In region 3 crystals of pure caffeine and the 1:1 cocrystal are in equilibrium with a solution of invariant composition E1, while region 4 describes the stability zone of the 1:1 cocrystal. Its skewed nature is indicative of the problems encountered in using solvent evaporation to crystallize the 1:1 cocrystal.19 The hatched area in Figure 5 shows the region where the 2:1 cocrystal appeared. The data points shown represent the compositions of dry and drop grinding experiments (one with 5 drops and one 10 drops of acetone), which gave the 2:1 cocrystal in around 50% of experiments. This cocrystal could not be accessed by solvent crystallization
in acetone. On the other hand if acetone is added to solid 2:1 cocrystal, a mixture of 1:1 cocrystal and caffeine can be found. With further addition of acetone, it is possible to put the overall mixture in domain 2 resulting in a suspension of pure caffeine and its saturated solution. These observations suggest that the stability region for the 2:1 cocrystal is metastable and submerged within region 3. These results indicate that while the 1:1 cocrystal is a stable phase, the 2:1 complex is metastable, at 25 C, explaining why solution crystallization has not been successful. Interestingly, it appears that this situation is not true at the melting points since in the binary phase diagram the 2:1 complex is a stable form. This raises the possibility that at higher temperatures in the ternary system the 2:1 cocrystal may be stable if a solvent of higher boiling point were used