments of Gv.
    These extensions facilitate the treatments of modestly perturbed states.
  8. "Robust" least-squares fits may now be performed: see J.K.G. Watson, J. Mol. Spectrosc. 219, 326-328 (2003), and references therein.
  9. Parameter PRCMO input on data line #3 has been replaced by PRINP which allows a user to print out a summary of the input data immediately after it is read in (rarely useful). The compact [calc.- obs.] fit summary now is always printed at the end of each run, and the full [calc.- obs.] output listing is always written to Channel-8.
Beginning with version 3.0, there were a number of minor modifications of and additions to the channel-5 "instruction" data file, which are outlined on the link: Version-3 Data File Modifications

The earlier upgrade from Version 1.0 of this code incorporated a number of enhancements, including a re-organization of the required form of the main experimental data input file. For current users of Version 1.0, a summary of, and partial explanation for these changes found in the link Version 2.0 Updates. This "Updates" link also provides access to a script for converting input data files prepared for use in Version 1.0 to the modified format expected by the current version.

I would be pleased to receive information regarding apparent or real malfunctions of this code, or weakness in the documentation, by post [to the address on the manual], by voice [at (519) 888-4051] or by e-mail to the address given below. However, the program should be fairly robust, and is fairly extensively commented internally. If you encounter problems obtaining a copy of the manual from the www, a postscript (or if necessary, paper) copy may be obtained by sending me a note by e-mail.

Site owned by R.J. Le Roy: leroy @ UWaterloo.ca

This page was last updated on