Transition State Search (TS) Module

The Transition State (TS) Search module is used to find first-order saddle point structures on the potential energy surface. This module can be combined with different search algorithms to handle various types of reaction path problems. Currently, MAPLE supports the following four TS methods:

• prfo (default method): RS-PRFO single-point saddle point optimization
• neb: Nudged Elastic Band path search + (optional) CI-NEB + RS-PRFO
• string: Reaction path-based GSM/String method
• dimer: Dimer TS search (requires only one initial structure)

Each method's input format is set through the #ts(...) command.

I. TS Module Unified Input Parameters

Default parameters for the TS module are as follows, and the adjustable parameters include:

method

Choose the TS search algorithm:

• = prfo (default)
• = neb
• = string
• = dimer

maxiter

Maximum number of iterations, shared across all methods.

default: maxiter=200

neb_image

Number of intermediate images ("replicas") used to discretize the reaction path for TS searches that employ multiple images along a path. This parameter has no effect for single-structure methods such as prfo or dimer.

default: neb_images=7

level

Defines the overall aggressiveness or rigor of the transition-state (TS) search strategy.

=medium (default)

refine

Specifies whether additional post-processing is applied after the main NEB/TS path search to improve the quality of the transition state.

• =None (default)
• =cineb
• =nebts

II. Input Structure Requirements

PRFO and Dimer require only one initial structure because they are local transition-state search methods. These algorithms explore the local curvature of the potential-energy surface around the starting geometry and ascend toward the nearest saddle point. As a result, a single starting geometry is sufficient (e.g., initial_structure.xyz).

NEB and STRING, in contrast, require two structures—the reactant and the product—because they are path-based methods. They construct and optimize a reaction pathway between two known minima on the potential-energy surface. Therefore, both endpoint geometries must be provided (e.g., reactant.xyz and product.xyz), and the transition state is identified as the highest-energy point along the optimized path.