Charge Density and Optical Properties of Multicomponent Crystals Containing APIs

Technological progress forces improved material performance therefore controlling synthesis of new crystal phases requires moving from the trial-and-error method to comprehensive solutions. Active Pharmaceutical Ingredients (API) are of particular interest in crystal engineering [1]. Molecular flexibility reflected in many polymorphic forms and appropriate spatial distribution of hydrogen bond donors and acceptors make many known drugs useful for designing optically active materials [2]. The effective correlation between properties and structural features of a given material is possible through quantitative crystal engineering combined with in silico crystal engineering. Quantitative crystal engineering utilizes modern charge density analysis and properties calculations, whereas in silico crystal engineering assesses synthon formation capability probing weak interactions existing within the crystal phases. Optical properties of a crystal strongly depend upon spatial distribution of molecules in the crystal structure, as well as on the electronic properties of molecular building blocks (dipole moments, polarizabilities, hyperpolarizabilities)[3]. Recently we have investigated materials based on pharmaceutically active ingredients: barbiturates, antiarrhythmic drugs, alkaloids, combined with organic molecules and/or transition metal salts. Partial results of our research have already been published [2]. Factors that contribute to molecular recognition in the selected polar/chiral crystal phases (derived through charge densityand Hirshfeld Surfaces Analysis) have been determined. The predicted values of refractive indices were confirmed experimentally using the immersion oil method. Second Harmonic Generation efficiency was assessed using a modified Kurtz-Perry technique [2].

QM/electrostatics scheme for predicting optical properties of molecular crystals

In this contribution it is shown that modest calculations combining first principles evaluations of the molecular properties with electrostatic interaction schemes to account for crystal environment are reliable for predicting and interpreting the experimentally-measured electric linear and second-order nonlinear optical susceptibilities within the experimental error bars. This is illustrated by considering two molecular crystals, namely: 2-methyl-4-nitroaniline (MNA) and 4-(N,N-dimethylamino)-3-acatamidonitrobenzene (DAN) [1]. A good agreement between theory and experiment (see figure below for DAN) is achieved providing the electric field effects originating from the electric dipoles of the surrounding molecules are accounted for. The presentation will also i) highlight the key role of the geometry on the χ(1) and χ(2) responses, ii) demonstrate the impact of electron correlation on the molecular and crystal properties, iii) assess the performance of exchange-correlation functionals, and iv) address the amplitude of the zero-point vibrational energy contributions [2]. A second illustration will deal with the χ(1) and χ(2) responses of two anil crystals, [N-(4-hydroxy)-salicylidene-amino-4-(methylbenzoate) and N-(3,5-di-tert-butylsalicy-lidene)-4-aminopyridine, which can switch between a enol (E) and a keto (K) form [3].

Modeling of the optical properties of molecular crystals

In this contribution we present our current findings in the calculations of the linear and second-order nonlinear electric susceptibility tensor components of organic crystals. The methodology used for this purpose is based on a combination of the electrostatic interaction scheme developed by Hurst and Munn (Hurst & Munn, 1986) with electronic structure calculations for the isolated molecules. Our modification of the method consists in i) running periodic boundary condition (PBC) calculations for an adequate chromophore geometry (either experimental or optimized) to obtain atomic charges and in ii) performing the calculations of the molecular properties within a non-uniform embedding field generated by point charges located spherically around the reference molecule. Using this approach good accuracy is achieved on the electric susceptibility tensor components in comparison with the uniform dipole electric field (Seidler et al., 2013). We extend here the application of this method to other molecular crystals as well as we present the first attempt to predict the chi(1) and chi(2) components of two-component organic crystals (Gryl et al., 2014).