PGOPHER

<Prev Next>

Known Problems and Limitations

• Spherical Tops
• Electron and Nuclear Spin effects in Symmetric Tops
• Magnetic dipole transitions in Symmetric Tops
• Internal rotation effects
• Vibrational dependence of rotational constants. (The current design requires the constants for each vibronic state to be entered separately.)
  

Known problems

• The Help menu only works on Microsoft Windows (but the documentation can be read using any web browser).
• The Linux and Mac versions are less polished than the windows version. Any problems are, however, likely to affect the user interface only, rather than the calculations. (A comprehensive test suite for the calculations pases on all systems.)

Upgrading from previous versions

• The default for display of quantum numbers from 5.2.343 is now to display, and accept as input, true values throughout, so a quantum number of 1/2 is now displayed as 0.5 rather than 1. Associated with this the variable setting the display style has changed from DoubleQnos to QuantumNumberFormat at the Mixture level. To revert to the previous style of working set this to 2J rather than the default of J. Some care needs to be taken with data files when moving between versions; while the .pgo files always have quantum numbers written as doubled values line list files used for fitting may well use the wrong format when reading as they will default to the format determined by QuantumNumberFormat (or DoubleQnos). This can be adjusted as required by adding one of the following directives to the line list file:

  QuantumNumberFormat 2J
  QuantumNumberFormat J

• The window positions and other settings can now be written to the .pgo file. If this is done the files can't be read by versions of PGOPHER before 5.2.343 - an error message along the lines of "Can't find FormSettings" will be displayed. Saving this additional information can be disabled by turning off the "Save" settings under File, File Options. Alternatively, these are separate objects in the .pgo file and can be removed with a text editor if necessary.
  
• In version 5.2 of PGOPHER a different numbering scheme is introduced to label the rovibronic symmetry for asymmetric tops. This has a more straightforward relationship with the various ways of denoting symmetry, and the schemes are as described in Axis systems and symmetries for asymmetric tops. This only affects molecules with C_2v, D₂ and D_2h symmetry, but does imply that for these the input and display of symmetry for transitions has necessarily changed. This is perhaps most obvious for line list files for fits; files produced for the old version can only be read by the new version if the following directive is added at the top of the line list file:

pgopherversion 5.1

• In the course of full implementation of the PseudoC2v setting in version 5.2 the numbering scheme for some of the symmetries has changed; see the documentation for the new numbering scheme. Some input line list files may require manual editing to work with the new version as as a result.
• For asymmetric top molecules the default value of LimitSearch in manifold is now true rather than false. Old files loaded into the new version will normally use the new setting rather than the old, which may affect the values displayed for the Ka and Kc quantum numbers in some circumstances. It does not affect the energy levels or intensities calculated, merely the algorithm used to generate quantum numbers for display. See Manifold for a discussion of this.
  
• The default IntensityUnits in Simulation are now (after 5.1.44) Normalized, rather than Arbitrary. As default values for settings are not written to .pgo files, files saved by older versions of the program with IntensityUnits=Arbitrary will now load with it set as IntensityUnits=Normalized. This change was made as a setting of IntensityUnits=Arbitrary gives misleading relative intensities for isotopologues or isotopomers with different symmetry or statistical weights, such as ³⁵Cl₂ and ³⁵Cl³⁷Cl. (The calculation is "correct", in that the partition function is excluded by design when IntensityUnits=Arbitrary but it means this setting makes a poor default.) No action is required by this change unless abundances or transition moments have been set to work around this issue.
• Statistical weights were treated incorrectly for D_2d, D_4d, D_6d, D_8d symmetric tops in versions before 5.2.513. Files prepared with older versions using these point groups may give a warning on loading, and the statistical weights will typically need correcting.
• After version 5.3.137 a "Parameter" directive is normally required at the start of a line in an observation file adding a parameter as an input value for fit. (Without this directive otherwise valid input files could be misread.)
• In versions before 5.3.330 the relative intensity of magnetic dipole transitions was not always correct with respect to electric dipole transitions. For the units used for the Strength of transition moments (Debye for electric dipole, Bohr Magneton for magnetic dipole) the intensities of magnetic dipole transitions need to be scaled by 10⁴²μ_B² = 8.6007252×10^-5 to give the correct relative intensity. This factor was included when the simulation had units (IntensityUnits = EinsteinA, nm2MHzperMolecule, cm2WavenumberperMolecule or HzperMolecule) but not otherwise. The appearance of simulations with no units will not change unless both electric and magnetic dipole transitions are present in the same simulation. To obtain the same absolute intensities scale the magnetic dipole moments by 10^-21/μ_B.
• In versions before 5.3.335 intensities of transitions arising from a rotational magnetic moment in a linear molecule were a factor of 3 too large. The units for rotational magnetic moments have also been switched to nuclear magnetons, μ_N, from Bohr magnetons, μ_B so the Strength will need scaling by m_e/m_p.
• The Rank property was been removed from some TransitionMoment objects in version 5.3.335, as it was superfluous.
• The branch label format for symmetric tops has changed in versions after 5.3.335. All spaces have been removed, and the kl labels now show as '+', '-' or nothing, rather than '+1', '-1' and space. In addition a problem with degenerate states in symmetric tops has been fixed; the kl label was sometimes taken from the upper, rather than the lower state in the transition.
  

Fixed problems and other changes

• Fixed problems in 6.0.111
• Changing the rank of a spherical transition moment did not take effect until saved and loaded from a .pgo file.
  
• New Features in 6.0.110 (Note that not all changes have been listed for this major revision)
  
• New vibrational structure mode.
• Effects of static external electric and/or magnetic fields can now be simulated.
  
• Wavefunctions can now be complex if necessary, though this is disabled by default.
  
• State dependent widths implemented for symmetric tops.
  
• Variable character size on printer, clipboard and exported plots can now be disabled.
  
• Convolute option added to experimental plots to allow them to be convoluted with the simulation linewidth(s).
  
• "Add all levels" to prepare custom populations can now be applied to a state as well as a manifold.
• Add InterpolatedPartitionFunction objects to allow partition functions to be generated from external tables as a function of temperature, such as those available from HITRAN.
  
• Partial implementation of C_3h symmetry.
• Add (partial) import of OceanOptics .ProcSpec files.
• Symmetry directive in observations file to select symmetry to use on BranchFormat input for cases such as symmetric tops where two otherwise identical transitions are found.
• Add centrifugal distortion of qplus term for symmetric tops.
• Implement 'auto' component for transition moments in symmetric tops.
• Add symmetric top to the choice available in File, New.
• Pick destination form switched to using radio buttons to select destination channels and add "None" button.
• Spin terms added as perturbation types for asymmetric tops, allowing higher centrifugal distortion terms in spin-rotation coupling to be added.
• Add "ScalePrev" setting to perturbation types to allow fixed combinations of operators to be expressed, such as that required for B_bar in asymmetric tops.
• "Split", "As Linelist" , "Hide Out of Range" operations added to overlay menu on plot.
• "Copy Branch Table" added to "More" menu on line list form
• "Spline" option and dialog box in calibration form.
• Log output control ("PrintLevel") added to log window
• Detailed printout of parameters and matrix elements to log if PrintLevel set appropriately.
  
• Initial External Molecule implementation.
• State eigenvector can now be limited to displaying the n largest contributors to any state, rather than all basis states.
  
• Fixed problems in 6.0
  
• kl label not always generated from the correct state for symmetric top transitions involving degenerate states.
  
• Intensities of transitions arising from a rotational magnetic moment in a linear molecule were a factor of 3 too large. The default units have also been switched to nuclear magnetons, μ_N, from Bohr magnetons, μ_B.
• Plots exported as SVG can now be read directly by Firefox.
  
• For transitions with contributions from multiple ranks (e.g. rank 0 and rank 2 together) some low J transitions could be excluded
• Jmax for upper manifold was being taken from lower manifold
• Multiphoton transitions in symmetric tops had been broken by previous change.
• HITRAN line list input can now cope with states with vibrational angular momentum.
• Accept .5 as well as 0.5 for quantum number input.
• Allow "$$" at start of line for comment in .jdx files.
• Allow "," at end of free format numbers in .jdx files.
• Adding new transition moment for symmetric top molecules failed in tree view.
• Adjust branch format (such as pQ(10)) to make use on input easier.
• Bitmap overlays would not save after being compacted (needed fix to pngimage library).
• Edit, Copy Picture and File, Export, Plot menu items split into separate items for Fortrat and main plot window. Also fix right click on plots to copy correct plot.
• Etalon peak finding mode works with partially calibrated plot; speed up.
• ComboBox form settings were not restored corectly from a .pgo file if previous value was not the default.
• Speed up selecting states by index number; this was very slow if no J selection.
• Add choice of methods for timing; note that no method is guaranteed to give sensible values, particularly on multiple CPU systems.
• Empty line list not always saved as empty.
• JCAMP-DX files with units of "MICROMETERS" not loaded correctly.
• Transitions with Colour = clHide now not used for contour fitting.
• Some default values written to .pgo files.
• Use "FrequencyScale" on plots adjusted so that centre point of plot does not shift in frequency.

• New Features in 5.2.524
• Import of selected .svw files from Ohio SpecView program.
• Right click in Fortrat plot only shows lines close to selected J. (Previously only the horizontal position of the mouse was used to select lines.)
• Speed up of importing spectrum data files.
• Paste of file from Windows Explorer implemented.
• Initial implementation of spline fitting for calibration.
• Option to print constants used in I₂ calibration simulation.
• Implemented equivalent nuclei in linear molecules, and initial implementation for asymmetric tops (C_2v, pairs only).
• General constraints on parameter values in fits.
  
• Fixed problems in 5.2.524
• Right click in Fortrat diagram was not always giving lines under mouse.
• Read of linear molecule JPL .cat files when J not specified in quantum numbers in .cat file was failing.
• Drag and drop of .pgo file on to plot window discarded limits.
• Save Overlay and Save Overlay As menu items in main window failed if Overlays window was not open.
• Display of partition function sum was not using QuantumNumberFormat.
• Calculate Boltzmann factor for cm^-1 and eV by conversion so that identical populations are obtained when compared with energy levels calculated using MHz. Previously a value tabulated to the (relatively low) precision of k was used, giving slightly different populations with notionally identical constants
• Unit conversion incorrectly treated C and D in a transition.moment as having units.
• Intensity incorrectly scaled for some IntensityUnits when the PlotUnits were different to the calculations Units.
• Crash on hitting F1 for help in main window.
• Adjust height of individual items in the constants window to make it less likely to show scroll bars when not really needed, and make sure selected item in tree is always visible.
• In line list displays, make column heading "Position" rather than "Frequency" (as line position is often not in frequency units).
• Fix column header alignment in line list export.
• Line list for IntensityUnits = PopDist was blank, and fix related problems when right clicking on a population distribution plot.
• Optimize logic for not saving default form settings.
• Force legal symmetries following point group changes.
• Possible I/O error 105 in Slave Editor.
• Width conversion to/from velocity scale wrong.
• Default Kmin and Kmax display as "all" for asymmetric tops.
• Foffset on simulation incorrectly taken into account when peak measuring on a simulation in calibration mode.
• Adding a new linear state to a manifold in the tree view defaults to the same g/u value as the other states in the manifold.
• Close and re-open transition selection window does not loose manifold/state selection.
• "Auto" checkbox in linelist dialog was always treated as checked.
• Measure using mouse of peak position on simulated spectrum does not result in automatic assignment in line list window.
• Right click in vertical axis region of simulation should not find lines.
• TransitionMoments with more than 2 components and at least 2 different ranks may not simulate correctly.
• TransitionMoment selection with the transition selection window not working.
• Various fixes to calibration methods.
• Potential crash when form closed when editing a value in a grid.
• Statistical weights wrong for D_2d, D_4d, D_6d, D_8d. As a consequence old data files with these point groups will give a warning on loading.
• Symbolic matrix element output now honours QuantumNumberFormat.
• Transition strength calculation fixes for symmetric tops.
  

• New Features in 5.2.343
  
• Slave Editor added to distribution.
  
• "Hide others/Show All" menu items in overlay context menu on main and overlay forms.
• Colour = clHide lines now excluded from consideration in more contexts, including line count, scaling and right click on plot.
  
• Choose random overlay colour (rather than black) if no unused colours available.
• Upgrade to Eigenvalue routines from version 3.1 of LAPACK.
• Implement Branch style input in .lin files for symmetric tops.
• Add velocity (in units of cm^-1/2) as a horizontal scale in spectrum plots.
• Hyperfine quantum number display for basis states has been adjusted - now always shows J.
• Read .sp files from (at least some) Perkin-Elmer spectrometers.
• Prompt when clearing line list window when there are some assigned lines.
• Swap frequency and intensity, discard intensity operations added to experimental plots.
• Mouse over on observed-calculated line on main plot shows linelist number on both ends.
• Small adjustments for Microsoft Windows Vista.
  
• Temperature in I₂ calibration spectrum independent of overall simulation temperature.
• Simulation parameters (such as linewidth and temperature) are now recorded in a separate Simulation object rather than the base Mixture object. This allows for multiple Simulation objects corresponding to multiple simulation windows. For back compatibility if there is only one Simulation object its values and settings are written to (and read from) the base Mixture object. .pgo files written with more than one simulation object can't be read by previous versions.
  
• Optionally save and restore more settings, including parameters for energy level and transition windows, and form positions.
• Constants and overlay windows will rearrange their contents depending on their aspect ratio.
  
• Replace DoubleQnos with QuantumNumberFormat, Mixture, and set the default to undoubled quantum number display.
  
• Symmetry selection in the transition (and other) windows now uses symmetry names, rather than numbers.
• Write version to .pgo files. (It is ignored on input for the time being.)
• SMargin now implemented in simulation, which defines how many additional points to calculate off either end of a spectrum.
  
• Fixed problems in in 5.2.343
• Do not permit JAdjustedSym when hyperfine structure present in asymmetric tops.
  
• Allow .pgo files that trigger errors to load so they can be fixed.
  
• Hyperfine quantum numbers not displayed for asymmetric tops.
  
• Compilation fixes for 64 bit and Mac/Intel.
• Fix crash on pasting too many items into parameters in constants form; allow copy/paste in settings grid.
• Load of some overlay files could fail for a if the frequency was not sufficiently varying.
• Cancel on "File Changed - discard?" prompt not always effective.
• Fix error handling for failures in calculating eigenvalues.
• Data file name shown in task bar.
• Detect spurious xml entries in .pgo file, rather than silently ignoring them.
• Wrong sign in documentation for spherical transition moments.

• New features in 5.2.149
  
• Experimental support for multiple CPU cores is now available for selected calculations; this is most useful for fitting. It is disabled by default, as it is only helpful for larger problems and can reveal obscure bugs. To try this, select the number of threads using the View, Threads menu item - twice the actual number of processors is a good starting point.
  
• PseudoC2v is now implemented for all appropriate point groups and full documentation prepared. The numbering scheme for some of the symmetries has changed so some line list input files may need manual adjustment.
• Change in numbering scheme for labellings rovibronic symmetry for asymmetric tops - now directly related to standard symmetry labels as described in Axis systems and symmetries for asymmetric tops
  
• Adjust overlay scaling algorithm - for IntensityUnits = Normalized now uses current Baseline and Scale.
• The least squares fit routine now displays Watson's sensitivity parameter (J. K. G. Watson, J. Mol. Spectrosc. 66, 500 (1977)) for each floated variable, and uses this to determine the number of figures to display in the estimated error for each parameter.
• The least squares fit routine will display the results of a singular value decomposition if Mixture.FitDetails is set to True. Singular value decomposition eigenvalues less than Mixture.SVDThresh will be discarded. This is normally 0, but small values can be used to limit changes to combinations of parameters that are poorly or completely undetermined.
• Parameters can now be used as "observed" values in fits as a method of applying constraints.
  
• Overall rovibronic symmetry selection has now been generalized so it applies to all types of states and perturbations.
  
• The number of lines displayed in the Line List window on right clicking on a simulation can now be adjusted (click on More,  Options in the line list window.) The default is now to display the 10 strongest lines with intensity more than 5% of the strongest, and to assign lines that are more that 25% of the intensity of the strongest.
  
• Level plot can now select by other quantum numbers, including K_a and K_c (Asymmetric tops) and Ω for Linear molecules.
  
• Better handling of overflow in Boltzmann case - see J Range and Partition Functions.
  
• Several processing commands for numerical (as opposed to bitmap) overlays, including smoothing and joining spectra together.
• Selected calibration spectra are now built into PGOPHER (I₂, Ne, Fe) for use with the new calibration system.
  
• A full calibration process is now available for overlaid spectra.
  
• IntensityUnits in overlays now implemented
• IntensityUnits = PopDist, NormPopDist for simulating time of flight spectra or similar cases where energy levels are seen, rather than transitions.
• Fitting and associated plots now work for fitting to energy levels, as well as transitions. (Use blank upper manifold name and J = '-' in fit files to indicate this).
• Colour coded names in constants and overlay tree view.
• Several improvements to the linelist window, and fix some bugs causing range check errors.
• Symbolic Matrix elements now available for Linear Molecules and Asymmetric tops
• Plate tracing has been updated to allow accurate determination of the plate angle, and selection of the area to trace.
• Float setting for parameters can now be toggled by single (rather than double) mouse click or space bar.
• Ctrl + Left mouse now also offsets spectra as well as Alt + Left mouse, to allow for Linux window managers that use Alt+Left to drag windows.
• Updates to non-Boltzmann populations.
• Add K selection to asymmetric top perturbations.
• Fixed problems in 5.2.149
  
• Asymmetric tops with C_i symmetry have approximately 1/4 of the rotational states missing.
  
• Transitions with exactly zero frequency within a Manifold are now discarded; previously these could appear twice in a line list (with zero intensity).
  
• Plot of overlay with negative intensity scale could have scale problems.
  
• Peak find on a line list including more than one peak took the first line in the range, rather than an average when all the peaks in the range had the same intensity.
  
• Asymmetric tops with C_2h symmetry have half the rotational states missing for states with A_u or B_u vibronic symmetry.
• Auto detection of frequency channel of overlay can fail if intensity channel is flat.
  
• Parameters with an absolute value less than about 10^-18 are not saved in the pgo file.
  
• Minor adjustment to phase choice for linear molecule LS perturbation matrix element for the case that Λ' = Λ. Now chooses parameter to be matrix element such that largest S is on the left.
• Right click on first spectrum from File, New, Linear Molecule did not show assignments.
• Some asymmetric top perturbation matrix elements not calculated correctly for operators containing J_xⁿJ_y^m with m ≠ n and m ≠ 0 and n ≠ 0.
• Fixed problem with Normalized intensities introduced in earlier development version
• Quantum number label is always shown as J in Fortrat plots, even if not appropriate.
• Mouse over on Level plot could indicate wrong line in observations file on status bar.
  

• New features in 5.1.222
• Convert for compilation with Delphi2006.
• JPL line lists (.cat) as overlay.
• K selection in symmetric tops.
• Cartesian transition moments now accepted for symmetric tops.
• Perturbations in symmetric tops now implemented.
• Fixed problems in 5.1.222
  
• ζ constant in symmetric tops was incorrectly converted when switching units.
• Copy with linked items, then paste fails when there is more than one transition moments object.
• Linux/Mac version: Linelist export no longer crashes
• Mac version: fixed problems due to compiler.
• State assignment in Cs or lower symmetry asymmetric tops.
• Count in exported linelists wrong.
• Intensity fits of transitions within a manifold could be to wrong state.

• New features in 5.1.164
  
• Line fits can now be to intensities as well as positions.
• Non-Boltzmann population distributions can now be entered and fit
  
• Residuals from fit can now be plotted in a variety of ways in the Log window.
  
• Multiple views of the same spectrum now available from View, Another. The plot range is independently selectable, and the frequency offset, temperature and linewidths can be synchronized or independent.
• Saturation parameter.
  
• Add Quadrupole and Nuclear Spin - Rotation matrix elements to asymmetric tops.
• Options to exclude some or all spins when line fitting to some or all lines.
• More options for energy level plots.
• More options for Fortrat plots.
• Linear molecule states can now be restricted by Ω and parity (This is particularly designed for Hund's case (c) or other cases where there are many perturbations.)
  
• Linear molecule perturbations can now be symmetry and omega selective
• Least squares fit to line positions now gives file and line number in the output log.
• HITRAN format line lists (.par files) can now be used as overlay files or as line position or intensity input to least squares fits.
  
• Fixed problems in 5.1.164:
  
• Preliminary fix for C_3v perpendicular band intensities.
  
• Change to default method for displaying quantum numbers for asymmetric top molecules (see discussion above).
  
• Drag and drop of multiple files of the same type forces merge mode
• Improve prompt when loading files with two or more spectra in the same file.
• Mis-pelt colour names in observation file now gives error message, rather than prompting for colour.
• Fixed problem where save (rather than save as) of overlay file sometimes goes to wrong directory.
  
• Intensities more consistent when switching from spectra with width to "stick" spectra where width does not show.
  
• Fix problem with vertical axes on multiple plots.
  
• Fix incorrect printing of weighted overall residual after fit.
• Negative and zero fit increments not displayed in constants window
  
• Allow units in observed line file to differ from units in constants window.
  
• Export of individual overlay of multiple fixed
  
• File including itself gives error message rather than crashing
  
• Adjust partition function convergence algorithm and add AutoQConverge setting.
  
• Copy and paste of linked items in constants window now renames transition moments
  
• Decimal separator now forced to be "." regardless of system setting.
  
• Default IntensityUnits now Normalized (See comment about upgrading above)
• Switch Delphi version to use FastMM memory manager
  
• mkpgopher.bat build script fixed
• Exclude *.cfg files from source distribution
• Minor change to number output style
  

• In 5.1.44
• Cancel in rename of new matrix element deletes new object
• Remove possibly confusing test for ungerade states in symmetric linear molecules
• Add Suncouple as perturbation type for linear molecules
• Allow state number selection in transition dialog
• Correct error in formula given in the documentation for linear molecule statistical weights

• In 5.1.42
• In level plots:
• Fix observed level plot problem with emission spectra
• Simple symmetry selection
• Observed level plot can now be both upper and lower states simultaneously
• Fix observed level plot problem where transition is within a single manifold
  
• .pgo files (these changes are backward compatible)
• "Name" of matrix element style items no longer written to file ("Bra" and "Ket" are sufficient).
• Display name in tree of Mixture is just filename part of full path
  
• Sample data files:
• Add ¹⁸O¹⁶O to show different isotopes and symmetry
  
• Add ¹⁷O¹⁶O to show nuclear hyperfine structure
• Clear observed line marks menu item
• Colour of sum spectrum is now always black
• Fix plotting glitch with multiple spectra
  
• Add "Hide" colour; "None" will show as black.
• Documentation: Add "forbidden transitions in linear molecules" section
  
• Fix display glitch in bitmap overlays
• Fix state assignment problem when EigenSearch = True and LimitSearch = False
• In 5.1.34
• CFBr sample data file corrupt
  
• Octic centrifugal distortion in symmetric tops
• States form display fix
• In 5.1.34
• C_3v symmetric tops now working
• Fixed crash when deleting overlays with overlays property window open
• In 5.1.32
• File, Print and Edit, Styles now working
• Version numbers now consistent