Identification of key issues in the transition from laboratory to pilot plant scale

Thread Content

The transition from laboratory research to scale-up in a factory is a process that every manufacturing process must go through, and one that every process researcher must consider and experience. The ultimate goal of laboratory research is to pave the way for production; the scale-up process serves as a link between laboratory research and actual factory production. This scale-up process has two important purposes: analyzing and resolving the process-related issues encountered in the laboratory, as well as enabling large-scale production. However, due to the risks associated with the scaling-up process, every researcher involved in the project should follow these principles when preparing for and carrying out process scaling: A: Every researcher involved in the project must have a thorough understanding of the process ； B: Every R&D personnel involved in the project must ensure the safety of process scaling ； C: Researchers should discuss with those responsible for process scaling the potential issues that may arise during scaling, including control points and extended processing times in the process, mixing issues, as well as problems that may occur outside of the process itself ； D: R&D personnel and those responsible for process scaling must ensure that the quality and quantity of raw materials are sufficient to produce a sufficient amount of qualified products ； E: The equipment available in the factory must be capable of meeting the requirements of the production process ； F: Develop production processes based on laboratory research, and strictly scale up these processes while closely monitoring the entire production process ； G: By observing the critical state of the reaction online, this is particularly important when doing it for the first time, in order to observe the differences between the amplified phenomenon and those observed in laboratory reactions ； Skilled personnel and good equipment are key for laboratories to address scaling issues. It is much less stressful to resolve problems in a pilot scale reaction than to do so on a larger scale. When moving from the laboratory to scale-up, various factors need to be taken into consideration in order to avoid experimental failures caused by technical issues. In scale-up reactions, it is often necessary to extend the duration of the process operation, which is related to the operation time in pilot tests; transferring large amounts of solvent and solid raw materials requires more time. The rate of heat transfer is proportional to the (surface area/volume) of the reactor. Therefore, larger reaction vessels require more time to reach the reaction temperature. Most reaction vessels are cylindrical; their surface area is proportional to the radius, while their volume is inversely proportional to the square of the radius. This will ultimately lead to an increased amplification time. If the yield of the reaction is low or its rate is slow, extending the reaction time often leads to an increase in by-products. For example, in the esterification reaction, the reaction is catalyzed by carboxylic acids, and the removal of water through azeotropy can accelerate the reaction, as an increase in acid concentration favors an increased reaction rate. By extending the reaction time, the product may be contaminated by high concentrations of impurities, which can result in a slower crystallization rate than expected. If the operation time is extended excessively, it may also lead to the formation of thermodynamically stable but unintended isomers. The third problem in the amplification reaction is difficulty in mixing. Achieving a completely uniform mixture is not something that happens in an instant. Putting the raw materials into the reaction vessel results in either homogeneous mixing or phase separation. If stirring is uneven, it leads to excessively high local concentrations of the reactants, which in turn causes faster local reaction rates and quicker heat release, resulting in abnormally high local temperatures. This often leads to an increase in by-products. Special attention must be paid to heterogeneous conditions during the post-treatment and separation processes; if mixing is insufficient, it is difficult for impurities to be extracted from one phase to another. The residues remaining in the reactor after extraction may contaminate the product. Insufficient processing time can prevent the two phases from separating completely, leading to emulsification and contamination of the product with an emulsion. Undetectable cleaning agents may remain suspended in the extraction phase, causing contamination. The discussion covers the practical issues to consider when moving from laboratory development to scale-up in a factory: A: Defining the goals for scale-up and safety factors. The first step in preparing for process scale-up is to have clear objectives, knowing the quality and quantity of the product required as well as which is more important, and whether the process meets the requirements for scale-up. To ensure the safety of laboratory research and scale-up, it is necessary to comprehensively assess the hazards of chemical reactions in the process; not all reactions require thorough analysis. For example, the hydrolysis reaction of lipids does not pose a risk in moderately alkaline solutions; tests can be used to determine how much heat is released during the reaction and whether such release could lead to hazards. The reduction reaction is an exothermic reaction; in laboratory-scale tests, no significant exothermic effect is observed. However, on a larger scale, it often results in a intense exothermic reaction. If the temperature is not properly controlled, the heat released during the reaction can cause the solvent to boil violently, and there is even a risk of explosion if the heat generated by the reaction cannot be dissipated in time. Furthermore, the safety of the operators is also an important factor to consider, especially the safety of those who handle the materials and separate the final products. For example, in industrial processes for extracting alkaloids from plants (such as aconitine), ethanol is often used instead of methanol, as methanol poses a much greater risk to human health than ethanol. B: Identify the key process steps. Before drafting the process specifications, it should be discussed with the researchers involved in the process development. The process specification should take into account every aspect of the manufacturing process. If the feeding speed is important, then how can it be controlled? ； Should the reagent be placed above or below the reaction solution? ； Will the reagent solidify into a solid when added to a reaction at low temperature? ； What ranges should the reaction temperature and moisture content be kept within? ； How to control the end point of the reaction? By observing the reaction phenomena ； TLC monitoring or developing a final HPLC method ； It is possible to exert strict control over the recrystallization of intermediates and establish relatively strict and reasonable quality standards ； Issues such as whether specialized equipment is needed to separate and dry the final product must be considered and discussed. C: Limiting the scope of device use. Most equipment used for amplification in factories is multi-purpose, and specialized equipment is rarely used. In fact, the amplification equipment used in factories all has limited usage ranges. For example, in a certain project we have a step that requires operation at low temperatures in an oxygen- and moisture-free environment, and n-butyllithium is used in this process; therefore, the equipment we choose must be able to withstand low temperatures as well as corrosion from strong alkalis. Special reaction equipment is needed for this purpose. It is important to ensure that the pipes used for transferring the n-butyllithium solution, along with the sealing plugs and detectors, meet the requirements of production ； Another example is that hydrogenation reactions often require high-pressure hydrogenation reactors; in strong acid environments, metal reactors cannot be used for such reactions, as this can lead to corrosion. D: Using online monitoring in amplification reactions. Online monitoring is important for effective amplification; however, it is difficult to collect representative samples from the reaction vessel due to the need to avoid harm to the operator’s health, impacts on the reaction, and contamination of the product as well as constraints related to the reaction vessel itself. Especially when the reaction is heterogeneous, more factors need to be taken into account to obtain representative samples. After the sample is removed from the reaction system, appropriate separation and analysis of the reaction must be carried out. The sample may contain key components such as raw materials, products, impurities, and solvents that need to be quantified. E: Designing emergency response plans for incomplete reactions and uncontrolled reactions. As part of process development, it is necessary to design emergency plans to deal with situations where a reaction does not complete within the specified time. Under normal circumstances, the reaction time needs to be extended until the reaction is complete. It is also necessary to proceed without reagents, so a safe scheme for adding reagents during amplification must be designed. For safety reasons, when a safe operating condition cannot be developed in a timely manner or when the reaction involved releases large amounts of heat, emergency response plans must be devised for potential out-of-control reactions. In this case, it is best to use a safety bottle; no matter what reaction occurs, any liquid that is ejected from the reactor will be absorbed by the safety bottle. F: Determine the impacts of extending or interrupting the process. Solving route issues in the laboratory is one of the key steps for successful scale-up. The factors affecting magnification can be detected in laboratory pilot tests. For example, a portion of the reaction mixture is taken from the original reactor, kept at the original reaction temperature for a time that is twice as long as the usual reaction time, and analyzed using online analysis methods to determine the end point of the reaction. To simulate the time required during an extended concentration process, take the volume or water content needed to concentrate a portion of the extract, and then use data at that concentration temperature to predict the time required for the scaling-up operation. To determine whether impurities will crystallize together with the product, this can be done by keeping a portion of the slurry adiabatically and then analyzing the crystals. The effects of the feeding process are similar to those in the laboratory, where nodes in the process are determined using such a gradual approach. G: Research approach for specific chemicals Before proceeding with scale-up reactions that involve significant resources and expensive chemicals, it is necessary to examine the reaction to determine whether the desired outcomes can be achieved, along with products of the expected yield and quality. Reactive components include all the compounds used in the reaction: starting materials, reagents, and solvents. Reagents and solvents of different grades, commercial raw materials of different batches, and intermediates of different batches should all be tested. There are two basic methods for detection: analyzing the substance to be detected and through a single experiment. Chemical analysis is a method for detecting reactants and products. Simple analyses include pH measurement, titration, comparison of infrared spectra with standard spectra, determination of melting point, determination of refractive index, etc. These data are compared with standards to determine the mass of the compound. The quality of products produced locally should be above the minimum standard. The content of such impurities should also be kept at the lowest possible level. Finally, the data obtained at each step of the preparation are summarized. The danger that arises during the analysis of the product is mainly due to the fact that this compound has not been classified as a hazardous substance. It can be detected in experiments, ensuring the quality of the product. If the testing method for one of the steps is not satisfactory, the method from the previous step can be used, the procedure can be modified, or wait until an alternative method becomes available that meets the standards for subsequent batches. Intuition is often helpful, but one cannot rely on it entirely. If there are doubts, experiments must be conducted, data must be collected, and let the data speak. H: To determine the maximum tolerance for the amplified reaction. If time permits, certain parameters of the reaction need to be monitored, including acceptable operating conditions, yield, and product quality. In the factory, the volume error of the reagents added is within 5%, while the quality error is between 1% and 2%; this range is acceptable for future experiments. If this range is not achieved, the precise method of adding the reactants needs to be explored. If time permits, destructive experiments can be conducted. In these tests, the reagent is added in excess or below the required amount to ensure the allowable error margin. To ensure conditions similar to those in normal experiments, including pH value, temperature, and other parameters, useful data can be obtained through monitoring, analysis, and comparison. I: Ensure that analytical methods for the final product are established. Once the amplification reaction is complete, a series of product analyses must be carried out: purity, impurity content, residual solvents, ash residue, heavy metal residues, crystal form testing, etc. Yield, purity, and quality can all be used to calculate reaction equilibrium. J: Define the cleaning processes and waste treatment procedures. Cleaning processes are becoming increasingly important, as they can be used to reduce contaminants in products across different batches, especially when various products are produced in multi-functional reactors. The best method for the cleaning process is to dissolve the reaction product and then transfer it. The reaction solution needs to be treated properly; acidic or alkaline solutions must first be neutralized before further processing can take place. The solvents used in the reactions are generally recycled, and it is then determined whether these recycled solvents can be reused. If they do not affect the reaction, appropriate quality standards are established for them. K: Guidelines for drafting a technology transfer document. Often, the main goal of process development is not to achieve a certain scale, but rather to explain the process clearly so that others can replicate it successfully and obtain the final product. Between R&D laboratories, kilogram-scale laboratories, test workshops, and factories ； Between the R&D laboratory and the custom manufacturing plant ； Technology transfer can occur among peers. A technology transfer document should transfer process information efficiently and effectively, allowing the recipient to easily replicate the process. The optimization of any process must be based on repeating the original process. The technology transfer document should clearly outline the detailed process, specifying the key operational steps. Appendix: Suggestions for pilot-scale scaling up: Just because a process succeeds in laboratory tests does not mean that it will work smoothly at the pilot scale or on an industrial scale; in fact, it is even possible that such scaling up may not be feasible at all. There is a big difference between pilot tests and pilot plants. 1. For pilot tests, we generally use three-necked or four-necked flasks, which are all made of glass; this allows us to observe any phenomena clearly. The amount of material used is small, making it easy to control. But production isn’t possible; glass-lined and stainless steel reactors are used, and they are opaque, so it’s not possible to see what’s inside. What to do? It is necessary to observe closely the changes in the materials during pilot and scale-up tests, so as to have a clear understanding of what happens at each stage during industrial production; try to increase the number of measurement points as much as possible. Example: There is a product (I’m sorry, but due to technical confidentiality, the names of all materials and products cannot be disclosed; please understand), for which a water-soluble solvent is used in the raw material stage. After synthesis, part of the solvent is removed, and water is added to cause the product to precipitate. During the pilot scale testing, there was concern that safety accidents might occur after the solvent dried out, and since the solvent receiving tank did not have any measuring devices, the yield for the first three batches was low. The reason was found to be too little solvent evaporation, which resulted in insufficient water addition; some of the products were still in the water. Solution: Specify the amount of solvent to be removed. 2. The pilot glass bottle thermometer can be raised and lowered, allowing for reading even when it is out of reach. But the reaction kettle is different; that one is fixed. So you need to be aware of your equipment. For example, in a 1000L reactor, what is the minimum amount of material that can be stirred by the anchor, and how much material can reach the thermometer probe? What is the viscosity of the material? Will it be necessary to increase the motor power when a larger volume is used? Example: There was a product for which, during the recovery of the mother liquor after crystallization, no attention was paid to the amount of material remaining after the solvent had evaporated. The temperature of the liquid phase could not be measured, so steam heating was continued, which led to an explosion; fortunately, no one was injured. Solution: Heat with hot water. Products with low concentration are desolventized by multiple inhalations using small reactors. 3. During pilot testing, the amount of material is small, so the post-treatment time is very short. But pilot-scale industrialization is different; the processing time has to be extended significantly, and this must be taken into account. Example: There is a product for which an aqueous solution of triethylamine is used as a solvent, and filtration is carried out after the reaction. During suction filtration, the material is placed in the filtration tank; over time, the temperature of the material rises, causing triethylamine to separate out, and the product is carried away along with the filtrate. Solution: After cooling, perform suction filtration in small amounts multiple times to reduce the residence time of the material. 4. During pilot tests, the glass bottles did not have insulation, and the amount of material used was small; therefore, the heat absorption and release during the reaction were not significant. However, this is different in the case of reaction vessels, as they have insulation and a larger amount of material is used. Therefore, during the pilot testing phase, it is important to observe whether the reaction is endothermic or exothermic. If this is not certain, it is best to connect all available utility pipes during the pilot scale testing; and if such conditions are not available, a pipe should still be reserved in the jacket, ready to supply steam or cooling water to the tank at any time. Example: There was a product for which, during pilot testing, no heat was observed when it was added drop by drop; in the pilot-scale production process, no cooling pipes were installed, and as a result, it took over ten hours to add the material in batches. Solution: Install cooling water pipes. 5. When carrying out industrial design, it is necessary to consider all aspects comprehensively; one should not act arbitrarily, but rather listen to the opinions of others. A professor came up with a design in which, considering the significant exothermic reaction, coils were added inside the reactor vessel. However, the material is particularly viscous; adding a coil does not help stir the material on the outside of the coil, resulting in a low yield. Solution: Remove the coil and use cryogenic brine instead. Titanium coils – I wonder how much silver it will cost. 6. When selecting a production process, it is necessary to take into account the ease of industrialization, the owner’s preferences, as well as the surrounding environment. The surrounding environment mainly refers to environmental protection pressures. Some manufacturing processes face greater environmental pressure, making it difficult to implement them in some developed regions. Example: There was a product where, considering the cost at the time, a high-voltage route was chosen. When it was time to proceed with pilot testing, I reported it to my boss, who believed that lacking experience in high-pressure environments meant a lot of effort had been wasted. As the saying goes, the position determines one’s perspective; different positions inevitably lead to different ways of thinking about issues. A more conservative boss certainly won’t choose processes that involve high risks, as his wealth and even his life depend on it – this needs to be understood. Solution: Took a different route. 7. If all the work is done by technical staff during the pilot test, it will be troublesome. The skills required of laboratory technicians and workshop operators are different; a good laboratory technician is not necessarily a good worker. If it is a technician leading less experienced workers, the technician needs to take extra care to conduct inspections to ensure that the valves that should be opened are indeed open and those that should be closed are indeed closed; it is necessary to verify this on-site. If it is a technician who leads the skilled workers and operators responsible for the work, congratulations – all you need to do is ensure that the standard operating procedures are followed, chat with them in the control room, and address their questions as they arise. When problems occur, take responsibility yourself rather than shifting it onto the operators. Example: There was a product for which, during the pilot production phase, the person in charge was not on site; someone else supervised a few workers who had just been hired to carry out the work. When the materials were poured into the sedimentation tank, no check was done on the valve at the bottom of the tank. All that money just went down the drain. 8. During pilot-scale and industrial design, it is necessary to determine the dimensions of the equipment clearly, including its outer diameter, height, weight, etc. It is also important to consider whether the valve installation is easy to operate, and whether the vent openings are positioned facing the operation surface. With these problems, after installation you’re just waiting to get scolded by the technicians. Example: When building a steel platform for the first time, I didn’t read the instructions for the reactor carefully; as a result, the holes in the reactor were made too small. To lift all those reactors, it was necessary to cut away their jackets first – and those are pressure vessels after all. I felt like I didn’t have the face to see anyone at that time. 9. One should be good at summarizing experiences and lessons. When faced with something unfamiliar, everyone might make mistakes, but once a mistake is made, one should remember to try not to fall into the same pit twice. 10. In short, pilot tests are used to establish the process route. The process route is working, and there are no major issues anymore; what remains to be considered for pilot-scale industrialization are all the details. Think more and ask more questions. It’s normal to encounter minor problems during pilot testing; if everything went smoothly, that would be unusual. When something goes wrong, think carefully – there are many reasons, but in my experience there is only one real cause; however, don’t doubt everything. 11. Under no circumstances should you stand beside the reactor to modify the process procedures; any ideas should be tested in the laboratory first before being taken to the workshop. Accidents are likely to happen. 12. Be good at communicating with others. Since ancient times, scholars have looked down on one another. Scholars, you know – as soon as they can recognize a few characters and write their own name, they easily think they’re great and look down on others. But don’t forget what our ancestors told us: every tool has its limits, the wise consider many things, and so on – others also have strengths that surpass ours. For example, a workshop operator has skills that you may not possess; of course, it’s not required that you have them either. But you can’t use your own strengths to compare yourself to him just because of that and look down on him, right? Throughout the entire process of implementing the process, it is important to communicate more with relevant personnel and listen to their opinions; even if those opinions are wrong, they still represent a possible approach.