Stephan Sacher

Optical coherence tomography as a novel tool for in-line monitoring of a pharmaceutical film-coating process

Optical coherence tomography (OCT) is a contact-free non-destructive high-resolution imaging technique based on low-coherence interferometry. This study investigates the application of spectral-domain OCT as an in-line quality control tool for monitoring pharmaceutical film-coated tablets. OCT images of several commercially-available film-coated tablets of different shapes, formulations and coating thicknesses were captured off-line using two OCT systems with centre wavelengths of 830nm and 1325nm. Based on the off-line image evaluation, another OCT system operating at a shorter wavelength was selected to study the feasibility of OCT as an in-line monitoring method. Since in spectral-domain OCT motion artefacts can occur as a result of the tablet or sensor head movement, a basic understanding of the relationship between the tablet speed and the motion effects is essential for correct quantifying and qualifying of the tablet coating. Experimental data was acquired by moving the sensor head of the OCT system across a static tablet bed. Although examining the homogeneity of the coating turned more difficult with increasing transverse speed of the tablets, the determination of the coating thickness was still highly accurate at a speed up to 0.7m/s. The presented OCT setup enables the investigation of the intra- and inter-tablet coating uniformity in-line during the coating process.

Optimized continuous pharmaceutical manufacturing via model-predictive control.

This paper demonstrates the application of model-predictive control to a feeding blending unit used in continuous pharmaceutical manufacturing. The goal of this contribution is, on the one hand, to highlight the advantages of the proposed concept compared to conventional PI-controllers, and, on the other hand, to present a step-by-step guide for controller synthesis. The derivation of the required mathematical plant model is given in detail and all the steps required to develop a model-predictive controller are shown. Compared to conventional concepts, the proposed approach allows to conveniently consider constraints (e.g. mass hold-up in the blender) and offers a straightforward, easy to tune controller setup. The concept is implemented in a simulation environment. In order to realize it on a real system, additional aspects (e.g., state estimation, measurement equipment) will have to be investigated.

In-line monitoring of a pharmaceutical pan coating process by optical coherence tomography

This work demonstrates a new in-line measurement technique for monitoring the coating growth of randomly moving tablets in a pan coating process. In-line quality control is performed by an optical coherence tomography (OCT) sensor allowing nondestructive and contact-free acquisition of cross-section images of film coatings in real time. The coating thickness can be determined directly from these OCT images and no chemometric calibration models are required for quantification. Coating thickness measurements are extracted from the images by a fully automated algorithm. Results of the in-line measurements are validated using off-line OCT images, thickness calculations from tablet dimension measurements, and weight gain measurements. Validation measurements are performed on sample tablets periodically removed from the process during production. Reproducibility of the results is demonstrated by three batches produced under the same process conditions. OCT enables a multiple direct measurement of the coating thickness on individual tablets rather than providing the average coating thickness of a large number of tablets. This gives substantially more information about the coating quality, that is, intra- and intertablet coating variability, than standard quality control methods.

A Review of PAT Strategies in Secondary Solid Oral Dosage Manufacturing of Small Molecules

Pharmaceutical solid oral dosage product manufacturing is a well-established, yet revolutionizing area. To this end, process analytical technology (PAT) involves interdisciplinary and multivariate (chemical, physical, microbiological, and mathematical) methods for material (e.g., materials, intermediates, products) and process (e.g., temperature, pressure, throughput, etc.) analysis. This supports rational process modeling and enhanced control strategies for improved product quality and process efficiency. Therefore, it is often difficult to orient and find the relevant, integrated aspects of the current state-of-the-art. Especially, the link between fundamental research, in terms of sensor and control system development, to the application both in laboratory and manufacturing scale, is difficult to comprehend. This review compiles a nonexhaustive overview on current approaches from the recognized academia and industrial practices of PAT, including screening, selection, and final implementations in solid oral dosage manufacturing, through a wide diversity of use cases. Finally, the authors attempt to extract a common consensus toward developing PAT application guidance for different unit operations of drug product manufacturing.

Automated pharmaceutical tablet coating layer evaluation of optical coherence tomography images

Film coating of pharmaceutical tablets is often applied to influence the drug release behaviour. The coating characteristics such as thickness and uniformity are critical quality parameters, which need to be precisely controlled. Optical coherence tomography (OCT) shows not only high potential for off-line quality control of film-coated tablets but also for in-line monitoring of coating processes. However, an in-line quality control tool must be able to determine coating thickness measurements automatically and in real-time. This study proposes an automatic thickness evaluation algorithm for bi-convex tables, which provides about 1000 thickness measurements within 1 s. Beside the segmentation of the coating layer, optical distortions due to refraction of the beam by the air/coating interface are corrected. Moreover, during in-line monitoring the tablets might be in oblique orientation, which needs to be considered in the algorithm design. Experiments were conducted where the tablet was rotated to specified angles. Manual and automatic thickness measurements were compared for varying coating thicknesses, angles of rotations, and beam displacements (i.e. lateral displacement between successive depth scans). The automatic thickness determination algorithm provides highly accurate results up to an angle of rotation of 30°. The computation time was reduced to 0.53 s for 700 thickness measurements by introducing feasibility constraints in the algorithm.

Automatic correction for window fouling of near infrared probes in fluidised systems

Near infrared (NIR) spectroscopy is a versatile, non-invasive and non-destructive tool that is often used for process monitoring in the pharmaceutical industry. Often, equipment window fouling or probe fouling of in-situ NIR probes occurs, leading to biased spectra and wrong interpretations (e.g. process-state estimation). Physical counter measures, including self-cleaning probes and geometrical considerations, are called for. This paper presents a mathematical solution to the problem of window fouling for an NIR-monitored process: by determining the distance to the particles, we established which part of the signal was missing owing to the coating accumulation on the probe window. The proposed approach is illustrated with the example of hot-melt coating in a fluidised bed, during which coating buildup on substrate particles was monitored despite window fouling.

RTD Modeling of a Continuous Dry Granulation Process for Process Control and Materials Diversion

Disturbance propagation during continuous manufacturing processes can be predicted by evaluating the residence time distribution (RTD) of the specific unit operations. In this work, a dry granulation process was modelled and four scenarios of feeding events were simulated. We performed characterization of the feeders and developed RTD models for the blender and the roller compactor based on impulse-response measurements via color tracers. Out-of-specification material was defined based on the active pharmaceutical ingredient (API) concentration. We calculated the amount of waste material at various diversion points, considering four feeder-related process-upset scenarios and formulated considerations for the development of a control concept. The developed RTD models allow material tracking of materials that may be used for following the spread contaminants within the process and for batch definition. The results show that RTD modeling is a valuable tool for process development and design, as well as for process monitoring and material tracking.

Monitoring of a Hot Melt Coating Process via a Novel Multipoint Near-Infrared Spectrometer

The aim of the present work was to develop a PAT strategy for the supervision of hot melt coating processes. Optical fibers were placed at various positions in the process chamber of a fluid bed device. Experiments were performed to determine the most suitable position for in-line process monitoring, taking into account such requirements as a good signal to noise ratio, the mitigation of dead zones, the ability to monitor the product over the entire process, and reproducibility. The experimental evidence suggested that the position at medium fluid bed height, looking towards the center, i.e., normal to particle movement, proved to be the most reliable position. In this study, the advantages of multipoint monitoring are shown, and an in-line-implementation was created. This enabled the real-time supervision of the process, including the fast detection of inhomogeneities and disturbances in the process chamber, and the compensation of sensor malfunction. In addition, a model for estimating the particle size distribution via NIR was successfully created. This ensures that the quality of the product and the endpoint of the coating process can be determined correctly.

Use of mechanistic simulations as a quantitative risk-ranking tool within the quality by design framework.

The purpose of this study is to evaluate the use of computer simulations for generating quantitative knowledge as a basis for risk ranking and mechanistic process understanding, as required by ICH Q9 on quality risk management systems. In this specific publication, the main focus is the demonstration of a risk assessment workflow, including a computer simulation for the generation of mechanistic understanding of active tablet coating in a pan coater. Process parameter screening studies are statistically planned under consideration of impacts on a potentially critical quality attribute, i.e., coating mass uniformity. Based on computer simulation data the process failure mode and effects analysis of the risk factors is performed. This results in a quantitative criticality assessment of process parameters and the risk priority evaluation of failure modes. The factor for a quantitative reassessment of the criticality and risk priority is the coefficient of variation, which represents the coating mass uniformity. The major conclusion drawn from this work is a successful demonstration of the integration of computer simulation in the risk management workflow leading to an objective and quantitative risk assessment.

Continuous monitoring of API content, API distribution and crushing strength after tableting via near-infrared chemical imaging

Near-infrared chemical imaging (NIR-CI) with high-speed cameras based on the push-broom acquisition principle is a rapidly-evolving and can be used for a variety of purposes, from classification (and sorting) of products to mapping spatial distribution of materials. The present study examined if NIR-CI is suitable for tablet manufacturing. To that end, the tablets were introduced into the CI system via a flat belt conveyor. A formulation, which consisted of 4wt.%-6wt.% caffeine, 5wt.% crospovidone as a disintegrant, 88wt.%-90wt.% lactose as a filler and 1wt.% magnesium stearate as a lubricator, was tableted at compression forces ranging from 5kN to 30kN. The intra- and inter-tablet homogeneity of caffeine and the tablets hardness were analyzed via NIR-CI. For the homogeneity evaluation, two methods were applied: standard deviation (SD) and distributional homogeneity index (DHI). The results showed that the SD of caffeine in a single tablet increased with an increase in the caffeine content. This was attributed to natural variations in a binary mixture of caffeine and excipients. Overall, the chosen NIR-CI setup has strong potential to be transferred to the production scale to monitor all tablets in a production stream.