Interfacial Interactions of Faceted Organic Crystals - An in-silico study with Atomic Force Microscopy

As the most popular form of drug delivery, solid form formulations, such as tablets, are utilised throughout the world, and their development is crucial to delivering new medicines to market. However, the solid form drug product does not only require focus on its active pharmaceutical ingredient (API) but also its interactions with other ingredients (excipients) within the formulation. The compatibility between the API and excipients plays a critical role in the final drug product performance. Currently, drug products are being developed using a trial and error approach to determine the best formulation. The ability to pre-screen API-Excipient particle interactions would allow formulators to make a more informed decision on the experimental studies to be carried out, thus reducing the development time and resources required to get a molecule from discovery to product. This body of work presents two methodologies for assessing the inter-particulate interactions between faceted organic crystals, such as those used for drug product formulations. Previous studies have utilised single probe molecules to computationally calculate the interactions between the probe and surface, in order to relate those to the inter-particulate forces. Within this project, an in-silico methodology has been developed which calculates interactions of opposing crystal slabs, taking into account approach distance, atomic roughness induced steric repulsions, displacements and rotation. Maps of the interaction space were produced allowing identification of the mechanism of surface interaction, all a marked improvement on single molecular probe methods. This development could aid pre-screening efforts, and inform solid-state and formulation scientists, of the impact the surface chemistry can have interactions. Atomic force microscopy adhesion data has previously been shown to correlate well with formulation performance. In this study, the facet specific interactions were measured for the first time and shown to agree with the in silico method. It was found, not unexpectedly, that micro-surface roughness controlled surface adhesion, although in silico measurements showed conversely that atomic roughness increased surface adhesion. The materials analysed were Paracetamol Form I (Para), � - L-Glutamic acid (LGA) and � - D-Mannitol due to their varying crystal morphology, surface chemistry, and ease with which it can be crystallised. Multiple paracetamol facets were used for adhesive measurements, and a cleaning study was carried out determining the most effective protocol of cleaning paracetamol surfaces from a topographical and surface chemistry point of view. For the in-silico approach, a molecular mechanics (MM) workflow has been developed and an accompanying analysis tool, allowing high-throughput analysis. It was found that Para shows higher cohesive behaviour in the presence of DMAN compared to LGA due to the higher interactions between the hydrophilic surfaces of DMAN. Surface roughness was found to be inversely correlated to the interaction energy, except for a few cases where higher roughness equated to more energy due to the slabs interlocking. Facets (10¯1) and (11¯1) of Para were found to be the most interactive across Para and DMAN/LGA substrate facets respectively. Finally, a strong case has been made for using energy distributions to describe facet interactions, over conventional use of the lowest interaction energy. Experimentally, for the first time, the adhesion between faceted organic crystals was measured. Small crystals were mounted onto the end of AFM cantilevers, and with the orientation of the substrate crystals (also faceted) adjusted, the two crystals were brought into contact. Forces were measured for the fully indexed Para crystal giving a ranking of most interactive facets Para/Para : F10¯1 >> F101 > F11¯1. Similarly to the in-silico results, the (10¯1) was found to be the most adhesive for Para/Para systems followed by the (101) and (11¯1). As with the in-silico Para was found to be more cohesive in the presence of DMAN and more adhesive with LGA, showing a correlation with the model. The cohesive/adhesive balance analysis and interactive probe ranking were within agreement between the experimental and computation technique, thus validating the model and increasing confidence that it can be used to assist in pre-screening API-Excipient interactions.