Understanding matrix-assisted continuous co-crystallization using a data mining approach in Quality by Design (QbD)

Thumbnail
Publication date
2020-07
End of Embargo
2021-06-09
Author
Chabalenge, Billy
Korde, Sachin A.
Kelly, Adrian L.
Neagu, Daniel
Paradkar, Anant R.
Keyword
Co-crystallization process
Decision tree algorithms
Hot melt extrusion
Quality by Design (QbD)
Rights
©2020 ACS. This document is the Accepted Manuscript version of a Published Work that appeared in final form in Crystal Growth & Design, copyright © American Chemical Society after peer review and technical editing by the publisher. To access the final edited and published work see https://doi.org/10.1021/acs.cgd.0c00338.
Peer-Reviewed
Yes

Metadata
Show full item record
Abstract
The present study demonstrates the application of decision tree algorithms to the co-crystallization process. Fifty four (54) batches of carbamazepine-salicylic acid co-crystals embedded in poly(ethylene oxide) were manufactured via hot melt extrusion and characterized by powder X-ray diffraction, differnetial scanning calorimetry, and near-infrared spectroscopy. This dataset was then applied in WEKA, which is an open-sourced machine learning software to study the effect of processing temperature, screw speed, screw configuration, and poly(ethylene oxide) concentration on the percentage of co-crystal conversion. The decision trees obtained provided statistically meaningful and easy-to-interpret rules, demonstrating the potential to use the method to make rational decisions during the development of co-crystallization processes.
URI
http://hdl.handle.net/10454/17941
Version
Accepted manuscript
Citation
Chabalenge B, Korde S, Kelly AL et al (2020) Understanding matrix-assisted continuous co-crystallization using a data mining approach in Quality by Design (QbD). Crystal Growth & Design. 20(7): 4540-4549.
Link to publisher’s version
https://doi.org/10.1021/acs.cgd.0c00338
Type
Article
Notes
The full-text of this article will be released for public view at the end of the publisher embargo on 09 June 2021.
Collections