(462h) Statistical Approaches to Crystal Engineering
AIChE Annual Meeting
2015
2015 AIChE Annual Meeting Proceedings
Computational Molecular Science and Engineering Forum
Data Mining and Machine Learning in Molecular Sciences I
Wednesday, November 11, 2015 - 10:45am to 11:00am
Historical publications have left us with a wealth of crystallographic data which dates back for over a century. Further, the collation efforts of the CCDC have permitted the application of big data techniques to the modelling of the behaviour of molecules which may potentially form crystalline systems.
We have been working on using tools championed from applications such as Quantitative Structure Activity Relationships in the sphere of crystalline materials to aid understanding and also to attempt to form predictive models.
Co-crystalline systems are extremely useful in this regard, since the response is easy to test for, and is readily amenable to partition modelling. This will hopefully permit us to tease apart the avalanche of data which has been produced and can be calculated from the experimental data available, and begin to identify which factors are truly important in terms of crystalline material formation.
Such methods are not without their drawbacks, and details will be given on the consequences of entire regions of crystallographic space that have gone unexplored or otherwise unreported- in particular experiments which to most chemists would have yieled unexpected or apparently uninteresting results.