Michael Choi Stuart C. Porter, and Axel MeisenPCTS, Inc.
The old saying that the only certainties in life are death and taxes is not quite true. The need for change is also a certainty in coated tablet manufacturing. Operations must adjust to varying conditions caused by changing markets, regulations, and business environments; improve production efficiency and competitiveness while ensuring and improving product quality; and modify operating conditions with minimum risk and production cost.
Making these changes triggers even more questions, like, do changes in production methods justify costs, and are the benefits of shifting from batch to continuous manufacturing worth the disruptions?
In-house experts and external advisors answer these questions based on their expertise and process knowledge but, in many cases, the answers hold still more uncertainties. Small and large-scale experimentation can provide greater certainty but is typically costly and time consuming. If the experimentation is conducted within a manufacturing plant, the costs and potential production losses are high.
What other approaches permit examining a wide range of possibilities and operating conditions quickly and cost-effectively before implementing them? Mathematical modeling may provide answers. In the past, solving mathematical models for film coating included non-deterministic approaches – e.g., discrete element method (DEM), computational fluid dynamics (CFD), and the Monte Carlo method (MCM). These required powerful computers, even with process simplifications and a staff of process modeling experts to estimate the outcome of coating processes. We devised a deterministic approach that integrates key elements of processing science and gives results almost instantaneously. The benefits of this approach are that executives, engineers, formulators, technicians, and operators can learn and use our Apps easily to improve cost and quality of film coating processes.
Our Approach
Our approach involves three main ‘elements’ of film-coating processes: process thermodynamics, process efficiency, and coating uniformity. Mathematical models for each element offer useful information on their own but, when combined, they also permit overall (i.e., multivariate) optimization of capacity, quality, and cost. These models and their benefits can be briefly described as follows:
1. Process Thermodynamics. This refers to the gas-liquid equilibrium, and mass and energy conservation relationships between the air and solvent under steady-state conditions. The ‘tuning’ variable for the thermodynamics model is unaccounted energy or ‘heat loss’. Our research1 showed that, for a given piece of equipment, heat loss can be assumed constant for wide ranges of typical operating conditions. Our approach uses the thermodynamics relationships with this tuning variable to bridge coating conditions between scales and equipment. It helps identify unfeasible process conditions during scale-up and optimization considerations, and in implicating or eliminating process variables as root causes for batch deviations.
Tuning Variables – coating efficiency, heat loss, and effective spray zone are the tuning variables for process efficiency, thermodynamics, and uniformity models, respectively. These model variables are assumed to be constant for model predictions but can be farther tuned as more information becomes available for different operating conditions and equipment.
2. Process Efficiency and Coating Efficiency. Process efficiency – often measured as ‘batch yield’ – is based on material accountability calculations. Batch yield and coating efficiency are not the same. Using the yield to characterize the effectiveness of coating processes can be misleading, because yield accounts for both tablet and coating weights, while coating efficiency only accounts for coating weight. Since a tablet core typically weighs significantly more than the coating on the tablet, any change in the coating weight can be overshadowed in the yield calculation. Therefore, coating efficiency should always be used to determine the coating process effectiveness.
Our approach uses coating efficiency as the tuning variable for the material accountability calculations. Our research2,3 showed that for a given nozzle setup/operation, tablet and coating suspension formulation, coating efficiency can be assumed constant. This tuning variable is then used to predict the coating weight gains with reasonable accuracy for different batch sizes and processing conditions. The tuning variable is updated whenever the process is changed to provide more precise predictions as the digital twin for process troubleshooting and improvements.
Digital Twin - a virtual representation that serves as the real-time digital counterpart of actual film-coating process and associated operations.
3. Coating Uniformity. This describes the consistency of distribution of coating material within (i.e., on the surface of) an individual tablet and between tablets. Prior research has shown that within and between distributions are closely related in film-coating processes,4,5 i.e., improving the coating variation between tablets also improves it within variation. Here, coating uniformity refers to the between distribution, which is typically measured using the coefficient of variation (CV) or relative standard deviation. Our research2,6 showed that, for a given nozzle setup/operation, the size of the spray zone can be assumed constant provided allowance is made for changes in tablet shape, size, and bulk density. Our approach uses the size of the spray zone as the tuning variable to bridge CV between processes, formulations, and equipment. This model (in effect a digital twin of the real process) can help build strategy and provide justifications for variable batch sizes, batch deviations, and tech transfers involving batch and continuous manufacturing technologies.
Combining and solving the above three models provides the means for finding optimal conditions when one or more process requirements change. Our initial focus on utilizing this capability was to build computer Apps that consist of these mathematical models and that relate material and processing parameters to production capacity, quality, and cost. These Apps illustrate the utility of the combined mathematical models for quickly visualizing these relationships and arriving at optimum conditions.
Application Examples
Changes to manufacturing operations are most often driven by changing markets, regulations, and business environments. However, the end goals are usually the same—to deliver a lower cost of goods by improving production efficiency while ensuring the same or better product quality. The journey in achieving these goals may be just as important in that the goals must be met with minimal guesswork, risk, and disruption to current production. This is where mathematical models come into play. Models can provide quick visualization of production rates within possible ranges of operating conditions. Let’s take an example where an increase in market demand calls for increased production.
Batch Size Change
Increasing batch size with existing equipment provides one way to boost production. But larger batch sizes also result in less uniform coating—unless the coating time is also lengthened. Therefore, increasing the batch size does not result in proportionate increases in production. The overall effect of batch size on production can be quantitatively determined with the aid of mathematical models. The modeling steps are:
1. Determining the values of the tuning parameters—coating efficiency, heat loss, and spray zone size for the currently operating equipment. Examples of process inputs and model results using our Apps are shown in Figure 1.
2. Calculating the production rates for the same CV and tuning parameters determined in Step 1 for a range of batch sizes. The Batch Capacity App shown in Figure 2 utilizes the combined models, inputs and calculations from Figure 1, and operational data (e.g., batch changeover time) to calculate the production rate as a function of batch size. Figure 3a shows the corresponding production costs. This App also calculates the conditioning of processing air needed to maintain the same coating environment—see Figures 3b and c, respectively, for process air flow rates and temperatures as functions of batch size.
Tech Transfers to Batch and Continuous Manufacturing
Opting for another piece of equipment, such as a batch process with larger pan size or a continuous manufacturing process, can also boost capacity. In either case, the same steps as Batch Size Change are used, except that Step 1 must also be repeated with the new equipment, since the tuning variables are equipment-specific. The analysis in Step 2 above is then carried out with the tuning variables for the new equipment.
Back to the earlier question ‘should an existing batch process be replaced by a continuous process?’, can the models be used to help decide between batch and continuous manufacturing? The answer is ‘yes’, even though continuous manufacturing technologies for film coating can vary significantly. For example, L.L. Bohle GmbH uses batch technology, Driam GmbH uses semi-continuous technology, and O’Hara Technologies Inc. and Thomas Processing LLC use truly continuous technologies. Despite these differences, equipment vendors claim to have taken steps to significantly reduce or eliminate the changeover times between batches. Based on these claims, we can run the models to see how reducing the changeover times affects the quality and cost of film-coating operation.
Figures 4a and b show model predictions of production rates and corresponding costs for the scenario where CV is fixed, which eliminates changeover time. Material and testing costs were removed from the analysis since these do not apply in the same way for continuous manufacturing. Two immediate observations arise: 1. operating cost decreases due to elimi- nating changeover time; and 2. production rate increases and cost decreases with smaller equipment fill volumes, opposite to the results shown in Figure 3a. The latter can be explained by higher spray coverage area per volume in shallower or smaller diameter equipment, resulting in more uniform coating per bed cycle that, in turn, leads to shorter coating time to achieve the same coating uniformity.6 It should be no surprise that all commercially available continuous manufacturing technologies use smaller diameters and longer beds compared to their batch counterparts. These observations point to significant savings in manufacturing cost by moving to continuous operation.
Once continuous manufacturing is chosen, one must make another decision regarding the technology. Our earlier work using these models6 showed that the lowest CV (i.e., best quality) could be obtained in batch mode while the lowest production cost is possible with the truly continuous mode. Therefore, the choice should first be made according to the product quality requirements. If more than one operating mode meets the quality requirements, the production cost is more favorable with technologies using semi-continuous or continuous mode of operation.
Formulation Optimization Based on Production Cost
In the examples thus far, the coating weight relative to tablet weight was kept constant. During formulation optimization or scaleup of aesthetic or protective coating processes where the coating weight is small relative to the tablet, coating weight is considered one of the optimization variables. In this case, the models can be used to analyze quality (i.e., CV) as a function of cost. The steps involved for this case are the same as those in Tech Transfers to Batch and Continuous Manufacturing but, rather than analyzing the effect of batch size on cost, the effects of coating weight gain and CV on cost are analyzed. Figure 5 shows an example optimization case where the coating weight gain of 3.5% resulted in the lowest production cost.
In cases where you change coating formulation, for example a switch to a higher solids coating solution, which can impact both coating uniformity and productivity,7 the models also predicted such formulation effect.6
Conclusion
Despite how well we develop and optimize film-coating processes, process changes during the product life are inevitable due to changing markets, regulations, and business environments. These changes represent not only opportunities to achieve the end goals of improving production efficiency and cost competitiveness, but also opportunities to arrive at these goals with minimal project risk and disruption to production. As we demonstrated, achieving both goals is possible by mathematical modeling for process thermodynamics, coating efficiency, and coating uniformity. The models’ capabilities to bridge and optimize processes under various process- and formulation-change scenarios require simple steps as demonstrated by the given examples. Once optimized, the models become digital twins for future troubleshooting and improvements.
Models also enable assessments of differences between batch and continuous manufacturing as well as the different modes of operation for continuous manufacturing. The model analyses showed the cost benefit of continuous manufacturing over batch manufacturing.
In addition to addressing capacity needs, the models, as digital twins, can be a powerful visualization tool for
- training operators, technicians, engineers, and scientists working on film-coating processes,
- building procedures, metrics, utility requirements, safety and environmental controls to support the film-coating manufacturing operation.
- building the control strategy for Regulatory filings for drug-layering processes.8
- executives whose concerns are focused on economics and profitability.
Details of the Apps and other wide-ranging uses not discussed in this article are available at www.particlecoating.com. Details on test-driving the Apps are also given at the site.
References
1. Choi, M. Applications of process thermodynamics in pharmaceutical coating. Tablets and Capsules. (2007) 4:1-10.
2. Choi, M., Porter, SC., Meisen, A., Interrelationships between coating uniformity and efficiency in pan coating processes. AAPS PharmSciTech, (2021) 22:265.
3. Choi, M. Determining the manufacturability of drug-layered tablets. Pharmaceutical Manufacturing. (2007) 6(4):34-42.
4. Freireich, B., Wassgren, C. Intraparticle coating variability: analysis and Monte-Carlo simulations. Chem Eng Sci. (2010) 65:1117-1124.
5. Brock, D., Zeitler, J.A., Funke, A., Knop, K., Kleinebudde, P. Evaluation of critical process parameters for intra-tablet coating uniformity using terahertz pulsed imaging. Eur J Pharm Biopharm. (2013) 85(3):1122-1129.
6. Choi, M., Porter, SC., Macht, B., Meisen, A. Novel coating uniformity models for tablet pan coaters. AAPS PharmSciTech. (2021) 22(7).
7. Porter, S.C. Trends in continuous film-coating processes. Tablets and Capsules. (2021) 24:5. 8. Choi, M., Porter, SC., Meisen, A., Application of mathematical models to determine the feasibility of amorphous drug layering in pan coaters. Pharmaceutics, (2022), 14(1), 149.
Corresponding author michael.choi@particlecoating.com