Methods for improving crystallization of small molecules

Abstract

Aspects of the present disclosure include methods for improving crystallization of small molecules by understanding the effect of hypergravity conditions on particle size distributions (PSDs). Aspects of the present disclosure also include a small molecule product produced by the methods of the present disclosure.

Description

METHODS FOR IMPROVING CRYSTALLIZATION OF SMALL MOLECULES1. BACKGROUND[1] Pharmaceutical R&D and manufacturing can widely benefit from a microgravity environment. Although the use of microgravity to improve control over the size of protein crystals has been demonstrated, very little is known about the effect of gravity on small molecules in a hypergravity state.[2] While recent advances achieved in space-based drug development build on foundational knowledge of protein crystallization developed onboard space stations, microgravity crystallization of small molecules has been less explored. Variable gravity platforms can serve an important role in generating datasets that measure the impact of gravity to better inform both space and conventional crystallization development programs.[3] There is a need to develop rapid screening tools in different gravitational environments to understand and optimize crystallization of small molecule pharmaceuticals.2. SUMMARY[4] Aspects of the present disclosure include methods for improving crystallization of small molecules by understanding the effect of hypergravity conditions on particle size distributions (PSDs)In contrast to biologies, small molecule therapeutics rely on crystallization as one of the most important unit operations at the intersection of both drug substance and drug product manufacturing. Control over crystal and particle properties are critical for both manufacturability and bioavailability of drug compounds.[5] Increased access to space has presented a renewed opportunity to conduct pharmaceutical development in orbit, largely due to the commercialization of reusable rockets. With in-space drug development increasingly feasible, rapidly advancing knowledge and understanding of how gravity impacts drug development processes such as crystallization can create great benefit. Variable gravity platforms can serve an important role in generating datasets that measure the impact of gravity to better inform both space and conventional crystallization development programs.[6] While gravitational forces do not directly impact thermodynamic properties of systems, kinetic and hydrodynamic processes are significantly altered. Reducing gravity suppressesconvection and sedimentation resulting in diffusion-driven transport and a reduction of crystallization rates.[7] In contrast, hypergravity environments achieved through increased centrifugal forces often increase crystallization rates through gravity-induced concentration gradients that alter local supersaturation to overcome kinetic barriers. In contrast to previous hypergravity studies that examine the crystallization of proteins and inorganic materials at high g-levels, the examples of the present disclosure provides the effect of crystallization of small molecules in hypergravity conditions in the 1-g to 5-g range.[8] An aspect of the present disclosure includes a method comprising: applying, for each of a plurality of crystallization operations, a different gravitational force greater than or equal to 1g to a small molecule solution during crystallization to produce a corresponding set of crystallized small molecules; measuring, for each set of crystallized small molecules, a particle size distribution (PSD) within the set of crystallized small molecules; selecting a desired PSD for a target set of crystallized small molecules; determining a target gravitational force to apply to a target small molecule solution to produce the target set of crystallized small molecules based on the measured PSD within each set of crystallized small molecules and the corresponding gravitational forces applied during the plurality of crystallization operations; and producing the target set of crystallized small molecules by applying the target gravitational force to the target small molecule solution during crystallization.[9] In some embodiments, at least one crystallization operation comprises spinning a centrifuge at a desired rpm corresponding to the gravitational force. In some embodiments, the plurality of crystallization operations is run on a crystallizer. In some embodiments, at least one crystallization operation comprises stirring the small molecule in solution at one or more desired rotations per minute (rpm) in the crystallizer. In some embodiments, at least one crystallization operation comprises stirring the small molecule in solution at one or more desired temperatures.

[0010] In some embodiments, the crystallization operation comprises performing cooling crystallization experiments.

[0011] In some embodiments, the method further comprises determining a cooling rate that generates a supersaturation required to crystallize the small molecule solution. In some embodiments, the method further comprises, before, during, and / or crystallization of the smallmolecule, taking optical images of the small molecule. In some embodiments, the crystallization operation further comprises, during crystallization, one or more of: measuring a metastable zone width (MSZW) of the small molecule in solution; measuring solute concentration of the small molecule solution at one or more temperatures; determining a nucleation rate of the small molecule in solution; measuring a crystallization growth rate of the small molecule solution; determining an induction time of the small molecule solution; calculating a supersaturation ratio of the small molecule solution; and measuring chord length distribution (CLD) of particles in the small molecule solution.

[0012] In some embodiments, the CLD is length-weighted CLD and / or cube-weighted CLD. In some embodiments, the solute concentration of the small molecule solution during crystallization is measured using Raman spectroscopy.

[0013] In some embodiments, the PSD within the set of the crystallized small molecules is measured using a laser diffraction instrument. In some embodiments, the PSD is calculated from raw data received from the laser diffraction instrument using a volume fraction (volume % of particles per m) of each set of crystallized small molecules, the volume % distribution, and the particle size.

[0014] In some embodiments, the different gravitational forces are selected from: 1g or more, 1.5 g or more, 2g or more, 2.5 g or more, 3g or more, 3.5g or more, 4g or more, 4.5g or more, and 5g or more. In some embodiments, each of the different gravitational forces are 1g, 2g, 3g, 4g, and 5g. In some embodiments, the rpms for a gravitational force of: 2g ranges from 30-35 rpm; 3g ranges from 40-45 rpm; 4g ranges from 45-50 rpm; and / or 5g ranges from 50-55 rpm.

[0015] In some embodiments, the rpm for a gravitational force of: 2g is 31 rpm; 3g is 40 rpm; 4g is 47 rpm; and 5g is 53 rpm.

[0016] In some embodiments, selecting the desired PSD for the target set of crystalized small molecules comprises selecting a target particle size range of the target set of crystallized small molecules.

[0017] In some embodiments, the method further comprises determining the target gravitational force that produces a PSD that maximizes the number of particles within the selected particle size range. In some embodiments, the desired PSD ranges from 1 pm to 2000 pm. In someembodiments, the desired PSD ranges from 5-110 pm. In some embodiments, the desired PSD ranges from 1-5 pm, 5-10 m, 10-15 pm, 15-20 pm, 20-25 pm, 25-30 pm, 40-50 pm, 50-60 pm, 60-70 m, 70-80 pm, 80-100 pm, 100-150 pm, or 150-200.

[0018] In some embodiments, determining the target gravitational force to apply to the target small molecule solution comprises extrapolating a relationship between PSDs and the applied gravitational forces. In some embodiments, the relationship between the PSDs and the applied gravitational force is the effects of different applied gravitational forces on particle size distribution (PSD) of the small molecules in the small molecule solution. In some embodiments, the determined target gravitational force is different from the applied gravitational forces. In some embodiments, the determined target gravitational force is less than 1g.

[0019] In some embodiments, the method further comprises determining a target time duration for which the target gravitational force is applied. In some embodiments, the method further comprises, before applying the different gravitational forces to the small molecule solution during crystallization, determining a cooling rate (°C / min) of the small molecule solution to promote desired crystallization nucleation and growth rates. In some embodiments, the method further comprises, before applying the different gravitational forces to the small molecule solution during crystallization, determining the solubility of the small molecule solution. In some embodiments, the method further comprises determining the correlation between the different applied gravitational forces and the crystal growth kinetics of the small molecule in the small molecule solution.

[0020] In some embodiments, the method further comprises determining the correlation between the different applied gravitational forces and the crystal nucleation (e.g., secondary nucleation) of the small molecule solution. In some embodiments, the method further comprises determining the correlation between the different applied gravitational forces and the de-supersaturation rate and / or supersaturation rate of the small molecule solution. In some embodiments, the method further comprises determining a correlation between the different applied gravitational forces and the polymorphism of the small molecules in the small molecule solution. In some embodiments, the method further comprises determining a target gravitational force to apply to a target small molecule solution to produce the target crystallized small molecule based on at least one of: the correlation between the different applied gravitational forces and the crystal growth kinetics of the small molecules in the small molecule solution; the correlation between the different applied gravitational forces and the crystal nucleation of the small molecules in the small molecule solution; thecorrelation between the different applied gravitational forces and the de-supersaturation rate of the small molecules in the small molecule solution; the correlation between the different applied gravitational forces and the solubility of the small molecules in the small molecule solution; and the correlation between the different applied gravitational forces and the polymorphism of the small molecules in the small molecule solution.

[0021] In some embodiments, the particles within each set of crystalized small molecules ranges from 1x101 crystals to 1x109 crystals. In some embodiments, the particles within each set of crystalized small molecules ranges from 1 x 103to 1 x 106crystals or 1 x 103to 1 xlO9crystals.

[0022] An aspect of the present disclosure includes a small molecule produced by the methods of the present disclosure.3. BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS

[0023] These and other features, aspects, and advantages of the present disclosure will become better understood with regard to the following description, and accompanying drawings, where:

[0024] FIGs. 1A-1C provide a schematic of the centrifuge setup during crystallization of small molecules. FIG. 1A provides an illustration of net force experienced during crystallization, which is the as a vector sum of the gravitational force and the centrifugal force. FIG. IB provides photographs of the centrifuge which features a tiltable basket at rest and (FIG. 1C) in motion.

[0025] FIG. 2 shows the experimental protocol used for seeded crystallization experiment with Crystal 16 on centrifuge.

[0026] FIGs. 3A-3B shows characterization of MSZQ of L-histidine for cooling crystallization in water. FIG. 3 A shows solubility and metastable zone width (MSZW) of L-histidine cooling crystallization in water. FIG. 3B shows temperature, HDR turbidity and nucleation temperature in the crystallizer for metastable zone width determination experiments.

[0027] FIG. 4A shows the Raman spectra of L-histidine in water with major peaks. FIG. 4B shows tracking peaks in the Raman spectra with time.

[0028] FIG. 5A shows peak height from L-histidine spectra. FIG. 5B shows experimental L- histidine concentration and FIG. 5C shows the temperature. FIG. 5D shows predicted vs actual concentrations from the chemometric model.

[0029] FIG. 6A shows temperature, peak height (cm'1) vs time for validation experiment. FIG. 6B shows predicted L-histidine concentration vs actual L-histidine concentration for validation experiment.

[0030] FIGs. 7A-7C provides concentration and temperature parameters of L-histidine in solution. FIG. 7A shows evolution of L-histidine concentration and temperature evolution. FIG. 7B shows evolution of total counts / s with y-axis on a logarithmic scale, and FIG. 7C shows evolution of concentration in the phase diagram for seeded cooling crystallization experiments with L-histidine.

[0031] FIGs. 8A-8F provide in-situ images obtained from the Blaze probe and FIGs. 8G and 8H show evolution of (FIG. 8G) length-weighted CLD and (FIG. 8H) cube weighted CLD throughout the seeded cooling crystallization process.

[0032] FIG. 9 show volume-based PSD for different g-levels at cooling rate of 0.1 °C / min

[0033] FIGs. 10A-10E provide optical micrographs of seeded cooling crystallization of L-histidine at (FIG. 10A) 1g, (FIG. 10B) 3g, (FIG. 10C) 5g, (FIG. 10D) Seed crystals. Scale bars, 500 pm, (FIG. 10E) Bar plots indicating mean mode of PSD along with individual modes overlayed as a function of g-level. Error bars indicate standard deviation.

[0034] FIGs. 11A-11E show supersaturation rates of the API in solution. FIG. 11A shows an illustration of vial used for centrifuge cooling crystallization experiments. The color indicates a gradient in concentration. Evolution of system during cooling crystallization (FIG. 11B) the case of a uniform concentration profile across the vial and (FIG. 11C) for a non-uniform concentration profile. Evolution of the system during a dissolution experiment for (FIG. 11D) a spatially homogenous solution concentration and (FIG. HE) when the solution concentration has a spatial gradient.

[0035] FIGs. 12A-12B provide volume-based PSD at (FIG. 12A) after 20 minutes and (FIG. 12B) 40 minutes for different g-levels.

[0036] FIG. 13 provide volume % vs size distribution for seed crystals and different g-levels.

[0037] FIG. 14 shows X-ray powder diffraction patterns for seed and product samples, along with reference patterns for Form A and Form B.

[0038] FIG. 15 shows counts / s from FBRM and temperature from unseeded cooling crystallization experiment implemented with same temperature profile as seeded cooling crystallization.

[0039] FIG. 16 shows rarnan spectra for L-histidine in solution used to develop the chemometric model.

[0040] FIGs. 17A-17B shows Cumulative distributions for (FIG. 17A) Circularity, (FIG. 17B) L / D ratios for the seed and product crystals.

[0041] FIGs. 18A-18E shows disappearance temperature of L-histidine in water on heating at concentration of (FIG. 18A) 0.033g / g, (FIG. 18B) 0.039 g / g, (FIG. 18C) 0.047 g / g, (FIG. 18D) 0.054 g / g, (FIG. 18E) 0.063 g / g.

[0042] FIG. 19 shows PSD from breakage experiments run with (a) bottom stirrer, (b) overhead stirrer.

[0043] FIG. 20 shows a histogram of the predicted concentration from the chemometric model of equation (5).4. DETAILED DESCRIPTION

[0044] Aspects of the present disclosure includes methods for improving crystallization of small molecules by understanding the effect of hypergravity conditions on PSDs.

[0045] One aspect of the present disclosure includes a method comprising: applying, for each of a plurality of crystallization operations, a different gravitational force greater than or equal to 1g to a small molecule solution during crystallization to produce a corresponding set of crystallized small molecules; measuring, for each set of crystallized small molecules, a particle size distribution (PSD) within the set of crystallized small molecules; selecting a desired PSD for a target set of crystallized small molecules; and determining a target gravitational force to apply to a target small molecule solution to produce the target set of crystallized small molecules based on the measured PSD within each set of crystallized small molecules and the corresponding gravitational forces applied during the plurality of crystallization operations. In some embodiments, the method comprises producing the target set of crystallized small molecules by applying the target gravitational force to the target small molecule solution during crystallization.

[0046] Another aspect of the present disclosure includes a small molecule product produced by the methods of the present disclosure.

[0047] Another aspect of the present disclosure includes a system for carrying out the methods of the present disclosure.4.1. Summary of experimental observations

[0048] The methods described herein were used to understand the effect of gravity on the particle size distribution, nucleation, and growth kinetics of a small molecule such as L-histidine in order to improve small molecule drug manufacturing. The present inventors constructed a large diameter centrifuge for hypergravity studies, which aimed to measure the effect of g-level on product particle size distributions. Seeded cooling crystallization experiments were designed and conducted on a Crystal 16 (Technobis Crystallization Systems) reactor at different g-levels. This was followed by further targeted experiments in an EasyMax reaction calorimeter fitted with Process Analytical Technology (PAT) tools to understand the fundamental mechanisms underlying the sensitivity of the cooling crystallization to increasing g-levels.4.1.1. Crystallization Operations

[0049] An aspect of the present methods include applying, for each of a plurality of crystallization operations, a different gravitational force greater than or equal to lg to a small molecule solution during crystallization to produce a corresponding set of crystallized small molecules.

[0050] In some embodiments, at least one crystallization operation comprises spinning a centrifuge at a desired rotations per minute (rpm) corresponding to the gravitational force. In some embodiments, the rpm corresponding to the gravitational force depends on the size parameters of the centrifuge.

[0051] In some embodiments, at least one crystallization operation comprises stirring the small molecule in solution at one or more desired rpms in a crystallizer. In some embodiments, at least one crystallization operation comprises stirring the small molecule in solution at one or more desired temperatures.

[0052] In some embodiments, at least one crystallization operation comprises performing cooling crystallization experiments. During cooling crystallization, temperature is used to generate thesupersaturation necessary to crystallize the molecules from solution. The solubility of most pharmaceutical molecules decreases with temperature. Thus, cooling the solution reduces the solubility, thereby increasing the supersaturation which is the driving force for crystal nucleation and growth. There are two major modes of cooling crystallization - unseeded and seeded cooling crystallization. In contrast to unseeded crystallization, seeded cooling crystallization is more robust and well-controlled, and is thereby widely used in the pharmaceutical industry. Various micromeritic properties such as particle shape and size can be controlled by implementing desired cooling profiles during batch cooling crystallization.

[0053] In some embodiments, one or more of the plurality of crystallization operations is performed using a crystallization platform.4.1.1.1 Crystallization platform

[0054] Crystallization operations can be performed using a crystallization platform. In some embodiments, the crystallization platform is a commercially available crystallization platform. In certain embodiments, the crystallization platform is a crystallizer. In certain embodiments, the crystallization platform is a multi-reactor crystallizer. In certain embodiments, the crystallization platform is a commercially available crystallizer (e.g., Crystal 16) attached to one or more centrifuges. In certain embodiments, the crystallization platform is a commercially available crystallizer (e.g., EasyMax Reactor) attached to one or more centrifuges. In some embodiments, the crystallization platform includes two or more commercially available crystallizers (e.g., EasyMax Reactor and Crystal 16) attached to one or more centrifuges. In some embodiments, the crystallizer is housed in a basket with an adjustable tilt to align the sample vial axis with a gravity vector. The gravity vector is a function of the rotations per minute (rpm), the radius of the arm, and the earth’s gravity vector, as described in FIG. 1.

[0055] In some embodiments, the crystallization platform is a hypergravity crystallization payload system. In certain embodiments, the crystallization platform is a hypergravity thermal payload system. In certain embodiments, the crystallization platform is a hypergravity antisolvent crystallization payload system. In some embodiments, the hypergravity antisolvent crystallization payload system enables solution-based antisolvent crystallization, cooling crystallization, slurry crystallization, and / or evaporative crystallization. In some embodiments, the payload system is designed to ensure consistent performance over a wide range of effective gravity levels, enablingcomparison in performance as a function of gravity. In some embodiments, the payload system is gravity-independent, ensuring uniform mixing independent of convection at a large enough scale for producing significant material. Non-limiting examples of hypergravity crystallization platforms are described in U.S. Patent Application No.: 18 / 361,615, which is hereby incorporated by reference in its entirety.

[0056] In some embodiments, the payload system includes more than one growth chamber. In some embodiments, the payload system includes growth chambers of 20 mL. In some embodiments, the payload system includes a tuning mechanism that allows an antisolvent ratio to be tunable between 0 and 1. In some embodiments, the payload system is configured to operate at room temperature. Alternatively, or in addition, the payload system may incorporate thermal control. For example, it may be desirable to heat both the growth chamber and the mixing chip to the same temperature to avoid any undesired changes in supersaturation due to temperature differences. In some embodiments, thermal control may be achieved by a TEC and a heatsink. In some embodiments, a convective cover may also be used to achieve thermal control.

[0057] In some embodiments, the crystallization platform also includes one or more in-situ analytical sensors for infrared spectroscopy, Raman spectroscopy, turbidity sensing, optical imaging, and video microscopy. These sensors can provide information about the crystallization process that can be used to trigger endpoints or drive feedback control to over supersaturation or other relevant process control parameters. In some embodiments, the sensors may also interface with the mixing chip, with flow cells in between the major subsystems of the hardware, or directly with the growth chambers and / or filters.

[0058] In some embodiments, the crystallization platform for crystallization of a small molecule under hypergravity comprises a solution well or vial configured to hold a liquid solution containing a chemical substance such as a small molecule and / or active pharmaceutical ingredient.4.1.1.2 Gravitational Force

[0059] An aspect of the present methods include applying, for each of a plurality of crystallization operations, a different gravitational force greater than or equal to 1g to a small molecule solution during crystallization to produce a corresponding set of crystallized small molecules.

[0060] In some embodiments, the gravitational force is applied using a centrifuge. In some embodiments, applying the gravitational force comprises spinning a centrifuge at a desired rpm corresponding to the gravitational force.

[0061] In some embodiments, the different gravitational forces are selected from 1g or more, 1.5 g or more, 2g or more, 2.5 g or more, 3g or more, 3.5g or more, 4g or more, 4.5g or more, and 5g or more. In certain embodiments, the different gravitational forces comprise lg, 2g, 3g, 4g, and 5g.

[0062] The rpm used to create a desired gravitational force in the centrifuge depends on the centrifuge diameter. For example, the net acceleration in the centrifuge is given by equation (1) where is the gravitational acceleration, co is the rotation speed of the centrifuge and r is the radius of the centrifuge arm. The angle for the effective gravitational force is given by equation(2).

[0063] a = ^g2+ cor)2(1) 1 g

[0065] The rotation speed (rpm) is changed to obtain a desired effective gravitational force. The rotation speed can be calculated from equation (1) and the angle of the effective gravitational force is given by equation(4). An overview of the net acceleration vector and the effective angle is shown in FIG. 1A.

[0068] In some embodiments, the rpm for a gravitational force of 2g ranges from 30-35 rpm. In some embodiments, the rpm for a gravitational force of 3g ranges from 40-45 rpm. In some embodiments, the rpm for a gravitational force of 4g ranges from 45-50 rpm. ranges from 45-50 rpm. In some embodiments, the rpm for a gravitational force of 5g ranges from 50-55 rpm.

[0069] The term “different gravitational force”, as used herein, refers to applying, a first gravitational force (e.g., 1g, 2g, 3g, 4g, or 5g) on a small molecules solution during crystallization, followed by applying a second (e.g., different from the first gravitational force) gravitational force on a second small molecule in solution, to understand how different gravitational forces each play a role in the crystallization process of a set of small molecules.

[0070] For example, applying a different set of gravitational forces to a small molecule solution can include applying a first gravitational force of 1g to a first small molecule solution during crystallization, applying a second gravitational force of 2g to a second small molecule solution during crystallization, applying a third gravitational force of 3g to a third small molecule solution during crystallization, applying a fourth gravitational force of 4g to a fourth small molecule solution during crystallization, and / or applying a fifth gravitational force of 5g to a fifth small molecule in solution during crystallization.

[0071] The particle size distribution (PSD) measurement occurs following each gravitational force applied.

[0072] In some embodiments, applying a different set of gravitational forces to a first small molecule solution can include applying a first gravitational force of 1g to a set of small molecules in solution during crystallization, measuring the PSD for each set of crystallized small molecules, and following measurement of the PSD, applying a second gravitational force (e.g. different from the first gravitational force, e.g., 2g) to the a second small molecule solution during crystallization, and measuring the PSD for each set of crystallized small molecules, etc. These steps can be repeated for different gravitational forces (e.g., 3g, 4g, and / or 5g), and PSD of the set of crystallized molecules is measured for each gravitational force.

[0073] In some embodiments, each gravitational force is applied for 1 minute or more, 2 minutes or more, 3 minutes or more, 4 minutes or more, 5 minutes or more, 6 minutes or more, 7 minutes or more, 8 minutes or more, 9 minutes or more, 10 minutes or more, 15 minutes or more, 20 minutes or more, 25 minutes or more, 30 minutes or more, 35 minutes or more, 40 minutes or more, 45 minutes or more, 50 minutes or more, 55 minutes or more, 60 minutes or more, 65 minutes or more, 70 minutes or more, 75 minutes or more, 80 minutes or more, 85 minutes or more, 90 minutes or more, 95 minutes or more, 100 minutes or more, 105 minutes or more, 110 minutes or more, 115 minutes or more, or 120 minutes or more.

[0074] In some embodiments, each gravitational force is applied for 1 minute, 2 minutes, 3 minutes, 4 minutes, 5 minutes, 6 minutes, 7 minutes, 8 minutes, 9 minutes, 10 minutes, 15 minutes, 20 minutes, 25 minutes, 30 minutes, 35 minutes, 40 minutes, 45 minutes, or 50 minutes.

[0075] In some embodiments, each gravitational force is applied between 1-120 minutes, 1-50 minutes, 1-25 minutes, 1-20 minutes, 5-20 minutes, 5-50 minutes, 1-5 minutes, 5-10 minutes, 10-15 minutes, 15-20 minutes, 20-25 minutes, 25-30 minutes, 30-35 minutes, 35-40 minutes, 40-45 minutes, or 45-50 minutes.4.1.1.3 Metastable Zone Width (MSZW)

[0076] In some embodiments, at least one crystallization operation comprises measuring metastable zone width (MSZW) of the small molecules in solution. The MSZW is one of the critical process parameters that define a crystallization process. In some embodiments, wherein measuring MSZW of the small molecules in solution is measured during crystallization of the small molecules in solution.

[0077] Conventional crystallization processes of small molecule organic compounds are divided into three main regions: labile region, metastable zone, and the stable region. In the stable region, the concentration of the dissolved molecule is below that of the solubility limit, hence no crystallization can occur. As the temperature is reduced (cooling) or the loss of solvent occurs (evaporation), the solution will move to the saturation limit, the maximum thermodynamically stable concentration. Further cooling or concentration results in a metastable state (metastable zone) in which the solution is supersaturated but kinetically stable. The third region of the solubility crystallization diagram is the labile region, an area of supersaturation where spontaneous precipitation (or other phase changes such as “oiling out”) occurs. Using this method, crystals of a molecule are typically grown from an initially subsaturated solution. Cooling or evaporation can be used to move to the saturation limit and beyond, into the metastable zone. At this stage, following either spontaneous nucleation or the addition of nucleation sites (e.g., seed crystals), crystal growth can begin. Once this has commenced further cooling or evaporation results in further crystal growth, with the crystalline material and molecules in solution in an equilibrium at the saturation limit. The solubility crystallization process is described in Metherall et al., (Metheral et al. “Advanced crystallisation methods for small organic molecules.” The Royal Society of Chemistry. 2023, 52, 1995-2010), which is hereby incorporated by reference in its entirety.

[0078] In some embodiments, the metastable zone is identified by using a probe-based technique (e.g., Blaze probe). In some embodiments, the Blaze probe monitors the particle counts / second in the small molecule solution. In some embodiments, the method comprises measuring the MSZW at different active pharmaceutical ingredient (API) concentrations of the small molecule corresponding to different saturation temperatures.

[0079] In some embodiments, the method further comprises characterizing the MSZW of the small molecules in solution for cooling crystallization in water.

[0080] For example, in order to measure the MSZW, a saturated small molecule solution can be cooled slowly until the High Dynamic Range (HDR) turbidity provided by the Blaze software shows an inflection indicating the formation of fine particles in solution. In some embodiments, the solution is heated to 5°C above its saturation temperature, and then is cooled down. The process was repeated at different API (active pharmaceutical ingredient) concentrations corresponding to different saturation temperatures.4.1.1.4 Solubility & Concentration

[0081] In some embodiments, at least one crystallization operation comprises measuring the solubility of the small molecules in solution.

[0082] In some embodiments, at least one crystallization operation comprises measuring solute concentration of the small molecules in solution at one or more temperatures. In some embodiments, the solute concentration of the small molecules in solution is measured before and / or during crystallization of the small molecules in solution.

[0083] In some embodiments, at least one crystallization operation comprises measuring the supersaturation rate of the small molecules in solution. In some embodiments, at least one crystallization operation comprises measuring the de-supersaturation rate of each set of small molecules in solution.

[0084] In some embodiments, the solute concentration of the small molecule solution during crystallization is measured using Raman spectroscopy. In some embodiments, the Raman spectra are a function of the solute concentration and temperature. In some embodiments, the method comprises building a chemometric model to calculate the small molecule concentration as a function of the Raman spectra and the small molecule solution temperature. In some embodiments, thechemometric model can be built based on the solubility and nucleation threshold of the small molecules in solution and the Raman spectra of the small molecules in solution as shown in FIGs. 3 and 16.

[0085] In some embodiments, a chemometric model is calculated using equation (5), where h is the peak height from the Raman spectra and T is the temperature. This model describes the peak height in the Raman spectra as a function of the small molecule concentration and solution temperature:

[0086] where C is the predicted concentration of the small molecule in solution, T is the temperature, h is the peak height of the Raman peak at a specific wavelength, and ai, a2, a3, a4, and as are the concentration coefficients are empirically determined with experimental data (e.g., calibration experiments) for each small molecule solution. Calibration experiments capture peak height response as a function of temperature. An example of this is shown in FIG. 5. The calibration experiments of FIG. 5A captured the peak height response as a function of temperature. FIG. 5B shows the concentration for each experiment, and FIG. 5C shows the temperature during each experiment (ramping down). For example, to fit the coefficients ai, a2, a3, a4, and as, experiments are run where the input variables are temperature and the small molecule concentration and the response variable is the peak height (e.g., 1286 cm'1for the 1-histidine sample in Example 1). The coefficients are fit to the experimental data by solving a system of linear equations to get to a best fit model. Once the coefficients are calculated, the model can be applied to predict the small molecule concentration from only Raman spectra and temperature.

[0087] In some embodiments, the model calibration is carried out using a written script in MATLAB software. In some embodiments, linear regression is carried out in MATLAB using the function mldivide which solves a system of linear equations from which the coefficients in equation (5) are determined, followed by determination of goodness of fit for the model.

[0088] In some embodiments, at least one crystallization operation comprises measuring chord length distribution (CLD) of particles of the small molecules in solution. Focused beam reflectance measurement (FBRM) is an analytical tool to study crystallization processes and to measure the chord length distribution of particles in real-time during the crystallization process. Additionally, a Blaze probe can provide information of how the CLD changes in the solution with time. FBRM utilizes a focused laser beam that scans in a circular path. As the light scans across the particlespassing in front of the probe window, light is scattered in all directions. The light that is scattered back towards the probe measures a chord length distribution (CLD) off the given particle. An advantage of using FBRM is that it can be inserted directly into the crystallizer with dense slurry and the measured data count can be used to isolate the size range in which a change occurred.

[0089] In some embodiments, the CLD is length-weighted CLD and / or cube-weighted CLD.

[0090] In some embodiments, the evolution of the Cube weighted (CW) CLD in the crystallization process is a representation of the evolution of the coarser particles in the crystallization process.4.1.1.5 Nucleation

[0091] In some embodiments, at least one crystallization operation comprises determining a nucleation rate of the small molecules in solution. The point of nucleation is registered as a sudden increase in the counts / s, which corresponds to the primary nucleation threshold.4.1.1.5.1 Primary nucleation

[0092] Primary nucleation is the spontaneous appearance of nuclei in a supersaturated solution. The critical temperature needed to cause primary nucleation is called the nucleation threshold, and the region of the phase diagram between the solubility curve and the nucleation threshold is called the metastable zone.4.1.1.5.2 Secondary nucleation

[0093] In some embodiments, at least one crystallization operation comprises measuring a crystal nucleation of the small molecules in solution. In some embodiments, the crystal nucleation is secondary nucleation.

[0094] Within the metastable zone, secondary nucleation can occur via surface breeding, shear, and attrition. The rates of secondary nucleation and growth are both dependent on instantaneous supersaturation. Since the rate of cooling determines instantaneous supersaturation, it also determines the relative rates of secondary nucleation. Simultaneously, breakage and attrition of crystals in solution can occur due to particle-impeller, particle-wall or particle-particle interactions. Breakage and attrition can both lead to a shift in the particle size distribution, however breakage / attrition conserves the total mass of solids during the crystallization while secondary nucleation and growth leads to increased solids mass.4.1.1.6 Growth kinetics

[0095] In some embodiments, at least one crystallization operation comprises measuring a crystallization growth rate of the small molecules in solution. In some embodiments, wherein at least one crystallization operation comprises determining an induction time and / or nucleation rate of the small molecules in solution.4.1.1.7 Supersaturation Ratio

[0096] In some embodiments, at least one crystallization operation comprises calculating a supersaturation ratio of the small molecules in solution.

[0097] In some embodiments, at least one crystallization operation comprises determining a supersaturation ratio of the small molecules in solution. In some embodiments, the supersaturation ratio (SS) is calculated according to equation (6):

[0098] (6), where C (gsoiute / gsoivent) is the solute concentration and Csat(gsoiute / gsoivent) is the saturation concentration.4.1.2. Particle size distribution (PSD)

[0099] An aspect of the present methods further includes measuring, for each set of crystallized small molecules, a PSD within the set of crystallized small molecules.

[0100] In some embodiments, for each gravitational force applied, the set of crystallized small molecules are withdrawn from the crystallization platform and analyzed using laser diffraction. In some embodiments, PSD is measured using a laser diffraction instrument (e.g., LS 13320, Beckman Coulter). For example, laser diffraction measurements can be used to measure and / or quantify the particle size distribution of the obtained set of crystallization small molecule products at a range of g-levels, and the observed peaks from the processed to determine PSD.

[0101] In some embodiments, the laser diffraction instrument uses a 5mW laser diode with a wavelength of 750 nm as the main illumination source. Reagent alcohol (anhydrous ethanol 90%, methanol 5%, 2-propanol 5% v / v), can be used as the dispersant for the laser diffraction measurements. The raw data obtained from the laser diffraction instrument is the fractional volume% vs particle size (e.g. volume fraction as a function of particle size). This distribution is post processed to calculate the particle size distribution (vol% / pm) following equation (7).

[0102] Equation (7) is used normalize the laser diffraction data to improve interpretability. fv(vol% / pm) is the normalized volume fraction, Nv(vol%) is the volume fraction measured by laser diffraction, and Ax is a bin width. To calculate fv(vol% / pm), Nv(vol%) is divided by the bin width.

[0103] In some embodiments, the bin width is 0.1 pm. In some embodiments, the bin width is 0.2 pm, 0.3 m, 0.4 pm, 0.5 pm, 0.6 m, 0.7 m, 0.8 m, 0.9 pm, 1 pm, 2 pm, 3 pm, 4 pm, or 5 pm.

[0104] In some embodiments, the laser diffraction instrument outputs N directly. N is volume fraction (vol%) vs particle size. A non-limiting example of N for L-histidine as described in Example 1 showing the volume fraction (volume %) vs particle is provided in FIG. 13. To normalize this data, every value of N is divided by the bin width. A non-limiting example of the normalized volume fraction of L-histidine is shown in FIG. 9.4.1.3. Small molecules

[0105] An aspect of the present methods includes applying a different gravitational force to a small molecule solution during crystallization to produce a corresponding set of crystallized small molecules. In some embodiments, the corresponding set of crystallized small molecules is produced for each gravitational force applied.

[0106] In some embodiments, the small molecule includes one or more active agents. In some embodiments, the small molecule is a small molecule active pharmaceutical ingredient (API). An API is typically a highly purified chemical compound or mixture of compounds. Some common chemical classes of APIs include small molecules such as peptides, carbohydrates or nucleotides and nucleic acids. APIs can exist in different chemical forms, including the free base or free acid (e.g., the neutral form of the molecule), salt forms (e.g., such as hydrochloride or sodium salts) that improve solubility or stability, and hydrates or solvates that incorporate water or other solventmolecules into their crystal structure. Additionally, APIs can have polymorphs, which are different crystalline forms of the same chemical compound that can affect solubility and bioavailability.

[0107] The particle size and distribution of API particles can significantly affect the drug's properties. For instance, micronized APIs are reduced to very small particle sizes to improve dissolution and bioavailability, while some APIs are formulated as nanoparticles to enhance delivery. Chirality is another important aspect, as many APIs are chiral molecules that can exist in two mirror-image forms (enantiomers). A racemic mixture contains equal amounts of both enantiomers, whereas a single enantiomer formulation contains only one, often chosen for its superior efficacy or safety profile.

[0108] In some embodiments, the small molecule API is formulated as co-crystals, which are crystalline structures containing the API molecule along with another neutral molecule. Co-crystals can improve the solubility, stability, or bioavailability of the API. In some embodiments, APIs are prodrugs. These are inactive compounds that are metabolized in the body to produce the active drug, allowing for improved absorption, distribution, or reduced side effects.

[0109] In some embodiments, a small molecule solution includes one or more small molecules (e.g., 1 or more, 2 or more, 3 or more, 4 or more, 5 or more, 6 or more, 7 or more, 8 or more, 9 or more, or 10 or more small molecules) in a solution. In some embodiments, a small molecule solution includes 10 or more small molecules (e.g., 10 or more, 20 or more, 30 or more, 40 or more, 50 or more, 60 or more, 70 or more, 80 or more, 90 or more, or 100 or more small molecules) in a solution. In some embodiments, the small molecules are in an aqueous based-solution. In some embodiments, the small molecules are in a nonaqueous based-solution. In some embodiments, a small molecule solution includes a saturated solution of one or more small molecules. In some embodiments, the small molecules are in slurry-based solution. In some embodiments, a small molecule solution comprises 10 or more small molecules in solution, 15 or more small molecules in solution, 20 or more small molecules in solution, 25 or more small molecules in solution, 30 or more small molecules in solution, 35 or more small molecules in solution, 40 or more small molecules in solution, 45 small molecules in solution, or 50 or more small molecules in solution.[HO] In some embodiments, the method further comprises adding seed crystals into the small molecule solution during seed cooling crystallization experiments.[HI] In some embodiments, the number of particles within each set of crystalized small molecules ranges from 1x101crystals to 1x109crystals. In some embodiments, the particles within each set of crystalized small molecules ranges from 1 x 103to 1 x 106crystals or 1 x 103to 1 xlO9crystals. In some embodiments, the number of particles within each set of crystalized small molecules ranges from 1 x 103to 1 xlO6crystals. In some embodiments, the number of particles within each set of crystalized small molecules ranges from 1 x 102to 1 xlO6crystals. In some embodiments, the number of particles within each set of crystalized small molecules ranges from 1 x 102to 1 xlO9crystals. In some embodiments, the number of particles within each set of crystalized small molecules ranges from 1 x 104to 1 xlO9crystals. In some embodiments, the number of particles within each set of crystalized small molecules ranges from 1 x 105to 1 xlO9crystals. In some embodiments, the number of particles within each set of crystalized small molecules ranges from 1 x106to 1 xlO9crystals. In some embodiments, the number of particles within each set of crystalized small molecules ranges from 1 x 107to 1 xlO9crystals. In some embodiments, the number of particles within each set of crystalized small molecules ranges from 1 x 108to 1 xlO9crystals. In some embodiments, the number of particles within each set of crystalized small molecules ranges from 1 x 101to 1 xlO10or more crystals. In some embodiments, the number of particles within each set of crystalized small molecules comprises 1 x 101or more crystals, 1 x 102or more crystals, 1 x 103or more crystals, 1 x 104or more crystals, 1 x 105or more crystals, 1 x 106or more crystals, 1 x107or more crystals, 1 x 108or more crystals, 1 x 109or more crystals, 1 x IO10or more crystals, 1 x 1011or more crystals, 1 x 1012or more crystals, 1 x 1013or more crystals, 1 x 1014or more crystals, 1 x 1015or more crystals, 1 x 1016or more crystals, 1 x 1017or more crystals, 1 x 1018or more crystals, 1 x 1019or more crystals, 1 x 1020or more crystals, 1 x 1021or more crystals, 1 x 1022or more crystals, 1 x 1023or more crystals, 1 x 1024or more crystals, or 1 x 1025or more crystals.4.1.4. Optical Imaging

[0112] In some embodiments, the method further comprises taking optical images after crystallization of the set of small molecules. Optical images can be taken using any known optical imaging modality, e.g., using stereo microscope, scanning electron microscopy, and the like.4.1.5. Selection of PSD for target set of crystallized small molecules

[0113] An aspect of the present methods includes selecting a desired PSD for a target set of crystallized small molecules. In some embodiments, selecting the desired PSD for the target set ofcrystallized small molecules comprises selecting a target particle size range of the target set of crystallized small molecules.

[0114] Determining a target particle size range of a target small molecule depends on a number of factors, including, but not limited to: target delivery location of the small molecule in the body, the route at which the small molecule drug is delivered, the bioavailability of the small molecule drug, and the absorption of the small molecule drug. Such factors are taken into consideration when determining a target particle size distribution of a small molecule during small molecule drug formulation and manufacturing.

[0115] The particle size of the small molecule drug can directly affect the solubility and dissolution rate of the drug, thereby affecting the clinical efficacy. For example, in the development of solid preparations, reducing the particle size of small molecule drug particles can significantly increase the dissolution rate of insoluble drug tablets.

[0116] Moreover, additional considerations of target particle size distributions of a target set of crystallized small molecule can depend on the target areas of interest in the body (e.g., stomach and intestines, lungs, skin, brain, cancer tumor, kidneys, livers, muscles, eyes, nose, ears, etc.), as well as delivery mechanism (e.g., drug particles or carriers in m or nm, from passively targeting its site of action, crossing the blood-brain barrier, to nanoparticle encapsulation and surface modification for optimal active uptake), and route of administration including gastrointestinal tract, lungs, skin, kidney, liver, eye, nose, and ear, and into the blood-brain barrier, the cancer tumour matrix, intramuscular administration, subcutaneous administration, intravenous administration, oral administration, and administration via inhalation).

[0117] In some embodiments, the desired PSD ranges from 1 pm to 2000 pm. In some embodiments, the desired PSD ranges from 5-110 pm. In some embodiments, the desired PSD ranges from 1-5 pm, 5-10 pm, 10-15 pm, 15-20 pm, 20-25 pm, 25-30 pm, 40-50 pm, 50-60 pm, 60-70 pm, 70-80 pm, 80-100 pm, 100-150 pm, 150-200 pm, 200-250 pm, 250-300 pm, 300-350 pm, 350-400 pm, 400-450 pm, 450-500 pm, 500-600 pm, 600-700 pm, 700-800 pm, 800-900 pm, 900-1000 pm, 1000-1500 pm, or 1500-2000 pm.

[0118] In some embodiments, the target particle size ranges from 1 pm to 2000 pm. In some embodiments, the target particle size ranges from 5-110 pm. In some embodiments, the targetparticle size ranges from 1-5 pm, 5-10 pm, 10-15 pm, 15-20 pm, 20-25 pm, 25-30 pm, 40-50 pm, 50-60 pm, 60-70 pm, 70-80 pm, 80-100 pm, 100-150 pm, 150-200 pm, 200-250 pm, 250-300 pm, 300-350 pm, 350-400 pm, 400-450 pm, 450-500 pm, 500-600 pm, 600-700 pm, 700-800 pm, 800- 900 pm, 900-1000 pm, 1000-1500 pm, or 1500-2000 pm.

[0119] In some embodiments, the desired PSD depends on the bioavailability of the small molecule drug. In some embodiments, the desired PSD depends on the route of administration of the small molecule drug. In some embodiments, the desired PSD depends on the absorption of the small molecule drug. In some embodiments, the desired PSD depends on the drug delivery mechanism or dosage form of the small molecule (e.g., tablet, nanoparticle, emulsions, capsule, ointment, etc.). In some embodiments, the desired PSD depends on the bioavailability of the small molecule drug. In some embodiments, the desired PSD depends on the dissolution rate of the small molecule.4.1.6. Other Crystallization Operations

[0120] An aspect of the present methods further includes before applying the different gravitational forces to the small molecule solution during crystallization, determining a cooling rate (°C / min) of the small molecule solution to promote desired crystallization nucleation and growth rates of the small molecules in solution.

[0121] In some embodiments, the method further comprises, before applying the different gravitational forces to the small molecule solution during crystallization, determining the solubility or solute concentration of the small molecule solution.

[0122] In some embodiments, the method further comprises, before applying the different gravitational forces to the small molecule solution during crystallization, determining the MSZW of the small molecule solution.

[0123] In some embodiments, the method further comprises, before applying the different gravitational forces to the small molecule solution during crystallization, determining the nucleation rate of the small molecule in solution. In some embodiments, the method further comprises, before applying the different gravitational forces to the small molecule solution during crystallization, determining the growth rate of the small molecule in solution. In some embodiments, the method further comprises, before applying the different gravitational forces to the small molecule solution during crystallization, determining the supersaturation ratio of the small molecule in solution. Insome embodiments, the method further comprises, before applying the different gravitational forces to the small molecule solution during crystallization, determining the de-supersaturation rate of the small molecule in solution. In some embodiments, the method further comprises, before applying the different gravitational forces to the small molecule solution during crystallization, determining the CLD and / or LWCLD of the small molecule in solution.4.1.7. Determining a target gravitational force

[0124] An aspect of the present methods includes determining the target gravitational force that produces a PSD that maximizes the number of particles within the selected particle size range. The target gravitational force depends on the target or desired PSD of the small molecule of interest. In some embodiments, determining the target gravitational force to apply to the target small molecule solution comprises extrapolating a relationship between PSDs and the applied gravitational forces.

[0125] For instance, once a desired PSD or target particle size range for the target set of crystallized small molecules is determined, the method determines which applied gravitational force from the measured PSDs would lead to the desired PSD or target particle size range. The method then applies that applied gravitational force (e.g., target gravitational force) that would lead to the desired PSD or target particle size range for the target set of crystallized small molecules. In some embodiments, the relationship between the PSDs and the applied gravitational force is the effects of different applied gravitational forces on particle size distribution (PSD) of the small molecules in the small molecule solution.

[0126] Alternatively, in some embodiments, the determined target gravitational force is different from the applied gravitational forces. For example, after extrapolating the relationship between PSDs and the different applied gravitational forces, the method determines that the target gravitational force to obtain a desired PSD or target particle size range is different from the applied gravitational force, e.g., such as 1.5g, 2.5g, 3.5g, 4.5g, less than lg, or more than 5g. Another nonlimiting example where the determined target gravitational force is different from the applied gravitational force includes methods where the target gravitational force is a gravitational force under microgravity conditions or on earth, but the applied gravitational forces were used to understand crystallization conditions in hypergravity conditions in order to inform the number of experiments or manufacturing capabilities needed to be performed in microgravity conditions or on earth. For example, in some embodiments, if the small molecules are manufactured in low earthorbit, the target gravity condition is less than lg, and as a result, the method provides an understanding of the effect microgravity will have on crystallization.

[0127] In some embodiments, the determined target gravitational force is less than 1g. In some embodiments, the determined gravitational force is 1g. In some embodiments, the determined gravitational force is 2g. In some embodiments, the determined gravitational force is 3g. In some embodiments, the determined gravitational force is 4g. In some embodiments, the determined gravitational force is 5g. In some embodiments, the determined gravitational force is 1g or more, 2g or more, 3g or more, 4g or more, or 5g or more. In some embodiments, the methods of the present disclosure can include at least a first gravitational force followed by a second gravitational force. In some embodiments, the method comprises one or more target gravitational forces. In some embodiments, the method comprises two or more, three or more, four or more, or five or more target gravitational forces (e.g., applied sequentially).

[0128] In some embodiments, the method comprises determining a time duration for applying the target gravitational force. In some embodiments, the target gravitational force is applied for 1 minute or more, 2 minutes or more, 3 minutes or more, 4 minutes or more, 5 minutes or more, 6 minutes or more, 7 minutes or more, 8 minutes or more, 9 minutes or more, 10 minutes or more, 15 minutes or more, 20 minutes or more, 25 minutes or more, 30 minutes or more, 35 minutes or more, 40 minutes or more, 45 minutes or more, 50 minutes or more, 55 minutes or more, 60 minutes or more, 65 minutes or more, 70 minutes or more, 75 minutes or more, 80 minutes or more, 85 minutes or more, 90 minutes or more, 95 minutes or more, 100 minutes or more, 105 minutes or more, 110 minutes or more, 115 minutes or more, or 120 minutes or more.

[0129] In some embodiments, the target gravitational force is applied for 1 minute, 2 minutes, 3 minutes, 4 minutes, 5 minutes, 6 minutes, 7 minutes, 8 minutes, 9 minutes, 10 minutes, 15 minutes, 20 minutes, 25 minutes, 30 minutes, 35 minutes, 40 minutes, 45 minutes, or 50 minutes.

[0130] In some embodiments, the target gravitational force is applied between 1-120 minutes, 1-50 minutes, 1-25 minutes, 1-20 minutes, 5-20 minutes, 5-50 minutes, 1-5 minutes, 5-10 minutes, 10-15 minutes, 15-20 minutes, 20-25 minutes, 25-30 minutes, 30-35 minutes, 35-40 minutes, 40-45 minutes, or 45-50 minutes.

[0131] In some embodiments, the method further comprises determining the correlation between the different applied gravitational forces and the crystal growth kinetics of the small molecules in the small molecule solution.

[0132] In some embodiments, the method further comprises determining the correlation between the different applied gravitational forces and the crystal nucleation (e.g., secondary nucleation) of the small molecules in the small molecule solution.

[0133] In some embodiments, the method further comprises determining the correlation between the different applied gravitational forces and the de-supersaturation rate and / or supersaturation rate of the small molecules in the small molecule solution.

[0134] In some embodiments, the method further comprises determining a correlation between the different applied gravitational forces and the polymorphism of the small molecules in the small molecule solution.

[0135] In some embodiments, the method further comprises determining a target gravitational force to apply to a target small molecule solution to produce the target crystallized small molecule based on the correlation between the different applied gravitational forces and the crystal growth kinetics of the small molecules in the small molecule solution.

[0136] In some embodiments, the method further comprises determining a target gravitational force to apply to a target small molecule solution to produce the target crystallized small molecule based on the correlation between the different applied gravitational forces and the crystal nucleation of the small molecules in the small molecule solution.

[0137] In some embodiments, the method further comprises determining a target gravitational force to apply to a target small molecule solution to produce the target crystallized small molecule based on the correlation between the different applied gravitational forces and the de-supersaturation rate of the small molecules in the small molecule solution.

[0138] In some embodiments, the method further comprises determining a target gravitational force to apply to a target small molecule solution to produce the target crystallized small molecule based on the correlation between the different applied gravitational forces and the solubility of the small molecules in the small molecule solution.

[0139] In some embodiments, the method further comprises determining a target gravitational force to apply to a target small molecule solution to produce the target crystallized small molecule based on the correlation between the different applied gravitational forces and the polymorphism of the small molecules in the small molecule solution.4.1.8. Producing the target set of crystallized small molecules

[0140] In one aspect of the present disclosure, the method comprises producing the target set of crystallized small molecules by applying the target gravitational force to the target small molecule solution during crystallization. In some embodiments, the method comprises producing the target set of crystallized small molecules by applying the target gravitational force at a target time duration to the small molecule solution during crystallization.

[0141] Another aspect of the present disclosure includes a small molecule produced by the methods described in the present disclosure. The small molecule product will have the desired PSD or particle size range depending on the desired parameters of the small molecule drug. Such parameters include route of administration, delivery mechanism and location in the body, bioavailability, biodistribution, dissolution rate, pharmacodynamics, and / or pharmacokinetics.5. EXAMPLES5.1. Example 1: The Effect of g-level Conditions on Product Particle Size Distributions

[0142] A large diameter centrifuge was constructed for hypergravity studies, which aimed to measure the effect of g-level on product particle size distributions. Seeded cooling crystallization experiments were designed and conducted on a Crystal 16 (Technobis Crystallization Systems) reactor at different g-levels.

[0143] This was followed by further targeted experiments in an EasyMax reaction calorimeter fitted with Process Analytical Technology (PAT) tools to understand the fundamental mechanisms underlying the sensitivity of the cooling crystallization to increasing g-levels.5.1.1. METHODS

[0144] Chemicals: L-histidine (>99% TLC), obtained from Sigma Aldrich (CAS Number 71-00- 1). Deionized water was obtained by using an in-house purification system. Reagent alcohol(anhydrous ethanol 90%, methanol 5%, 2-propanol 5% v / v) was obtained from Fischer Scientific as a solvent for particle size distribution (PSD) analysis for laser diffraction.

[0145] Seed preparation: L-histidine as received from the supplier was sieved using a stack of sieves (500 pm, 250 pm, 125 pm, 63 pm, pan). After sieving, the sieve fraction obtained from the 125 pm sieve was used as the seed crystals for the hypergravity experiments. The PSD of the seed crystals was analyzed using the laser diffraction method described in the section titled “Particle size distribution measurement by laser diffraction”.

[0146] Solubility: The solubility of L-histidine in water was measured using a known polythermal technique. Focused beam reflectance measurement (FBRM) is an online tool to measure the chord length distribution of particles in real-time during the crystallization process. A slurry of a known concentration was prepared in the EasyMax reactor with the FBRM probe inserted and equilibrated at constant temperature. The slurry was then heated at a constant rate until the crystals disappeared which was detected by a reduction in counts / s with the FBRM. The process was repeated at different heating rates (0.1, 0.25, 0.4, 0.55 °C / min) and the temperature of disappearance at each of the heating rates was recorded. The saturation temperature was calculated by extrapolating the temperature of disappearance data to a 0 °C / min heating rate. The entire process was repeated at different solids concentration to obtain the solubility temperature at different solids concentration. The solubility curve was then determined from regression analysis in Excel.

[0147] Metastable zone width (MSZW) using EasyMax reactor: Experiments to determine the metastable zone width were conducted in an EasyMax reactor (Mettler Toledo Inc) equipped with a pitched blade turbine impeller. The impeller was stirred at 200 rpm for all the experiments. The reactor was equipped with a Blaze900 Basic probe which was used to measure the evolution of the chord length distribution, perform real-time microscopy during cooling crystallization and determine the metastable zone width. The reactor was also equipped with a Raman probe which was used for measurement of concentration in real-time during the crystallization.

[0148] For the metastable zone width measurements, a saturated solution was cooled slowly until the High Dynamic Range (HDR) turbidity provided by the Blaze software showed an inflection indicating the formation of fine particles in solution. The solution was typically heated to 5 °C above its saturation temperature, and then cooled down to 10 °C at 0.05 °C / min. The process was repeatedat different API (active pharmaceutical ingredient) concentrations corresponding to different saturation temperatures.

[0149] Optical imaging: Optical images were taken using an optical stereo microscope (Model M205C from Leica Microsystems). The field of view of the microscope is in the range 1.44 mm - 29.5 mm, and the resolution is in the range 525 Ip / mm and 1050 Ip / mm. The zoom range of the microscope is 20.5:1 and a K3C CMOS camera from Leica is used to capture optical micrographs.

[0150] Particle size distribution measurement by laser diffraction: The PSD of the particles was measured using laser diffraction (LS 13320, Beckman Coulter). The instrument uses a 5mW laser diode with a wavelength of 750 nm as the main illumination source. Reagent alcohol (anhydrous ethanol 90%, methanol 5%, 2-propanol 5% v / v), in which L-histidine is insoluble, is used as the dispersant for the laser diffraction measurements. The raw data obtained from the laser diffraction instrument is the volume % vs particle size. However, this distribution is post processed to calculate the particle size distribution (vol% / pm) following equation (7).Nv(vol%) (7) fv(vol% / pm) =Ax

[0151] Seeded cooling crystallization experiments on EasyMax: A saturated solution of L- histidine in water at 40 °C was prepared by dissolving the requisite amount of L-histidine in water. The solution was heated to 60 °C to ensure complete dissolution, which was also confirmed by the counts / s measured by the Blaze probe. The solution was then cooled down to 32.8 °C as fast as possible to bring the system into the metastable zone, corresponding to a supersaturation of 1.2. The supersaturation ratio (SS) is calculated according to equation (6), where C (gsoiute / gsoivent) is the solute concentration and Csat (gsoiute / gsoivent) is the saturation concentration.Seed crystals (10% by weight of solute in solution) were then added to the reactor, and the solution was kept at 32.8 °C for 5 minutes for slurry equilibration. The solution was then cooled to 18.0 °C at 0.1 °C / min. After the experiment, the solution was filtered using a 0.2 pm Nylon filter (ColePalmer) followed by washing with a wash solvent (95% ethanol and 5% water). This was followed by drying the filtrate at 40 °C over 12 hours.

[0152] Centrifuge Experiments: Experiments were conducted on a centrifuge built in-house, with an arm length of 1.574 m. All cooling crystallization experiments were performed on a Crystall6 reactor (Technobis Crystallization Systems). The Crystall6 was mounted at the end of the arm on the centrifuge. The centrifuge rotates along its axis at defined rotations per minute (rpm). During the centrifuge rotation, the Crystal 16 reactor is housed in a basket with an adjustable tilt to align the vial axis with the apparent gravity vector, which is a function of the rpm, the radius of the arm and the Earth’s gravity vector. A picture of the setup is shown in FIG. 1.

[0153] The net acceleration in the centrifuge is given by equation (1) where g is the gravitational acceleration, (O is the rotation speed of the centrifuge and r is the radius of the centrifuge arm. The angle for the effective gravitational force is given by equation(2).

[0156] The rotation speed is changed to obtain a desired effective gravitational force. The rotation speed can be calculated from equation (1) and the angle of the effective gravitational force is given by equation(4). An overview of the net acceleration vector and the effective angle is shown in FIG. 1A. Since the solution is stirred during the process, the risks of vibrations influencing the process are very low and are not considered while analyzing the resulting PSD of the process.

[0159] Seeded cooling crystallization on centrifuge: Three of the Crystal 16 slots within a reactor were used for each experiment for obtaining repeats for each experiment. Cylindrical bottommagnetic stirrers with diameters of 0.03 in and length 0.27 in were used for the experiments unless otherwise mentioned. The Crystal 16 reactors were stirred at 800 rpm to match the tip speed with the experiments conducted in the EasyMax reactor. The volume of the API solution dispensed into each vial was 2 mL. A saturated API solution at 40°C was first prepared by dissolving the requisite quantity of L-histidine in water in a vial in the lab, followed by pipetting 2 mL of the saturated API solution into the HPLC vial and then inserting the magnetic stirrer into the bottom of the vial. The capped HPLC vials were then transferred to the slots of the Crystal 16 reactor for the crystallization experiments.

[0160] The solution was then heated in the Crystal 16 to 60.0 °C to ensure complete dissolution of the solids. The solution was stirred at 800 rpm using a magnetic stir bar kept at the bottom of the vial. The transmittance in the Crystal 16 vial was recorded to ensure all the particles were dissolved. Simultaneously the centrifuge was rotated at the desired rotations per minute (rpm) corresponding to the effective g-level desired in the experiment. The relationship between rpm and the effective g- level is provided in Table 1. At the end of the experiment, the vials were removed from the Crystal 16 reactor for particle size analysis. An overview of the entire process is shown in FIG. 2.

[0161] Table 1. Calculated g-levels and corresponding rpm.>

[0162] Dissolution experiments on centrifuge: An L-histidine solution saturated at 24 °C was prepared and pipetted into 2 mL vials for the Crystal 16 experiments. The vials were then inserted into the sample holders in the Crystal 16 reactor and were equilibrated at 24 °C. The solution was stirred at 800 rpm throughout the experiment. Seed crystals (10% w / w) were then poured into the vials, followed by slow heating from 24 °C to 28 °C at 0.1 °C / min over 40 minutes. Simultaneously the centrifuge was spun at the desired rpm corresponding to the g-level in Table 1. The centrifuge and the Crystal 16 apparatus are simultaneously stopped at 20 minutes to withdraw the first sample. This step is followed by restarting the centrifuge and the Crystal 16 reactor and running theexperiment for another 20 minutes. The PSDs of the samples withdrawn at 20 and 40 minutes during the dissolution experiment are analyzed using laser diffraction.5.1.2. RESULTS5.1.2.1 Experiments on EasyMax Reactor

[0163] The process window for the seeded cooling crystallization experiments in the EasyMax reactor are mapped with a top-down pitch blade turbine (PBT) impeller. First, the MSZW and solubility were tested, followed by development of a chemometric model to monitor concentration. Lastly, seeded de-supersaturation experiments were run to distinguish between primary nucleation, secondary nucleation, and breakage.Solubility and metastable zone width (MSZW):

[0164] Slurries of L-histidine were prepared at 8°C in water in the EasyMax reactor and stirred at 300 rpm. Five different slurries of different solids concentrations (0.035, 0.0415, 0.0479, 0.0545, 0.0601 g / g) were used for the experiments.

[0165] The disappearance temperature of the slurry was calculated using the polythermal method outlined under the “Solubility” section of the methods. FIGs. 18A-18E shows the disappearance temperatures of the slurries at different solids concentration for polymorph A of L-histidine during slow heating measured with FBRM probe. A 2nd order polynomial is used to calculate the solubility curve from the experimental measurements of the solids concentration vs disappearance temperature and is shown in FIG. 3A. The solubility calculated from these experiments agree well with known solubility calculations of L-histidine.

[0166] The MSZW is one of the critical process parameters that define the crystallization process. The metastable zone was identified by using a probe-based technique (Blaze probe). 100 mL of a saturated L-histidine solution was stirred in the EasyMax reactor (Mettler Toledo) at 200 rpm. Since the agitator used in the EasyMax has a larger diameter, the rpm needs to be lower to maintain similar tip speed compared to the bottom stirrer in the Crystal 16. The solution was heated to 10 °C above its saturation temperature to ensure complete dissolution. This was followed by slow cooling at 0.1 °C / min to 18 °C. The particle counts / s in the solution was monitored using the Blaze probe throughout the process. The point of nucleation is registered as a sudden increase in the counts / s, which corresponds to the primary nucleation threshold. It should be noted that while the laserdiffraction data in the results section titled “Effect of apparent g-level” shows a normalized volumetric distribution, this section discusses counts / s, and readers comparing the visual appearance of the plots should not expect a 1 : 1 relative peak height due to the nature of the axes being plotted. The process is repeated at different L-histidine concentrations to find the metastable zone over the entire range of temperatures. The increase in the High Dynamic Range (HDR) turbidity with cooling at the nucleation threshold (18.78 °C) for L-histidine at a concentration of 0.06 mg / mL is shown in FIG. 3A. The nucleation temperature is the temperature at which the HDR curve records an inflection. The metastable zone as a function of the temperature is plotted in FIG. 3B, along with the solubility curve in the phase diagram. There have not been previous reports in literature that have characterized the MSZW of L-histidine for cooling crystallization in water.Chemometric modeling:

[0167] To fully characterize the crystallization process, the evolution of the particle size and the solute concentration during the crystallization process was tracked. The Blaze probe provides information of how the chord length distribution (CLD) changes in the solution with time, which can provide information about fundamental mechanisms. The evolution of the system in the phase diagram was tracked to quantify whether the system crosses the metastable zone width during the operation.

[0168] The Raman spectra are a function of the solute concentration and temperature, so a chemometric model was built to calculate the solution concentration from the in-situ measurements of Raman spectra at variable temperatures.

[0169] The experimental data used for MSZW determination was used directly for building the chemometric model and is shown in FIG. 3B. The Raman spectra is shown in FIG. 16. A zoomed-in version of the spectra is shown in FIG. 4A with the major spectral peaks.

[0170] Both univariate and multivariate methods were considered for building a chemometric model. The time evolution of all the peaks was thoroughly analyzed, and a subset of peaks were chosen which are most sensitive to change around the time when nucleation occurs. FIG. 4B indicated that the peaks around 1286 cm-1, 1440 cm-1, 1160 cm-1, 990 cm-1 have a reduction in peak height around the time nucleation occurs (12 h), which were indicative of peaks which correspond to L-histidine molecules in solution. However, peaks around 1323 cm-1, 1088 cm-1showed an increase in peak height around 12 h, which is indicative of peaks corresponding to the solid phase. For monitoring the concentration of L-histidine in solution, the first subset of peaks was chosen for further analysis.

[0171] The Raman spectra were postprocessed using ICRaman software from Mettler Toledo before exporting to Matlab for building the chemometric model. SNV scatter correction and Pearson’s- like baseline correction, which adjusts spectra to a straight baseline, was applied to the spectra, followed by smoothing over a 10 cm'1window. A total of 1356 spectra were used for the analysis. No spectral outliers were detected during the model calibration. The chemometric model calibration pipeline was run for all the 4 peaks and the goodness of fit for the 4 peaks are shown in Table 2. The goodness of fit is lower for peaks with a Raman shift of 1160 cm'1and 990 cm'1and are hence not used for further analysis. The strategy employed in this work was to proceed with model calibration with only one peak in the chemometric model. The Raman peak with the highest R2in Table 2 (1286 cm'1) was chosen for further analysis. The chemometric model is shown in equation(5), where h is the peak height and T is the temperature. This model describes the peak height in the Raman spectra as a function of L-histidine concentration and solution temperature.

[0173] The model calibration is carried out using a custom-written script in Matlab R2023a. Linear regression is carried out in Matlab using the function mldivide44which solves a system of linear equations from which the coefficients in equation (5) are determined, followed by determination of goodness of fit for the model. The coefficients for the chemometric model calibration (for a Raman shift of 1286 cm'1) in equation (5) are shown in Table 3.

[0174] The time-course evolution of the peak height, temperature, L-histidine concentration are shown in FIG. 5. L-histidine concentrations are constant for each of the 5 experiments in FIG 5B since the region pre-nucleation has been analyzed. The peak height (FIG. 5A) increases with decrease of temperature (FIG. 5C) for each of the 5 experiments. This indicates the peak height is a function of the L-histidine concentration and the temperature, which adds validity to the functional form of the chemometric model in equation (5).

[0175] FIG. 20 shows the distribution of the predicted concentration for each of the data points used in the chemometric model. It can be seen that the distributions are very narrow around their meansand there is no overlap in the distributions. The coefficient of variation (standard deviation / mean) of the predicted concentrations for the 5 different L-histidine concentrations are 1.2%, 2.4%, 1.3%, 1.5%, 1.6% respectively. The average coefficient of variation (1.6%) is less than 5% for all experiments which illustrates that the chemometric model is accurately fitting the experimental data. However, if the COV of the predicted L-histidine concentration was not less than 5%, PLSR models could have been used alternatively.

[0176] Table !. Goodness of fit for chemometric model calibration for different peaks in the Raman spectra.Raman shift (cm-1) R2(Calibration)990 0.9142

[0177] Table 3. Coefficients from chemometric model in equation 5 (Raman shift of 1286 cm'1).Coefficient Value ai 0.0267 a2 0 a3-0.0013 a40.0032 as 0.0320

[0178] To validate the goodness of fit, a validation experiment was run with an L-histidine concentration of 54.5 mg / mL L-histidine following a similar experimental protocol. The saturated L-histidine solution was cooled until nucleation and the experimental data before nucleation were chosen for model validation. The coefficients evaluated in model (5) were used to calculate the predicted L-histidine concentration. FIG. 6A shows the time-course evolution of the temperature and the peak height from the L-histidine spectra before nucleation occurs. The peak height decreases with increase of temperature, similarly to the calibration experiments (FIGs. 5A-5C).

[0179] FIG. 6B compares the evolution of the predicted L-histidine concentration with the actual L- histidine concentration during the validation experiment. The mean of the predicted concentration value for the validation experiment was 54.7 mg / mL and the adjusted R2was 0.897. The coefficient of variation for the model predictions in FIG. 6B is 2.54%.Seeded de-supersaturation experiments (Easy Max):

[0180] Seeded de-supersaturation experiments were run in Mettler Toledo EasyMax 102 reactor.100 mL of L-histidine solution saturated at 40.0 °C was heated to 60.0 °C to ensure complete dissolution. The solution was stirred at 200 rpm throughout the process. The solution was then cooled to 32.8 °C to bring the system into the metastable zone for seeding. This corresponded to a supersaturation ratio of 1.2 at seed addition.

[0181] Seeds crystals (10% w / w of total mass of solute) were added to the solution and stirred for 5 minutes to form a spatially homogeneous seed bed. This is followed by implementing the temperature profile, which causes de-supersaturation of the L-histidine solution.

[0182] The Raman spectra and temperature profiles collected during the experiment were postprocessed using Matlab. The chemometric model built in Section titled “chemometric modeling” is then applied to the Raman spectra to estimate the L-histidine concentration. The evolution of the L- histidine concentration, along with the solution temperature is shown in FIG.7A. The L-histidine concentration reduces quickly until 1.1 hours, followed by a slower decrease in concentration until 2.5 hours. The crystal growth rate coefficient of many APIs follows Arrhenius dependence with temperature, leading to faster kinetics with increasing temperature. As a result, cooling crystallization can exhibit high initial growth rates and de-supersaturation rates which decrease over time as the reactor temperature is lowered. The evolution of the counts / s recorded in the Blaze probe is shown in FIG. 7B. The y-axis of FIG. 7B is plotted on a logarithmic scale to exemplify the differences in the evolution of counts / s for the 3 different size ranges (1-10, 50-120, 140-280 microns). The fine particle counts (1-10 pm) increase rapidly during the crystallization, while the coarse particle counts (140-280 pm) remain relatively stable. The stability in particle counts for large particles corresponding to the seed crystals (140-280 pm) suggests breakage is unlikely to be playing a role. The evolution of the system in the concentration phase diagram is shown in FIG. 7C. The system stays within the metastable zone which rules out primary nucleation and points to secondary nucleation and growth as the dominant mechanisms during the de-supersaturationexperiment. The pH of the solution was the same before and after the experiment (Table 4), which shows L-histidine did not undergo deprotonation. pH of solution before and after crystallization

[0183] The pH of the solution was measured after complete dissolution using a Mettler Toledo pH routine pH electrode. The pH of the solution was also measured after the crystallization experiment and is shown in Table . There is not much difference in the pH of the solution before and after the crystallization experiment indicating that L-histidine did not undergo any deprotonation in solution during the crystallization experiment.Table 4. pH of the solution before and after the seeded crystallization experiment Status pHBefore crystallization experiment 7.884After crystallization experiment 7.882

[0184] The chord length distribution data from the Blaze probe was post-processed with a Matlab script. The recorded length-weighted and cube-weighted chord length distributions were extracted at defined time points and plotted with the Matlab script and are shown in FIG. 8G-8H. Images from the Blaze probe were also analyzed to elucidate the dominant mechanism during de-supersaturation. At 0.57 hours (FIG. 8B), secondary nuclei on the surface of the crystals can be seen in the image obtained from the Blaze probe. The images from 0.87-2.5 hours (FIGs. 8C-8F) also indicate secondary nuclei are present and do not indicate growth of the needles along the lateral direction. The evolution of the length-weighted chord length distribution (LWCLD) in FIG. 8G also shows that the rate of increase of the counts / s is much slower during 2.18-2.5 hours compared to the earlier parts of the experiment. This indicates slower de-supersaturation during the last part of the experiment, which is corroborated with the slower concentration decrease in FIG. 8 from 2.2-2.5 hours. As exhibited in FIG. 8G, seed crystals are still present in the final LWCLD, but they are overshadowed by the high proportion of smaller crystals in the product samples.

[0185] FIG. 8H also shows the evolution of the cube-weighted chord length distribution (CWCLD) throughout the experiment. The CWCLD in the Blaze probe biases the chord length distribution towards coarser particles. It is to be noted that the CWCLD distribution is equivalent to the square- weighted chord length distribution (SWCLD) in the FBRM. Thus, the evolution of the CWCLD in the process is a representation of the evolution of the coarser particles in the process. If breakage had been the only dominating mechanism, the larger particles would have broken down into smaller particles. Breakage is characterized by a shift of the mode of the CWCLD to the left with time, which was not observed in FIG. 8G. In contrast, during secondary nucleation, the generation of many fine particles increases the total counts / s in the LWCLD with time. The evolution of the LWCLD in the process (FIG. 8G) is most representative of secondary nucleation, which confirms that secondary nucleation is the dominant mechanism in the EasyMax experiments. In addition, the final mass of the filtered crystals (1.45 g) is much higher than the seed crystal mass (0.6 g), which cannot be attributed to breakage alone and can only be caused by secondary nucleation and / or growth. FIG. 8 is typical of a crystallization process dominated by secondary nucleation and growth where both processes occur with similar driving force, e.g. the supersaturation consumption due to nucleation and growth is comparable.

[0186] To examine the evolution of the seed particles during the experiment, the average circularity of the seed and product crystals was calculated from dynamic image analysis and indicated little change in the shape of the seed particles (FIGs. 17A-17B).Dynamic Image Analysis (DIA)

[0187] Dynamic image analysis was carried out on the seed crystals and the product crystals from the EasyMax experiments. A dry version of the BeVision D2 dynamic image analyzer was used for the experiments. 1 gram of the crystals were taken with a spatula and poured into the funnel of the DIA instrument and the vibratory feeder was simultaneously started. The circularity and the L / D ratios of the particles were analyzed directly with the image analysis software associated with the instrument.

[0188] Dynamic Image Analysis results: FIGs. 17A-17B shows the cumulative distributions of the circularity and the L / D ratios of the seed population, and the product obtained from the cooling crystallization experiment. The average values of the circularity and the L / D ratio are shown in Table 5, which shows the difference in the average L / D ratio is < 10% (9.13%).

[0189] Table 5. Average values of the circularity and L / D ratios of the seed and product crystals.Average Seed ProductAverage 0.82 0.81 circularity Average 1.99 2.19L / D ratioUnseeded nucleation experiments (EasyMax)

[0190] Unseeded cooling crystallization experiments were also run following the same protocol as the seeded cooling crystallization experiments but without adding seeds. The total counts / s from the FBRM probe and the temperature profile are shown in FIG. 15. No increase is observed in the counts / s from the FBRM probe after the seeding temperature (33 °C), which indicates that no primary nucleation occurs during the unseeded crystallization experiments. This further indicates that the changes in the LWCLD and concentration seen in FIG. 8 was due to secondary nucleation as the dominant mechanism.5.1.2.2 Experiments on Crystall6Effect of apparent g-level

[0191] The L-histidine experiment developed on the EasyMax at the 2mL scale was run on the Crystal 16 crystallizer.

[0192] Laser diffraction measurements were used to measure the particle size distribution of the obtained crystallization products at a range of g-levels and the observed peaks were interpreted. Additional experiments were run to rule out breakage, primary nucleation, and other effects.

[0193] FIG. 9 shows the evolution of the volume-based PSD with changes in g-levels at a cooling rate of 0.1 °C / min. All the samples, except the seed crystals, are those obtained at the end of the de- supersaturation experiment. The apparent g-level has a marked effect on the PSD of the samples. There are three distinct peaks to consider across the seed and product particle size distributions - one around 0.7 pm, a second peak around 20-30 pm, and a third peak at 180 pm. It should be notedthat FIG. 9 is a volume based distribution of the filtered samples and cannot be directly compared to FIG. 8G which is a number based distribution of samples measured in-situ.

[0194] The peak at 0.7 pm is present in the seed PSD, as well as the product samples. The appearance of the 0.7 pm peak could be attributed to the presence of fine particles which were either (a) stuck on the screen during the sieving process along with the larger seed crystals, possibly due to screen blinding or (b) were attached to the seed particles during sieving and got detached when seed crystals were dispersed into the solution.

[0195] Alternatively, the 0.7 pm peak may correspond to the shortest dimension of crystals present in the sample population. However, when the un-normalized volume fraction is plotted vs the particle size (FIG. 13), the total volume fraction of the fines is low (less than or equal to 1.8%) for each seed and product sample. Since the total volume fraction of the peak around 0.7 pm does not change significantly between the seeds and the product (FIG. 13), it is unlikely to be a result of crystal growth or attrition during the experiment.

[0196] Though the peak around 180 pm is not evident in the normalized volume-based product PSD, particles in the size range of 180 pm are expected to be present as confirmed with optical imaging, and previously in the section titled “Seeded de-supersaturation experiments (EasyMax)”. The volume fraction of the seed crystals is overshadowed by the volume fraction of the product crystals in the normalized distributions plotted in FIG. 9. For further clarification, the volume fraction distribution measurements from the laser diffraction instrument are shown in FIG. 13.

[0197] The mass percentage of the seed crystals is only 10% of the total mass of the L-histidine in the solution (14% of the total mass of final product mass considering theoretical yield). Thus, most of the crystals at the end of the crystallization are expected to be formed during de-supersaturation. The appearance of the peak at 20 pm is attributed to secondary nucleation, which is explained in detail in the section titled “seeded de-supersaturation experiments (EasyMax)”.

[0198] As the g-force on the reactor increases, the mode of the PSD increases from 19 pm at 1g to 30 pm at 3g. However, when the g-force is further increased, the mode of the distribution decreases to 22 pm at 4g and 21 pm at 5g. The change in the mode of the distribution at different g-levels is shown in Table 6. Table 6 also shows that the standard deviation of the PSD increases from 8.5 pm at 1g to 11 pm at 3g, before reducing to 10.5 pm at 5g. This indicates that with increasing g-levelthe particle size distribution initially broadens, but the effect is non-monotonic. This shows that by crystallizing molecules in different gravity environments, it may be possible to achieve shifts in the PSD, with the purpose of improving drug product performance.

[0199] Table 6. Mode and standard deviations of the PSDs at different g-levels g-level Mode of distribution (pm) Standard deviation of distribution (pm)

[0200] FIGs. 10A-10D shows optical micrographs of samples grown at 3 different g-levels (1g, 3g and 5g). The optical micrographs confirm the broad features visible in the laser diffraction data across the sample e.g. peaks at 20-30 pm, and 180 pm. In addition, the appearance of the seed crystals is markedly different from the product particles generated at different g-levels. Optical micrographs are not as quantitative as laser diffraction data since a very small fraction of particles is sampled by the field of view of an optical image. Because of the limitations in quantifying PSD using static image analysis, the PSD is quantified exclusively using laser diffraction data.

[0201] The hypergravity experiments at 1g, 3 g and 5g were independently repeated on different days to assess repeatability of the process.

[0202] The results are shown in FIG. 10E and indicate that the measured PSD is consistent at several g-levels. The coefficient of variation of the modes of the PSDS are 4.5%, 7.6% and 7.4% respectively for lg, 3g, and 5g. Statistical analysis was also carried out on the modes of the PSD at each g-level using SAS JMP 17.0. The p-value for the pairwise difference of means of the modes across different g-levels is less than 0.05. Thus the null-hypothesis that the modes are equivalent across g-levels can be rejected with >95% confidence. This indicates that a statistically significant shift in mode with g-level was observed.

[0203] Additional characterization of the seed and product crystals was performed to assess the potential for polymorphic changes, and unexpected changes in the seed crystal particle size due to breakage. The XRPD diffractograms of the seed and the product samples, together with reference diffraction patterns are shown in FIG. 14. The diffractograms of both the seed and the product crystals are consistent with Form A, and Form B was not detected, indicating polymorphic changes did not influence the observed particle size distribution.X-ray powder diffraction (XRPD) of seeds and product samples

[0204] The crystal structure of the samples was analyzed using a Rigaku Miniflex 6G x-ray powder diffraction instrument. The instrument uses a Cu Ka radiation source (40kV, 15 mA, X=0.15406 nm). Data was recorded in the range 20 = 3- 60° with a scanning rate of 27min with a step width of 0.02°. Diffraction patterns were collected and analyzed using SmartLab studio II provided by Rigaku. XRPD diffractogram of the seed and product samples are shown in FIG. 14. PDF 02-063- 2351 is used for the reference for Form A, and PDF 02-100-850 is used as the reference for form Bl.Investigation of breakage of seed crystals in Crystal 16 experiments.Breakage experiments

[0205] Breakage experiments were conducted with L-histidine seed crystals dispersed in ethanol. As shown in FIGs. 19A-19B, the shift of the PSD due to mechanical forces is small compared to the shift observed in the crystallization experiments during the centrifuge studies. Thus, breakage can be ruled out as a major contributing factor to the shift in the observed PSD for the Crystall 6 centrifuge experiments. Thus, the fundamental mechanistic understanding of the crystallization phenomena observed in EasyMax experiments is still transferable to analyzing the experiments conducted in Crystal 16.

[0206] Seed crystals (0.125 pm) were prepared, as described in the methods section under “seed preparation”, and the PSD of the seed crystals was measured using laser diffraction. To examine if breakage occurred, 0.1 grams of the seeds were dispersed in 2mL of ethanol in an HPLC vial, together with a magnetic stirrer. The vial was placed in a Crystal 16 mounted on a centrifuge. The temperature of the vial was maintained at 20 °C and the solution was stirred at 800 rpm for 2h 20mins to match the duration of the crystallization experiments run on the Crystal 16. The vials were then removed, and the PSD was measured by laser diffraction to examine the modes of the PSD.

[0207] The PSD of the seed crystals and the product crystals are shown in FIG. 19A. For comparison, the PSD of the product crystals from the crystallization experiments run at 3g are also shown in FIG. 19A. It is evident that some breakage and / or attrition occurs as shown by the slight shift in the mode of the PSD after breakage experiments compared to the seed crystals. However, the PSD of the product after crystallization is very different from the PSD after the breakage experiments. The modes of the distributions are shown in Table 7. The mode of the PSD after breakage experiment at 3g is 110 gm and the mode of the PSD after crystallization experiment at 3g is 30.1 m. The difference in PSD after crystallization experiments cannot be attributed to breakage alone.

[0208] The mode of the distribution for the breakage experiment with bottom stirrer reduces to 101 m at 1g, followed by an increase to 110 gm and 121 gm at 3g and 5g respectively. At 1g, most of the particles are suspended, so there are higher number of particle-wall contacts leading to a higher reduction in the mode of the PSD. However, at 3g and 5g, most of the particles are concentrated towards the bottom of the vials. This leads to less particle-wall contacts thereby leading to a reduced change in the PSD of the seeds due to mechanical effects.

[0209] Next, the effect of the stirrer type and location was investigated on the breakage experiments. Seeded crystallization experiments were run with the same protocol in Crystal 16, but with an overhead stirrer (Hastelloy hook). The PSD of the crystallization experiments with the overhead stirrer along with the seed PSD is shown in FIG. 19B. Due to the significant change in PSD at 1g with the overhead stirrer, the bottom stirrer is preferred for the crystallization study.Table 7. Mode of distribution for breakage experiments for bottom stirrer and overhead stirrerType Bottom stirrer Overhead stirrerSeed 161 161 lg breakage 101 39.83g breakage 110 1335g breakage 121 1333 g Crystallization 30.1 21.2Mechanistic hypothesis of effect of gravity on seeded cooling crystallization of L-histidine

[0210] The EasyMax experiments in the section titled “Seeded de-supersaturation experiments (EasyMax)” revealed that the crystallization process was dominated by secondary nucleation, which is also the most plausible mechanism for Crystal 16 experiments.[2H] Since the EasyMax experiments confirmed that the system lies within the metastable zone during the entire duration of the experiment, we do not need to consider primary nucleation as a possible mechanism. The hypothesis presented here considers secondary nucleation and crystal growth as the dominant mechanisms during the de-supersaturation. As crystals grow, they consume solute molecules, changing the local L-histidine concentration in the solution. Because transport is not instantaneous, spatial inhomogeneities arise, with non-uniformities in concentration across the reactor driven by density differences. With increasing g-level, the lower and higher density regions undergo stratification resulting in a spatial gradient in supersaturation across the container. This leads to the formation of a concentration boundary that moves to the bottom of the vial with time. This results in a spatial gradient of secondary nucleation and crystal growth rates, with higher rates of nucleation and growth at higher supersaturation. At higher g-levels, the supersaturation is higher at the bottom of the solution and lower at the top of the solution (FIG. 11A-11C). Increasing the g- level to 3g increases the size of the supersaturation gradient, forming regions of increased supersaturation at the bottom of the vial, and regions of lower supersaturation at the top of the vial. In the higher supersaturation regions, both secondary nucleation rates and growth rates are increased, allowing the crystals generated to reach larger sizes, shifting the PSD to 30 pm.

[0212] Surprisingly, the trend reverses while further increasing g-level from 3g to 5g. This may be attributed to the spatial distribution of the supersaturation gradient. While at an apparent g-level of 5 g the peak super saturation may be expected to be higher than at 3 g, the supersaturated region is more tightly confined to a limited volume of the container. In addition, particles tend to concentrate towards the bottom of the vial due to the high g-forces. Since at higher supersaturations, secondary nucleation supersedes crystal growth, the mode of the final PSD is lower at 5g compared to 3g.

[0213] The ultimate balance of the change of the nucleation and growth rate is a complex function of the density of the solvent, the apparent g-level, the PSD of the seed crystals and the secondary nucleation and growth kinetics. This needs to be considered on a case-by-case basis since it is highly dependent on the API and the solvent, and no generalized rule can be proposed for a new molecule and solvent system.Dissolution experiments

[0214] Seeded dissolution experiments were performed to further validate the hypothesis that spatial concentration gradients drive changes in the observed particle size distribution. If concentration is homogeneous throughout the vial, changes in the PSD during dissolution would be expected to be independent of g-level. However, if a spatial variation in concentration forms differently at different g-levels, then shifts in the PSD during dissolution will be different at different g-levels, FIG. 11C illustrates the case where the concentration is homogeneous spatially across the vial. The rate of crystal dissolution is given by equation (8), where kd , d are dissolution constants, Csat is the saturation concentration and C is the concentration of the L-histidine in solution. If the solution concentration is spatially uniform, the rate of dissolution will be uniform throughout the solution.Rd= kdcsat- c (8)

[0215] However, if the concentration is not spatially homogeneous throughout the solution, the rate of dissolution will also have a spatial gradient. FIG. 1 ID shows three operating points in the solution distributed spatially - one near the top of the vial, one at the middle of the vial and one at the bottom. The solution concentration has a spatial gradient, with lowest concentration at the top and highest concentration at the bottom. Since the rate of dissolution is dependent on the difference between the solution concentration and the solubility limit of L-histidine, particles are expected to experience the fastest rate of dissolution at the top of the vial and the slowest rate of dissolution at the bottom of the vial. Thus, a spatial gradient in the supersaturation can lead to different particle sizes at the end of the experiment.

[0216] FIG. 12 shows the volume-based particle size distributions for the dissolution experiments described in Section titled “Dissolution experiments on centrifuge”. The modes of the distributions are provided in Table for readability. After 20 minutes of dissolution, the mode of the volumebased PSD shifts to the left from 185 pm to 47 pm, 44 pm, and 39 pm at lg, 3g, and 5g,respectively as shown in FIG. 12A. The trends with increasing g-level indicate that the rate of dissolution increases with increasing centrifugal forces, with the enhancement explained by the formation of low concentration regions that promote faster dissolution. At 40 minutes of dissolution (FIG. 12B), the modes are 40 pm, 22.5 pm, and 39 pm at 1g, 3g, and 5g, respectively. While the left-shift of the first two modes are consistent with the notion that increasing g-levels increases spatial gradients and promotes enhanced dissolution, the sample that undergoes dissolution at 5g undergoes little shift from 20-40 minutes, which is attributed to sedimentation effects brining the crystals in equilibrium with localized high concentration region at the bottom of the vial.

[0217] The observed modes of the volume-based PSD at 40 minutes as a function of g-level (Table 8), exhibit the reverse of the trend observed in Table 6 during the seeded de-supersaturation experiments. The observation of non-monotonic behavior across both crystallization and dissolution studies helps support the hypothesis that observed changes in particle size distributions are driven by concentration gradients, and that the behavior is a result of multiple competing dynamics.

[0218] Table 8. Observed modes during dissolution studies.Condition Prior to dissolution 20 minutes 40 minutes(jim) (jim) (jim)

[0219] It is to be noted that although the hypothesis of supersaturation-mediated changes in crystallization under hypergravity has been proposed in literature before, there was limited experimental supporting evidence available. This work confirms the hypothesis through a combination of (a) quantitative measurements of particle size distributions with laser diffraction (b) crystallization experiments with in-situ PAT tools for enhanced process understanding, and (c) dissolution experiments at variable g-levels to further probe supersaturation-mediated changes in crystallization outcomes in hypergravity.

[0220] In addition, while previous hypergravity studies primarily focused on protein molecules at g-levels often in the 1,000-g to 10,000-g range, and often found no effects of gravity at less than100-g, this work shows that at the 2 mL scale small molecule crystallizations have different outcomes depending on the effective gravity level. This study revealed that gravity-driven supersaturation gradients can arise even while stirring at 800 rpm at the 2 mL scale, which impacts a wide range of studies on typical screening platforms such as the Crystal 16.

[0221] It is to be noted that the transferability of crystallization mechanisms across the Crystal 16 reactor and EasyMax is case-dependent i.e. it depends on the mechanical properties of the product crystals and the sensitivity of the crystallization mechanisms to the reactor hydrodynamics. Based on the experiments performed, we can conclude that secondary nucleation is the primary mechanism for L-histidine cooling crystallization that leads to the observed PSDs for both the EasyMax and Crystal 16 experiments.5.1.3. CONCLUSIONS

[0222] Seeded cooling crystallization of L-histidine was performed at different g-levels using a centrifuge. The measured shifts in the PSD were non-monotonic with increasing g-levels.Experiments were run in the EasyMax reaction calorimeter with in-situ PAT tools to quantify the metastable zone width and the sensitivity of the Raman spectra to changing L-histidine concentration and temperature. A chemometric model was built to calculate the L-histidine concentration as a function of the instantaneous Raman spectra and the solution temperature. In-situ PAT data from the Blaze probe revealed secondary nucleation to be the dominant mechanism of crystallization. The evolution of the system in the phase diagram showed rapid de-supersaturation initially, followed by slower de-supersaturation at the end due to the Arrhenius nature of the growth kinetics. A hypothesis was developed which explains changes in the product PSD as a function of g- level through the formation of spatial variations in concentration sand supersaturation. Dissolution experiments were designed to support the hypothesis that concentration gradients drive changes in crystallization outcomes. The possibility of particle breakage to significantly influence the measured PSD was ruled out by experiments discussed herein. The hypergravity experiments also show that gravity affects the crystallization process even when the solution is stirred at a high rpm, highlighting that gravity likely plays a significant role in many small molecule crystallization processes. Since particle size has a significant impact on dissolution rate and oral bioavailability, as has been well documented in literature, accounting for the effect of gravity may be especially important for APIs whose crystallization behavior is sensitive to spatial concentration gradients. Variable gravity platforms are well-positioned to rapidly generate datasets that show how gravityimpacts the crystallization of small molecule pharmaceuticals, paving the way to achieve better control over crystallization and highlighting opportunities for improved process control in microgravity.6. EQUIVALENTS AND INCORPORATION BY REFERENCE

[0223] While the invention has been particularly shown and described with reference to a preferred embodiment and various alternate embodiments, it will be understood by persons skilled in the relevant art that various changes in form and details can be made therein without departing from the spirit and scope of the invention.

[0224] All references, issued patents and patent applications cited within the body of the instant specification are hereby incorporated by reference in their entirety, for all purposes.

Claims

WHAT IS CLAIMED IS:

1. A method comprising : applying, for each of a plurality of crystallization operations, a different gravitational force greater than or equal to lg to a small molecule solution during crystallization to produce a corresponding set of crystallized small molecules; measuring, for each set of crystallized small molecules, a particle size distribution (PSD) within the set of crystallized small molecules; selecting a desired PSD for a target set of crystallized small molecules; determining a target gravitational force to apply to a target small molecule solution to produce the target set of crystallized small molecules based on the measured PSD within each set of crystallized small molecules and the corresponding gravitational forces applied during the plurality of crystallization operations; and producing the target set of crystallized small molecules by applying the target gravitational force to the target small molecule solution during crystallization.

2. The method of claim 1, wherein at least one crystallization operation comprises spinning a centrifuge at a desired rpm corresponding to the gravitational force.

3. The method of any one of claims 1-2, wherein the plurality of crystallization operations is run on a crystallizer.

4. The method of claim 3, wherein at least one crystallization operation comprises stirring the small molecule in solution at one or more desired rotations per minute (rpm) in the crystallizer.

5. The method of any one of claims 1-4, wherein at least one crystallization operation comprises stirring the small molecule in solution at one or more desired temperatures.

6. The method of any one of claims 1-5, wherein the crystallization operation comprises performing cooling crystallization experiments.

7. The method of claim 6, wherein the method further comprises determining a cooling rate that generates a supersaturation required to crystallize the small molecule solution.

8. The method of any one of claims 1-7, wherein the method further comprises, before, during, and / or crystallization of the small molecule, taking optical images of the small molecule.

9. The method of any one of claims 1-8, wherein the crystallization operation further comprises, during crystallization, one or more of: measuring a metastable zone width (MSZW) of the small molecule in solution; measuring solute concentration of the small molecule solution at one or more temperatures; determining a nucleation rate of the small molecule in solution; measuring a crystallization growth rate of the small molecule solution; determining an induction time of the small molecule solution; calculating a supersaturation ratio of the small molecule solution; and measuring chord length distribution (CLD) of particles in the small molecule solution.

10. The method of claim 9, wherein the CLD is length-weighted CLD and / or cube-weighted CLD.

11. The method of claim 9, wherein the solute concentration of the small molecule solution during crystallization is measured using Raman spectroscopy.

12. The method of claim any one of claims 1-11, wherein the PSD within the set of the crystallized small molecules is measured using a laser diffraction instrument.

13. The method of claim 12, wherein the PSD is calculated from raw data received from the laser diffraction instrument using a volume fraction (volume % of particles per pm) of each set of crystallized small molecules, the volume % distribution, and the particle size.

14. The method of any one of claims 1-13, wherein the different gravitational forces are selected from: 1g or more, 1.5 g or more, 2g or more, 2.5 g or more, 3g or more, 3.5g or more, 4g or more, 4.5g or more, and 5g or more.

15. The method of claim 14, wherein each of the different gravitational forces are 1g, 2g, 3g, 4g, and 5g.

16. The method of claim 14 or 15, wherein the rpms for a gravitational force of: 2g ranges from 30-35 rpm;3g ranges from 40-45 rpm;4g ranges from 45-50 rpm; and / or 5g ranges from 50-55 rpm.

17. The method of claim 16, wherein the rpm for a gravitational force of:2g is 31 rpm;3g is 40 rpm;4g is 47 rpm; and 5g is 53 rpm.

18. The method of any one of claims 1-17, wherein selecting the desired PSD for the target set of crystalized small molecules comprises selecting a target particle size range of the target set of crystallized small molecules.

19. The method of claim 18, wherein the method further comprises determining the target gravitational force that produces a PSD that maximizes the number of particles within the selected particle size range.

20. The method of claim 19, wherein the desired PSD ranges from 1 pm to 2000 pm.

21. The method of claim 20, wherein the desired PSD ranges from 5-110 pm.

22. The method of claim 20, wherein the desired PSD ranges from 1-5 pm, 5-10 pm, 10-15 jim, 15-20 pm, 20-25 pm, 25-30 pm, 40-50 pm, 50-60 pm, 60-70 pm, 70-80 m, 80-100 m, 100-150 pm, or 150-200.

23. The method of any one of claims 1-22, wherein determining the target gravitational force to apply to the target small molecule solution comprises extrapolating a relationship between PSDs and the applied gravitational forces.

24. The method of claim 23, wherein the relationship between the PSDs and the applied gravitational force is the effects of different applied gravitational forces on particle size distribution (PSD) of the small molecules in the small molecule solution.

25. The method of any one of claims 1-24, wherein the determined target gravitational force is different from the applied gravitational forces.

26. The method of any one of claims 1-25, wherein the determined target gravitational force is less than 1g.

27. The method of any one of claims 1-26, wherein the method further comprises determining a target time duration for which the target gravitational force is applied.

28. The method of any one of claims 1-26, wherein the method further comprises, before applying the different gravitational forces to the small molecule solution during crystallization, determining a cooling rate (°C / min) of the small molecule solution to promote desired crystallization nucleation and growth rates.

29. The method of any one of claims 1-28, wherein the method further comprises, before applying the different gravitational forces to the small molecule solution during crystallization, determining the solubility of the small molecule solution.

30. The method of any one of claims 1-29, wherein the method further comprises determining the correlation between the different applied gravitational forces and the crystal growth kinetics of the small molecule in the small molecule solution.

31. The method of any one of claims 1-30, wherein the method further comprises determining the correlation between the different applied gravitational forces and the crystal nucleation (e.g., secondary nucleation) of the small molecule solution.

32. The method of any one of claims 1-31, wherein the method further comprises determining the correlation between the different applied gravitational forces and the de- supersaturation rate and / or supersaturation rate of the small molecule solution.

33. The method of any one of claims 1-32, wherein the method further comprises determining a correlation between the different applied gravitational forces and the polymorphism of the small molecules in the small molecule solution.

34. The method of any one of claims 29-33, wherein the method further comprises determining a target gravitational force to apply to a target small molecule solution to produce the target crystallized small molecule based on at least one of: the correlation between the different applied gravitational forces and the crystal growth kinetics of the small molecules in the small molecule solution; the correlation between the different applied gravitational forces and the crystal nucleation of the small molecules in the small molecule solution; the correlation between the different applied gravitational forces and the de- supersaturation rate of the small molecules in the small molecule solution; the correlation between the different applied gravitational forces and the solubility of the small molecules in the small molecule solution; and the correlation between the different applied gravitational forces and the polymorphism of the small molecules in the small molecule solution.

35. The method of any one of claims 1-34, wherein the particles within each set of crystalized small molecules ranges from lxlO1crystals to IxlO9crystals.

36. The method of claim 35, wherein the particles within each set of crystalized small molecules ranges from 1 x 103to 1 x 106crystals or 1 x 103to 1 xlO9crystals.

37. A small molecule produced by the method of any one of claims 1-36.

Images