Crystal structure prediction of small organic molecules: a second blind test

Area of Science:

• Crystallography
• Computational Chemistry
• Materials Science

Background:

• The Crystal Structure Prediction (CSP) workshops aim to assess computational methods for determining crystal structures from molecular information.
• Following CSP1999, CSP2001 was held to further evaluate and compare different crystal structure prediction programs.

Purpose of the Study:

• To test the efficacy of various computational programs in predicting the crystal structures of three organic molecules.
• To assess the reliability of current crystal structure prediction methodologies based on lattice energy minimization.

Main Methods:

• Participants used their computational programs to predict crystal structures for given organic molecules within specified space groups.
• The primary prediction strategy involved minimizing calculated lattice energy using molecular models.
• A secondary test explored the utility of non-indexed powder diffraction data.

Main Results:

• Correct crystal structure predictions were achieved for two out of three molecules.
• The molecule with torsional flexibility proved more challenging to predict accurately.
• Correct structures were often found within the top-ranked low-energy predictions, but typically beyond the allowed top three.

Conclusions:

• Current crystal structure prediction methods demonstrate potential but are not yet fully reliable, particularly for flexible molecules.
• The workshop provided an objective benchmark for the successes and limitations of existing prediction techniques.
• Further development is needed to improve the accuracy and ranking of predicted crystal structures.