Analyzing Hydration Differences in Cocrystal Polymorphs: High-Resolution X-ray Investigation of Caffeine–Glutaric Acid Cocrystals

Abstract

Two polymorphic forms of caffeine (CAF)–glutaric acid (GLU) cocrystals have been studied via high-resolution X-ray crystallography and Bader’s quantum theory of atoms in molecules (QTAIM). For both the monoclinic, 1, and triclinic, 2, systems the experimental charge density distributions of the 1:1 ratio of CAF and GLU polymorphs have been determined and compared. Previous studies have determined that 1 is less stable than 2, in relative humidity (RH) testing. A topological analysis of the electron density distribution (EDD) revealed little difference between the two polymorph internal systems. The packing densities (0.76 vs 0.74) and lattice energies (−101.1 vs −107.1 kJ mol–1) of 1 and 2, respectively, are nearly equivalent, implying that the differences in hygroscopicity between the two polymorphs are not due to crystal lattice porosity or stability. A topological analysis of the number and strength of hydrogen bonds for 1 and 2 revealed nine hydrogen bonds in both polymorphs. “Classical” (O–H···X) hydrogen bonds were similarly present in both polymorphs, stabilizing the cocrystals. However, the sum of the stability produced from the “nonclassical” (C–H···X) bonds is higher in 2: −27.6 vs −38.2 kJ mol–1 for 1 and 2, respectively. One of the nine hydrogen bonds in 1 and 2 varies from the others, caused by the torsional rotation of the aliphatic carbon chain in GLU. This bond is critical for packing stabilization, creating a parallelogram-like packing arrangement in 2 in comparison to ribboning in 1. A Hirshfeld surface analysis found that the percentages of O–H···X hydrogen bonds were nearly identical in 1 and 2 (23.9% vs 22.1%); however, the H···H contacts were higher in 2 (61.4% vs 65.8% for 1 and 2, respectively), suggesting that more hydrogen-based contacts require competitive displacement by water in the hydration of 2 in comparison to 1. Additionally, a stabilizing aromatic cycle stack between CAF molecules is present in 2 due to the varied parallelogram packing arrangement, which was absent in 1; this provided ∼11.3 kJ mol–1 of stability to the system of 2. The solid-state entropies and molecular dipole moments (MDMs) of 1 and 2 supported the relative stability of the individual polymorphs, with 1 having a higher entropy and dipole moment in comparison to 2 (123.2 vs 112.8 J K–1 mol–1 and 7.45 and 4.93 D for 1 and 2, respectively), implying that it has the potential to hydrate more rapidly. These findings are in good agreement with previous experimental RH stability studies, giving further insight into the information gained from thermally averaged ground-state crystal electron density data.

An experimental and theoretical charge density study of theophylline and malonic acid cocrystallization

“…A study by Hawkins et al showed an underestimation of the EXP ∇ 2 ρ bcp compared to the theoretical value but this did not affect ρ bcp of the internal covalent bond, hydrogen bond or synthon. 12 As such the derived products of hydrogen bond strength were not affected in this environment. 12 Similarly, the study of Wagner et al which was at a higher level of experimental resolution found the ∇ 2 ρ bcp around the nitrogen species had near identical ρ bcp but much lower ∇ 2 ρ bcp compared to the theoretical, which also did not affect the derived hydrogen bond strength.…”

“…For example, in CAF:GLU polymorphs there was a variation of just a single hydrogen bond and this was the cause of the slower hydration rates seen between polymorphs. 12 This suggests that a comparison of the topological features in hydrogen bonding between reactants and products may reveal why THEO:MA formation did not enhance the hygroscopic profile of THEO. This review can be done in many ways, the ability to identify hydrogen bonds and other interactions is a now standard procedure, e.g.…”

“… 7 In (1), 22% of the hydrogen bonds are ‘non-classical’, this is less than what was seen in CAF:GLU (∼70%, in both polymorphs). 12 In the case of CAF:GLU a single weak ‘non-classical’ hydrogen bond was the cause of the stability differences seen in the polymorphs noting the importance of the weak hydrogen bonds to give depth in the packing arrangement. 12 As discussed in geometry, the planar shape of MA and THEO in (1) limits the number of these staggered hydrogen bonds able to form.…”

“… 12 In the case of CAF:GLU a single weak ‘non-classical’ hydrogen bond was the cause of the stability differences seen in the polymorphs noting the importance of the weak hydrogen bonds to give depth in the packing arrangement. 12 As discussed in geometry, the planar shape of MA and THEO in (1) limits the number of these staggered hydrogen bonds able to form. Hence, it results in limited packing diversity, which is why THEO:MA did not improve the hygroscopic stability of THEO in (1).…”

Two anhydrous forms and one monohydrate of a cocrystal of axitinib and glutaric acid: characterization, property evaluation and phase transition study

“…16–19 Thus, it is necessary to carry out more research on cocrystal polymorphism including synthesis strategies, solid-state characterization, performance evaluation and phase transition processes, which is of great significance for better application of cocrystallization technology in the pharmaceutical field. 20–25…”