Advanced continuous-flow microfluidic device for parallel screening of crystal polymorphs, morphology, and kinetics at controlled supersaturation

Automated Growth Rate Measurement of the Facet Surfaces of Single Crystals of the β-Form of l-Glutamic Acid Using Machine Learning Image Processing

Precision measurement of the growth rate of individual single crystal facets (hkl) represents an important component in the design of industrial crystallization processes. Current approaches for crystal growth measurement using optical microscopy are labor intensive and prone to error. An automated process using state-of-the-art computer vision and machine learning to segment and measure the crystal images is presented. The accuracies and efficiencies of the new crystal sizing approach are evaluated against existing manual and semi-automatic methods, demonstrating equivalent accuracy but over a much shorter time, thereby enabling a more complete kinematic analysis of the overall crystallization process. This is applied to measure in situ the crystal growth rates and through this determining the associated kinetic mechanisms for the crystallization of β-form l-glutamic acid from the solution phase. Growth on the {101} capping faces is consistent with a Birth and Spread mechanism, in agreement with the literature, while the growth rate of the {021} prismatic faces, previously not available in the literature, is consistent with a Burton-Cabrera-Frank screw dislocation mechanism. At a typical supersaturation of σ = 0.78, the growth rate of the {101} capping faces (3.2 × 10-8 m s-1) is found to be 17 times that of the {021} prismatic faces (1.9 × 10-9 m s-1). Both capping and prismatic faces are found to have dead zones in their growth kinetic profiles, with the capping faces (σc = 0.23) being about half that of the prismatic faces (σc = 0.46). The importance of this overall approach as an integral component of the digital design of industrial crystallization processes is highlighted.

Snap-on Adaptor for Microtiter Plates to Enable Continuous-Flow Microfluidic Screening and Harvesting of Crystalline Materials

Microtiter plate assay is a conventional and standard tool for high-throughput (HT) screening that allows the synthesis, harvesting, and analysis of crystals. The microtiter plate screening assays require a small amount of solute in each experiment, which is adequate for a solid-state crystal analysis such as X-ray diffraction (XRD) or Raman spectroscopy. Despite the advantages of these high-throughput assays, their batch operational nature results in a continuous decrease in supersaturation due to crystal nucleation and growth. Continuous-flow microfluidic mixer devices have evolved as an alternate technique for efficiently screening crystals under controlled supersaturation. However, such a microfluidic device requires a minimum of two inlets per micromixer to create cyclonic flow, thereby creating physical limitations for implementing such a device for HT screening. Additionally, the monolithic design of these microfluidic devices makes it challenging to harvest crystals for post-screening analysis. Here, we develop a snap-on adapter that can be reversibly attached to a microtiter plate and convert it into a continuous-flow microfluidic mixer device. The integration of the snap-on adapter with a flow distributor and concentration gradient generator provides greater control over screening conditions while minimizing the number of independent inlets and pumps required. The three-dimensional (3D)-printed snap-on adaptor is plugged into a 24-well plate assay to demonstrate salt screening of naproxen crystals. Different naproxen salts are crystallized using four different salt formers (SFs)-sodium hydroxide, potassium hydroxide, pyridine, and arginine-and four different solvents-ethanol, methanol, isopropyl alcohol, and deionized water. The wells are further inspected under an optical microscope to identify their morphological forms and yields. The crystals are then harvested for solid-state characterization using XRD and Fourier transform infrared spectroscopy, followed by measurement of their dissolution rates. The flexibility of the snap-on adapter to fit on a wide range of microtiter plates and the ease in harvesting and analyzing crystals postscreening are two significant advantages that make this device versatile for various applications.

In-line measurement of liquid-liquid phase separation boundaries using a turbidity-sensor-integrated continuous-flow microfluidic device

Liquid-liquid phase separation (LLPS), also known as oiling-out, is the appearance of the second liquid phase preceding the crystallization. LLPS is an undesirable phenomenon that can occur during the crystallization of active pharmaceutical ingredients (APIs), proteins, and polymers. It is typically avoided during crystallization due to its detrimental impacts on crystalline products due to lowered crystallization rate, the inclusion of impurities, and alteration in particle morphology and size distribution. In situ monitoring of phase separation enables investigating LLPS and identifying the phase separation boundaries. Various process analytical technologies (PATs) have been implemented to determine the LLPS boundaries prior to crystallization to prevent oiling out of compounds. The LLPS measurements using PATs can be time-consuming, expensive, and challenging. Here, we have implemented a fully integrated continuous-flow microfluidic device with a turbidity sensor to quickly and accurately evaluate the LLPS boundaries for a β-alanine, water, and IPA mixture. The turbidity-sensor-integrated continuous-flow microfluidic device is also placed under an optical microscope to visually track and record the appearance and disappearance of oil droplets. Streams of an aqueous solution of β-alanine, pure solvent (water), and pure antisolvent (IPA or ethanol) are pumped into the continuous-flow microfluidic device at various flow rates to obtain the compositions at which the solution becomes turbid. The onset of turbidity is measured using a custom-designed, in-line turbidity sensor. The LLPS boundaries can be estimated using the turbidity-sensor-integrated microfluidic device in less than 30 min, which will significantly improve and enhance the workflow of the pharmaceutical drug (or crystalline material) development process.

Machine Learning-Driven, Sensor-Integrated Microfluidic Device for Monitoring and Control of Supersaturation for Automated Screening of Crystalline Materials

Integrating sensors in miniaturized devices allow for fast and sensitive detection and precise control of experimental conditions. One of the potential applications of a sensor-integrated microfluidic system is to measure the solute concentration during crystallization. In this study, a continuous-flow microfluidic mixer is paired with an electrochemical sensor to enable in situ measurement of the supersaturation. This sensor is investigated as the predictive measurement of the supersaturation during the antisolvent crystallization of l-histidine in the water-ethanol mixture. Among the various metals tested in a batch system for their sensitivity toward l-histidine, Pt showed the highest sensitivity. A Pt-printed electrode was inserted in the continuous-flow microfluidic mixer, and the cyclic voltammograms of the system were obtained for different concentrations of l-histidine and different water-to-ethanol ratios. The sensor was calibrated for different ratios of antisolvent and concentrations of l-histidine with respect to the change of the measured anodic slope. Additionally, a machine-learning algorithm using neural networks was developed to predict the supersaturation of l-histidine from the measured anodic slope. The electrochemical sensors have shown sensitivity toward l-histidine, l-glutamic acid, and o-aminobenzoic acid, which consist of functional groups present in almost 80% of small-molecule drugs on the market. The machine learning-guided electrochemical sensors can be applied to other small molecules with similar functional groups for automated screening of crystallization conditions in microfluidic devices.

Selective desolvation in two-step nucleation mechanism steers crystal structure formation

The two-step nucleation (TSN) theory and crystal structure prediction (CSP) techniques are two disjointed yet popular methods to predict nucleation rate and crystal structure, respectively. The TSN theory is a well-established mechanism to describe the nucleation of a wide range of crystalline materials in different solvents. However, it has never been expanded to predict the crystal structure or polymorphism. On the contrary, the existing CSP techniques only empirically account for the solvent effects. As a result, the TSN theory and CSP techniques continue to evolve as separate methods to predict two essential attributes of nucleation - rate and structure. Here we bridge this gap and show for the first time how a crystal structure is formed within the framework of TSN theory. A sequential desolvation mechanism is proposed in TSN, where the first step involves partial desolvation to form dense clusters followed by selective desolvation of functional groups directing the formation of crystal structure. We investigate the effect of the specific interaction on the degree of solvation around different functional groups of glutamic acid molecules using molecular simulations. The simulated energy landscape and activation barriers at increasing supersaturations suggest sequential and selective desolvation. We validate computationally and experimentally that the crystal structure formation and polymorph selection are due to a previously unrecognized consequence of supersaturation-driven asymmetric desolvation of molecules.

Patterned microfluidic devices for rapid screening of metal-organic frameworks yield insights into polymorphism and non-monotonic growth

Metal-organic frameworks (MOFs) are porous crystalline structures that are composed of coordinated metal ligands and organic linkers. Due to their high porosity, ultra-high surface-to-volume ratio, and chemical and structural flexibility, MOFs have numerous applications. MOFs are primarily synthesized in batch reactors under harsh conditions and long synthesis times. The continuous depletion of metal ligands and linkers in batch processes affects the kinetics of the oligomerization reaction and, hence, their nucleation and growth rates. Therefore, the existing screening systems that rely on batch processes, such as microtiter plates and droplet-based microfluidics, do not provide reliable nucleation and growth rate data. Significant challenges still exist for developing a relatively inexpensive, safe, and readily scalable screening device and ensuring consistency of results before scaling up. Here, we have designed patterned-surface microfluidic devices for continuous-flow synthesis of MOFs that allow effective and rapid screening of synthesis conditions. The patterned surface reduces the induction time of MOF synthesis for rapid screening while providing support to capture MOF crystals for growth measurements. The efficacy of the continuous-flow patterned microfluidic device to screen polymorphs, morphology, and growth rates is demonstrated for the HKUST-1 MOF. The effects of solvent composition and pH modulators on the morphology, polymorphs, and size distribution of HKUST-1 are evaluated using the patterned microfluidic device. Additionally, a time-resolved FT-IR analysis coupled with the patterned microfluidic device provides quantitative insights into the non-monotonic growth of MOF crystals with respect to the progression of the bulk oligomerization reaction. The patterned microfluidic device can be used to screen crystals with a longer induction time, such as proteins, covalent-organic frameworks, and MOFs.

On-the-spot quenching for effective implementation of cooling crystallization in a continuous-flow microfluidic device

Illustrated is a microfludic cooling crystallization device that can effectively screen polymorphs, growth rates, and morphology of crystalline materials.