Step Size and Hessian Updates¶
As Gaussian follows the Intrinsic Reaction Coordinate (IRC), it must determine how far to move between successive points and how the local curvature of the Potential Energy Surface changes along the reaction pathway.
These two tasks are controlled by
- the step size, which determines the distance between consecutive IRC points,
- the Hessian matrix, which describes the local curvature of the Potential Energy Surface.
Together, they allow Gaussian to generate a smooth and accurate reaction pathway.
Step Size¶
The reaction pathway is generated by taking a series of small displacements away from the transition state.
If the current point on the IRC is
the next point is approximately
where
is the step size.
The value of \( \Delta s \) determines the spacing between successive IRC points.
Choosing the Step Size¶
The step size must be chosen carefully.
A large step size
- reduces computational time,
- generates fewer IRC points,
- may skip important structural changes,
- may move away from the true Minimum Energy Path.
A small step size
- follows the reaction pathway more accurately,
- captures gradual structural changes,
- increases the number of IRC points,
- requires more computational time.
Gaussian selects a default step size that provides a good balance between efficiency and accuracy, although it can be modified by the user if needed.
Local Curvature Changes¶
The Potential Energy Surface is rarely uniform.
As the reaction proceeds,
its curvature changes continuously.
Because of these changes,
the reaction direction also changes.
Gaussian must therefore continually update its description of the local Potential Energy Surface.
Hessian Matrix Along the IRC¶
The Hessian matrix describes the curvature of the Potential Energy Surface.
During the IRC calculation,
the Hessian is used to
- determine the local shape of the energy surface,
- improve the prediction of the next IRC point,
- maintain the correct reaction direction.
Unlike the initial Transition State calculation,
Gaussian generally does not recompute the exact Hessian at every IRC point.
Instead,
the existing Hessian is updated as the calculation progresses.
Hessian Update Procedure¶
The reaction path can be viewed as
IRC Point
│
▼
Gradient Evaluation
│
▼
Update Hessian
│
▼
Predict Next Point
│
▼
Correct Geometry
│
▼
Next IRC Point
Updating the Hessian is much less expensive than calculating a completely new matrix at every step.
This makes long IRC calculations computationally practical.
Why are Hessian Updates Important?¶
If the Hessian were never updated,
Gaussian would assume that the curvature of the Potential Energy Surface remains unchanged.
For real chemical reactions,
this assumption is incorrect.
Updating the Hessian allows Gaussian to
- follow curved reaction pathways,
- maintain numerical stability,
- improve the accuracy of each predictor–corrector step,
- reduce accumulated numerical errors.
Balance Between Accuracy and Efficiency¶
The IRC algorithm balances computational cost with numerical accuracy.
Exact Hessian Every Step
│
High Accuracy
High Cost
──────────────
Updated Hessian
│
Good Accuracy
Lower Cost
This strategy enables Gaussian to follow long reaction pathways without the expense of repeatedly calculating exact second derivatives.
Summary¶
During an IRC calculation, Gaussian generates the reaction pathway using a sequence of small steps along the reaction coordinate. The step size determines the spacing between successive IRC points, while the Hessian matrix describes the local curvature of the Potential Energy Surface. By updating the Hessian throughout the calculation rather than recomputing it exactly at every point, Gaussian efficiently follows curved reaction pathways while maintaining both numerical stability and computational efficiency.
Next Section¶
The final chapter presents the Complete Intrinsic Reaction Coordinate Algorithm, combining all of the concepts developed throughout this section into a single workflow that illustrates how Gaussian constructs and verifies the complete reaction pathway.