Orientational Disorder in Crystalline Disubstituted Benzenes and Its Implications for Sublimation and Polymorphism

Abstract

Derivatives of benzene with multiple similar substituents possess a large potential to exhibit static disorder of their crystalline phases. Since multiple orientations of these molecules in such crystals can lie close in terms of the electronic energy, appreciable contributions to the thermodynamic properties of the crystals and their phase behavior, such as polymorphism or sublimation, can arise due to related configurational effects. Four target materials, always containing two ethynyl, cyano, fluoro, carbonyl, or nitro substituents attached to a benzene ring, are selected for a combined computational and experimental study. Calorimetry, vapor pressure measurements, density functional theory, and ab initio calculations are employed to derive and to interpret the residual and configurational entropy terms relevant for these materials and to trace the disorder to particular molecular interactions. Impact of the local disorder on the sublimation equilibrium and polymorphism is demonstrated computationally. Comparison of calculated and experimental sublimation data reveals that interactions of ethynyl- and cyano-derivatives of benzene can be described satisfactorily within the current PBE-D3 framework. On the other hand, such an accuracy is not reached for the nitro-derivative of benzene.

First-Principles Models of Polymorphism of Pharmaceuticals: Maximizing the Accuracy-to-Cost Ratio

“…Note that the typical interlaboratory scatter of the crystal densities is significant and using only few isolated density points recorded with different experimental setups (sample characteristics, temperature control, etc.) at different temperatures may lead to misleading thermal expansivity data. , Isothermal compressibility data could be derived from variable-pressure entries only for IBU RS1 . More information on experimental density data assessment is given in Figures S9–S12.…”

“…at different temperatures may lead to misleading thermal expansivity data. 6 , 32 Isothermal compressibility data could be derived from variable-pressure entries only for IBU RS1 . More information on experimental density data assessment is given in Figures S9–S12 .…”

“…To obtain reliable reference data, flufenamic acid and sulfathiazole samples (description provided in Table S5 ) were subject to calorimetric investigations of their phase behavior (using heat-flux differential scanning calorimetry) 32 and determinations of heat capacities (using Tian–Calvet microcalorimetry 64 and the continual heating method 65 ) of selected crystalline phases in the temperature range of 240–350 K. Identity of the treated samples was verified through all of the experimental stages using X-ray powder diffraction. More details on the experimental setup and data curation are provided in Section S2 of the SI.…”

“…Nevertheless, the main interest in the development of first-principles models for molecular crystals is now focused on active pharmaceutical ingredients (API) or organic semiconductors (OSC), not forgetting, however, also other applications in areas of food processing, dyes, explosives, fertilizers, etc. Molecules of such materials possessing a real significance are frequently significantly larger, containing typically dozens of nonhydrogen atoms, imposing significant computational requirements in cases of real API and OSC. , Modeling properties of such materials at finite temperatures from first principles remains challenging due to the computational cost, related numerous conformational degrees of freedom, or possible occurrence of local disorder . Development of novel methods, enabling high-throughput in silico ranking of candidate structures, is also highly desirable from the perspective of crystal structure prediction …”

“… 29 , 30 Modeling properties of such materials at finite temperatures from first principles remains challenging due to the computational cost, 31 related numerous conformational degrees of freedom, 27 or possible occurrence of local disorder. 32 Development of novel methods, enabling high-throughput in silico ranking of candidate structures, is also highly desirable from the perspective of crystal structure prediction. 33 …”