Having started my career in pharmaceutical industry, my research focuses on the control and use of solid-state forms of API. The solid state of API has an important impact on properties of solid drug forms such as tablets, influencing hygroscopicity, biodisponibility. My research focuses on engineering the solid-state of API adding multiple components to the same solid phase. We showed how multi-component crystallization (co-crystallization) can be used to control the solid-state properties of drug compounds (hygroscopicity, melting point stabilization, …)1 but can be used as well as to expand the patent life-time of a given drug.2 In parallel, we also specialize in developing multi-component crystallization processes, upscaling it to the kg scale.3 Doing so, requires a careful control of underlying thermodynamics and kinetics. In a final part of my research, we go one step further, using the specific properties of multi-component systems to develop novel crystallization-based applications. In particular my group has developed a novel type of resolution based on enantiospecific co-crystallization from solution.4 We recently expanded this to resolve mandelic acid using preferential co-crystallization.5 In our latest ground-breaking work, we used the underlying thermodynamics of these systems, to pull them towards a transformation of the racemate into an enantiopure drug in a so-called deracemization process.6
actually this is to my personal opinion still one of the biggest challenges from an academic point of view.
For now, we mainly just do a 'trial-and-error' approach trying to identify a suitable co-former.
I guess this is also related to the small free energy of cocrystal formation
A+B-->AB should have a ΔG which is negative. But what you do is you break hydrogen bonds (let us start from the idea they are the main interactions) in both parent compounds, to form new type of hydrogen bonds in the cocrystal. The couple of theoretical articles out there, predict this indeed to be a couple of kcal.mol-1, so just by changing the nature of B slightly you can have either a negative or postive free energy of reaction.
This makes predicting a suitable co-former so difficult. I think nice work in crystal structure prediction is currently under way, and even in using database analysis, so I am hoping to see this change in the future.
For now, we let ourselves be guided by 'structural resssemblance'. So eg. if one dicarboxylic acid works ... well we try a whole series of dicarboxylic acids in our screen.
Tom
