Taming Tautomerism in Organic Crystal Structure Prediction

Area of Science:

• Computational chemistry
• Solid-state chemistry
• Pharmaceutical sciences

Background:

• Tautomerism significantly affects drug properties like solubility, stability, efficacy, and toxicity.
• Computational solid-form screening aids in identifying optimal drug formulations and mitigating risks from undesirable crystal forms.
• Accurate prediction of tautomeric crystal polymorph energetics is crucial for reliable drug development.

Purpose of the Study:

• To evaluate the accuracy of widely used density functionals in predicting tautomeric crystal polymorph energetics.
• To investigate the impact of density-driven delocalization error on DFT predictions.
• To develop an improved computational method for accurate prediction of tautomeric polymorphs.

Main Methods:

• Application of various density functionals (DFT) to model tautomeric crystal polymorphs.
• Analysis of density-driven delocalization error in DFT calculations.
• Development and application of a hybrid DFT approach combined with intramolecular coupled-cluster (CC) theory corrections.
• Solid-state Nuclear Magnetic Resonance (NMR) spectroscopy for experimental validation.

Main Results:

• Standard DFT models exhibit significant errors in predicting tautomeric polymorph energetics, underestimating solid-state tautomerization barriers.
• Tautomeric polymorph energy differences can exceed 10 kJ/mol in ~40% of cases, with DFT errors reaching tens of kJ/mol.
• The refined hybrid DFT-CC method shows substantially improved agreement with experimental data.
• The study identified an incorrect tautomer in the experimentally reported crystal structure of mebendazole form B, confirmed by solid-state NMR.

Conclusions:

• Widely used DFT functionals are unreliable for predicting tautomeric crystal polymorph energetics due to delocalization errors.
• Accurate prediction of tautomeric polymorphs requires advanced computational methods, such as hybrid DFT with CC corrections.
• The findings necessitate re-evaluation of computational predictions for drug formulations involving tautomerism and highlight potential errors in existing crystal structure assignments.