tako.reactionPath Settings

This page is generated from code-owned Doxygen-style metadata in src/script/takoScriptDocs.ts. Update that file first, then regenerate this page so runtime capability descriptions, docs, and agent guidance stay aligned.

tako.reactionPath

Runs NEB/reaction-path calculations between an initial and final structure.

Signature: await tako.reactionPath(initialStructure, { calculator, reactionFinalStructure, reactionImageCount, reactionInterpolation, reactionNebMethod, reactionOptimizer, ... })

Theory

• Nudged elastic band represents a reaction path as images connected by springs. True forces are projected perpendicular to the path, while spring forces act along the path to maintain image spacing.
• The highest-energy image approximates the transition region. Climbing image NEB removes the spring component on that image and drives it uphill along the path to refine the saddle estimate.
• IDPP interpolation builds a better initial path by preserving pair-distance changes between endpoints. It often reduces atom crossing and improves convergence compared with pure linear interpolation.
• Barrier heights are model- and endpoint-dependent. Endpoints should be relaxed consistently, atom ordering must match between endpoints, and reported barriers should state whether zero-point, entropy, or solvent effects are omitted.

Usage Notes

• Always provide reactionFinalStructure with the same atoms in the same order as the initial structure.
• Use reactionInterpolateOnly: true to inspect generated images before committing to a full NEB run.
• Keep endpoint, path, and optional saddle vibration checks under the same calculator/settings for consistent barrier bookkeeping.
• Prefer the default bfgs optimizer for reaction paths. Any optimizer can reach a numerically converged unintended path, so inspect the image trajectory and mechanism.
• For periodic paths, verify matching endpoint cells and consistently wrapped coordinates yourself; the runtime endpoint validator checks atom count and element order, not cell equality.
• Do not blindly pre-optimize both endpoints before calling reactionPath, especially for weakly-bound or degenerate-exchange systems: independent endpoint relaxation can collapse reactant and product into the same minimum and destroy the path.
• Use tako.structure.fix(structure, indices) to freeze anchor atoms (for example a shared spectator atom in an exchange reaction) when the mechanism needs a constrained coordinate.

Settings

Example

const initial = await tako.input("input/reactant.extxyz");
const final = await tako.input("input/product.extxyz");
const calculator = tako.calculator.gfn2({ dispersion: "d4" });
const result = await tako.reactionPath(initial, {
  calculator,
  reactionFinalStructure: final,
  reactionImageCount: 7,
  reactionInterpolation: "idpp",
  reactionNebMethod: "improvedTangent",
  reactionNebClimb: true,
  reactionFmaxEvPerAngstrom: 0.05,
});
await tako.output("output/reaction-path.json", result);

Result

Reaction-path output contains NEB/path/barrier/saddle data; it does not return the common charges, DOS/PDOS, density, NCI, WFN, or stress channels. Validation, engine failure, timeout, and cancellation reject the promise; cancelled bands are not restartable.