Gaussian learning pathways
MODEL · CONVERGE · VALIDATE · INTERPRET
Build a molecular calculation around a chemical question: define charge and spin, choose a method and basis, validate electronic and structural evidence, and interpret energies or spectra with their limitations.
Inputs are unexecuted teaching examples. Record executable versions, data provenance and units; establish convergence and scientific validity for your own system. External tools are identified separately. Use a licensed Gaussian installation.
Choose a calculation question
1 Getting a calculation right
Define charge, spin, geometry, resources and the chemical question.
1.1 A licensed executable and a readable input
A Gaussian job is a scientific specification written as a plain-text document.
Study the workflow →1.2 Charge multiplicity and the oxygen atom
Counting nuclei is not the same as specifying an electronic state.
Study the workflow →1.3 Water single point and the meaning of energy
A single-point calculation asks for electrons at fixed nuclear positions.
Study the workflow →1.4 File provenance and checkpoint branching
A checkpoint is a state container, not a self-explanatory result.
Study the workflow →1.5 Access, environment and input anatomy
Explain every input block and reconcile program and scheduler resources.
Study the workflow →1.6 First water single-point calculation
Prepare a complete fixed-geometry input and validate its electronic result.
Study the workflow →1.7 Shared inputs, conventions and evidence
Complete parent inputs, unit conventions and reproducibility evidence for this course.
Study the workflow →2 Numerical control and molecular structure
Inspect electronic states, SCF behavior and wavefunction evidence.
2.1 Basis sets as a controlled experiment
A basis set is the vocabulary available to the electronic wavefunction.
Study the workflow →2.2 General basis input and ECP bookkeeping
Built-in basis names are convenient, but a real project may mix basis quality across elements or use an effective core potential.
Study the workflow →2.3 SCF convergence is an inner loop
In a self-consistent-field procedure, the density creates an effective field and the orbitals obtained in that field create a new density.
Study the workflow →2.4 Optimize a water molecule and read all four criteria
Geometry optimization uses energies and gradients to move nuclei toward a stationary point.
Study the workflow →2.5 Method, basis and numerical convergence
Separate chemical-model accuracy from numerical convergence.
Study the workflow →3 Conformers, thermochemistry and solvent
Verify stationary points, conformers and thermodynamic conventions.
3.1 Mapping peroxide torsion with a relaxed scan
How does a molecule's energy change when one bond rotates while the rest of its geometry is allowed to adapt?
Study the workflow →3.2 Proving that an optimized water structure is a minimum
Does “optimization completed” prove that a structure is stable?
Study the workflow →3.3 Keeping thermochemical corrections and standard states consistent
Why do electronic energy, zero-point-corrected energy, enthalpy, and Gibbs energy differ for the same optimized molecule?
Study the workflow →3.4 Comparing gas phase and SMD solution consistently
What changes when a polar molecule is placed in a polarizable environment?
Study the workflow →3.5 Optimization, frequencies and thermochemistry
Verify a stationary point and label thermal corrections without double counting.
Study the workflow →4 Reaction energies, transition states and pathways
Connect minima and transition states before interpreting barriers.
4.1 Computing a balanced reaction energy at a consistent level
Is ethane more stable than ethene?
Study the workflow →4.2 Finding the ammonia inversion transition state
How can an optimizer locate a mountain pass instead of falling into a valley?
Study the workflow →4.3 Following both IRC directions and validating endpoints
A correct-looking transition-state vibration suggests a reaction direction, but where does the path actually go?
Study the workflow →4.4 Reporting an activation free energy for ammonia inversion
What separates a convincing reaction-path report from a attractive curve?
Study the workflow →5 Spectroscopy
Relate vibrational, magnetic and electronic observables to a defined model.
5.1 From a Hessian to IR and Raman spectra
How can the same vibration produce different IR and Raman signals?
Study the workflow →5.2 Isotopes change nuclear motion and thermochemistry
What changes when H₂O becomes D₂O if the electrons see the same nuclear charges?
Study the workflow →5.3 GIAO NMR: shielding is not chemical shift
Gaussian computes a nuclear magnetic shielding tensor: the electronic response that changes the field felt by a nucleus.
Study the workflow →5.4 Vertical TDDFT excitations and a UV–visible stick spectrum
A photon can rearrange electrons faster than nuclei appreciably move.
Study the workflow →5.5 Orbitals, populations and excited states
Interpret model-dependent analyses and a labeled vertical excitation.
Study the workflow →6 Excited structures, electronic interpretation and interactions
Analyze density, populations and interaction energies with explicit references.
6.1 Excited-state optimization and state identity
After vertical absorption, nuclei can relax on an excited-state surface.
Study the workflow →6.2 Populations, orbitals, and density cubes
What does a molecular “charge picture” actually show?
Study the workflow →6.3 Water-dimer interaction energy and BSSE
Why can a small basis make two molecules look too strongly bound?
Study the workflow →6.4 Open-shell O₂: SCF convergence, stability, and spin
Can an SCF calculation converge to the wrong electronic solution?
Study the workflow →7 Electronic-model reliability
Separate numerical error, basis limitations and chemical-model accuracy.
7.1 HF and DFT compare questions before numbers
Comparing two methods is an experiment about an approximation, not a contest to produce the most negative total energy.
Study the workflow →7.2 Correlation energy with MP2 and CCSD(T)
Mean-field electrons respond to an average environment.
Study the workflow →7.3 Stretch H2 and discover a broken reference
H2 looks easier than water, yet its dissociation exposes a central failure of a restricted single determinant.
Study the workflow →7.4 ONIOM understand the subtraction before trusting the partition
Two-layer ONIOM constructs E≈Ehigh(model)+Elow(real)−Elow(model).
Study the workflow →8 Reproducibility, recovery and a complete study
Preserve checkpoints, execution settings and reproducible analysis.
8.1 Restart the calculation you actually have
“Restart” can mean three different things: continue an optimizer's saved state, start a new optimization from saved coordinates, or repeat an electronic calculation with a saved orbital
Study the workflow →8.2 One node resources scratch and a safe Slurm launch
A calculation consumes resources through two contracts: the scheduler reserves them, and Gaussian decides how to use them.
Study the workflow →8.3 Parse logs without manufacturing success
A reproducible calculation should leave enough evidence for someone else to reconstruct the model and audit each reported quantity.
Study the workflow →8.4 A formaldehyde study with an uncertainty ledger
A complete molecular study connects a question to a validated structure, an appropriate electronic model, a property calculation, and an honest limitation statement.
Study the workflow →8.5 Troubleshooting, checkpoints and Slurm
Classify failures, preserve restart data and match shared-memory resources.
Study the workflow →Reference guides
- 1.5 Access, environment and input anatomy — Explain every input block and reconcile program and scheduler resources.
- 1.6 First water single-point calculation — Prepare a complete fixed-geometry input and validate its electronic result.
- 2.5 Method, basis and numerical convergence — Separate chemical-model accuracy from numerical convergence.
- 3.5 Optimization, frequencies and thermochemistry — Verify a stationary point and label thermal corrections without double counting.
- 5.5 Orbitals, populations and excited states — Interpret model-dependent analyses and a labeled vertical excitation.
- 8.5 Troubleshooting, checkpoints and Slurm — Classify failures, preserve restart data and match shared-memory resources.
What counts as a completed calculation?
Explain the model, input choices, units, parent/restart data, convergence of the target quantity and acceptance criteria. Preserve raw inputs and outputs with version and dataset provenance. A successful process exit, numerical convergence and physical interpretation require separate evidence.