When chemists are developing processes in the laboratory on small scale the reactions will often be selective for the desired product, partly as a result of the reaction conditions chosen by the chemist and partly because the reactions are usually well mixed. During development, it is important that work is carried out that is relevant to the likely operational conditions at scale so issues such as rate of reagent addition, point of addition and impact of agitation are evaluated to guide the development of the process and not just the chemistry
Many reactions show differences in profile upon scaling up and mixing can be a very significant factor in the differences between lab and plant, so chemists would be well-advised to have some understanding of the impact of mixing and the options that may be available, thereby enabling design and development of truly scalable processes.
For a competitive consecutive reaction, to use engineer’s terminology, ie one in which the product of reaction can react further with the reagent (a common reaction with complex molecules) it is not unusual for byproduct formation to increase from a few % in the lab to 10-30% and beyond on scale up, particularly if the reaction is fast, and the reaction is not well mixed. So rather than getting a 95% yield and 5% byproduct, a reaction on scale may give 75% yield and 25% byproduct, which may then cause further yield losses as the work up and product isolation has to be adjusted to enable the large amount of byproduct to be removed.
For parallel reactions, where, for example, the main byproduct is formed at the same time as the desired product, the ratio of product to byproduct will usually be sensitive to temperature, concentration or reagent stoichiometry, and this can easily be controlled accurately in the laboratory. But as the reaction is scaled up, particularly if the reaction is exothermic and the exotherm is controlled by careful dosing of a limiting reagent, then control of both effective stoichiometry and temperature will be affected by the dosing regime in the reactor and how well the reactor is mixed. For large reactors the temperature gradient in the reactor will be affected by the mixing and the design of the reactor/agitator/baffling in the process.
Mixing is often crucial in heterogeneous systems, whether solid/liquid, liquid/liquid or, as in catalytic hydrogenations, gas/liquid/solid systems and expert advice is needed to scale up these processes. For example, enantioselectivity can be affected in both heterogeneous and homogeneous catalytic hydrogenations by mixing effects.
Even when a reaction seems to appear homogeneous and all reactants are miscible in the solvent, density differences between the bulk solution and the added reagent can cause inhomogeneity in the system, where poor mixing can lead to loss of selectivity on scale up. This might be caused by simply adding a cold reagent in solvent to a hot solution of the same solvent, or more obviously in a bromination by adding neat bromine to a solution of the substrate. Clearly in the latter case the dense bromine needs to be well dispersed in the solvent after dosing to avoid over-bromination, so adding the bromine as a solution in the solvent is the preferred scale-up option to avoid mixing-related selectivity issues.
Mixing of course is not just important for chemical reactions, but also in work up too. I well remember my surprise during scale up of a process to make a keto-ester many years ago, when during work up of the strong base suspension in aprotic solvent by adding water slowly, the ester hydrolysed extensively because of the extended time the ester was in contact with the strong aqueous base/ solvent (an ideal hydrolysing medium) before dilution. I found that by adding the extraction solvent (toluene) first and quenching and extracting at the same time, there was much less hydrolysis. In fact in this method of work up, the extent of hydrolysis was then controlled by the mixing in the vessel, as well as the quench rate and temperature.
Liquid-liquid extractions are rarely a problem in the laboratory since the two layers can be well mixed by shaking vigorously. As a process is scaled up the process may need to be adjusted for optimum performance to suit the mixing characteristics of the agitated vessel, and performance (eg settling and emulsion formation) may be affected by whether the solvent is added to aqueous or vice-versa, as well as the position of the interface in relation to the position of the agitator in the vessel. (An inappropriate position of the agitator (for a particular batch-size) can also be a common problem in phase-transfer catalysed reactions and other processes run with two immiscible liquid phases).
During crystallisations, the mixing in the vessel will affect the extent of secondary nucleation as well as crystal growth, and so determine the particle size distribution (and hence the filtration characteristics) of the solid product. Crystallisations in which an anti-solvent is used to reduce the solubility of the product in a solvent (which chemical engineers often term “drown-out crystallisations”) are particularly prone to mixing-related issues, and can even affect the crystal form or level of hydration as well as the particle size distribution. So it is wise to understand these effects (or talk to a friendly chemical engineer) before going to kilogram scale. Remember a product which filters badly will not only retain residual solvent, but also the impurities which reside in that solvent, so that washing out of the impurities may also be a problem. When this product is dried, the impurities from the solvent are then deposited on the product.
So in summary development chemists, particularly those who are taking processes into kilogram and tonne manufacture, as well as chemical engineers, need to understand the rudiments of mixing and how it affects process selectivity both from a reaction and work up outlook.
Attendance at one of Scientific Update LLP’s courses on ‘Chemical Development and Scale Up, Secrets of Batch Processes or similar course, or at conferences such as The Scale Up of Chemical Processes, where detailed unpublished Case Studies are presented, should increase the chemist’s knowledge of mixing related issues and help the chemist to avoid scale up issues. See the website www.scientific update.co.uk for details. Following the next Scale Up conference in Boston, USA in July 2011, there is also a special event on Mixing ‘The Influence of Mixing in Your Process’ organised by VisiMix, where chemists would gain an insight into the details of how mixing can be visualised/simulated in reactors and thus how this knowledge can be used in process control, particularly on scale up. This conference also contains a special training course given by Dr Moshe Bentolila on ‘Application in Different Unit Operation Systems’.
Scientific Update’s consultants and associates (who are chemists, chemical engineers and other disciplines) can advise on all aspects of chemical route selection, process R&D, development and scale up, crystallisation and polymorphism etc and have an outstanding track record in problem solving and troubleshooting, whether in laboratory, pilot plant or in “routine” manufacture.
For more details see our Consultancy Website: www.scientificupdate.com