From powder to tablets: Investigation of residence time distributions in a continuous manufacturing process train as basis for continuous process verification

Statistical data treatment for residence time distribution studies in pharmaceutical manufacturing

Residence time distribution (RTD) method has been widely used in the pharmaceutical manufacturing for understanding powder dynamics within unit operations and continuous integrated manufacturing lines. The dynamics thus captured is then used to develop predictive models for unit operations and important RTD-based applications ensuring product quality assurance. Despite thorough efforts in tracer selection, data acquisition, and calibration model development to obtain tracer concentration profiles for RTD studies, there can exist significant noise in these profiles. This noise can make it challenging to identify the underlying signal and get a representative RTD of the system under study. Such concerns have previously indicated the importance of noise handling for RTD measurements in literature. However, the literature does not provide sufficient information on noise handling or data treatment strategies for RTD studies. To this end, we investigate the impact of varying levels of noise using different tracers on measurement of RTD profile and its applications. We quantify the impact of different denoising methods (time and frequency averaging methods). Through this investigation, we see that Savitsky Golay filtering turns out to a good method for denoising RTD profiles despite varying noise levels. The investigation is performed such that the key features of the RTD profile (which are important for RTD based applications) are preserved. Subsequently, we also investigate the impact of denoising on RTD-based applications such as out-of-specification (OOS) analysis and RTD modeling. The results show that the degree of noise levels considered in this work do not significantly impact the RTD-based applications.

Challenges encountered in the transfer of atorvastatin tablet manufacturing - commercial batch-based production as a basis for small-scale continuous tablet manufacturing tests

As is the case with batch-based tableting processes, continuous tablet manufacturing can be conducted by direct compression or with a granulation step such as dry or wet granulation included in the production procedure. In this work, continuous manufacturing tests were performed with a commercial tablet formulation, while maintaining its original material composition. Challenges were encountered with the feeding performance of the API during initial tests which required designing different powder pre-blend compositions. After the pre-blend optimization phase, granules were prepared with a roller compactor. Tableting was conducted with the granules and an additional brief continuous direct compression run was completed with some ungranulated mixture. The tablets were assessed with off-line tests, applying the quality requirements demanded for the batch-manufactured product. Chemical maps were obtained by Raman mapping and elemental maps by scanning electron microscopy with energy-dispersive X-ray spectroscopy. Large variations in both tablet weights and breaking forces were observed in all tested samples, resulting in significant quality complications. It was suspected that the API tended to adhere to the process equipment, accounting for the low API content in the powder mixture and tablets. These results suggest that this API or the tablet composition was unsuitable for manufacturing in a continuous line; further testing could be continued with different materials and changes in the process.

Material Transport Characteristics in Planetary Roller Melt Granulation

Melt granulation for improving material handling by modifying particle size distribution offers significant advantages compared to the standard methods of dry and wet granulation in dust reduction, obviating a subsequent drying step. Furthermore, current research in pharmaceutical technology aims for continuous methods, as these have an enhanced potential to reduce product quality fluctuations. Concerning both aspects, the use of a planetary roller granulator is consequential. The process control with these machines benefits from the enhanced ratio of heated surface to processed volume, compared to the usually-applied twin-screw systems. This is related to the unique concept of planetary spindles flowing around a central spindle in a roller cylinder. Herein, the movement pattern defines the transport characteristics, which determine the energy input and overall processing conditions. The aim of this study is to investigate the residence time distribution in planetary roller melt granulation (PRMG) as an indicator for the material transport. By altering feed rate and rotation speed, the fill level in the granulator is adjusted, which directly affects the average transport velocity and mixing volume. The two-compartment model was utilized to reflect these coherences, as the model parameters symbolize the sub-processes of axial material transport and mixing.

Optimal quantification of residence time distribution profiles from a quality assurance perspective

Residence time distribution (RTD) has been widely applied across various fields of chemical engineering, including pharmaceutical manufacturing, for applications such as material traceability, quality assurance, system health monitoring, and fault detection. Determination of a representative RTD, in principle, requires an accurate process analytical technology (PAT) procedure capturing the entire range of tracer concentrations from zero to maximum. Such a wide concentration range creates at least two problems: i) decreased accuracy of the model across the entire range of concentrations, relating to limit of quantification, and ii) ambiguity associated with the detection of the tracer for low concentration levels, relating to limit of detection (LOD). These problems affect not only the RTD profile itself, but also RTD-based applications, which can potentially lead to erroneous conclusions. This article seeks to minimize the impact of these problems by understanding the relative importance of different features of RTD on the detection of out-of-specification (OOS) products. In this work, the RTD obtained experimentally was truncated at different levels, to investigate the impact of the truncation of RTD on funnel plots for OOS detection. The main finding is that the tail of the RTD can be truncated with no loss of accuracy in the determination of exclusion intervals. This enables the manufacturing scientist to focus entirely on the peak region, maximizing the accuracy of chemometric models.

Automatic system dynamics characterization of a pharmaceutical continuous production line

Continuous Manufacturing (CM) of drug products is a new approach in the pharmaceutical industry. In the presented paper, a GMP continuous wet granulation line for production of solid oral dosage forms was investigated in order to assess the system dynamics of the line and to define the best control and diversion strategy. The following steps were involved in the continuous process: dosing / feeding, blending, twin-screw wet granulation, fluid-bed drying, sieving and tableting. Two drug products with two different drug substances were compared during this study: one drug substance as model drug compound and one formulation of a currently evaluated commercial drug product. Several step tests in API concentration were performed in order to characterize the process flow and assess the process dynamics. API content was monitored in real time by Process Analytical Technologies (PAT) thanks to three Near Infrared (NIR) probes located along the process and measuring the API content after blender, after dryer and in the tablet press feed frame. The process parameter values were changed during production in order to detect the impact on the quality of the final product. An automatic residence time distribution (RTD) computation method has been developed in order automate the RTD calculation on the basis of process data to further define and monitor the system dynamics with the final aim of out of specification material diversion during the continuous production. The RTD has been seen as a process fingerprint: a change in the RTD values implies a change in the process.

Challenges in tracing material flow passing a loss-in-weight feeder in continuous manufacturing processes

Loss-in-weight feeders are an integral part of most continuous manufacturing processes, ensuring a constant mass flow. The feeders cause a significant degree of back-mixing in such lines. Understanding back-mixing is essential for the treatment of disturbances. However, feeders refilled semi-continuously contradict the common theory assuming steady-state. This study aims at understanding dynamic back-mixing and related phenomena. Low filling levels of a feeder are investigated using a fluorescent tracer. These investigations prove an impact of the addition of material probably caused by a non-uniform draw-in of the screws and dead material in the hopper. In turn, the dead material accounts for up to 50 % of the material in the hopper. Possible evidence of dead zones at higher filling levels and in feeders from literature are discussed additionally. Steady-state models from literature are extended to represent the observations and back-mixing at all filling levels. This extension reduces the root-mean-squared deviation of the model from the experimental data by 41%. The model predicts different responses to similar disturbances depending on the filling. This state-dependent back-mixing and the observed dead zones are challenging for diverting non-conforming material and material traceability. Therefore, these phenomena should be considered in selecting and operating feeders.

Tracking raw material flow through a continuous direct compression line Part I of II: Residence time distribution modeling and sensitivity analysis enabling increased process yield

Continuous manufacturing (CM) offers advantages in quality and space-time yield compared to common batch manufacturing. However, higher yield losses due to the start-up procedure make a broad application uneconomical. This work discusses the possibility of reducing yield losses by adjusting the degree of back-mixing. Back-mixing of nonconforming material from disturbances or start-up will result in the contamination of subsequent material. Therefore, higher degrees of back-mixing cause the discharge of additional material. Choosing an advantageous setting of operational parameters may be a simple way to change the degree of back-mixing. Based on direct compression, this work demonstrates the identification of promising parameters. Therefore, step-change experiments using color-marked material in the feeder, blender, and tablet press quantify the impact of three operational parameters per device. Models for the devices and the entire process result from those measurements. Subsequently, a global variance-based sensitivity analysis identifies the most influential parameters. As a result, adjusting the minimal filling level of the feeder and the rotational feed frame speed of the tablet press reduces back-mixing by more than 30%. At high costs of the raw materials, the resulting savings can significantly improve the economic performance of CM compared to batch manufacturing.

Using residence time distribution in pharmaceutical solid dose manufacturing - A critical review

While continuous manufacturing (CM) of pharmaceutical solid-based drug products has been shown to be advantageous for improving the product quality and process efficiency in alignment with FDA's support of the quality-by-design paradigm (Lee, 2015; Ierapetritou et al., 2016; Plumb, 2005; Schaber, 2011), it is critical to enable full utilization of CM technology for robust production and commercialization (Schaber, 2011; Byrn, 2015). To do so, an important prerequisite is to obtain a detailed understanding of overall process characteristics to develop cost-effective and accurate predictive models for unit operations and process flowsheets. These models are utilized to predict product quality and maintain desired manufacturing efficiency (Ierapetritou et al., 2016). Residence time distribution (RTD) has been a widely used tool to characterize the extent of mixing in pharmaceutical unit operations (Vanhoorne, 2020; Rogers and Ierapetritou, 2015; Teżyk et al., 2015) and manufacturing lines and develop computationally cheap predictive models. These models developed using RTD have been demonstrated to be crucial for various flowsheet applications (Kruisz, 2017; Martinetz, 2018; Tian, 2021). Though extensively used in the literature (Gao et al., 2012), the implementation, execution, evaluation, and assessment of RTD studies has not been standardized by regulatory agencies and can thus lead to ambiguity regarding their accurate implementation. To address this issue and subsequently prevent unforeseen errors in RTD implementation, the presented article aims to aid in developing standardized guidelines through a detailed review and critical discussion of RTD studies in the pharmaceutical manufacturing literature. The review article is divided into two main sections - 1) determination of RTD including different steps for RTD evaluation including experimental approach, data acquisition and pre-treatment, RTD modeling, and RTD metrics and, 2) applications of RTD for solid dose manufacturing. Critical considerations, pertaining to the limitations of RTDs for solid dose manufacturing, are also examined along with a perspective discussion of future avenues of improvement.

Digital twin of low dosage continuous powder blending - Artificial neural networks and residence time distribution models

In this paper we present a thorough description of the digital twin development for a continuous pharmaceutical powder blending process in accordance with the Process Analytical Technologies (PAT) and Quality by Design (QbD) guidelines. A low-dosage system of caffeine active pharmaceutical ingredient (API) and dextrose excipient was examined via continuous blending experiments. Near infrared (NIR) spectroscopy-based process analytics were applied; quantitative evaluation of spectra was achieved using multivariate data analysis. The blending system was represented with mechanistic residence time distribution (RTD) models and two types of recurrent artificial neural networks (ANN), experimental datasets were used for model training or fitting and validation. Detailed comparison of the two modelling approaches, the optimization of the model-based digital twin, and efficiency of the soft sensor-based process monitoring is presented through several validating simulations. Both RTD models and nonlinear autoregressive neural networks demonstrated excellent predictive power for the low dosage blending process. RTD models can prove to be more advantageous in industrial development as they are less resource-intensive to develop and prediction accuracy on low concentration levels lacks dependency from the precision of chemometric calibration. Reduced material costs and limited development timeframe render the digital twin an efficient tool in technological development.

Pharmaceutical application of multivariate modelling techniques: a review on the manufacturing of tablets

The tablet manufacturing process is a complex system, especially in continuous manufacturing (CM). It includes multiple unit operations, such as mixing, granulation, and tableting. In tablet manufacturing, critical quality attributes are influenced by multiple factorial relationships between material properties, process variables, and interactions. Moreover, the variation in raw material attributes and manufacturing processes is an inherent characteristic and seriously affects the quality of pharmaceutical products. To deepen our understanding of the tablet manufacturing process, multivariable modeling techniques can replace univariate analysis to investigate tablet manufacturing. In this review, the roles of the most prominent multivariate modeling techniques in the tablet manufacturing process are discussed. The review mainly focuses on applying multivariate modeling techniques to process understanding, optimization, process monitoring, and process control within multiple unit operations. To minimize the errors in the process of modeling, good modeling practice (GMoP) was introduced into the pharmaceutical process. Furthermore, current progress in the continuous manufacturing of tablets and the role of multivariate modeling techniques in continuous manufacturing are introduced. In this review, information is provided to both researchers and manufacturers to improve tablet quality.

Determination of Residence Time Distribution in a Continuous Powder Mixing Process With Supervised and Unsupervised Modeling of In-line Near Infrared (NIR) Spectroscopic Data

Successful implementation of continuous manufacturing processes requires robust methods to assess and control product quality in a real-time mode. In this study, the residence time distribution of a continuous powder mixing process was investigated via pulse tracer experiments using near infrared spectroscopy for tracer detection in an in-line mode. The residence time distribution was modeled by applying the continuous stirred tank reactor in series model for achieving the tracer (paracetamol) concentration profiles. Partial least squares discriminant analysis and principal component analysis of the near infrared spectroscopy data were applied to investigate both supervised and unsupervised chemometric modeling approaches. Additionally, the mean residence time for three powder systems was measured with different process settings. It was found that a significant change in the mean residence time occurred when comparing powder systems with different flowability and mixing process settings. This study also confirmed that the partial least squares discriminant analysis applied as a supervised chemometric model enabled an efficient and fast estimate of the mean residence time based on pulse tracer experiments.