A specific MNDO parameterization for water.

Area of Science:

• Computational Chemistry
• Quantum Chemistry
• Materials Science

Background:

• Predicting the properties of water clusters is crucial for understanding various chemical and physical processes.
• Existing computational methods may struggle with accuracy or efficiency for large water aggregates.

Purpose of the Study:

• To develop and parameterize a modified neglect of differential overlap (NDDO) Hamiltonian for accurate calculations of water clusters.
• To assess the performance of the new method in reproducing experimental and high-level computational data for water oligomers.

Main Methods:

• Parameterization of a modified NDDO Hamiltonian including polarization functions, Feynman dispersion, and adjusted hydrogen nucleus treatment.
• Utilizing the Benchmark Energy and Geometry Database water-cluster data for parameterization.
• Validation using MP2/aug-cc-pVDZ optimized geometries and CCSD(T)/CBS oligomerization energies.

Main Results:

• The parameterized method accurately reproduces oligomerization and rearrangement energies with a root-mean-square error (RMSE) of 0.79 kcal mol⁻¹.
• Geometries of 38 water oligomers are reproduced with an RMSE of 0.17 Å.
• The Feynman dispersion term unexpectedly influences atomic polarizability.

Conclusions:

• The modified NDDO approach provides a computationally efficient and accurate method for studying hydrogen-bonded water aggregates.
• The findings have implications for developing future NDDO Hamiltonians and force-field models for water systems.