Continuous Flow Case Study – Efficiency in API Manufacturing
Synopsis: The webinar will be about reaction design and scale-up for a larger production. Attendees will explore the benefits of continuous flow reactions in pharmaceutical API & Intermediate manufacturing. Raybow has a long history in the field with experience in R&D as well as manufacturing. Raybow has multiple Continuous Flow reactors, the largest one of which produces more than 140 tons API per year.
Our guest, Alessandra Vizza from Corning, will share her knowledge of flow chemistry along with the multiple benefits it can provide as part of an API & Intermediate manufacturing strategy. She will also highlight the industrial unit type.
Scroll down to see answers to questions asked during the live event!
Presenters
Raybow's Dirk Hütten welcomes guest presenter Alessandra Vizza from Corning.
Dr. Dirk Hütten holds a PhD in Chemistry from the University of Cologne, Germany. He has more than 20 years of experience in global sales and marketing alongside the supply chain of pharmaceuticals and APIs including positions at GSK, Altana and PharmaZell. In 2019, he joined Raybow Pharmaceutical as Senior Director Business Development. In addition, Dirk is passionate about high performance sailing.
Alessandra Vizza is the Regional Business Director for Corning® Advanced-Flow™ Reactor (AFR) Technologies. She has been with Corning for the last twelve years managing business continuity for Advanced-Flow Reactors and ensuring technical and commercial contact with customers, as well market strategy and business development. Over the course of her career in flow technology, Alessandra has co-authored several articles and delivered numerous talks and workshops. She is currently focusing her activity on the successful implementation of industrial cases using AFR Technology.
Alessandra earned a master’s degree in Food & Technology from Università Statale di Milano. Alessandra is passionate about travel and understanding of other cultures.
Questions and Answers
The following questions were submitting during our live webinar. If you have further questions please contact Dr. Hütten at dirkh@raybow.com.
Corning: In general, the maintenance procedure and frequency are process dependent. Usually, the most critical equipment that requires maintenance are the pumps. For the reactor a regular check of the status of the gaskets can help avoiding potential issues and emergency failures.
Data like pump flow rate, dosing system pressure, and system temperature are recorded in the DCS (Distributed Control System) and the customer can get all the historical data on a regular basis. Our customers are also using DCS data analysis to make decisions regarding maintenance. The frequency of data analysis and maintenance depends on projects and customers procedures.
The customers will find some regularity after the system runs for a few months. Corning can assist customers to complete the procedure for maintenance by providing training based on our experience in other projects.
The most critical parameter to follow in a flow installation is the pressure at the pressure side of the pump head as any change of the pressure indicates a change of the system downstream. Quick changes will usually indicate a change in the process conditions, while a slow evolution usually indicates a change on the equipment, like fouling or corrosion.
As a general rule, it is good practice to clean and rinse the equipment before and after use. Of course, it is important to make sure the equipment is fully clean after the use of reagents. If the idle time lasts for a few days, no rinsing and keeping the instrument under a stable liquid may be OK. If in use on a daily basis, it is important to clean it regularly but, in this case, it is possible to leave it under solvent if the gaskets and other more sensitive parts can handle it well (e.g. no swelling). To choose the appropriate solvent for short term storage make sure there can be no fouling, precipitation, scaling or otherwise degradation of the solvent. For long term storage (longer than 1 or 2 weeks) it is recommended to dry out the equipment.
Raybow: At Raybow we tend to isolate in batch mode as we usually have subsequent work up procedures resulting discrete batches and lots, usually defined by the raw material input and the capacity of the collecting vessel. In general, the definition is also open to other factors, e.g. the FDA definition is to have uniform character and quality, within specified limits, i.e. production period, quantity of material processed, quantity of material produced or production variation while material traceability is a key element.
Corning: There are multiple possible answers as it really depends on the chemistry involved. If there is gas involved, the situation might be a bit more complicated to asses. However, as the reactors have been characterized, the specifications given are here to ensure that users are making most of the equipment capability.
Corning: In order to check if clogging might be an issue, the most critical parameter to check is the pressure. When reaching the steady state, it should remain constant. Otherwise, something is happening, and it is important to check this. It is also better to start working diluted or at higher temperatures and identify beforehand a good solvent to clean the equipment.
Corning: As for in batch, many spectroscopic methods can be used link to each specific process. Today device such as IR, Uvvis, Raman, Low field NMR are commonly used with flow chemistry system. HPLC and GC are slowly making their way but are still more often used as a more accurate offline analytics.
Corning: The choice of the pumps will depend on the type of fluid to be pumped, as rheological parameters of the fluid can play an important role on the flow. Also, the chemical compatibility and the flow rate range are very important when choosing a pump.
Therefore, the first step in the pump selection process is to check a material compatibility chart for all the materials in contact with the fluid. Use a strict compatibility chart that takes in account the localized corrosion and not only the general corrosion (as a chart dedicated for Coriolis flowmeters or other sensitive equipment). Care should be taken as the compatibility with different materials can vary with concentration and temperature. Also, the fluid velocity can play a role in the corrosion.
Some of the common materials for pumps that THF is compatible with are SS, Hastelloy, and PTFE. For the O-rings (if the pump contains some), FFKM would be a good choice. Care should be taken regarding PEEK .
Regarding the type of pump, in general for our reactors we are using gear pumps, multiplex membrane and multiplex piston pumps combined with precise flow metering and pressure sensors. For THF specifically we have used gear pumps.
Corning: Dealing with solids depends on many parameters, and the feasibility of a process involving solids depends mainly on the concentration of solids, the size and nature: if solids tend to be big, concentrated or eager to aggregate, the applicability might be very low.
Solids can be generated during the reaction or injected as slurries as raw material. Different technical solutions will be offered in the two cases, as the dosing system will have to be specially designed for slurries in the second case.
Many examples exist of multiphasic systems (liq/sol and even liq/gas/sol) which greatly benefit from flow technology. To give you a sense of how our mixer performed in presence of solids, you can check related articles available on our website (https://www.corning.com/worldwide/en/innovation/corning-emerging-innovations/advanced-flow-reactors.html).
Corning: Yes, it can. Corning AFR reactors up to G1 size have been designed to be easily configurable within a couple of hours at the maximum. When dealing with a G4 such as the one presented by Raybow, this is still possible but requires longer and more complex work. Still we are speaking of one or two days of work.
Corning: Material compatibility are very high with our Corning AFR reactor. All reactors up to the G3 are fully non-metallic systems and very few materials are involved, such as glass/SiC (link to the reactor model), PFA and some polymers for the o-rings (perfluoroelastomer and Kalrez most of the time). For the G4/G5 reactors inlet/outlet can be metallic (SS, Hy, Ti, Tl) as well as in PFA and even in SiC.
Raybow: Actually, this is primarily a question for your RA & quality department and of course the authorities. In my opinion, it depends primarily on the by-products / reactions that could occur. Typically, you will have an analysis protocol in order to be able to assess the quality of the product. GMP / regulatory concerns only arise in the event of deviations from this defined quality. If there are fluctuations in flow rate that affect quality, you may need to work on the stability of your process.
Corning: Yes, you can. Many publications exist about nano/micro particles synthesis using flow chemistry. We have already made some internal work with microspheres done to 10µm (not published results).
Corning: Cleaning is very simple using flow reactors as internal volume is low and Corning AFR reactor behave as plug-flow systems. A G4 reactor have an internal volume from 1 to 5 liters in general. In average using 10 times the internal volume of the reactors in cleaning solvent is enough to get rid of impurities. As for batch you need to go through a cleaning development and validate your procedure.
Corning: Usually, if the conditions allow it, a simple mixture with water/glycol can be excellent in term of heat exchange. Otherwise, there are a numbers of different silicon oils specialized in this. This is especially true in the case of cryogenic and high temperature conditions where water is not an option anymore.
Raybow: In addition to the above answer, for this reaction we used a heat transfer fluid called Unistat Thermal Fluid from Huber.
Corning: Cleaning of the equipment, as in the rest of the pharmaceutical world, can be performed similar to a batch. A cleaning solvent is used in the reactor and collected. Then a validated cleaning analytical method is run on the cleaning solvent. This can be used and has been FDA approved. Additionally, extra measures can also be performed such as control of the internal volume of the fluidic modules.
Corning: There is the possibility to add a back pressure regulator. This can keep the system down to a high pressure and thus minimize the volume of gas in the reactor itself. This limits the drawback from potential flush effects. Semi permeable membranes can also be an option although we do not have much experience with it.
Corning: The real specification is the limitation in pressure drop from the equipment, which is 18 bars. So the flowrate can be as high as it is below this limit. This strongly depends on the viscosity of the medium. For higher, we have bigger reactors designed for it. The flow rate can go as low as 2 ml/min (for a plug flow behaviour) in the Lab Reactor and up to the range of 20 L/min in the G5.
Corning: This really depends on the number of parameters involved. But, once the system is started, changing the conditions (except for the temperature which takes longer) and obtaining new samples can be performed very quickly at the lab scale (up to 1 sample collected every 5/10 minutes). Then the bottle neck tends to be the analytics. So this depends but can be performed in a few days in some cases.
Raybow: In most cases, you can also automate some of these changes and run the experiments around the clock. This then has also an impact on the number of experiments that you can carry out. Of course, this depends heavily on the chemistry studied.
Corning: The concept behind flow chemistry is process intensification. The change really depends on a case by case basis. We have customers who managed to get rid of the solvent and use the reagents neat. Additionally, in many cases, the concentration can be increased from 1 to 10 times as long as there is no issues with solids.
Raybow: Of course, this depends on the chemistry involved. However, we have seen typically savings of solvents in the region of 50-60%. If you can combine this with recycling solvents for the same reaction the impact can be significantly higher.
Corning: The maximal temperature is 200°C. The minimal one is in the range of -50°C. We have 2 sets of gaskets to cover the full range.
Raybow: For this reaction we had a maximum temperature of -30 °C and a minimum set at -50 °C.
Corning: Yes, AFR reactors can handle slurries. Solid handling should be addressed on a case by case basis. Dealing with solids depends on many parameters, including the size of the solid. This is reaction-dependent although the main trend is that these types of reactions may not be the best for flow reactors if solids tend to be big, concentrated or eager to aggregate. The particles should not stick to the reactor wall. When considering slurry solutions, the pumping system will be the critical point of the installation. This part of the unit will control how homogenous your solution will be for your reaction in the flow reactor. The design of the complete line needs to be adapted, from the choice and potential agitation needed in the storage vessel, to the sizing of the lines to keep a high speed, to the choice and design of the pump head, usually a membrane pump. There is no universal equipment configuration for pumping slurries, but each case will have to be treated separately.
Corning: Corning AFR equipment is provided with all requested material certificate in order to comply with FDA regulation (like CFR 21, §177.1780). Also software is in compliance with CFR 21 part 11 recommendations.
Corning: By knowing the flowrate from the pumps and the volume of the reactor, the residence time can be calculated. With gas-liquid systems, it is slightly different as the gas volume varies over time, in which case approximations have to be carried out and such approximations can influence the system.
However, various phenomena make this residence time difficult to estimate: the gas may be consumed if it’s a reactant, pressure gradient along the reactor, vaporization phenomenon, sliding factor between gas and liquid phase. Thus, another way to estimate residence time could be the apparent residence time.
For gas-liquid applications, the installation of a back pressure regulator at the reactor output to increase the global pressure of the system is highly recommended in order to increase the residence time, as the real gas flow rate is pressure and temperature dependent.
Corning: Typically, a scale-up entails a larger piece of equipment. Traditionally, the surface to volume ratio changes, impacting the heat exchange transfer capability of the system. The mixing properties tend to be strongly impacted as well.
Corning reactors have been designed to keep heat exchange, mass transfer coefficients and residence time distribution (RTD) consistent across all reactors, ensuring a fast-pace scale-up process. Hydrodynamic parameters and dimensionless numbers (such as Nusselt-Reynolds-Prandtl correlation for heat exchange coefficients) were considered when the plates were characterized and validated both through modeling and characterization in laboratory experiments. The flowrates range of each type of reactor was defined so that the heat exchange, mass transfer coefficients and RTD remains consistent. Thus, there is no need to calculate dimensionless numbers when scaling-up. The most important thing is to have the same residence time per plate in order to keep the same mixing and heat exchange capacity at larger scale. Additionally, one has to ensure working at the steady state where the conditions (pressure, flowrate, etc.…) are reproducible. In the scale-up phase, some parameters can be slightly tuned, but typically they do not involve a new optimization, hence a swift scale-up.
What is very important is to work in the reactors specifications to have a good mixing and good heat exchange and to ensure seamless scale-up (plug flow system, thus within the specified flowrate range).
Corning: reactor@corning.com
Raybow: dirkh@raybow.com