Mechanochemical synthesis, i.e. chemical synthesis by milling or grinding, is a leading methodology for environmentally-friendly, ‘green’ synthesis. However, one of the most established and valuable applications of milling is screening for new solid forms of active pharmaceutical ingredients (APIs): cocrystals, polymorphs or salts [1]. Discovery of new cocrystals is one of the principal pursuits in modern drug development, as cocrystals represent an exciting opportunity to improve physicochemical properties of the API (solubility, bioavailability, tableting) and represent tremendous value in intellectual property development. The application of mechanochemistry in screening for cocrystals is based on the ability of neat and liquid-assisted grinding (LAG) to rapidly explore cocrystal formation without interference from problems of solubility, solvolysis or thermal instability, inherent to other cocrystal discovery strategies.

Despite its broad synthetic potential and the widespread presence of milling in industrial processing, mechanisms of mechanochemical reactions are almost unknown. This is most likely due to a general inability to directly observe the course of chemical reactions and phase transformations. Until recently, the only means to follow changes in composition of mechanochemical reactions has been by stepwise analysis, i.e. by periodically interrupting the milling process and analysing the reaction mixture. Such an approach is, however, very poorly informative or even misleading as the samples transform before analysis due to relaxation, recrystallisation, desolvation or simply moisture absorption.

In collaboration with ESRF scientists at beamline ID15B, now ID31, we have recently developed the first technique for direct, real-time and in situ monitoring of structural and chemical transformations during milling using X-ray powder diffraction. This technique was first introduced for monitoring the transformations of strongly diffracting inorganic oxides into metal-organic materials [2].

In continuation, we extend the experimental design to monitor mechanochemical complexation of poorly scattering organic solids, as encountered in the formation of pharmaceutical cocrystals. The required signal-to-noise ratio was achieved by performing the experiment in a milling vessel made of non-crystalline, poorly scattering Perspex (poly(methyl)metacrylate). The first application of this new technique for monitoring pharmaceutical cocrystallisation was on an archetypal pharmaceutical cocrystal consisting of carbamazepine and saccharin (Figure 133). Reaction monitoring demonstrated their amorphisation upon neat milling, while milling in the presence of a catalytic liquid additive (liquid-assisted grinding, LAG) led to rapid cocrystal formation observable by changes in intensities of characteristic reflections for each phase. This revealed a clear difference between neat and liquid-assisted approaches for solid-state molecular complexation, and demonstrated the unexpectedly high rate of mechanochemical reactions, leading to complete conversions in 2 minutes (Figure 133).

Fragment of the crystal structure of the pharmaceutical cocrystal of carbamazepine and saccharin

Fig. 133: a) Fragment of the crystal structure of the pharmaceutical cocrystal of carbamazepine and saccharin. b) Time-resolved X-ray diffractogram for the LAG cocrystallisation of saccharin and carbamazepine. c) reaction course obtained via Rietveld analysis.

We also explored the mechanochemical cocrystallisation of nicotinamide (vitamin B3) with suberic acid, a model pharmaceutical system known to encompass materials with different stoichiometric composition. In this system, in situ cocrystallisation monitoring revealed complex, multi-step mechanisms involving short-lived and metastable materials not observable by other means. Milling cocrystallisation of nicotinamide and suberic acid in a 2:1 respective stoichiometric ratio yielded the expected cocrystal over 40 min, but only through the formation of an intermediate cocrystal containing the two components in a 1:1 ratio. Replacing the slower neat milling with much faster LAG revealed the formation of the final product cocrystal within about 5 minutes. Again, the reaction involved the 1:1 cocrystal as an intermediate which, this time, lasted only for 3 minutes (Figure 134).

Time-resolved diffractogram for LAG cocrystallisation of suberic acid with nicotinamide

Fig. 134: Time-resolved diffractogram for LAG cocrystallisation of suberic acid with nicotinamide, indicating a multi-step mechanism involving the formation of different cocrystals.

Monitoring of the LAG reaction also revealed another short-lived intermediate appearing about 1 minute into milling. This novel intermediate has never previously been observed in studies of this well-known model pharmaceutical system and, although it could not be characterised structurally, in situ monitoring now allowed it to be prepared under conventional laboratory conditions.

The presented methodology provided the first in situ and real-time insight into the course of mechanochemical formation of pharmaceutical cocrystals, immediately demonstrating the unexpected speed of the process, possibility of multi-step reaction mechanisms, as well as the formation of intermediate phases that have previously not been observed. The latter is a particularly exciting development, as it demonstrates the value of in situ monitoring not only as a tool for the academic investigation of reaction mechanisms, but also for the discovery of new, potentially commercially valuable, solid forms of APIs.

 

Principal publication and authors
I. Halasz (a), A. Puškaric (a),  S.A.J. Kimber (b), P.J. Beldon (c),  A.M. Belenguer (c), F. Adams (d),  V. Honkimäki (b), R.E. Dinnebier (d), B. Patel (c), W. Jones (c), V. Štrukil (e) and T. Friščić (e), Angewandte Chemie, International Edition 52, 11538-11541 (2013).
(a) Ruđer Bošković Institute, Zagreb (Croatia)
(b) ESRF
(c) Department of Chemistry, University of Cambridge (UK)
(d) Max-Planck-Institute for Solid State Research, Stuttgart (Germany)
(e) Department of Chemistry and FRQNT Centre for Green Chemistry and Catalysis, McGill University (Canada)

References
[1] A. Delori, T. Frišcic and W. Jones, CrystEngComm 14, 2350-2362 (2012).
[2] T. Frišcic, I. Halasz, P. J. Beldon,  A. M. Belenguer, F. Adams,  S. A. J. Kimber, V. Honkimäki and  R. E. Dinnebier, Nature Chem. 5, 66-73 (2013).