SEAMM Gaussian Plug-in

A SEAMM plug-in for A SEAMM plug-in for Gaussian

• Free software: BSD-3-Clause

• Documentation: https://molssi-seamm.github.io/gaussian_step/index.html

• Code: https://github.com/molssi-seamm/gaussian_step

History

2026.10.29 – The double hybrid B2PLYP; the dispersion correction in the model

• Added the double-hybrid functional B2PLYP (Grimme 2006), with the dispersion corrections GD3 (Grimme’s D3 with zero damping) and GD3BJ. Gaussian computes its MP2 part with the core frozen, as it does for MP2, and has analytic frequencies for it, so Thermodynamics runs an optimization and analytic frequencies at, for example, B2PLYP/cc-pVTZ with EmpiricalDispersion=GD3, as in the CBH-ANL work.

• The dispersion correction is now part of the model, as for ORCA’s D3ZERO and D3BJ, since it changes the energy: B2PLYP-D3(0)-FC/cc-pVTZ for GD3, B3LYP-D3(BJ)/6-31G(d) for GD3BJ, and -D2 and -PFD for GD2 and PFD. This changes the names of the properties of the functionals that already offered a dispersion correction (B3LYP, PBE, PBE0 with GD3BJ, GD3, GD2 or PFD), which were stored under the model without it, e.g. B3LYP/6-31G(d); calculations without a dispersion correction keep their names. The energy of formation still uses the atoms’ references of the model without the correction, which are the same, since a lone atom has no dispersion energy.

2026.10.28.1 – Fingerprints of the settings of calculations; standard names

• The Energy, Optimization and Thermodynamics sub-steps store, with their results, a fingerprint of the settings of the calculation – settings fingerprint, element fingerprints and calculation (#Gaussian#<model>), and the frequency ones for Thermodynamics – so that one can tell whether results on different structures, such as the energies of the species of a reaction, come from the same kind of calculation. The fingerprint is a hash of the route, less the task, output, guess and Link 0 lines and the structure’s title, charge, multiplicity and geometry, and less the R or U that the step chose from the multiplicity (one the user chose stays). The release and revision letter of Gaussian count (G09 E), not the revision’s number.

• The energy is also stored as total energy (beside E) and, at 298.15 K, the absolute enthalpy as H(298.15 K), the names the quantum-chemistry steps and the Thermochemistry step share. These and ZPE, H and T are always stored as properties, whether or not they are selected as results.

• Requires seamm 2026.10.28 and seamm-util 2026.10.28.1.

2026.10.28 – Bugfix: every Gaussian calculation failed at the start

• Since 2026.10.6 the Energy, Optimization, Thermodynamics (including the composite methods such as CBS-QB3) and Wavefunction Stability steps stopped before running Gaussian with “AttributeError: … has no attribute ‘_timing_data’”. The old timing code that 2026.10.6 replaced was still called in those steps; it is now removed, and the timing records are written as intended.

• New tests run the Energy step end to end, with a recorded Gaussian run and, where Gaussian is installed, with Gaussian itself.

2026.10.24 – The optimized structure keeps its name

• The structures that optimizations write are now named ‘{current name} optimized with {model}’ by default, keeping the name the structure had, such as a title read from an SDF file: ‘DMC_cc optimized with B3LYP/6-31G*’. Before, every structure was named ‘optimized with <model>’, so the titles of several molecules were lost. Flowcharts already saved keep their setting, and the previous choice remains in the list. Needs seamm 2026.10.24.

2026.10.11 – The step declares what its cost model is made of

• substep.TIMING_SPEC – basis functions, electrons and atoms as the size variables, the method as the class, the task, and SCF runs as the unit – is passed with each timing record (seamm-exec 2026.10.6.1 and later write it beside the records), so the cost model is fitted from the step’s own description.

• Removed the unused Docker option from the gaussian.ini template.

• Requires Python 3.12 or later.

2026.10.6 – Timing records that a cost model can be fitted to

• Each Gaussian run appends a record to ~/.seamm.d/timing/gaussian.csv through seamm_exec.timing: the machine class, threads, wall time and outcome, and the descriptors of the calculation – the model, method and basis, the route’s kind of task, symmetry detected and used, basis functions and electrons, the structure, and from the log the SCF runs and cycles and Gaussian’s own CPU and elapsed times. This replaces the step’s own CSV, which grew without bound. See seamm_exec’s campaign of 2026-10-05.

• Requires seamm-exec 2026.10.6.

2026.10.1 – Settings that depend on each other; naming choices fixed

• The dialogs show only the settings that apply with the current choices, and SEAMM’s flowchart tools use the same rules: the functional and integration grid only for DFT; a dispersion correction only from the functional’s own list (now also GD2 and PFD); freezing the core only for methods that can; the basis set only for methods that use one; Wavefunction Stability takes its method from ‘method’.

• Bugfix: Energy and Optimization listed their naming choice twice for the configuration name and dropped “keep current name”; each name now offers the step’s own choice alongside the usual ones. Defaults are unchanged.

• Flowcharts saved with the dispersion correction DG2, a misspelling of GD2 that Gaussian rejects, are read as GD2.

• Documented in the user guide. Needs seamm 2026.10.1.

• Internal: CI now installs the package’s declared dependencies with uv rather than a conda test environment.

2026.9.29 – Bugfix: an installer that writes gaussian.ini, and cube files without it

• Added gaussian-step-installer, which the SEAMM Manager runs when the step is installed, to write the commented template to ~/SEAMM/gaussian.ini for you to edit, rather than only on the first calculation.

• Making cube files (cubegen) read gaussian.ini without creating it or finding Gaussian, and its error named MOPAC’s mopac.ini. It now finds Gaussian the same way as the calculations do.

• When Gaussian cannot be found, the error says what to set, rather than reporting a missing section.

• The documentation describes installing with the SEAMM Manager and configuring gaussian.ini.

2026.7.28: Energy of formation via seamm_thermochemistry; initial checkpoint job:// support

• The energy of formation is now computed via the shared seamm_thermochemistry library: a 0 K, ZPE-free energy of formation referenced to the elements, computed for every job with an electronic energy rather than only the subset with a full harmonic thermochemistry calculation. Part of unifying the atomic reference-energy handling across steps.

• Initial checkpoint (on the Energy and Wavefunction Stability steps) can now reference another job’s checkpoint file, via job://<job number>/<name> (SEAMM’s Node.file_path gained read-only cross-job references) – useful for seeding from a checkpoint computed in a different job.

2026.3.1: Internal: switching from deprecated library pkg_resources to importlib

2025.8.20: Added the atomic reference energies for wB97XD and M062X functionals

• Added the atomic reference energies for wB97XD and M062X functionals for the 6-31G, 6-311G, and CBSB7 basis set families.

• Fixed a bug with the description of the model if the DFT functional was given as a variable.

• Improved error handling when Gaussian fails.

2025.8.12: Added Wavefunction Stability step

• Added the Wavefunction Stability step, which analyzes the stability of HF and DFT wavefunctions and optimially lowers the symmetry to find the stable wavefunction. This step can also explore other spin states to find the lowest energy state.

• Added M062X and wB97XD hybrid functionals to the available list.

2025.1.31: Bugfix: shape of gradients array

• Fixed a bug in the shape of the gradients array when reading from the punch file. They are now correctly output as a [natoms][3] array.

2025.1.29: Added an option to save the structure in the standard orientation.

• Added an option to save the structure in the standard orientation, which is the orientation used by Gaussian for the calculations. This is useful for visualizing the structure, and is also the orientation for e.g. the calculated orbitals.

• Fixed a bug in displaying the structure with the orbitals or density. The sructure is now always in the standard orientation, regardless of the option above, as it must be since the orbitals and density are calculated in the standard orientation.

2025.1.5: Added PBErev, HSE06, PBE0 and PBE0rev energies for thermochemistry

• Added the atom energies for the 6-31G and 6-311G basis set families for the PBErev (PBEhPBE), HSE06 (HSEH1PBE), PBE0 (PBE1PBE), and PBE0rev (PBEH1PBE) density functionals for the thermodynamics step.

• Added the electronic state to the output and properties

• Added refereneces for all the methods and DFT functional in the GUI.

• Improved the output to include the full name of the method and DFT functional

2024.12.24: Minor enhancement to Thermochemistry.txt report

• Added the method and basis set to the header of the Thermochemistry.txt report.

• Caught some cases where the system name was reported as None.

2024.12.23: Bugfix: thermochistry for DFT methods failed

• The code was not properly handling the DFT methods in the thermodynamics step. This has been fixed.

• Improved the reporting on why the atomization energy or enthalpy of formation could not be calculated.

2024.12.22: Added more methods and basis sets for thermo

• Added the atom energies for the 6-31G and 6-311G basis sets to the thermodynamics step for HF, B3LYP, all MP, CC, CBS-xx, and Gn composite methods

• Added detailed output of the thermochemistry calculations in Thermochemistry.txt

• Allowed composite methods to be access through the Thermodynamics Step.

• Added ability to print and/or save the basis set to a file.

• Added options for the numerical grids used in DFT calculations.

2024.12.11: Loosened tolerance on gradients from punch file

• The code checks that the gradients in the punch file are similar to those in the output. Apparently the check was too tight, so it has been loosened. A warning is printed if the gradients differ too much so that we can understand if it is an issue.

2024.12.9: Improvements to thermodynamics step

• Added B3LYP atom energies to the thermodynamics step.

• Expanded the properties to include essentially everything calculated by Gaussian.

• Added the ability to get the initial Hessian for an optimization from propertiesm or from the checkpoint file.

• Improved the handling of the checkpoint file so by default it is saved and the next substep uses it for initial guesses, etc.

• Added options to remove the checkpoint files at the end of the calculation.

2024.12.1.1: Bugfix: Typographical problems with the output.

2024.12.1: Added thermodynamics step

• Access the FREQ command in Gaussian to calculation the thermodynamic functions.

• Added code and data for calculating enthalpy of formation for many computational models including the 6-31G and 6-311G basis set families.

• Added output of timing data to the standard directory ~/.seamm.d/timing

• Cleaned up and standardized the names of result data

2024.11.18: Enhancement: added properties

• Added properties for the energy and enthalpy, etc. from composite models.

• Protected the code from crashing is Gaussian failed.

2024.10.15: Bugfix: errors using short names of methods and functionals

• There were bugs in the code that caused errors when using short names, e.g. “HF” or “CCD” for the method, or e.g. “B3LYP” for the density functional.

2024.10.10: Enhancement: added sempiempirical methods

• Added the various semiempirical methods supported by Gaussian so they can be used from SEAMM.

2024.8.23: Enhancements: added bond orders and improved transition state optimization

• Added ability to calculate the Wiberg bond order matrix and optionally use to set the bond orders of the structure.

• Enhanced the optimization for transitions states to capture the vibrational analysis if the second derivatives are calculated, and report on the stability of the structure.

2024.7.27: Bugfix: issues when used in loop

• Fixed a bug that caused the plug-in to fail when used in a loop.

• Improved the creation of the gaussian.ini file to both work better and keep the comments in the file.

2024.6.5: Cleaned up logging from code.

• The logging was a bit aggressive, so moved most logging down to the debug level. This will make the output of e.g. geomeTRIC more readable.

2024.5.31: Added optimization of transition states, and…

• Corrected implementation of composite methods (Gn, CBS-x) to handle optimization.

• Added target of the optimization to allow transition states and saddle points.

• Corrected a bug in handling the maximum number of optimization steps.

• Corrected bug determining if optimization completed properly.

• Corrected bug handling the composite method results.

2024.5.27: Added number of optimization steps to results

• Added the number of steps for the optimizations to the results that can be output to tables, variables, etc.

2024.5.8 General enhancements

• Updated to new calculation handling, with ~/SEAMM/gaussian.ini controlling access to the installed version of Gaussian on the machine.

• Added energy and gradients to results to support general use in e.g. energy scans.

2024.1.19: Switched to new way to run Gaussian, added option to just write input file

• Switched to using the new way to run executables, which supports containers.

• Added an option to just write the input file, without running Gaussian. This is useful for debugging, and for running Gaussian on a remote server.

2023.10.25 Bug fixes: variable for functional, and parsing FChk file

• Fixed a problem with handling the functional if it was a variable rather than a specific functional.

• Fixed a problem parsing the FChk file. For exponents > 99 the FORTRAN format used in Gaussian grops the “E”, resulting in numbers like 0.947-104 that caused a problem when trying to read them.

2023.10.22 Bug fixes: orbital plots and output

• The plots of the HOMO and LUMO were shifted by one orbital due to some code counting from 1 and other, from 0. Sigh.

• The output to Job.out was inadvertently truncated.

2023.10.7 Added structure file for plots of density and orbitals.

• Always write the current structure as ‘structure.sdf’ in the directory where the cube files for orbitals and densities are written. The Dashboard picks up this file to render the structure along with the surfaces.

2023.9.27 Added composite and other methods, DFT functionals

• Now support HF, DFT, MP4, CCD & CCSD, CBS-x, and Gn methods

• Added PBE, PBE-98, PBE0, and HSE06 functionals

• Added analysis of HOMO/LUMO gap energy

• Added plotting of orbitals and densities

• Added otuput of atomic charges and spins, and placing them on the configuration.

• Added ability to control the system/configuration update

2023.2.26.1 Moved Gaussian output to output.txt

• Capturing stdout prevent users from seeing the output during a calculation. This fixes that.

2023.2.26 Initial version with energy and optimization

• Support running the energy or optimization with HF, DFT, MP2 and MP3 though testing has not yet been thorough.

• The DFT functional supported are at the moment limited.

2023.2.24

• Plug-in created using the SEAMM plug-in cookiecutter.