Know if a reaction works - and how fast - in one step
Avoid spending days on reactions that are thermodynamically or kinetically infeasible. Provide reactant and product SMILES. Get ΔG, K_eq, and activation barriers - without two separate computational setups and two sets of outputs to reconcile.
“Is this Diels-Alder thermodynamically feasible at 298 K, and what's the activation barrier?”
How it works
Provide reactant and product SMILES
predict_reaction_thermodynamics returns ΔG, ΔH, TΔS, and K_eq in one call. Confidence tiering: high (all-organic), medium (non-standard elements), low (transition metals - flag, not block).
Add kinetics if needed
Provide pre-optimized XYZ geometries to find_transition_state. CI-NEB locates the activation barrier, minimum energy path, forward and reverse barriers, and transition state geometry.
Confirm with vibrational analysis
run_qm_hessian confirms the transition state (one imaginary frequency) or true minimum (none). Returns ZPE and Gibbs corrections. Imaginary frequencies flag structures not at true minima.
Proof
Butadiene + ethylene → cyclohexene: ΔG = -48.8 kcal/mol, K_eq = 5.7×10³⁵ (spontaneous). Methane combustion: ΔG = -289.7 kcal/mol (effectively irreversible).
Ethanol dehydration at 298 K: +19.7 kcal/mol (non-spontaneous - consistent with textbook requirement for heat).
HCN → HNC: 59.4 kcal/mol forward barrier, 39.4 reverse, TS geometry attached, 6.7 seconds.
Boundary: predict_reaction_thermodynamics answers can it happen. find_transition_state answers how fast. Neither is valid for transition-metal-catalyzed mechanisms (low confidence tier).
Use this when you need to
Validate a synthetic pathway before committing lab time
Compare competing reaction routes quantitatively
Screen for dead-end reactions early in the design cycle
Get thermodynamics and kinetics from one tool - not two
Don't waste days on infeasible reactions
ΔG + activation barrier. Confidence-tiered. One conversation.