Reaction Kinetics

Compute transition-state-theory rate constants from reactant conformer ensembles and transition-state conformers.

from atomiverse import ReactionKinetics
from atomiverse.levels import GFN2_XTB

job = ReactionKinetics(
    reactants=[
        [reactant_conf_1, reactant_conf_2],  # one ensemble per reactant species
    ],
    transition_states=[ts_conf_1, ts_conf_2],
    charge=0,
    multiplicity=1,
    level_of_theory=GFN2_XTB,
    temperature_start=280.0,
    temperature_stop=340.0,
    temperature_step=10.0,
)
job.submit()

result = job.require_result()
for point in result.kinetics:
    print(point.temperature, point.rate_constant, point.rate_constant_unit)

Inputs

• reactants: list of reactant ensembles. Each ensemble is a list of Atoms conformers for one reactant species. Bimolecular reactions use two ensembles.
• transition_states: list of transition-state conformer Atoms.
• charge, multiplicity: charge and spin multiplicity used for every structure.
• level_of_theory: method used for single-point energies and Hessians.
• temperature_start, temperature_stop, temperature_step: temperature grid in K.

Workflow

For each reactant and transition-state conformer, Atomiverse evaluates a single-point energy, vibrational analysis, and MSRRHO thermochemistry. The conformer free energies are combined with a Boltzmann ensemble sum, then an Eyring transition-state-theory rate constant is reported at each temperature.

Results

result.kinetics is a table with:

• temperature
• activation_gibbs_kj_per_mol
• rate_constant
• rate_constant_unit

result.steps contains per-conformer energy, frequencies, zero-point energy, and Gibbs free energies used to build the rate table.

Thresholds and checks

• Temperature grids are limited to 2000 points.
• Vibrational modes with magnitude below 50 cm⁻¹ are treated as external/soft modes by the vibrational workflow.
• Transition-state conformers report the most negative retained frequency as imaginary_frequency_cm1; inspect this value to confirm the saddle point.
• Very large ensembles increase the requested resource tier automatically.