Probing the Molecular and Macroscopic Structure of Solid Solutions by Dynamic Nuclear Polarization (DNP) Enhanced 13C and 15N Solid-State NMR Spectroscopy

Crystallization is a widely used purification technique in the manufacture of active pharmaceutical ingredients (APIs) and precursor molecules. However, when impurities and desired compounds have similar molecular structures, separation by crystallization may become challenging. In such cases, some impurities may form crystalline solid solutions with the desired product during recrystallization. Understanding the molecular structure of these recrystallized solid solutions is crucial to devise methods for effective purification. Unfortunately, there are limited analytical techniques that provide insights into the molecular structure or spatial distribution of impurities that are incorporated within recrystallized products. In this study, we investigated model solid solutions formed by recrystallizing salicylic acid (SA) in the presence of anthranilic acid (AA). These two molecules are known to form crystalline solid solutions due to their similar molecular structures. To overcome challenges associated with the long 1H longitudinal relaxation times (T1(1H)) of SA and AA, we employed dynamic nuclear polarization (DNP) and 15N isotope enrichment to enable solid-state NMR experiments. Results of solid-state NMR experiments and DFT calculations revealed that SA and AA are homogeneously alloyed as a solid solution. Heteronuclear correlation (HETCOR) experiments and plane-wave DFT structural models provide further evidence of the molecular-level interactions between SA and AA. This research provides valuable insights into the molecular structure of recrystallized solid solutions, contributing to the development of effective purification strategies and an understanding of the physicochemical properties of solid solutions.

Sensitizing Explosives Through Molecular Doping

Cocrystallization assembles multicomponent crystals in defined ratios that are held together by intermolecular interactions. While cocrystals have seen extensive use in the pharmaceutical industry for solving issues with stability and solubility, extension to the field of energetic materials for improved properties has proven difficult. Predicting successful coformers remains a challenge for systems lacking well-understood synthons that promote reliable intermolecular assembly. Herein, an alternative method is investigated for altering energetic properties that operates in the absence of well-defined interactions by molecular doping. An impact sensitive primary explosive, cyanuric triazide (CTA), was selected as the dopant to test if less impact sensitive secondary explosives could gain increased sensitization to impact when CTA is inserted into their crystal lattices. Molecular doping was successful in sensitizing three melt-castable energetics: 2,4,6-trinitrotoluene (TNT), 2,4-dinitroanisole (DNAN), and 1,3,3-trinitroazetidine (TNAZ). CTA could also be incorporated as a stabilized inclusion to sensitize DNAN further. These results demonstrate how the judicious choice of dopant can lead to specific property improvements, providing a method for creating energetic materials with new properties to access metal-free primary explosives and physical hot spot models for explosive ignition.

Enhancement of Hydrate Stability through Substitutional Defects

Cytosine monohydrate (CM) and anhydrate crystal forms reversibly interconvert under high temperatures or high humidity conditions. Here, we demonstrate through defect engineering the ability to expand the thermal stability range of CM through the targeted creation of quantifiable defects in low-level concentrations. Twelve different molecular dyes with a variety of core structures and charges were screened as potential dopants in CM. CM-dye phases prepared with Congo red (CR), Evans blue (EB), and Azocarmine G (AG) exhibited the highest inclusion levels (up to 1.1 wt %). In these doped isomorphous materials, each dye is presumed to substitute for 4-7 cytosine molecules within the low-rugosity (102) planes of the CM matrixes, thereby creating a quantifiable substitutional defect and an impediment to the cooperative molecular motions which enable the transformation to the anhydrate. Dehydration of materials with these engineered defects requires significantly higher temperatures and proceeds with slower kinetics compared to pure CM. The CM-dye phases also exhibit a reduction in the thermal expansion along key crystallographic axes and yield dehydration products with altered particle morphologies.

NaCl Crystals as Carriers for Micronutrient Delivery

Iron deficiency leading to anemia is one of the most severe and important nutritional deficiencies in the world today. To combat this deficiency, the fortification of food products with iron is a natural way to increase the global iron uptake. Here, we report a novel strategy for iron encapsulation in NaCl crystals via microscopic inclusions containing dissolved iron salt. The liquid inclusions embedded in the crystal insulate the reactive iron salts from their environment while assuring that iron is in a soluble and bioavailable form. While the size distribution of inclusions remains independent of the evaporation conditions, their density increases during crystallization at lower relative humidity. Using Raman confocal microspectroscopy, we have been able to analyze the morphology, length/thickness ratio, of inclusions and show that inclusions evolve toward a plate-like structure with the increase in size. By growing a pure NaCl shell around the iron-containing NaCl crystals, the stability of the composite crystals can be even further enhanced. The role of halite crystals as a carrier for iron fortification opens the way for the delivery of other types of micronutrients by including them in table salt.

Modeling of the Crystallization Conditions for Organic Synthesis Product Purification Using Deep Learning

Crystallization is an important purification technique for solid products in a chemical laboratory. However, the correct selection of a solvent is important for the success of the procedure. In order to accelerate the solvent or solvent mixture search process, we offer an in silico alternative, i.e., a never previously demonstrated approach that can model the reaction mixture crystallization conditions which are invariant to the reaction type. The offered deep learning-based method is trained to directly predict the solvent labels used in the crystallization steps of the synthetic procedure. Our solvent label prediction task is a multi-label multi-class classification task during which the method must correctly choose one or several solvents from 13 possible examples. During the experimental investigation, we tested two multi-label classifiers (i.e., Feed-Forward and Long Short-Term Memory neural networks) applied on top of vectors. For the vectorization, we used two methods (i.e., extended-connectivity fingerprints and autoencoders) with various parameters. Our optimized technique was able to reach the accuracy of 0.870 ± 0.004 (which is 0.693 above the baseline) on the testing dataset. This allows us to assume that the proposed approach can help to accelerate manual R&D processes in chemical laboratories.