Using Hessian Updating To Increase the Efficiency of a Hessian Based Predictor-Corrector Reaction Path Following...

Area of Science:

• Theoretical Chemistry
• Computational Chemistry
• Chemical Dynamics

Background:

• The intrinsic reaction coordinate (IRC) is crucial for understanding chemical reaction mechanisms.
• Existing Hessian-based IRC algorithms are accurate but computationally expensive due to repeated energy second derivative calculations.

Purpose of the Study:

• To improve the efficiency of Hessian-based predictor-corrector algorithms for reaction path following.
• To evaluate the effectiveness of different Hessian updating schemes in accelerating IRC calculations.

Main Methods:

• Implementation and testing of three Hessian updating schemes: Murtagh and Sargent, Powell-symmetric Broyden, and Bofill.
• Integration of these updating schemes into a Hessian-based predictor-corrector IRC algorithm.
• Comparison of the computational cost and accuracy of the updated algorithms.

Main Results:

• Hessian updating significantly enhances the efficiency of the predictor-corrector algorithm.
• Bofill's updating scheme demonstrated the best performance among the tested methods.
• Excellent speed-up was achieved without compromising the accuracy of the reaction path following.

Conclusions:

• Hessian updating is a viable strategy to reduce the computational cost of accurate IRC calculations.
• The Bofill update scheme offers a substantial performance improvement for theoretical reaction path following.
• This work provides a more efficient computational tool for exploring chemical reaction mechanisms.