Real-time monitoring of multiparticulate coating processes at industrial-scale using ultra-high-resolution optical coherence tomography

Abstract

Optical Coherence Tomography (OCT) has been reported as a promising technology for in-line monitoring of pharmaceutical film-coating processes, providing real-time information on actual tablet coating thicknesses and thickness variability, together enabling correlation modeling of the dissolution behavior. However, an increasing interest in the in-line investigation of multiparticulates demands further evolution of the technology and new concepts for process interfacing. To achieve this, we present a novel way to interface OCT to the relevant coating processes using a cutting-edge industry-ready ultra-high-resolution OCT (UHR-OCT) for in-line monitoring. Based on the setup, real-time in-line data of multiparticulate coating processes was acquired, enabling an automated coating thickness evaluation starting at the low micron range. The system was tested with two different core sizes of 250- and 700-µm mean diameter, coated in two different coaters, namely a Glatt MultiLab (2  kg batch size) and a Glatt GPCG PRO 30 (30  kg batch size). Coating thicknesses were successfully measured directly in the process in real-time within a range from 2.5 to 20 µm. Reference measurements based on sprayed coating mass showed excellent agreement between OCT and the reference when agglomeration was neglectable. Additionally, conducted image-based photometry emphasizes the capabilities of UHR-OCT as a powerful technology for multiparticulate coating monitoring. The presented approach was found to be stable, batch size-independent, and translatable to an industrial setup. Moreover, it extends the capabilities of OCT toward a versatile process analytical technology to monitor even the thinnest coatings and small particles used in pharmaceutical coating processes.

Introduction

In recent years, Optical Coherence Tomography (OCT) has gained attention as a promising process analytical technology (PAT) for monitoring pharmaceutical coating processes. Already in 2011, the first OCT-based evaluation of a pharmaceutical tablet coating process was presented and validated by comparing OCT results to established reference methods (Koller et al., 2011). Consequently, the topic gained certain attention in the scientific community, with OCT being considered as an in-line PAT (Markl et al., 2014a) and assessment of the best suitable wavelength range to investigate pharmaceutical coating layer structures using OCT (Markl et al., 2014b). In parallel, Li et al. presented the first high-resolution (off-line) OCT images of film-coated pellets using full-field OCT, pushing the limits of spatial resolution on pharmaceutical coating structures at this time (Li et al., 2014). Several publications on OCT applications in a pharma context followed in the upcoming years, addressing in-line monitoring of pellet- (Markl et al., 2015b) and tablet coatings (Markl et al., 2015a), as well as enhanced evaluation algorithms for pharmaceutical coating layers (Dong et al., 2017, Lin et al., 2015). Several assessments of OCT as a PAT tool in comparison to other (established) technologies for pharmaceutical coating applications followed (Korasa and Vrečer, 2018, Lin et al., 2018, Wahl et al., 2019).

With the broadening of the scientific and application knowledge on OCT, and its capabilities for pharmaceutical coating monitoring, a certain interest from the industry evolved, leading to the first commercially available OCT system for pharmaceutical applications introduced in 2016 by Phyllon GmbH. Results from pharmaceutical coating process monitoring using these new OCT systems have been published for tablet coating (Sacher et al., 2019) and for fluid bed coating of granules (Pietsch et al., 2019). A concise summary of the capabilities and limitations of coating monitoring of tablet coatings using OCT including actual examples has been published by Sacher et al. (2021), and considerations to use OCT for real-time release testing by Markl et al. (2020). However, industrial use of any PAT sensor requires standardized approaches for validation, which has been addressed exemplarily for tablet coating monitoring using OCT by our working group in 2019 (Wolfgang et al., 2019). Obviously, coating thickness and coating variability of functional coatings can be considered critical parameters as they directly impact the release kinetics and the resulting product efficacy and quality (Priese et al., 2023). Previous studies have already shown the potential of OCT to predict dissolution behavior based on the coating thickness of tablets (Wolfgang et al., 2022) and even for uncoated tablets (Fink et al., 2023). However, for multiparticulate dosage forms, the situation is more complex: very thin coating layers (Turk et al., 2021) and multilayered structures (Wolfgang et al., 2020) make the direct in-line monitoring of these dosage forms challenging. Therefore, in-line OCT applications on pellet and multiparticulate coating processes seem still underrepresented in the available literature.

Documented attempts to use OCT for multiparticulate coating monitoring include the use of full-field OCT to investigate thin pellet coatings (Li et al., 2014), efforts to use scanning OCT sensors (Markl et al., 2015b) in-line, and adaptions of existing in-line OCT sensors to compensate for the distortions of the relevant process interfaces (Pietsch et al., 2019). However, those experiments were mere proof-of-concept and could not be translated to an industrial scale so far. The main reason was that full-field OCT is inherently not in-line capable and that the galvanometric mirrors used in the scanning OCT probes are prone to (process-induced) vibrations, distorting the OCT image acquisition procedure. The latest attempt by Pietsch et al. was only applicable for coating thicknesses starting at approx. 15 to 20 µm, which is too thick for several multiparticulate coatings used today. The main challenge for multiparticulate coating monitoring is the required high temporal and spatial resolution. The acquisition speed on the one hand is relevant because the number of scans per object is crucial to extracting meaningful coating layer information on fast-moving and small objects (Markl et al., 2014a). The spatial resolution on the other hand is crucial as the small objects have a high curvature, which would be smeared out as the focal point gets larger, as OCT likewise other tomographic techniques acquires volumetric information. Especially the high spatial resolution inherently introduces several constraints regarding an actual optical sensor setup and limits severely the achievable field of view (Fujimoto and Drexler, 2008, Hu and Rollins, 2008, Wojtkowski, 2010), impacting the demand to present the objects under test at a very small and defined area in front of the sensor.

Both coaters used in this study work according to the Wurster principle, first introduced by Dr. Dale Wurster in 1952 (Kumpugdee-Vollrath and Krause, 2011) and subsequently patented in 1966 by Wurster and Lindlof (1966). This modality of fluidized bed coating has the spray nozzles integrated into the base plate of the coater spraying toward the top of the coater and ensuring that the spray aerosol is always surrounded by the particles to be coated. This arrangement together with one (or more) tube(s) aligning the gas dynamics inside the coater forces the particles into a permanent movement along a toroidal trajectory with the symmetry axis in the middle of the coater (Jones and Godek, 2017). Consequently, this process provides excellent separation of particles, guaranteeing ideal conditions to coat smallest granules, micro-pellets, and mini-tablets. One main advantage of the process is that a uniform coating can be achieved, i.e. even with particles that are not spherical, a uniform layer thickness can be created by coating (Kumpugdee-Vollrath and Krause, 2011). Moreover, the process can be optimized in silico toward optimal coating yield and minimal coating time (Madlmeir et al., 2022), both of the highest interest to the industry, making this coater modality very attractive.

In addition to the considerations above and regardless of the considerable developments of OCT as a technology (and other technologies, for that matter), all previous attempts to do multiparticulate in-line coating monitoring struggled with the acquisition and automated evaluation of a sufficient number of objects per minute. Sufficient in this context seems a representative sampling of several hundred objects per minute to derive sound statistics for assessing actual coating thickness and thickness distribution. The main constraints regarding in-line measurements of the involved process dynamics in fluid bed coaters include the small particle sizes of less than 1  mm in diameter, the high speeds and arbitrary trajectories of the particles, and disturbances of the measurements by the processes themselves (vibration, dust, etc.). Moreover, the often extremely thin functional coating layers of below 10 µm require ultra-high spatial resolution of the involved technology in the range of a few microns.

Introducing ultra-high-resolution OCT (UHR-OCT) systems (Haindl et al., 2020) changed this situation and encouraged our group to re-engage in this topic (Haindl et al., 2023, Wolfgang et al., 2023). Unfortunately, the challenges concerning proper process interfacing persist, as another relevant aspect in fluid bed coating monitoring is the fact that previous studies showed only results from small lab scales. In contrast, the scientific and commercial relevance of OCT as the PAT of choice for multiparticulate coating monitoring can only be justified by examining actual industrial-scale processes. Hence, we aimed to fill these gaps by developing a setup to investigate multiparticulate coating processes at an industrial scale. Our approach is straightforward, with the development of a new process interface suitable for fluid bed coater integration, the translation of the UHR-OCT concept into a sensor system that can be used in an industrial environment, and the conduction of field tests.

Real-time monitoring of multiparticulate coating processes at industrial-scale using ultra-high-resolution optical coherence tomography

Abstract

Introduction

Visit our Excipient Landscape 2024 here:

Get more information on different excipients and their manufacturers