Growing-String DE-TS

A growing-string double-ended transition-state (GS-DE-TS) job locates the transition state connecting a reactant and a product geometry using the growing-string method. This method builds a chain of interpolated geometries (a "string") and grows it toward the minimum-energy path, taking the highest-energy node as the TS.

What GS-DE-TS Runs

For each submitted job, Atomiverse executes:

1. Atom mapping — reorders product atoms to match the reactant ordering.
2. Growing-string search — builds and grows the string between reactant and product using pyGSM.
3. String path is produced with energies of all nodes; the highest-energy node is taken as the TS.

Your First GS-DE-TS Job

import atomiverse
from atomiverse import Atoms, GrowingStringDETS, JobFailedError
from atomiverse.levels import GFN2_XTB

atomiverse.configure(api_key="your-api-key")

reactant = Atoms.from_smiles("CC")
product = Atoms.from_smiles("C=C")

job = GrowingStringDETS(
    reactant_atoms=reactant,
    product_atoms=product,
    level_of_theory=GFN2_XTB,
)
job.submit()

try:
    result = job.require_result()
except JobFailedError as exc:
    print(f"Job failed: {exc.failure.error}")
else:
    print(f"Converged: {result.converged}")
    print(f"TS energy: {result.ts_energy:.6f} Hartree")
    print(f"Barrier: {result.reaction_barrier:.6f} Hartree")
    print(f"String nodes: {len(result.gsm_path)}")

Input Parameters

Required:

• reactant_atoms: reactant geometry
• product_atoms: product geometry
• level_of_theory: electronic-structure method

Optional:

• charge (default 0)
• multiplicity (default 1)

The string resolution is determined adaptively — growth stops automatically once the geometric change between new nodes falls below a convergence threshold. No manual image count is needed.

The reactant and product must have the same number of atoms and the same elemental composition.

job = GrowingStringDETS(
    reactant_atoms=reactant,
    product_atoms=product,
    charge=0,
    multiplicity=1,
    level_of_theory=GFN2_XTB,
)

Result

job.require_result() returns a GrowingStringDETSResult.

Field	Type	Description
ts_atoms	Atoms	Converged TS geometry.
ts_energy	float	TS energy in Hartree.
gsm_path	list[TrajectoryStep]	Final string path with positions, energy, and forces.
ts_optimization_trajectory	list[TrajectoryStep]	TS refinement optimization trajectory.
converged	bool	Whether the growing-string search converged.
imaginary_frequency_cm1	float | None	Imaginary frequency of the converged TS.

Convenience accessors:

Property	Returns
gsm_path_atoms	list[Atoms]
ts_optimization_atoms	list[Atoms]
reaction_barrier	float

result = job.require_result()

for step in result.gsm_path:
    print(step.energy)
    print(step.forces)

print(f"TS: {result.ts_energy:.6f} Ha")
print(f"Barrier: {result.reaction_barrier:.6f} Ha")

Handling Failures

If the growing-string search crashes or cannot converge, the job transitions to FAILED with a user-facing error message. Partial GSM path frames may be available in the typed failure payload.

from atomiverse import JobFailedError

job.submit()

try:
    result = job.require_result()
except JobFailedError as exc:
    failure = exc.failure
    print(f"GS-DE-TS failed: {failure.error}")

    if failure.partial_gsm_path:
        print(f"Partial path frames: {len(failure.partial_gsm_path)}")
        last = failure.partial_gsm_path[-1]
        print(last.energy)
else:
    print(f"TS energy: {result.ts_energy:.6f} Ha")
    print(f"GSM nodes: {len(result.gsm_path)}")

Non-blocking Submission

from atomiverse import State

job.submit(wait=False)
job.refresh()

if job.state == State.DONE:
    result = job.require_result()
    print(f"TS energy: {result.ts_energy:.6f} Ha")
elif job.state == State.FAILED:
    failure = job.require_failure()
    print(failure.error)
    print(f"Partial path: {len(failure.partial_gsm_path)} frames")