Recommended Settings for Jaguar

The sections below provide recommendations for computational settings when using Jaguar to calculate chemical properties. These settings come from a wide range of calculations of structures and properties.

Recommended Basis Set

See the Molecule tab in any of the Jaguar task panels.

  • Ground state geometries: use 6-31G** (the default) or LACVP** (when metals are present). Note that LACVP actually uses 6-31G for the elements H-Ar, but with 5 spherical harmonic d orbitals rather than the 6 Cartesian functions. The def2-SVP basis set is a good alternative that provides consistent coverage across a wider range of elements.

  • Accurate single point energies: use at least a triple-zeta basis set such as cc-pVTZ(-f), def2-TZVP, or 6-311G-3df-3pd. The def2 basis sets are recommended for DFT calculations because they are efficient and consistently available across the periodic table.

  • Anions, excited states, intermolecular binding energies, polarizabilities, transition state geometries, or any other situation in which the electron density is likely to extend unusually far from the nuclei: include diffuse functions in whatever basis set you choose.

  • Atomic ESP charges: 3-21G may be used to save time when modeling large systems.

  • Studying convergence with respect to basis set size: use the series cc-pVDZ, cc-pVTZ, cc-pVQZ(-g) (you may include diffuse functions with each of these basis sets if desired), or the def2 series def2-SVP, def2-TZVP, def2-QZVP.

In most situations, the small basis sets STO-3G, 3-21G, and 4-31G should be avoided.

For more information on basis set families and choosing a basis set, see Basis Set Overview.

Recommended Theoretical Method for Different Jaguar Calculations

See the Theory tab in any of the Jaguar task panels.

  • Optimization of ground state geometries, thermochemical energies, zero-point energies: use B3LYP (the default). For a more efficient alternative, especially for larger molecules, use a 3c Composite Methods for Geometry Optimization such as r2SCAN-3c, which provides accurate structures with built-in corrections for dispersion and basis set errors at a fraction of the cost of a hybrid functional calculation.

  • Transition state geometries, conformational energies, intermolecular binding energies: use M06-2X or ωB97X-D (use M06 or MN15 if transition metals are present).

  • Tautomer energies: use M06-2X.

  • Electronic excitation energies of valence or Rydberg type: use M06-2X or MN15.

  • Charge transfer excitations: use M11.

  • Atomic ESP charges: use Hartree-Fock.

In most situations, LDA functionals (such as Xalpha or SVWN) should be avoided.

DFT methods like PBE and B97-D are actually faster than Hartree-Fock for energies and gradients, so if saving time is essential, use one of these functionals. But for second derivative properties like vibrational frequencies, HF is faster than these DFT methods.

Recommendations for Jaguar Solvation Energies

See the Solvation tab in any of the Jaguar task panels.

Use PBF or SM8 for energies. You must use PBF for solution-phase geometry optimizations because analytic gradients for SM8 are not available.

Recommendations for Jaguar Intermolecular Binding Energies

The binding energy for formation of the bimolecular complex A-B from molecule A and molecule B is defined as:

Ebinding = E(A-B) - E(A) - E(B)

where E is the total electronic energy. If the binding energy is required in solution, then E is the total electronic energy in solution (solution phase energy).

To obtain a binding energy which has been corrected for basis set superposition error (BSSE), you can use the Counterpoise Panel: click the Tasks button and browse to Quantum Mechanics → Counterpoise. Equivalently, you can use the script counterpoise.py from the command line: see counterpoise.py: Counterpoise-Corrected Binding of Two Molecules for details.

Do not apply the counterpoise correction if you are using LMP2 as the theoretical method.

If the complex A-B is bound by one or more hydrogen bonds, you can compute its binding energy by using the HBond Panel: click the Tasks button and browse to Quantum Mechanics → Hydrogen Bond Strength. This task includes the counterpoise correction. Alternatively, you can use the script hydrogen_bond.py from the command line (see hydrogen_bond.py: Binding Energy of a Hydrogen-Bonded Complex for details). The recommended functionals for hydrogen bonding are ωB97X-D or the computationally more expensive ωB97M-V.

If A and B are bound by other noncovalent interactions, use M06-2X, MN15, or computationally more expensive ωB97M-V.

Recommendations for Reaction Energies and Activation Barriers

Run the geometry optimizations, transition state search, and the vibrational frequency calculations for all structures using B97-D3/6-31++G**. Greater accuracy can be obtained by using a larger basis set, like cc-pVTZ(-f)+ or 6-311-3df-3pd.

Ionization Energy and Electron Affinity Recommendations

Use B97-D3/cc-pVTZ(-f)++ for these calculations. Use a spin-unrestricted formalism (UDFT, not RODFT) when modeling the open-shell systems.