Developing a Modular Continuous Drug Product Manufacturing System with Real Time Quality Assurance for Producing Pharmaceutical Mini-Tablets

Abstract

The pharmaceutical industry has shown keen interest in developing small-scale modular manufacturing systems for producing medicinal products. These systems offer agile and flexible manufacturing, and are well-suited for use in situations requiring rapid production of drugs such as pandemics and humanitarian disasters. The creation of such systems requires the development of modular facilities for making solid oral drug products. In recent years, however, the development of such facilities has seen limited progress. This study presents a development of a prototype modular system that uses drop on demand (DoD) printing to produce personalized solid oral drug products. The system’s operation is demonstrated for manufacturing mini-tablets, a category of pediatric drug products, in continuous and semi-batch modes. In this process, the DoD printer is used to generate molten formulation drops that are solidified into mini-tablets. These dosages are then extracted, washed and dried in a continuous filtration and drying unit which is integrated with the printer. Process monitoring tools are also incorporated in the system to track the critical quality attributes of the product and the critical process parameters of the manufacturing operation in real time. Future areas of innovation are also proposed to improve this prototype unit and to enable the development of advanced drug manufacturing systems based on this platform.

Introduction

For decades, the pharmaceutical industry has relied on mass manufacturing to make drugs. This technique has allowed the industry to produce vast quantities of medicines while ensuring stringent quality control and low production costs. This has resulted in a large section of the world’s population gaining access to safe and reliable medication. However, mass manufacturing faces several limitations, such as slow production and fragile supply chains, which make it challenging to effectively respond to shocks in product demand. This is particularly problematic for situations where there is an urgent need for medication, such as during pandemics, humanitarian disasters, and battlefield medicine etc.¹⁸

To tackle this problem, a promising solution is to develop small-scale modular manufacturing systems that can produce pharmaceutical products on demand.¹⁴ These systems are envisioned to incorporate the latest advances in manufacturing technology, such as end to end processing, continuous operation, modularization, and real time quality assurance, to address the complexities associated with this production challenge. End to end processing involves conducting all parts of the drug manufacturing process in an integrated fashion at the same location. It enhances the supply chain robustness and speeds up the process by reducing inter-unit holdup and transportation times. Continuous manufacturing is a technique where all unit operations in the process are executed without interruption. It provides for significantly higher ‘volume flexibility’, i.e. flexibility in production volumes, thereby enabling a rapid response to demand fluctuations. Additionally, it also allows for operating the process with a smaller equipment footprint, in turn reducing the size of the manufacturing plant and making it easier to modularize and transport.^4,18 Modularization involves designing the system to be easily reconfigured and transported. It enhances ‘process flexibility’, allowing for individual equipment modules to be rearranged as required to quickly shift production from one product to another.⁵ It also facilitates quick relocation of the production facility, by enabling modules of each subprocess to be easily detached, shipped, and re-assembled at the new site without requiring special expertise to handle the instrumentation, communication, and control of each unit. Real time quality assurance involves measuring and controlling the critical process parameters (CPPs) and critical quality attributes (CQAs) of the product during the manufacturing process.^15,21 It helps in ensuring that product quality can be guaranteed as the product leaves the last manufacturing step, and that it can be immediately dispensed for use. This is vital because traditional means of quality assessment for medications cannot be applied for small scale systems without causing significant delays in product delivery or consuming a substantial portion of the generated product in destructive tests.

Indeed, many researchers have attempted to develop such small-scale systems, ranging from refrigerator size units to shipping container size systems.^2,20,27 However, most of these units are limited to synthesizing the drug substance only, or focus solely on end-to-end manufacturing for liquid formulations. There is a scarcity of research on small-scale end-to-end manufacturing systems designed specifically for solid oral drug products.²⁰ This is partly because of the challenges associated with incorporating traditional powder-based drug product manufacturing techniques into flexible and modular systems. Conventional drug product manufacturing involves a series of powder processing operations like powder feeding, blending, granulation, and tableting, to create the final dosage form. Achieving continuous manufacturing, real time process monitoring, and modeling for such systems is challenging, largely due to a limited understanding of the properties of powder blends, powder rheology and compaction physics.¹¹ Additionally, achieving swift transition between products is difficult, as this system may require changes in tooling and operating conditions to process different types of formulations and produce dosages with different sizes and drug loadings. Therefore, emerging techniques like 3D printing are gaining attention for building modular drug product manufacturing systems. 3D printing provides for great ‘product flexibility’ as it allows for production of drug products with different active ingredients, release behavior, sizes, and drug loadings, with the same apparatus and minimal operational changes.¹³

This study proposes the design of a modular drug product manufacturing system that can be easily integrated with other process units. Foundation for this system is a 3D printing platform called drop on demand (DoD) printing. In DoD printing, drug products are created by printing one or more drops of a liquid-based formulation onto an inert substrate⁷. This technique is well-suited for modular drug product manufacturing due to several reasons:

1. Mitigates powder handling: DoD processes a liquid formulation, which helps avoid many of the challenges associated with powder processing.

2. Integrated continuous manufacturing: The DoD printer is compatible with continuous operation and can be integrated with other upstream unit operations to form an end-to-end process. A previous study has demonstrated its integration with a continuous crystallizer through a novel unit operation called three phase settling.²⁵

3. Dosage personalization: DoD printing provides an excellent mechanism to tailor different aspects of drug products based on the patient’s requirements including drug loading, dosage form (capsule, tablet, edible film, or mini-tablet) and release profile (immediate, extended etc.). Thus, it can support different range of dosage types while requiring minimal tooling adjustments or downtime.

4. Automation and process monitoring: The DoD platform incorporates on-line sensors that can provide an accurate picture of the printing operation in real time. Its operation is also fully automated via a LabVIEW based interface.^9,10

Following the development of the modular drug product manufacturing system with DoD, this study proposes to demonstrate its operation through a case study to continuously manufacture mini-tablets for a test active pharmaceutical ingredient (API). In a previous study the DoD platform was used in producing a category of pediatric drug products called mini-tablets.²⁶ Conducting this process using a modular continuous system is beneficial for expanding the use of mini-tablets. Continuous processing can enable post production steps to be carried out with minimal labor requirement, and modular systems coupled with real time monitoring of CQAs can facilitate production of mini-tablets in distributed manufacturing locations such as hospitals and compounding pharmacies, making them more accessible to patients.²² This study, therefore, aims to demonstrate the operation of the modular DoD system by using it to establish a continuous manufacturing route for pharmaceutical mini-tablets with real time quality assurance.

To make the manufacturing of mini-tablets fully continuous, the modular DoD system is integrated with a continuous filtration carousel (CFC) unit. The CFC unit performs the post-production processing steps (filtration, washing, and drying) for the mini-tablets and dispenses them as ready-to-use dosages. To monitor CPPs of the process, an on-line camera is installed in the DoD system to measure parameters such as size, position, solidification time, and settling velocity of the drop. Additionally, monitoring of the CQA drug loading is accomplished by combining measurements of drop size (made by the camera) and formulation concentration (made by an ultraviolet (UV) spectrophotometer probe installed in the reservoir section of the printer). The following sections provide a detailed description of the modular DoD platform, continuous production of mini tablets, and real time monitoring of CQAs and CPPs in the mini-tablet manufacturing process.

Drop on Demand Printing of Pharmaceutical Mini-Tablets

DoD printing is a pharmaceutical additive manufacturing technique that creates dosages by printing droplets of an API containing formulation ink onto inert substrates. It can process a variety of formulations like solutions, melts, and suspensions, and can manufacture different dosage forms such as tablets, capsules, edible films, and mini-tablets.⁸ A previous study conducted by the authors demonstrates the application of DoD printing in manufacturing pharmaceutical mini-tablets. These are a small-size dosage form that are easy to swallow and are well-suited for use in pediatric patients. In comparison to traditional pediatric drug products like liquid dosages, powders and split tablets, mini-tablets offer increased dosing accuracy. Additionally, they retain many advantages of conventional tablets, such as longer-term stability, customizable dissolution behavior, and taste masking etc. Mini-tablets also support dosage personalization as they can be dispensed as single or multiple units.^12,24,28

The manufacturing technique employed in the previous study is maintained for the continuous manufacturing case under evaluation in this study (Fig. 1). The process begins with the preparation of a melt-based formulation, where the API is dissolved in the molten excipient. The materials used in experiments presented in this study are polyethylene glycol 2000 (excipient) and atorvastatin (API). The DoD platform then generates drops of the molten formulation with precise volumes, and prints them into a solidification chamber containing silicon oil (Xiameter PMX-200) at room temperature. Silicon oil is chosen as the bath solvent for its inertness with respect to both API and excipients. As the drop settles in the silicon oil bath, convective heat transfer causes it to cool, resulting in it solidifying completely by the time it reaches the bath vessel’s bottom. After solidification, the mini-tablets are manually extracted from the bath and washed with a lighter silicon oil, hexamethyldisiloxane (HMDSO). Subsequently, the dosages are dried under a hood to obtain the final consumable product.

Figure 1.

Manufacturing process for mini-tablets using DoD printing developed in the previous study by the authors. A) Schematic, and B) Apparatus.

The previous study demonstrated that DoD printing exhibits comparable performance to other mini-tablet manufacturing techniques like direct compression, and fused deposition modeling etc., when it comes to consistency in drug loading across dosages. The key advantage of utilizing DoD printing lies in its ability to leverage modularization, continuous operation, and real-time monitoring, which this study seeks to demonstrate.

DoD printing also possesses many limitations, such as 1) limited capability for producing dosages with high drug loadings (pump operations are inconsistent at high API loadings). 2) Lack of control on the polymorphic form of API crystals in the printed drug product. For melt-based formulations, the API may undergo a transformation in its polymorphic form during dosage solidification. This is undesirable as it can lead to unwanted/unstable crystal habits. 3) Limited availability of approved excipients. DoD requires safe-to-consume, inactive excipients that have low melting points close to room temperature, such materials are not widely available. The cleaning procedure for this system involves passing a cleaning solvent (typically pure ethanol) through the entire apparatus (pump, reservoirs, connecting tubes, and nozzle) to flush out any residual API. In an industrial setting, completion of cleaning can be determined by monitoring API concentration in the spent cleaning solvent. Once this value passes below an acceptable threshold, production of the next drug product can be commenced.

Building a Modular DoD Printer

Modular systems are characterized as highly flexible manufacturing facilities that can be rapidly reconfigured and transported on demand. These systems are designed to respond to sudden changes in market demand. They emphasize the compartmentalization of operational functions, enabling rapid structural adjustments to accommodate varying production scales (by adding machines) or different product types (by changing tooling or altering the production process).^16,17 In the context of pharmaceutical manufacturing, modular systems are well-suited for use in emergency scenarios, where these systems can be transported to the impacted locations, and rapidly installed and deployed, to produce the required medication.¹⁴ Modular systems can also serve in enabling distributed manufacturing of pharmaceuticals. Distributed or decentralized manufacturing is a production technique that advocates for making the product close to the end consumer. This technique is more adept at responding to demand fluctuations and is better placed to deliver personalized products to patients.^3,22

In collaboration with Neff Automation, USA, a movable containment unit was designed to house the modular DoD system. The dimensions of the unit are 0.81 m in length, 0.81 m in width, and 1.95 m in height, with an aluminum frame construction (Figs. 2A and 3A). It features two removable and adjustable platforms with a chemical-resistant coating, and the walls are made of polycarbonate material. Additionally, the unit includes an external bench to accommodate computer peripherals. Fig. 3A shows the DoD printer housed in the modular containment unit. It is organized such that all working equipment and components occupy the top bracket, with all controllers being located in the middle bracket. The bottom section contains the CPU of the computer, communication cables for the printer components, and power strips for the various equipment. Primary user interaction during operation is with the computer peripherals placed on the shelf in the rear. The unit is mobile and can be opened for integration on two sides after removing the side walls.

Figure 2.

A] Conceptual design of modular DoD system, 2] Conceptual modular integrated containment unit (MICU).

Figure 3.

Pictures of the equipment used. A] Modular DoD unit. The computer peripherals for this unit are placed on a shelf on the rear side. B] CFC unit modified to process mini-tablets. New pinch valves are installed. C] Integrated DoD-CFC system, connected by the suction line.

The overarching goal is to replicate similar units to accommodate other upstream unit operations such as reactors, solvent switch distillation columns, and crystallizers. These units will be integrated with the DoD system, which can then pave the way for the construction of a modular integrated containment unit (MICU). The MICU will serve as a mobile manufacturing facility specifically designed for the production of solid oral drug products (Fig. 2B).

Continuous Manufacturing of Mini-Tablets

Continuous Filtration Carousel Unit

The continuous filtration carousel (CFC) unit is an innovative filtration and drying system developed by Alconbury Weston Ltd (AWL) ^1,19 (Fig. 3B). Designed as an integrated system, it is primarily used to process slurries generated in crystallizers and transform them into dry filter cake, consisting of relatively pure crystals. Its working principle is as follows: a vacuum suction line is employed to draw the crystal suspension into a charge vessel positioned above the main filter body. The main body comprises of five rotating ports, resembling a carousel, that sequentially receives the slurry from the charge vessel in a semi-batch fashion. The operation begins by dispensing the crystal slurry from the charge vessel into a port, where the crystallization solvent is filtered out under a vacuum. The resulting filter cake moves to the next port for a washing step. Following the washing process, the cake passes through two ports for air drying. Finally, the cake is discharged from the filter body through a retractable piston. The operation of this unit is described in detail by Liu et al.¹⁹ and Acevedo et al.¹

Integration of DoD Printer and CFC Unit

In this study, the CFC unit is employed for a novel application to perform the post-manufacturing steps for pharmaceutical mini-tablets. The suction inlet of the CFC unit is connected to the bottom of the solidification bath. During operation, the solidified mini-tablets and the bath oil are transported to the charge vessel and subsequently into the filtration port where the excess silicon oil is removed. Next, a washing process is conducted using HMDSO, followed by air drying at room temperature. The mini-tablets are then dispensed as finished drug products ready for use. To achieve this goal, certain modifications have been made to the original CFC system. Given that silicon oil has significantly higher viscosity compared to typical crystallization solvents, a new filter plate with a larger pore size of 1 mm has been incorporated. Additionally, the pinch valves used in the original system are seen to be unable to support the passage of mini tablets through them. This is primarily due to the larger size of the mini-tablets in comparison to typical crystal sizes (approximately 10–500 μm) that the unit is designed for. To address this new air-actuated pinch valves, developed by AWL, have been installed. These valves are designed without any moving parts and thus enable the smooth flow of the larger mini-tablets.

For executing continuous processing of mini-tablets, the modular DoD system is moved and located close to the CFC unit (Fig. 3C). The operating conditions used in the CFC system are summarized in Table 1 below. Two modes of operation are studied: semi-batch mode and continuous mode. In the semi-batch mode several mini-tablets are printed and solidified as the first step, they then undergo processing en masse in the CFC unit (multiple mini-tablets per port). Mini-tablet production in this mode is independent of the dynamics of the CFC unit and thus can offer higher production rates. In the fully continuous route, mini-tablets are processed in the CFC system sequentially (one mini-tablet per port). Although its production rate is limited by the CFC cycle time and is slower, it offers the advantage of dosage tracking. Coupled with real time monitoring, this system can identify individual mini-tablets, determine their CQAs, and can track them through the post-manufacturing steps. It can even support control by rejection, where the mini-tablets with off-spec CQAs are discarded after passing through the CFC unit. This study demonstrates the integrated operation for both modes of processing.

Table 1.

Operating conditions for the CFC unit for processing mini-tablets.

CFC attributes	Value
Charge time	30 s
Filtration time	45 s
Wash time	5 s
Dry time	40 s
Fill time	30 s
Total cycle time	2.5 min

Real Time Quality Assurance

Real time monitoring serves as the ‘eyes and ears’ of the manufacturing process, allowing for observing critical parameters and ensuring that the process continues to stay on-spec during production. It forms a core component of the ‘Quality-by-design’ initiative promoted by the United States Food and Drug Administration. As a result, researchers have looked at incorporating process monitoring tools across various manufacturing platforms, including small-scale modular systems, 3D printers, and conventional production systems.^6,15,21,23 This section outlines the work done in monitoring CQAs and CPPs of the mini-tablet manufacturing process. There are several CPPs that are important in ensuring that the dosages produced through DoD printing meet regulatory quality requirements. These can be broadly classified into two categories: critical parameters in drop generation and critical parameters in drop solidification. Previous studies conducted by this research group have focused on developing effective tools to monitor the former, such as detecting the number of satellite drops generated and measuring the volume variation between successive drops. In this study, the CPPs of drop solidification process are primarily targeted for monitoring.

During drop solidification, several aspects of the process need to be considered:

1. Complete solidification: It is crucial to ensure that the drops fully solidify before reaching the bottom of the settling chamber. This prevents formulation pooling, which is the accumulation of unsolidified or partially solidified formulation in the chamber.

2. Smooth drop settling: The smooth settling of drops is essential to prevent aggregation, or clumping together of multiple-mini-tablets that cannot be easily separated.

3. Minimizing vessel currents: Flow currents within the vessel can negatively impact the proper solidification of drops, leading to premature departure from the chamber or potential adherence to the walls of the vessel. Therefore, it is crucial to mitigate flow currents to ensure that the settling and solidification of drops follow desired paths.

These aspects can be ensured by observing the drop solidification process in real time. For melt formulations, a color change from transparent to white indicates a solidification point. Measuring settling velocity and drop position can further help determine the presence of flow currents. Measuring the size of mini-tablet also allows for identification of drop aggregates. To make these measurements, a camera is installed in the modular system to observe the solidification chamber during operation (Fig. 4A). A backlight is placed behind the bath vessel to enhance the quality of image obtained. The operation of the camera and the images taken by it are processed on LabVIEW using the vision acquisition and vision assistant packages. These programs track the drop as it falls and record its size (radius) and position in pixel units. To obtain measurements in real units, a calibration curve is built by dropping spheres of known dimensions into the solidification bath and recording their pixel measurements (Fig. 4B).

Figure 4.

Apparatus and calibration curves used for real time monitoring of drug loading.

For drug products, some of the primary CQAs are API content, content uniformity, dissolution behavior, residual solvent content, and long-term stability. Ensuring high quality standards in these aspects allows for a shift towards quality by design. This study primarily focuses on measuring the CQA of drug loading consistency. The objective is to measure the drug loading for each mini-tablet produced, enabling determination of whether the product falls within the acceptable range of variation in content. To accomplish this, a real-time measurement of the API concentration in the formulation being printed is crucial. This is accomplished by placing a UV spectrophotometer probe within the DoD printer’s reservoir. The UV probe measures the absorbance of the formulation, and thereby provides a continuous real-time assessment of its concentration (Fig. 4C). Combining this concentration measurement with the size measurement obtained from the camera allows for the determination of drug loading in each mini-tablet produced.

$$\begin{gathered} measureddrugloading(mg)=formconc\left(\frac{mgAPI}{gdose}\right) \\ \ast dropsize(ml)\ast density\left(\frac{g}{ml}\right) \end{gathered}$$ (1)

$$truedrugloading(mg)=formconc\left(\frac{mgAPI}{gdose}\right)\ast minitabletwt(g)$$ (2)

To evaluate the accuracy of this technique in measuring drug loading in mini-tablets, a validation run is conducted. Four sets of mini-tablets with varying sizes and API concentrations are manufactured, and the drug loadings are measured using the proposed method. The measured drug loadings are then compared with the true values, which are calculated by multiplying the measured API concentration with the post-manufacturing weight of the mini-tablets.

Materials and Methods

Materials

Xiameter PMX-200 silicone fluid 1000 cSt (dimethicone) (Dow Chemical company) (Silicon oil) is used as the solidification bath solvent. Hexamethyldisiloxane (HMDSO) (Thermo Scientific) is used as washing solvent. Polyethylene glycol (PEG) 2000 (Tokyo Chemical Industry, TCI, America) is used as the excipient. Atorvastatin (Dr. Reddy’s Laboratories) is used as the active ingredient. Hexane (Thermo Scientific) is used as the cleaning solvent for the CFC unit.

Instruments

A MSC621 Carl Zeiss UV/vis spectrophotometer and a Hellma ATR probe (type 661.822-UV) are used to determine API concentration in the formulation. A Waters ultra-pressure liquid chromatography (UPLC) system with a UV detector is used for calibrating the UV probe. A Nikon Eclipse E600 microscope is used to image the mini-tablets and the ImageJ software package is used to determine its size. An Ohaus Explorer Analytical balance is used to weigh the mini-tablets. An Allied Vision Mako 158B camera with a Moritex 25 mm lens 12 MP is used for real time monitoring of drop solidification. The camera’s operation and data collection are performed using a LabVIEW programming interface, National Instruments. The backlight used for the camera is a Gagne light panel 8.5 × 11 in.

Results and Discussion

Continuous Manufacturing Unit

This modular unit is transported near the CFC system and integrated with it (Fig. 3C). Mini-tablets of the drug atorvastatin are made following the procedure described previously. The integrated process is run for 20 min in the semi-batch mode and for 30 min in the continuous mode; the rate determining step in the process is the CFC unit’s cycle time (2.5 min). For continuous operation, the DoD platform prints 1 mini-tablet for each CFC cycle, and a total of 10 dosages are produced in this run (Table 2). For semi-batch operation the DoD platform prints 5 mini-tablets for each cycle of the CFC unit, and a total of 25 mini-tablets are produced in this run. Production is lower than the maximum rate due to the mini-tablets occasionally getting stuck in the suction line, thereby necessitating additional cycles to ensure their complete transfer to the CFC. Photographs from this process are shown in Fig. 5 below. Mini-tablets are discharged one by one, maintaining their order of printing, in the continuous manufacturing process (Fig. 5C) whereas they may be discharged in batches of mini-tablets, with their order of printing undiscernible, in semi-batch mode (Fig. 5D). Therefore, the mini-tablets manufactured via the continuous mode can be tracked throughout the process, but mini-tablets made via the semi-batch mode cannot. Startup time for the system is around 15 min and cleaning time is around 5 min. Thus, production can be shifted from one drug to another in under 20 min.

Table 2.

Summary of results from the DoD-CFC integrated operation.

Parameter	Semi-batch run	Continuous run
Operating time	20 min	30 min
# dosages	25	10
Dosages/cycle	5	1
Max Dosages/cycle	30	1
Dosage tracking	No	Yes

Figure 5.

Snapshots from the integrated DoD-CFC operation. A] Printing of 5 mini-tablets in each CFC cycle for the semi-batch mode. In continuous mode, only one tablet is printed per CFC cycle. B] Transport of mini-tablets via the suction line to the CFC unit. C] Release of mini-tablets post filtration, washing, and drying operations for the continuous mode. Note that the mini-tablets are released one-by-one in the same order as they are printed. D] Release of mini-tablets in the semi-batch mode. Note that the mini-tablets are released in groups, often the order of their generation cannot be discerned. E] Mini-tablets flattened during CFC operation.

Real Time Quality Assurance

Monitoring CPPs

A sample measurement made by the camera during mini-tablet solidification is shown in Fig. 6 below. Table 3 shows the measurement file obtained from the camera. The solidification of drops as captured by the camera is also shown in a video file attached as supplementary material. Initially, as the drop settles in the bath, it is fully liquid and is not detected by the camera. Gradually as the drop begins solidifying, its detected size increases until it plateaus at a maximum. At this point, the drop is fully solidified, and the time elapsed in the process is the solidification time for the drop. Fluctuation in the solidification time can indicate that the solvent bath is heating up or that the formulation printed into the vessel is not heated uniformly. The absence of a plateau can also indicate incomplete solidification of the mini-tablet (Fig. 6A, Incomplete drop solidification). The bottom part of the settler chamber is cropped out of the camera image to avoid reflection from the floor interfering with the measurement. Thus, as the drop leaves the frame, the size measured by the camera decreases (Fig. 6A, drop leaving the image frame). Sudden increases in drop volume can indicate aggregation of mini-tablets (Fig. 6A, large aggregated drop). The larger drop has a higher solidification time because it has a higher heat capacity compared to the regular sized drop and thus requires a longer time to cool by convection.

Figure 6.

Sample measurements of drop solidification, observed from the camera. A] Drop size, B] Drop Position, C] Drop velocity.

Table 3.

Sample readings from the measurement file generated by the camera. It records the radius of the drop detected, its x and y position, the number of drops it has detected, and the time elapsed in the experiment.

Drop radius (mm)	x-drop (mm)	y-drop (mm)	# circles detected	time (ms)
1.350	17.619	17.619	1	29,223
1.350	17.619	17.736	1	29,481
1.350	17.619	17.892	1	29,751
1.350	17.580	18.048	1	30,012
1.350	17.658	18.204	1	30,272
1.350	17.619	18.321	1	30,538
1.350	17.619	18.477	1	30,803

To assess if the size measurements made by the camera are accurate, a validation experiment is carried out (Table 4). In this experiment, the tablet radius measured by the camera is compared with the ‘true radius’ of the dosage. This is obtained by imaging the dosage under a microscope camera and then using the ImageJ software tool to compute its size. It is observed that the average error between the measured radius and true radius of mini-tablets is 2.3 % with the maximum error being 4 %, equal to a difference of nearly 50 μm. The error triples when it comes to measuring error in predicted volume due to their cubic relationship, resulting in an average error of 6.9 % with a maximum error of 12 %. This error is mainly caused by limits to the resolution of the camera, and the approximation made that the mini-tablets are perfectly spherical.

Table 4.

Results for drop size and drug loading measurements.

True radius (mm)	Measured radius (mm)	Error radius (%)	True vol (mm3)	Measured vol (mm3)	Error vol (%)	True API (mg)	Measured API (mg)	Error API (%)
1.188	1.167	1.76	7.03	6.67	5.19	0.684	0.777	11.93
1.207	1.182	2.12	7.38	6.92	6.22	0.71	0.785	9.52
1.279	1.227	4.07	8.78	7.74	11.74	0.745	0.802	7.08
1.225	1.201	2	7.71	7.26	5.89	0.745	0.785	5.07
1.265	1.228	2.92	8.49	7.76	8.53	0.747	0.817	8.52
1.267	1.233	2.69	8.52	7.85	7.87	0.756	0.817	7.5
1.238	1.21	2.21	7.95	7.43	6.48	0.762	0.8	4.73
1.271	1.244	2.1	8.61	8.08	6.17	0.778	0.846	8.13
1.296	1.25	3.56	9.13	8.19	10.31	0.788	0.832	5.27
1.186	1.181	0.39	6.98	6.9	1.18	0.793	0.835	5.29
1.376	1.323	3.84	10.91	9.7	11.1	0.884	0.997	11.3
1.404	1.388	1.2	11.61	11.2	3.56	1.02	1.045	2.39
1.427	1.377	3.49	12.18	10.95	10.12	1.084	1.258	13.76
1.477	1.514	2.48	13.52	14.55	7.65	1.326	1.213	9.35
1.553	1.532	1.39	15.75	15.01	4.12	1.372	1.366	0.45
1.518	1.517	0.08	14.67	14.63	0.26	1.449	1.515	4.37
1.517	1.548	2.04	14.63	15.55	6.27	1.54	1.515	1.65
1.55	1.563	0.85	15.68	16.05	2.6	1.586	1.606	1.28
1.533	1.591	3.81	15.09	16.88	11.89	1.673	1.538	8.74
1.599	1.648	3.11	17.12	18.78	9.64	1.711	1.603	6.72

The other measurements obtained by the camera include x and y position of the drop and its settling velocity (Figs. 6B and 6C). In a settling chamber with no flow currents, drop descent should be strictly vertical, that is in the y direction. Large changes in the x position of drops can indicate the presence of flow currents. Settling velocity of drops can be calculated as the derivative of y-position of the drop with respect to time. As the drop falls, its settling velocity should attain the constant value of terminal settling velocity. This can be used as a soft sensor for monitoring the viscosity of the bath solvent. Viscosity can be computed from terminal settling velocity through the following equation, assuming that Stokes’ law is valid. This is important if a blend of silicon oils with different viscosities is used as the solidification bath solvent and its viscosity is not experimentally known. For this sample measurement, the silicon oil viscosity is estimated to be around 1.10 Pas which is close to the true value of 0.95 Pas. These results thus demonstrate that the camera installed in the DoD system is effective in monitoring a variety of CPPs in the mini-tablet manufacturing process.

$${\eta }_{Sioil}=\frac{{gd}^2}{{18v}_t}\ast ({\rho }_{MT}-{\rho }_{Sioil})$$ (3)

Measurement of the solidification dynamics can also be used to control the solidification rate (and consequently API crystallization rate) of mini-tablets. Temperature of the solidification bath can be adjusted based on the dynamics desired, including the creation of multiple temperature zones in the bath. This in turn allows for controlling various critical quality attributes of the drug product such as polymorphic form of API in the final dosage, degree of crystallinity (ratio of API in crystalline and amorphous states), and even crystal size distribution of API in the manufactured mini-tablets. This is aligned with the approach taken by¹⁰ where crystallization of API is manipulated by controlling the cooling rate that the dosage is subjected to. Incorporating these controls can allow for individualizing the dissolution behavior of drug products based on the needs of the patients.

Monitoring CQAs

This section focuses on the measurement of drug loading in the mini-tablets. As discussed in previous sections, drug loading is determined by combining two measurements: drop size and formulation concentration. The actual drug loading in the dosages is calculated by multiplying the weight of the printed mini-tablets with the concentration obtained from the UV probe. The measured drug loadings and their comparison with the true drug loadings are summarized in Table 4. The results show an average error of 6.53 % between the measured and true drug loadings, with a maximum error of approximately 13.7 %. The majority of this error can be attributed to the inaccuracies in measuring drop size.

Though the error in drug loading measurement using the UV probe-camera system is relatively high, it can still be employed as an alarm system to verify if the mini-tablets are being produced within acceptable bounds for drug loading. Nevertheless, for more precise monitoring, the use of a higher-resolution camera would be necessary to reduce measurement errors.

Design Improvements to the System

This section discusses the design changes that can be implemented in the system to further enhance its operation. Starting with the design of the modular DoD system; currently, the main limitation in its design is the lack of a proper ventilation mechanism. This can be addressed by installing an exhaust fan at the top of the unit. It is however crucial to ensure that the printing apparatus is sufficiently isolated, in order to avoid air currents affecting the drop formation. Another area of improvement is in the organization of the unit. It is preferable to host electrical components on the higher brackets to afford protection from process solvents. This however leaves the process equipment in the bottom compartment which makes them difficult to access for cleaning and set-up. Automated clean-in-place procedures can be developed for the DoD system to overcome this challenge. The side shelf dedicated to computer peripherals can be eliminated if a touch screen monitor is installed on the front system wall. This modification would create an additional connection side to which the unit can be integrated, while also making the unit more compact. Lastly, to extend the mobility of these systems, the installation of a portable power source can be considered. This addition would allow for greater freedom of movement and flexibility in deploying the unit to various locations. Efficient tube heaters and solidification chambers with minimal dead volumes can be installed to prevent formulation solidification in-line and facilitate smooth tablet transfer respectively.

Some design changes can also be made to the CFC unit to improve its efficiency in processing mini-tablets. During operation, it is observed that in a few instances the mini-tablets are flattened in the CFC unit (Fig. 5E). This happens when the dosages are trapped between the filter plate and the rotating carousel. This problem can potentially be solved by reducing the gap between the rotating carousel body and the filter plates. The washing efficiency also needs to be improved to completely eliminate the silicon oil from tablet surface; this can be achieved either by optimizing the wash cycle or by using a less viscous oil in the bath that is easier to wash off. Additionally, a valve can be introduced in the system to separate the filtrate outlet from the washing outlet. The filtrate (which is largely pure silicon oil) can then be recycled back as clean bath solvent and can minimize the solvent waste generated by the process.

To improve the accuracy of drug loading measurements, a higher-resolution camera can be installed in the system. The camera currently employed has a resolution of 1.58 megapixels (MPs); to improve image quality cameras with a resolution of 12 MP or more can be installed in the system. Another source of error in this measurement stems from the assumption of spherical mini-tablets. This assumption is not entirely true as some surface imperfections can at times be observed on the dosage (Fig. 7). Accuracy in measuring drug loading can thus be further improved by installing two cameras to observe the bath vessel, thereby viewing drop formation from two angles and allowing for calculation of the 3D structure of the drop without assuming sphericity.

Figure 7.

Mini-tablet images taken under a microscope showing surface imperfections. A] small-protrusions on the tablet. B] Slightly elongated tablet with undulations on the surface. These imperfections reduce the accuracy of size measurements made by the camera.

Conclusions

In conclusion, this study presents the development of a modular manufacturing system for pharmaceutical drug products using drop-on-demand printing. The operation of this system has been demonstrated in a case study to continuously manufacture mini-tablets for the drug atorvastatin. Two modes of operation are evaluated, semi-batch manufacturing and continuous manufacturing. While the former offers higher production rates, the latter allows for dosage tracking and control by rejection. Additionally, the study explored the incorporation of real time quality assurance tools in the DoD system to observe the CPPs and CQAs of the mini-tablet production process.

Development of the modular DoD system opens up numerous exciting opportunities for future research and innovation in the field of pharmaceutical manufacturing. One promising avenue is the integration of this system with other modular units housing upstream operations for drug substance synthesis, enabling the creation of an end-to-end modular manufacturing unit. Investigating the production efficiency of this integrated unit under various manufacturing scenarios, such as optimizing campaigns for producing different drugs, and studying the responsiveness of a network of such facilities to demand fluctuations, can provide valuable insights for development of distributed manufacturing platforms. Furthermore, this integration can be extended to the virtual realm through the development of a digital twin for the modular system. The digital twin can be leveraged to develop control loops, guide operational decisions, and optimize performance. Additionally, development of a sensing tool capable of predicting the dissolution behavior of drug products is crucial. This can be in the form of a soft sensor that takes inputs like the API, excipient, and particle size distribution to estimate the dissolution behavior of the printed drug product. Quality assurance for all product CQAs needs to be provided to facilitate immediate use of the manufactured drug products by patients. Incorporation of these technological advancements will pave way for development of pharmaceutical manufacturing systems with enhanced efficiency, agility, flexibility, and quality assurance, that are capable of satisfying the requirements of patients in the future.

Abbreviations:

API

active pharmaceutical ingredient

DoD

drop-on-demand

PEG

polyethyleneglycol

HMDSO

hexamethyldisiloxane

Silicon oil

Xiameter PMX 200 (dimethicone)

UPLC

ultra pressure liquid chromatography

UV

ultraviolet

CFC

continuous filtration carousel

AWL

Alconbury Weston Limited

MICU

modular integrated containment unit