Creating a process to synthesize an organic compound at plant scale has several iterations. The guidelines vague as they are, differ depending upon the iteration within which you are operating. Since the molecule is to be made at scale, next to safety cost is the foremost consideration.
Route Scouting
The first iteration is termed route scouting. At this stage, you know the structure of your target molecule. This means that it has already been synthesized if the molecule is novel and not naturally occurring. Therefore, an original sequence of reactions exists but if the exploration was part of a wider screening study the route was not selected with this molecule alone in mind but rather for a family of molecules of which this molecule was a member.
If the molecule is a natural product and has been isolated from a natural source there is most likely no route at all.
Route scouting comprises two different activities. One is what is called ‘paper chemistry’. No laboratory is required- just paper, pencil, and in this computer age, appropriate database permissions. The second activity actually tries to conduct the experiments most promoted by the ‘paper’ search activity. Although some ‘paper chemistry’ needs to be done before laboratory experiments begin, it is most common that ‘paper chemistry’ proceeds first and then continues even as the early ideas are tested in the lab.
Paper Scouting
Route scouting of the ’paper’ variety is directed by clues that are found within the unique structure of the target molecule. The task is to take substances that can be bought or simply made by well-established paths and combine these in such a way that the target structure is constructed.
Chemical complexity cannot be usefully quantified, yet it is readily ranked intuitively by the experienced synthetic chemist. What is important to recognize at this scouting stage is that the complexity of commercial product offerings available as starting points is not proportional to price. That is to say, more complex chemical structures do not necessarily cost more than less complex ones. As a consequence, the structures chosen as starting materials matter a lot! Mapping substructures of the target molecule onto available inexpensive starting materials is one very useful exercise in paper chemistry that proceeds as part of route scouting.
An obvious example is the establishment of absolute chirality at a center in the target by incorporating a starting material sub-structure that already possesses that required chirality. The similarities of ring structure or substitution pattern between starting materials and products can similarly be useful.
Because no person can retain in memory all the different potential starting materials and their respective prices, this type of screening can only be done by computer search.
Retrosynthetic Routes
Retrosynthetic analysis is a product structure-based methodology for assembling promising ‘paper’ syntheses in route scouting. Retrosynthetic analysis was pioneered by Professor E. J. Corey. The method examines the target structure looking for bonds whose breaking leads to the largest reduction in structural complexity. The method uses the clues provided by the functionality already present n the target to rank the promise of these disconnections. The bond breaking must be done in accordance with the teachings of known chemical reaction transformations and preferably be compatible with the functionality already displayed by the target molecule.
Most preferably, the disconnections should lead towards the more preferred starting materials identified in the starting material-based searching.
The Service Synthesis Route
Throughout the development of an industrially suitable process for preparing a target structure, the other departments of the organization seeking to manufacture the target will have recurring needs for fresh supplies of the target compound for their own studies. These needs come from formulation, regulatory, sales …so many other departments. At first, their requests are modest but the amounts each is likely to need escalate as the proposed launch date for the product comes closer. These requirements must be met by the synthesis department making fresh product by what can be called the ‘service route’. The service route of synthesis evolves from the discovery route; the method by which the molecule was first prepared. It is the discovery route pragmatically adapted with go-arounds for aspects of the original method that could not be practiced as the scale increased. Removing steps that could not be scaled such as those requiring column chromatography, for example, or operations that were less safe in larger equipment or that could not be done at all. Nevertheless, the service route remains as close as possible to the discovery process to minimize the need for new chemical explorations.
Still at the outset of the entire project, a judgment has usually been made whether the benefits of a total rethink of the synthesis of the target involving starting material-based analysis and retro-synthetic analysis will be able to provide sufficient advantages to be eventually preferred over the evolving service route.
The service route will invariably retain certain advantages. It will become very familiar to operators. It will retain the closest impurity profile to the target. There will be more opportunities to improve the work-up, isolation and purifications of the steps in the service route with no additional research cost born by the synthesis department. The employees that are assigned to operating the service route will gain the most inter-silo integration with other company departments.
Still, the starting material based and the retrosynthetic-based routes will have their own advantages. The service route is least likely to improve costs. Expensive inputs are less likely to be replaced. Certain types of hazards are less likely to be addressed.
_________________________________________________________________
Because highly trained employees will be doing the work, safety, and toxicity hazards are more likely to be mitigated rather than avoided. Actual reaction optimization is unlikely to be advanced. Changes that could actually lose batches of material will not be risked. Remember the primary function of the service route is to provide promised amounts of product by promised dates. Regulatory concerns such as genotoxicity do not matter. A route created according to either the starting material-based or retrosynthetically based approaches may lead to patenting that is stronger and last longer than just obvious workman-like improvements in the discovery synthesis route.
The Ultimate Process Route
So substantial are the advantages of the discovery/service route that the final process is very likely to at least in part to follow it. Very often the starting material-based route and the retrosynthetic analysis may lead to solutions that provide alternate solutions for some of the unavoidable problems that existed in the service route. Certain steps may be totally replaced or reordered applying insights born in the ‘paper chemistry’ that comprised these analyses. Because these analyses from first principles are so much more cost focussed and less risk-averse these approaches are more likely to find major improvements.
Route Optimization
Route optimization takes the ultimate process route and without changing
the chemical transformations improves the overall yield or throughput while maintaining the quality. Although the stage is called optimization, optimization is a gross exaggeration of what is done. In fact, a team tries to make improvements until either it runs out of time or the value of its improvements is judged to be inferior to the benefit from alternative deployment of the scientific resources (ie. the people are reassigned to other projects).
The improvement that is being sought is just one thing— lower cost. The goal is the profitability of the corporation. Since this is so, the improvement priorities are established by the individual contributions to the overall cost of the product, and for this, the costing sheet is the guide.
Costs fall into categories of substrates (starting materials), reagents (reactants not incorporated into the product), and the twin efficiencies of assay yield and isolation yield. Categories are only useful if they provide a guide for action.
Starting materials cannot be changed but inferior grades of the same material which have a lower price can be tested. Sometimes a lower grade has equivalent performance or its reduced performance costs less than the monetary savings. The changed starting material must not cause undesirable changes for subsequent steps but most importantly, the change must not degrade the quality of the fill product of the entire process. Changes in the quality of starting materials are most often without consequences in the early reaction steps that are most removed from the target molecule.
Reagents are here defined as reactants that do not significantly supply atoms for the assembly of the target molecule. Reagents can be substituted and their quality varied. Changing reagents usually requires many other changes in the chemical transformation. The operation that is executed on the intermediate is maintained but the tool that is used gets substituted and all the reaction conditions change with it. Not only the reaction conditions have changed but the workup, separation, and often the purification of the intermediate product change. The reaction assay yield and the isolation yield will both change as must the step's overall yield. To justify so much fresh experimentation the cost saving must be prominent. Because of the inherent risk of a change in final product quality analysis, such an improvement should be closer to the beginning of the reaction sequence.
Reagent changes often are instead part of the service procedure development.
The overall yield fraction is the product of assay yield and isolation yield. This division of the overall yield into its components is rarely worked out during the setting of the ultimate process route but it is necessary for the efficient improvement of the steps. There is little to be gained by making changes in an area where the yield efficiency is already high. In the same way that the costing sheet should prioritize the above types of change, the determination of all of the assay and isolation yield prioritize the efforts at yield improvement.
Yield improvements in the reactions closer to the final product give rise to the greatest improvement in costs since an improvement reduces the amounts of reactants, reagents, equipment, labour, and other chemical processing costs in all of the prior steps. Improvements in the assay yields of these steps occur with concurrent reductions in residual starting materials and byproducts. This tends to improve product quality. Improvements in isolation yields often are accompanied by trapping more impurities in the isolated intermediate leading to lower purity in the final product. Improvements in isolation need to be changed in the mode of isolation and purification rather than just trying to up the recovery. A more rugged method of isolation can be expected to improve recovery yield while maintaining quality.
A Wider Range of Optimization Options
Because improvements in the yields of the late-stage steps have such a magnified influence on the overall cost, the range of options for achieving an improvement is enlarged. A more costly reagent that improves the yield somewhat is more easily made up for by the cascade of savings in the earlier steps. Increases in maximum storable volume that decrease the throughput of the overall process are easily compensated by the overall savings from a yield improvement. Thus higher dilutions and bulkier reagents become practicable. For example, steam distillation as a separation tool or polymer-supported reagents become more realistic.
The Importance of Phase-Switches and ‘Setting’ the Late-Stage Steps
Purifications require phase switches. Purification is effected when the desired substance makes it through a phase-switch more readily than the panoply of impurities. But phase-switches can be part of work-ups and isolations as well as purifications. The more phase-switches a compound negotiates the more dependable the degree of purification. There are many phase switches: crystallization, distillation, sublimation, dissolution, liquid-liquid extraction, adsorption, precipitation, and co-distillation …. The phase-switching that is built into the ultimate and penultimate steps of a reaction sequence is very important to the final quality of the target product. Therefore, the sooner in-process development patches are ‘set’ and no longer modified, the more likely that your final product will be identical to the samples submitted to potential customers and government regulators and to the batches of material that went into any clinical trials.
Pilot Plant Responsibilities
Besides process improvements, process optimization chemists are usually responsible for:
Forensic pilot plant sampling
In-process assays
Stopping points/ stability studies
Process equipment corrosion studies
Filtration flow rates
Special cleaning recommendations
Timing requirements (addition and transfer rates)
Minimum and maximum volume limits
These special topics are discussed in the final report that is the transmission document that transfers everything that is useful for the pilot plant personnel to know from the process development team.
InProcess assays
In-process assays are assays (usually quantitative tests) that are part of a batch sheet. They are performed by the analytical personnel from the analytical department. These results form the basis for decisions by the pilot plant and at a later stage the plant management about how to proceed in further executing the batch sheet. In-process assays are essential for executing the batch sheet which instructs the process step.
Forensic pilot plant sampling
Forensic pilot plant sampling is only part of a pilot plant batch sheet. Samples are taken at the prior, planned request of the process development chemists. They are taken by the pilot plant operators according to instructions in the pilot plant batch sheet and saved for the requesting scientists. Forensic pilot plant sampling normally never becomes part of the final process master batch sheet, because the purpose of these samples is to help the development team figure out what went wrong if either the expected yield or product purity fails. Once a pilot plant experiment is reported as being completely successful, these forensic samples are discarded by the scientists who requested them.
A pilot plant run is an expensive experiment. It qualifies as an experiment because no matter how many precautions are taken in advance the scale-up from laboratory to pilot plant scale involves risks that cannot be fully mitigated. Because it is too costly to repeat unnecessarily, the scientists responsible need to take whatever samples they might need to answer any question about what might have gone wrong at any stage in the experiment. The samples must be taken proactively- as a precaution in case a problem occurs.
By the way, remember process development chemists are visitors to the pilot plant (and the plant). They are observers only. They do not do anything. Neither do they instruct the operators. They can take notes for their own purposes. They can converse with the person managing the execution of the batch sheet. That is all!
Stopping Points/Stability Studies
At many points in chemical processing, chemical instability makes it essential to continue moving through the instructions in the batch sheet. Executing most batch sheets extends longer than 24 hours, however, and many pilot plants or even plant facilities do not operate around the clock, so it is useful for there to be tested advice on where the processing can be stopped and for how long that interruption can safely last. To that end, development chemists need to record at what points they have safely stopped processing.
Process Equipment Corrosion Studies
A rare but disastrous risk is that a chemical processing step will damage the chemical reactor, usually by corrosion of the reactor walls. The process development chemist needs to test whether this could occur. This is done by placing tared sample tiles of the reactor’s construction material in contact with the reaction mixture during the reaction period. This is normally done by attaching the tiles to some structure within the reactor that contacts the reaction mixture. When the processing is complete the tiles are recovered and weighed to see whether any material has been removed from them or any weight added.