The seventh blind test of crystal structure prediction: structure generation methods.

Area of Science:

• Crystallography and Materials Science
• Computational Chemistry and Molecular Modeling

Background:

• The Cambridge Crystallographic Data Centre organized the seventh blind test of crystal structure prediction (CSP) to evaluate current computational methods.
• The test included seven diverse systems, ranging from simple molecules to complex coordination compounds, cocrystals, and polymorphic materials.

Purpose of the Study:

• To assess the performance and limitations of various crystal structure prediction (CSP) methods across different molecular complexities.
• To explore the utility of CSP in conjunction with experimental data, such as powder X-ray diffraction (PXRD).
• To investigate the prediction of cocrystal stoichiometry and the handling of crystallographic disorder.

Main Methods:

• Participants employed a variety of crystal structure generation and refinement methods.
• A specific exercise involved using CSP to determine a crystal structure from a low-quality PXRD pattern.
• Methods were compared based on their ability to reproduce experimentally observed structures and predict properties like stoichiometry and disorder.

Main Results:

• Many CSP methods successfully predicted crystal structures for a small, flexible agrochemical compound.
• Fewer methods succeeded with systems of higher complexity, highlighting current challenges in the field.
• Two groups successfully and blindly predicted the existence of crystallographic disorder, a significant achievement.

Conclusions:

• Crystal structure prediction methods are effective for less complex systems but require further development for challenging targets.
• CSP can be a valuable tool for interpreting experimental data, including low-quality PXRD patterns.
• The blind test underscored the importance of addressing crystallographic disorder and highlighted the similarity in outputs from some prediction methods.