Detecting phase separation of freeze-dried binary amorphous systems using pair-wise distribution function and multivariate data analysis

Norman Chieng, Hjalte Trnka, Johan Boetker, Michael Pikal, Jukka Rantanen, Holger Grohganz

    8 Citations (Scopus)

    Abstract

    The purpose of this study is to investigate the use of multivariate data analysis for powder X-ray diffraction-pair-wise distribution function (PXRD-PDF) data to detect phase separation in freeze-dried binary amorphous systems. Polymer-polymer and polymer-sugar binary systems at various ratios were freeze-dried. All samples were analyzed by PXRD, transformed to PDF and analyzed by principal component analysis (PCA). These results were validated by differential scanning calorimetry (DSC) through characterization of glass transition of the maximally freeze-concentrate solute (Tg'). Analysis of PXRD-PDF data using PCA provides a more clear 'miscible' or 'phase separated' interpretation through the distribution pattern of samples on a score plot presentation compared to residual plot method. In a phase separated system, samples were found to be evenly distributed around the theoretical PDF profile. For systems that were miscible, a clear deviation of samples away from the theoretical PDF profile was observed. Moreover, PCA analysis allows simultaneous analysis of replicate samples. Comparatively, the phase behavior analysis from PXRD-PDF-PCA method was in agreement with the DSC results. Overall, the combined PXRD-PDF-PCA approach improves the clarity of the PXRD-PDF results and can be used as an alternative explorative data analytical tool in detecting phase separation in freeze-dried binary amorphous systems.
    Original languageEnglish
    JournalInternational Journal of Pharmaceutics
    Volume454
    Issue number1
    Pages (from-to)167-173
    Number of pages7
    ISSN0378-5173
    DOIs
    Publication statusPublished - 18 Jul 2013

    Fingerprint

    Dive into the research topics of 'Detecting phase separation of freeze-dried binary amorphous systems using pair-wise distribution function and multivariate data analysis'. Together they form a unique fingerprint.

    Cite this