A modified mechanistic approach for predicting ribbon solid fraction at different roller compaction speeds.

This research investigates the modeling of the pharmaceutical roller compaction process, focusing on the application of the Johanson model and the impact of varying roll speeds from 1 to 15 RPM on predictive accuracy of ribbon solid fraction. The classical Johanson's model was integrated with a dwell time parameter leading to an expression of a floating correction factor as a function of roll speed. Through systematic analysis of the effect of different roll speeds on the solid fraction of ribbons composed of microcrystalline cellulose, lactose, and their blends, corrective adjustment to the Johanson model was found to depend on both roll speed and formulation composition. Interestingly, the correction factor demonstrated excellent correlation with the blend's mechanical properties, namely yield stress (Py) and elastic modulus (E0), representative of the deformability of the powder. Validated by a multicomponent drug formulation with ±0.4-1.3 % differences, the findings underscore the utility of this modified mechanistic approach for precise prediction of ribbon solid fraction when Py or E0 is known for a given blend. Hence, this work advances the field by offering early insights for more accurate and controllable roller compaction operations during late-stage pharmaceutical manufacturing.

Machine learning modeling for ultrasonic quality attribute assessment of pharmaceutical tablets for continuous manufacturing and real-time release testing.

In in-process quality monitoring for Continuous Manufacturing (CM) and Critical Quality Attributes (CQA) assessment for Real-time Release (RTR) testing, ultrasonic characterization is a critical technology for its direct, non-invasive, rapid, and cost-effective nature. In quality evaluation with ultrasound, relating a pharmaceutical tablet's ultrasonic response to its defect state and quality parameters is essential. However, ultrasonic CQA characterization requires a robust mathematical model, which cannot be obtained with traditional first principles-based modeling approaches. Machine Learning (ML) using experimental data is emerging as a critical analytical tool for overcoming such modeling challenges. In this work, a novel Deep Neural Network-based ML-driven Non-Destructive Evaluation (ML-NDE) modeling framework is developed, and its effectiveness for extracting and predicting three CQAs, namely defect states, compression force levels, and amounts of disintegrant, is demonstrated. Using a robotic tablet handling experimental rig, each attribute's distinct waveform dataset was acquired and utilized for training, validating, and testing the respective ML models. This study details an advanced algorithmic quality assessment framework for pharmaceutical CM in which automated RTR testing is expected to be critical in developing cost-effective in-process real-time monitoring systems. The presented ML-NDE approach has demonstrated its effectiveness through evaluations with separate (unused) test datasets.

Interplay of Drug-Polymer Interactions and Release Performance for HPMCAS-Based Amorphous Solid Dispersions.

The interplay between drug and polymer chemistry and its impact on drug release from an amorphous solid dispersion (ASD) is a relatively underexplored area. Herein, the release rates of several drugs of diverse chemistry from hydroxypropyl methylcellulose acetate succinate (HPMCAS)-based ASDs were explored using surface area normalized dissolution. The tendency of the drug to form an insoluble complex with HPMCAS was determined through coprecipitation experiments. The role of pH and the extent of drug ionization were probed to evaluate the role of electrostatic interactions in complex formation. Relationships between the extent of complexation and the drug release rate from an ASD were observed, whereby the drugs could be divided into two groups. Drugs with a low extent of insoluble complex formation with HPMCAS tended to be neutral or anionic and showed reasonable release at pH 6.8 even at higher drug loadings. Cationic drugs formed insoluble complexes with HPMCAS and showed poor release when formulated as an ASD. Thus, and somewhat counterintuitively, a weakly basic drug showed a reduced release rate from an ASD at a bulk solution pH where it was ionized, relative to when unionized. The opposite trend was observed in the absence of polymer for the neat amorphous drug. In conclusion, electrostatic interactions between HPMCAS and lipophilic cationic drugs led to insoluble complex formation, which in turn resulted in ASDs with poor release performance.

Machine learning framework for extracting micro-viscoelastic and micro-structural properties of compressed oral solid dosage forms.

A compressed pharmaceutical oral solid dosage (OSD) form is a strongly micro-viscoelastic material composite arranged as a network of agglomerated particles due to its constituent powders and their bonding and fractural mechanical properties. An OSD product's Critical Quality Attributes, such as disintegration, drug release (dissolution) profile, and structural strength ("hardness"), are influenced by its micro-scale properties. Ultrasonic evaluation is direct, non-destructive, rapid, and cost-effective. However, for practical process control applications, the simultaneous extraction of the micro-viscoelastic and scattering properties from a tablet's ultrasonic response requires a unique solution to a challenging inverse mathematical wave propagation problem. While the spatial progression of a pulse traveling in a composite medium with known micro-scale properties is a straightforward computational task when its dispersion relation is known, extracting such properties from the experimentally acquired waveforms is often non-trivial. In this work, a novel Machine Learning (ML)-based micro-property extraction technique directly from waveforms, based on Multi-Output Regression models and Neural Networks, is introduced and demonstrated. Synthetic waveforms with a given set of micro-properties of virtual tablets are computationally generated to train, validate, and test the developed ML models for their effectiveness in the inverse problem of recovering specified micro-scale properties. The effectiveness of these ML models is then tested and demonstrated for a set of physical OSD tablets. The micro-viscoelastic and micro-structural properties of physical tablets with known properties have been extracted through experimentally acquired waveforms to exhibit their consistency with the generated ML-based attenuation results.

Does Media Choice Matter When Evaluating the Performance of Hydroxypropyl Methylcellulose Acetate Succinate-Based Amorphous Solid Dispersions?

Hydroxypropyl methylcellulose acetate succinate (HPMCAS) is a weakly acidic polymer that is widely used in the formulation of amorphous solid dispersions (ASDs). While the pH-dependent solubility of HPMCAS is widely recognized, the role of other solution properties, including buffer capacity, is less well understood in the context of ASD dissolution. The goal of this study was to elucidate the rate-limiting steps for drug and HPMCAS release from ASDs formulated with two poorly water soluble model drugs, indomethacin and indomethacin methyl ester. The surface area normalized release rate of the drug and/or polymer in a variety of media was determined. The HPMCAS gel layer apparent pH was determined by incorporating pH sensitive dyes into the polymer matrix. Water uptake extent and rate into the ASDs were measured gravimetrically. For neat HPMCAS, the rate-limiting step for polymer dissolution was observed to be the polymer solubility at the polymer-solution interface. This, in turn, was impacted by the gel layer pH which was found to be substantially lower than the bulk solution pH, varying with medium buffer capacity. For the ASDs, the HPMCAS release rate was found to control the drug release rate. However, both drugs reduced the polymer release rate with indomethacin methyl ester having a larger impact. In low buffer capacity media, the presence of the drug had less impact on release rates when compared to observations in higher strength buffers, suggesting changes in the rate-limiting steps for HPMCAS dissolution. The observations made in this study can contribute to the fundamental understanding of acidic polymer dissolution in the presence and absence of a molecularly dispersed lipophilic drug and will help aid in the design of more in vivo relevant release testing experiments.

End-point determination of heterogeneous formulations using inline torque measurements for a high-shear wet granulation process.

In this study, the torque profiles of heterogeneous granulation formulations with varying powder properties in terms of particle size, solubility, deformability, and wettability, were studied, and the feasibility of identifying the end-point of the granulation process for each formulation based on the torque profiles was evaluated. Dynamic median particle size (d50) and porosity were correlated to the torque measurements to understand the relationship between torque and granule properties, and to validate distinction between different granulation stages based on the torque profiles made in previous studies. Generally, the torque curves obtained from the different granulation runs in this experimental design could be categorized into two different types of torque profiles. The primary factor influencing the likelihood of producing each profile was the binder type used in the formulation. A lower viscosity, higher solubility binder resulted in a type 1 profile. Other contributing factors that affected the torque profiles include API type and impeller speed. Material properties such as the deformability and solubility of the blend formulation and the binder were identified as important factors affecting both granule growth and the type of torque profiles observed. By correlating dynamic granule properties with torque values, it was possible to determine the granulation end-point based on a pre-determined target median particle size (d50) range which corresponded to specific markers identified in the torque profiles. In type 1 torque profiles, the end-point markers corresponded to the plateau phase, whereas in type 2 torque profiles the markers were indicated by the inflection point where the slope gradient changes. Additionally, we proposed an alternative method of identification by using the first derivative of the torque values, which facilitates an easier identification of the system approaching the end-point. Overall, this study identified the effects of different variations in formulation parameters on torque profiles and granule properties and implemented an improved method of identification of granulation end-point that is not dependent on the different types of torque profiles observed.

Insights into the role of tooling characteristics on compressibility evolution in non-flat faced tablets.

The robustness of tablet manufacturability largely depends on compressibility behavior of a powder. The compressibility assessment is traditionally conducted on cylindrical flat-faced compacts in contrast to the fact that marketed tablets are majorly produced using non-flat faced or shaped toolings. The present work demonstrates the feasibility of quantifying average compressibility on shaped toolings through a proof-of-concept study by investigating the central band portion and the entire volume of the tablet, which led to several notable findings. Firstly, the yield stress (deformability) was found independent of type of tooling for a given powder in the in-die condition, but for the same tooling it conversely spanned over a wide range in the out-die condition due to characteristic elastic recovery. Secondly, the yield stress parameter correlated with the change in band volume of the shaped tablet with applied compaction pressure, thereby establishing an orthogonal approach to assess compressibility on non-flat faced toolings. The study emphasizes that tooling characteristics may affect compressibility and tablet robustness of a same powder, which should be practiced cautiously in drug product manufacturing.

Enabling direct compression tablet formulation of celecoxib by simultaneously eliminating punch sticking, improving manufacturability, and enhancing dissolution through co-processing with a mesoporous carrier.

The development of a high quality tablet of Celecoxib (CEL) is challenged by poor dissolution, poor flowability, and high punch sticking propensity of CEL. In this work, we demonstrate a particle engineering approach, by loading a solution of CEL in an organic solvent into a mesoporous carrier to form a coprocessed composite, to enable the development of tablet formulations up to 40% (w/w) of CEL loading with excellent flowability and tabletability, negligible punch sticking propensity, and a 3-fold increase in in vitro dissolution compared to a standard formulation of crystalline CEL. CEL is amorphous in the drug-carrier composite and remained physically stable after 6 months under accelerated stability conditions when the CEL loading in the composite was ≤ 20% (w/w). However, crystallization of CEL to different extents from the composites was observed under the same stability condition when CEL loading was 30-50% (w/w). The success with CEL encourages broader exploration of this particle engineering approach in enabling direct compression tablet formulations for other challenging active pharmaceutical ingredients.

Micro-viscoelastic Characterization of Compressed Oral Solid Dosage Forms with Ultrasonic Wave Dispersion Analysis.

Due to their constituent powders, the materials of advanced compressed oral solid dosage (OSD) forms are micro-composites and strongly visco-elastic at macro- and micro-length scales. The disintegration, drug release, and mechanical strength of OSD forms depend on its micro-texture (such as porosity) and micro-scale physical/mechanical properties. In the current work, an algorithmic ultrasonic characterization framework for extracting the micro-visco-elastic properties of OSD materials is presented, and its applicability is demonstrated with a model material. The proposed approach is based on the effect of visco-elasticity and granularity on the frequency-dependent attenuation of an ultrasonic wave pulse in a composite (granular) and viscous medium. In modeling the material, a two-parameter Zener model for visco-elasticity and a scattering attenuation mechanism based on Rayleigh scattering for long-wave approximation are employed. A novel linear technique for de-coupling the effects of micro-visco-elasticity and scattering on attenuation and dispersion is developed and demonstrated. The apparent Young's modulus, stress, and strain relaxation time constants of the medium at micro-scale are extracted and reported. Based on this modeling and analysis framework, a set of computational algorithms has been developed and demonstrated with experimental data, and its practical utility in pharmaceutical manufacturing and real-time release testing of tablets is discussed.

An extended macroindentation method for determining the hardness of poorly compressible materials.

Indentation hardness, H, is an important mechanical property that quantifies the resistance to deformation by a material. For pharmaceutical powders, H can be determined using a macroindentation method, provided they can form intact tablets suitable for testing. This work demonstrates a method for determining the hardness of problematic materials that cannot form suitable tablets for macroindentation. The method entails predicting the hardness of a given powder at zero porosity (H0) from those of microcrystalline cellulose and its binary mixture with the test compound using a power law mixing rule based on weight fraction. This method was found suitable for 13 binary mixtures. In addition, the H0 values derived by this method could capture changes due to different particle sizes of sucrose and sodium chloride. Furthermore, the derived H0 reasonably agreed with the single crystal indentation hardness of a set of 16 crystals when accounting for the effect of indentation condition and structural anisotropy. The mixture method thus extends the use of macroindentation for predicting indentation hardness of powders that cannot form intact tablets and, hence, their plasticity.

Ultrasonic characterization of complete anisotropic elasticity coefficients of compressed oral solid dosage forms.

In compacted materials, elastic anisotropy coupled with residual stresses could play a determining role in the manifestation of various types of defects such as capping and lamination, as it creates shear planes/bands and temporal relaxation. This internal micro-structure leads to time-delayed flaw initiation/formation, crack tip propagation under residual stresses, and ultimately product quality failures. Thus, their accurate characterization and variations are useful for understanding underlying failure mechanisms and to monitor variations in materials, processes and product quality during production prior to onset of failure. The extraction of tablet anisotropic elasticity properties is a challenging task, especially for commercial tablets with complex shapes, as shape often prevents the use of traditional destructive techniques (e.g., diametric compression testers) to produce accurate measurements. This study introduces and applies an ultrasonic approach to extracting the complete transverse isotropic elastic properties of compressed oral solid dosage forms to a commercial tablet product. A complete set of waveforms and the constitutive matrix for the compacted materials are reported. In addition, a perturbation analysis is carried out to analytically relate propagation speeds in various directions to the elastic coefficients. The proposed characterization approach is non-destructive, rapid, easy, and reliable in evaluating tablet anisotropy.

An insight into inter-relationships among tensile strength, elastic modulus and plasticity on tabletability of single components and binary mixtures.

The evolution of tablet strength is mainly influenced by deformability (bonding area) and strength of intermolecular interactions (bonding strength) from the intrinsic material properties and tableting process, respectively. Therefore, understanding of intrinsic material attributes is important for in-silico drug product designs. The present study shows that the separate effect of the above two factors can be better understood by systematic evaluation of pure APIs and their formulations. Using tensile strength, elastic modulus and yield stress as critical material attributes, a proof of concept shown in this work emphasizes that materials with greater deformability tend to possess greater tensile strength at comparable bonding strengths. In contrast, the influence of the deformability parameter is hidden when formulations are used, leading to a scenario where the effects of bonding area and bonding strength are more inseparable.

A semi-empirical model for estimation of flaw size in internally defective tablets.

Capping is a mechanical defect in tablets, which is attributed to multiple factors including intrinsic material properties and tableting conditions. A suitable non-destructive approach using acoustically derived elastic modulus has showed distinctive features between a defective tablet and a defect-free tablet. In this work, a semi-empirical model was developed to estimate flaw size in an internally defective tablet from the relationship among elastic modulus, tablet density, and time of flight (acoustic wave to traverse through the tablet). The model was found fundamentally consistent where the derived flaw size showed clear dependence on powder mechanical properties of seven diverse formulations studied. Furthermore, the flaw size was reasonably correlated with the internal tablet microstructure illustrated by X-ray micro-tomography findings, both qualitatively and quantitatively. This model could thus be efficiently implemented for risk-based evaluation of internal defects in visibly intact tablets to ensure robustness of drug products.

An insight into predictive parameters of tablet capping by machine learning and multivariate tools.

Capping is the frequently observed mechanical defect in tablets arising from the sub-optimal selection of the formulation composition and their robustness of response toward process parameters. Hence, overcoming capping propensity based on the understanding of suitable process and material parameters is of utmost importance to expedite drug product development. In the present work, 26 diverse formulations were characterized at commercial tableting condition to identify key tablet properties influencing capping propensity, and a predictive model based on threshold properties was established using machine learning and multivariate tools. It was found that both the compaction parameters (i.e., compaction pressure, radial stress transmission characteristics, and Poisson's ratio), and the material properties, (i.e., brittleness, yield strength, particle bonding strength and elastic recovery) strongly dictate the capping propensity of a tablet. In addition, ratio of elastic modulus in the orthogonal direction in a tablet and its variation with porosity were notable quantitative metrics of capping occurrence.

The impact of solid-state form, water content and surface area of magnesium stearate on lubrication efficiency, tabletability, and dissolution.

Magnesium stearate (MgSt) is a widely used pharmaceutical lubricant in tablet manufacturing. However, batch-to-batch variability in hydrate form and surface area can lead to inconsistency in tablet performance. In this work, several unique MgSt samples were studied: traditional monohydrate samples with high surface area, dihydrate forms with high and low surface area, and disordered forms with low and medium water content. The effects of solid-state form and particle properties on lubrication efficiency, tabletability and dissolution were studied for tablets in a model direct compression formulation. It was found that the monohydrate and dihydrate forms had good lubrication efficiency compared to the disordered form, while the disordered form had the best tabletability. The dissolution rate correlated with surface area, where slower dissolution rates corresponded with higher MgSt surface areas. The dihydrate sample with lower surface area had the best performance for this model formulation, in terms of lubrication efficiency, tabletability and dissolution. Overall, it is concluded that the choice of the most appropriate grade of MgSt for a particular formulation depends on a comprehensive evaluation of the impact of MgSt properties on lubrication efficiency, tabletability and dissolution.

How Does the Dissimilarity of Screw Geometry Impact Twin-screw Melt Granulation?

Using a model formulation of 80% gabapentin and 20% hydroxypropyl cellulose (KlucelTM), we investigate how differences in the geometry of mixing elements in the Leistritz Nano-16 and Micro-18 extruders affect granulation mechanisms and the properties of the resulting granules. Two extruders, Leistritz Nano-16 and Micro-18, commonly used in development and manufacturing, respectively, were used. The kneading blocks of the Nano-16 extruder are less efficient in dispersive mixing than the kneading blocks of the Micro-18 due to the thinner discs (2.5 mm wide) of the Nano-16. Therefore, our model formulation could be granulated only under a higher degree of fill (DF) by enhancing the axial compaction and heating of the barrel. In contrast, the thicker (5 mm wide) kneading blocks of the Micro-18 extruder provide efficient dispersive mixing that enables granulation without axial compaction and barrel heating. The higher specific mechanical energy (SME) achieved at higher screw speeds and lower feed rates led to more granulation. Because of the difference in granulation mechanisms between the two extruders, critical processing parameters also differed. Tabletability and degradant content of granules correlated positively with DF for the Nano-16 but with SME for the Micro-18 extruder.

Tabletability Flip - Role of Bonding Area and Bonding Strength Interplay.

Predicting tableting performance of mixtures from that of individual components is of practical importance for achieving efficient and robust tablet design. It has been commonly assumed that a solid form exhibiting better tabletability will result in better tabletability when formulated. However, we show that the rank order of tabletability between two powders can flip when mixed with another powder, a phenomenon termed tabletability flip. Using three examples, we show that the tabletability flip upon mixing with microcrystalline cellulose is activated by the switch of the dominating factor in the bonding area (BA) and bonding strength (BS) interplay that determines tablet tensile strength. A mechanistic understanding of this phenomenon can significantly improve the accuracy of predicted tableting performance of mixtures from that of individual powders.

Effect of Hydroxypropyl Cellulose Level on Twin-Screw Melt Granulation of Acetaminophen.

This study investigated the effect of binder level on the physicochemical changes and tabletability of acetaminophen (APAP)-hydroxypropyl cellulose (HPC) granulated using twin-screw melt granulation. Even at 5% HPC level, the tablet tensile strength achieved up to 3.5 MPa. A minimum of 10% HPC was required for the process robustness. However, 20% HPC led to tabletability loss, attributable to the high mechanical strength of APAP granules. The over-granulated APAP granules had thick connected HPC scaffold and low porosity. Consequently, these granules were so strong that they underwent a lower degree of fracture under compression and higher elastic recovery during decompression. HPC was enriched on the surface of APAP extrudates at all HPC levels. Amorphous APAP was also observed on the extrudate surface at 20% HPC level, and it recrystallized within 24 h storage. To achieve a robust process and optimal improvement in APAP tabletability, the preferred HPC level was 10 to 15%.

Reduction of Punch-Sticking Propensity of Celecoxib by Spherical Crystallization via Polymer Assisted Quasi-Emulsion Solvent Diffusion.

Punch-sticking during tablet compression is a common problem for many active pharmaceutical ingredients (APIs), which renders tablet formulation development challenging. Herein, we demonstrate that the punch-sticking propensity of a highly sticky API, celecoxib (CEL), can be effectively reduced by spherical crystallization enabled by a polymer assisted quasi-emulsion solvent diffusion (QESD) process. Among three commonly used pharmaceutical polymers, poly(vinylpyrrolidone) (PVP), hydroxypropyl cellulose (HPC), and hydroxypropyl methylcellulose (HPMC), HPMC was the most effective in stabilizing the transient emulsion during QESD and retarding the coalescence of emulsion droplets and the initiation of CEL crystallization. These observations may arise from stronger intermolecular interactions between HPMC and CEL, consistent with solution 1H NMR analyses. SEM and X-ray photoelectron spectroscopy confirmed the presence of a thin layer of HPMC on the surfaces of spherical particles. Thus, the sticking propensity was significantly reduced because the HPMC coating prevents direct contact between CEL and the punch tip during tablet compression.

Toward a Molecular Understanding of the Impact of Crystal Size and Shape on Punch Sticking.

Punch sticking during tablet manufacturing is a common problem facing the pharmaceutical industry. Using several model compounds, effects of crystal size and shape of active pharmaceutical ingredients (API) on punch sticking propensity were systematically investigated in this work to provide molecular insights into the punch-sticking phenomenon. In contrast to the common belief that smaller API particles aggravate punch sticking, results show that particle size reduction can either reduce or enhance API punch sticking, depending on the complex interplay among the particle surface area, plasticity, cohesive strength, and specific surface functional groups. Therefore, other factors, such as crystal mechanical properties, surface chemistry of crystal facets exposed to the punch face, and choice of excipients in a formulation, should be considered for a more reliable prediction of the initiation and progression of punch sticking. The exposure of strong electronegative groups to the punch face facilitates the onset of sticking, while higher plasticity and cohesive strength aggravate sticking.