(M1430-09-60) Assessing Stress-Induced Amorphization and Excipient Functionality in Tablet Matrices Using Correlative Microscopy Analysis

Purpose: Stress-induced amorphization is a common phenomenon among crystalline API when subjected to high shear forces during manufacturing process such as high shear granulation, roller compaction and tableting. The study aims to utilize advanced microscopic and imaging solutions to identify and assess stress-induced amorphization of crystalline API during tablet manufacturing process. The study further aims to correlate the occurrence of amorphization and pore distribution with high stress zones during tableting.

Methods: X-ray microscopy (XRM) and mosaic scanning electron microscopy (SEM) were the primary microscopy techniques utilized to visualize the internal structures of the tablet. XRM, was used to scan the entire tablet non-invasively at 3-micron resolution. Mosaic SEM was performed in a stepwise manner, beginning with a mechanical cut of the tablet after which an argon beam was then used to polish the cross-section, clearing debris from the initial cut and providing a continuous plane allowing visualization of the internal structure features. After cutting and polishing, 100s of SEM images were collected and stitched together to form a large mosaic SEM image with a resolution of 100 nm and total area of 3.9 mm by 4.8 mm. Complimenting the mosaic SEM, EDX spectroscopy (Energy Dispersive X-ray) map measurements were taken on a few select areas to correlate phase contrast with each ingredient, as well assess the potential to differentiate amorphous and crystalline domains. AI image analysis was applied on each imaging dataset respectively via a user trained machine learning session, classifying each phase based on contrast and morphologically unique features. The training, performed in an iterative process on one image in the dataset, was then applied to the other images in the dataset through an automated process. Once the entire dataset was classified, structural attributes including volume fractions within the tablet, size distribution, and spatial distribution were calculated.

Results: The X-ray microscopy scan performed across the tablet revealed two material domains with distinct grayscale contrasts, porosity and connected porosity, and a coating apparent by its bright white contrast. However, discrepancies were observed when comparing the imaging data to the actual weight load of two major ingredients. Furthermore, no correlations to the amorphous or crystalline domain could be drawn. The bright domain (~15%) was suspected to correspond to be rich in crystalline active ingredient and the darker domain (~ 84%) to a region rich with excipient and some active ingredient. Spatial distribution revealed a uniform distribution of all domains, including porosity, along the tablet height, whereas a cylindrical spatial distribution measurement suggested more active rich domains and higher porosity at the center of the tablet. Mosaic SEM imaging provided a much clearer identification of the two primary ingredients with the active domain exhibiting brighter contrast due to its elemental composition, while the excipient domain was darker and more porous. EDX mapping validated these suspect phase assignments, along with identification of additional functional excipients. EDX mapping was then applied to assess the distribution and amount of crystalline and amorphous active domains at different locations within the tablet. Mapped regions selected excluded excipient domains such that the EDX reading of active ingredient domain could be compared independently. Map percentage readings of oxygen, present in the crystalline API form, as well as additional unique elements to the active ingredient, where compared across each location to assess if crystalline vs amorphous forms could be distinguished. Beyond EDX, the mosaic SEM analysis provided valuable insight into the interactions of each ingredient, including the porosity behavior of excipients. Both Cartesian X and Cartesian Y spatial distributions found an increasing trend of porosity when going from the edge to the center of the tablet. This trend corresponded to greater porosity in the excipient domain, which is ultimately more protected from compaction at the core of the tablet compared to the edges.

Conclusion: A clear indication of amorphous vs crystalline domains was not observed in either XRM or SEM imaging experiments. Similarly, an immediate and clear distinction between amorphous and crystalline active regions was not apparent EDX mapping due to a combination of phenomena, including the scale of the interactions within the active lattice, the lack of sensitivity to pick up molecular interactions, and most importantly the lower degree of amorphization ( < 10%) observed in this case. While different forms within the active ingredient region could not be clearly identified, a variety of critical structural features such as porosity, interaction and distribution of ingredients were observed and quantified using XRM, SEM, and EDX. This can be helpful in optimizing the tablet tooling design, investigating dissolution or flowability or sticking challenges, and understanding plastic/brittle nature of materials during tableting.

EDX spectroscopy mapping to identify ingredients observed within the mosaic SEM cross-section.

Mosaic SEM analysis workflow including 1. mechanical cross-sectioning and polishing, followed by collection of 100s of images at nm resolution, 2. AI image analysis to classify material domains and porosity based on their grayscale contrast, 3. quantification of each classified domain, including their size distribution and spatial arrangement.