Reproduction of experimental data for stacked caffeine dimers using various computational methods

Area of Science:

• Computational chemistry
• Molecular modeling
• Supramolecular chemistry

Background:

• Aromatic molecule stacking influences biopolymer structure and function.
• Caffeine, a hydrophobic molecule, primarily self-associates via stacking interactions.
• Analysis of anhydrous caffeine crystal structures revealed five distinct stacking dimer types.

Purpose of the Study:

• To evaluate and identify reliable computational methods for describing caffeine stacking interactions.
• To compare geometric parameters from various methods against crystal data.
• To determine methods that accurately predict dimer energy and geometry.

Main Methods:

• Geometry optimization of caffeine dimers using molecular mechanics force fields.
• Ab initio calculations using Møller-Plesset perturbation theory of the second order (MP2) with Basis Set Superposition Error correction (MP2/CP).
• Density Functional Theory (DFT) with various functionals (PBE0-DH, SCAN, PBE-D3).

Main Results:

• MP2/CP, Poltev force field, PBE0-DH, SCAN, and PBE-D3 functionals were identified as reliable for describing caffeine stacking.
• Methods yielding dimer interaction energies close to MP2/CP also produced sublimation enthalpies near experimental values.
• MP2/CP, Poltev FF, and PBE0-DH accurately described both the energy and geometry of caffeine stacking dimers.

Conclusions:

• Specific computational methods, including MP2/CP and certain DFT functionals, provide accurate descriptions of caffeine stacking.
• Reliable stacking interaction models are crucial for understanding caffeine's role in biopolymers.
• The validated methods can be used for future studies on aromatic molecule interactions.