Study of the Crystallization Behavior of Organic Molecules: A Combined Computational and Experimental Approach

The present work investigates the morphological behavior of Xanthone, an aromatic heterocyclic compound from the polyphenolic xanthone family. In particular, the study examines how crystal morphology and growth kinetics depend on three different solvents (acetone, acetonitrile, and toluene). This aspect is of great relevance in the pharmaceutical, chemical, and engineering fields, as different crystal morphologies of the same compound can exhibit distinct properties and suitability for specific applications. The analysis was conducted through a dual approach: experimental crystallization and molecular dynamics (MD) simulations. The aim is to assess whether the internal energy at the crystal–solvent interface alone, obtained computationally, could serve as a reliable predictor of experimental crystal growth trends and interfacial properties. Experimentally, Xanthone was crystallized in the three solvents using identical supersaturation and seeding conditions. Crystals were sampled at regular time intervals over a 24-hour period and analyzed via image processing to determine characteristic size distributions. To evaluate the statistical representativeness of each sample, the number of crystals analyzed was examined through methods such as the cumulative running average and moment analysis of the size distributions: by comparing the calculated confidence intervals with the resolution of the optical microscope, it emerges that the discrepancy is minor and thus the results remain adequate to support the qualitative assessments performed in this work. Experimental results showed a clear solvent-dependent trend: growth was fastest and most anisotropicin Toluene, followed by Acetone, with Acetonitrile yielding the slowest and least anisotropic growth. From a computational perspective, MD simulations were conducted using the LAMMPS package. The simulations focused on evaluating the interfacial internal energy, the first term of the Helmholtz free energy equation, for the three orthogonal crystal faces (100), (010), and (001). After validating the simulation setup via density, Radial Distribution Functions, and diffusivity benchmarks obtained from literature, energy calculations were performed on both mono-component and crystal–solvent combined systems. The simulation results echoed the experimental hierarchy: Toluene yielded the lowest internal energy values, followed by Acetone and then Acetonitrile. However, discrepancies arose when comparing face-specific growth: in some cases, the faces that experimentally grew the least were those with the lowest internal energy, contrary to thermodynamic expectations. This mismatch indicates that internal energy alone is insufficient to predict directional growth: although the computed energies can reliably forecast overall solvent effects on crystal growth kinetics, they do not capture anisotropic development. In conclusion, the study demonstrates the viability of linking computational and experimental approaches to evaluate crystal morphology and grow kinetics, confirming the potential of molecular dynamics as a qualitative screening tool for the design of crystallization processes. Nonetheless, a more comprehensive energetic model including the entropic contribution is necessary for accurate morphological predictions. Future work should therefore focus on thermodynamic integration to quantify the entropic term and on crystal growth kinetics modeling to deepen the understanding of solvent-specific behavior and dependence parameters.