Skip to content

How Transition State Search Works

A Transition State (TS) Search is fundamentally different from a standard Geometry Optimization. While a Geometry Optimization searches for the lowest-energy structure on the Potential Energy Surface (PES), a Transition State Search aims to locate the highest-energy structure along the reaction pathway, known as the transition state.

This structure represents the point at which reactants are transformed into products and corresponds to a first-order saddle point on the Potential Energy Surface.

Gaussian accomplishes this by following a specialized optimization procedure that moves the molecular geometry toward the saddle point instead of the nearest energy minimum.


Step 1 — Reading the Initial Geometry

The calculation begins by reading the molecular geometry supplied in the input file.

Gaussian also reads

  • molecular connectivity,
  • charge,
  • multiplicity,
  • electronic structure method,
  • basis set.

A good initial guess is essential because transition states occupy only a very small region of the Potential Energy Surface.

Unlike minimum-energy optimizations, a poor starting geometry often causes the calculation to converge to an undesired minimum instead of the transition state.


Step 2 — Initial Electronic Structure Calculation

Using the supplied geometry,

Gaussian performs a Self-Consistent Field (SCF) calculation to determine

  • the electronic wavefunction,
  • molecular orbitals,
  • electron density,
  • total electronic energy.

These quantities are required before the optimization can begin.


Step 3 — Calculate the Initial Hessian

Because transition-state optimization requires information about the curvature of the Potential Energy Surface,

Gaussian first evaluates the Hessian matrix when the calcfc option is used.

The Hessian contains the second derivatives of the energy with respect to the nuclear coordinates.

This information allows Gaussian to distinguish between

  • directions leading toward an energy minimum,
  • and the direction corresponding to the reaction pathway.

The initial Hessian greatly improves the stability and convergence of Transition State optimizations.


Step 4 — Transition State Optimization

Gaussian now begins the optimization.

Unlike a normal optimization,

the program does not attempt to lower the energy in every direction.

Instead,

it searches for a geometry where

  • the energy is minimized in most directions,
  • but maximized along one specific direction corresponding to the reaction coordinate.

This process is repeated iteratively.

Current Geometry
Calculate Energy
Evaluate Gradient
Update Geometry
Closer to Transition State?
      │            │
     No           Yes
      │            │
      └────────────►

Each optimization step moves the molecule closer to the saddle point.


Step 5 — Locate the Stationary Point

The optimization continues until

  • the nuclear forces approach zero,
  • the geometry satisfies all convergence criteria.

At this stage,

Gaussian has located a stationary point on the Potential Energy Surface.

However,

the calculation is not yet complete.

A stationary point may correspond to

  • a minimum,
  • a transition state,
  • or a higher-order saddle point.

An additional verification step is therefore required.


Step 6 — Frequency Calculation

Immediately after the optimization,

Gaussian automatically performs a Frequency Calculation.

Using the optimized geometry,

the program

  • constructs the Hessian matrix,
  • calculates the normal modes,
  • determines the vibrational frequencies.

This calculation confirms the nature of the stationary point.


Step 7 — Verify the Transition State

The frequency calculation provides the definitive test for a successful transition-state search.

Three outcomes are possible.

Case 1 — No Imaginary Frequencies

All frequencies positive

The optimized structure is a local minimum, not a transition state.


Case 2 — One Imaginary Frequency

One negative frequency

This is the desired result.

The optimized structure is a first-order saddle point, corresponding to the transition state.


Case 3 — Multiple Imaginary Frequencies

Two or more negative frequencies

The optimization has converged to a higher-order saddle point.

In this case,

the transition-state search must usually be repeated using a better initial guess.


Step 8 — Examine the Reaction Coordinate

Finding one imaginary frequency alone is not sufficient.

The corresponding normal mode must also be inspected.

In GaussView,

the imaginary vibrational mode can be animated.

The animation should clearly show the expected chemical transformation,

such as

  • bond breaking,
  • bond formation,
  • proton transfer,
  • hydrogen migration,
  • molecular rearrangement.

If the animation does not correspond to the intended reaction,

the optimized structure is not the desired transition state.


Complete Workflow

The overall workflow performed by Gaussian is

Initial Molecular Geometry
SCF Calculation
Calculate Initial Hessian
Transition State Optimization
Stationary Point
Frequency Calculation
One Imaginary Frequency?
      │              │
     No             Yes
      │              │
      ▼              ▼
 Improve Guess   Transition State Confirmed

Why is Frequency Analysis Necessary?

A Geometry Optimization alone cannot distinguish between different types of stationary points.

Only the Frequency Calculation reveals the curvature of the Potential Energy Surface.

The number of imaginary frequencies determines the nature of the optimized structure.

Imaginary Frequencies Structure
0 Local minimum
1 Transition state
2 or more Higher-order saddle point

For this reason,

a Transition State Search should always be followed by a Frequency Calculation.


Summary

A Transition State Search combines electronic structure calculations with a specialized optimization algorithm to locate a first-order saddle point on the Potential Energy Surface. Starting from an initial guess, Gaussian iteratively updates the molecular geometry using both energy gradients and curvature information until a stationary point is reached. The optimized structure is then verified through a Frequency Calculation. A successful transition state is identified by the presence of exactly one imaginary vibrational frequency, whose associated normal mode corresponds to the expected reaction coordinate connecting the reactants and products.