(M1430-06-31) Feasibility of Synchrotron XRD for the Analysis of Mixture Sub-samples

Purpose: X-ray diffraction analysis (XRD) is the gold standard for identifying solid-state forms of pharmaceutical powders.  From form screening tests to quantifying components in mixtures, XRD has been used reliably for crystalline samples, but not without its limitations.  Generating reliable XRD data starts with preparing good samples, especially in the case of mixtures.  The quality of mixing becomes essential since it is not always possible to analyze the entire sample at once.  The purpose of this study is to demonstrate that synchrotron XRD (SXRD) can be effectively utilized to manage the risk of sub-sampling a large sample volume.  We used geometric mixing with the “coning and quartering” technique andconducted the analysis on sub-samples of the mixtures.  The power of SXRD was further highlighted by comparing the results from a conventional laboratory instrument (LXRD).

Methods: Geometric mixing of powders utilizing the “coning and quartering” technique was used to prepare three binary mixtures of Caffeine (Caff.) in Acetaminophen (APAP) at concentrations of 0.84%, 5% and 10%.  The coning and quartering technique was used at each step during mixing until the entire sample was prepared.  Pure samples of Caff and APAP were used to established specificity of Caff in the mixtures. Sub-samples were analyzed using both conventional LXRD and SXRD.  For LXRD, each subsample was analyzed using a zero-background holder in a Bragg-Brentano geometry generating a pattern.  For SXRD, sub-samples were packed into polyimide capillaries and tested by both a point and an area detector.  The beam was used to collect data from multiple spots on each capillary (sub-sample) generating a pattern per spot.  Patterns were examined, processed and the baseline corrected height of a main Caff peak (at 11.97° 2𝜃) was used when appropriate to build calibration curves.

Results: With both the point and area detectors, SXRD was able to detect Caff in the lowest-concentration mixtures at every tested spot.  By contrast, with LXRD, no Caff peaks were observed at that concentration.  Data from both the SXRD point and area detectors produced calibration curves with R² values as high as 0.99 depending on how the analysis was performed. When calibration curves of different pairing of sub-samples were created, resulting R² values ranged from 0.96 to 0.99 for the point detector data and 0.97-0.99 for the area detector data.  Given the essentially equivalent R² values obtained, the choice between using a point versus an area detector for data collection relies on the sample being analyzed and the experimental objectives.  Although both detectors generated highly reliable data, the data still varied in the quality (intensity and shape) of the produced patterns and the required pre-processing. The precision of peak intensity and shape produced by the point detector makes it a better option for quantification purposes, whereas the area detector is most appropriate for qualitative purposes and samples with grainy/broad peaked components. While SXRD data shows variations within each sub-sample (i.e., spots in each capillary), averaging the patterns of individual spots eliminated the variation between sub-samples of each preparation.  For LXRD, the variation between sub-samples of the 5 and 10% mixtures were apparent in both the intensity of peak heights and shape, while allowing the detection of only the most intense Caff peaks. 

Conclusion: SXRD offers better sensitivity and resolution of diffraction peaks than conventional LXRD because of high flux, tunable wavelength, and better alignment of the synchrotron ray, which improves the identification of minor components.  Our data show that with the proper sample preparation procedure, a single capillary (i.e., a sub-sample) is sufficient to generate high quality and reliable data from a mixture using SXRD, unlike with LXRD.  However, it is critical that multiple spots on that same capillary are scanned, and the data averaged to create a pattern that represents the entire capillary.  Lastly, different detectors for SXRD analysis can be used for different purposes depending on the sample to be tested and the experimental objective.