The term process optimization is a misnomer. There are too many variables to ever completely investigate a process to be sure that it is the best process possible. There are even different meanings of ‘best’. Are we interested in the lowest cost? Most dependable? Highest throughput? Providing the fewest unidentified impurities? Greenest?
Then there is the practical matter of the most cost-effective use of a research chemist's time. For a corporation, the improvement of a particular process should stop, when employees can be used in more productive ways attacking other problems. So process improvement must stop when the work no longer provides or is likely to provide benefits comparable to alternative investigations.
With this understanding then we can ask, ‘How can we identify the best aspects of a process to attempt to improvement?”
How much improvement do you need? As much as possible is the usual answer but it is much more useful if the answer can be amplified to include a specification of 'at least as much as’ because this can direct your search more towards rapidly introduced tweaks or substantive changes which may have a longer time frame to realize.
The most common process improvements aim to reduce the cost of the final product. The purpose of production for an economic entity is to make money. A process that does create something that makes money is very unlikely to be put into effect.
A chemical process comprises one to many steps. A step is a series of operations that transform starting materials to some easily identifiable intermediate stage. The final step converts the last intermediate stage to the final commercial product. The term step is therefore loosely but practically defined. Intermediates are defined as clearly defined stopping points where yields and purities are usually checked to ascertain whether the process is operating as prescribed.
Very often intermediates in a process can be accumulated and stored. This provides time to clean process equipment; order required reagents, order more starting materials and other processing chemicals; and schedule steps of different batches for the manufacture as well as other different production that may be proceeding in the same time period and sharing the plant’s facilities.
Starting Material and Reagent Sourcing
Usually, the most cost-effective improvement for any process is better sourcing. At what point your manufacturing begins depends upon the price and the structure of your starting materials. If your organization can buy a starting material more inexpensively than you can make it yourself you’ve saved both time and money! If you can purchase even a slightly different starting material and adjust your process so you can use that new material this too can lead to a major improvement.
When developing a process the best available qualities of starting materials are often used. This is because you do not want an experimental process step to fail simply because the highest quality of input material was not used. Different grades of a particular starting material may be available at different prices. Furthermore, your process step may behave indifferently no matter the input quality. This is particularly likely for a processing chemical that is not incorporated into the intermediate’s structure or if the workup or isolation of your intermediate is particularly robust with respect to phase switches.
On the other hand, input substructures that get incorporated into your product are very sensitive to the presence of homologs, isomers, and geometric isomers as impurities in a starting material.
Regulated or Unregulated Final Product
An important difference must be distinguished among processes depending upon their final use. If the output of the process you are working to improve is regulated by the purchaser or by the government on behalf of the general public, your product may need to meet very specific and demanding specifications. Take, for example, the active ingredient for a drug product. The government will have specified the final purity of the product and may have actually specified the profile of impurities that can exist. The same situation can exist when your product is the starting material for another company’s process. That customer can have distinct inflexible specifications. In these situations making changes in the final steps in your process can cause that final product to fail either government or customer specifications. Making pharmaceuticals provides a particularly severe example. One of the government regulations covering marketable human medicines is that a representative product must be tested in a clinical trial in which healthy persons are dosed with the medicine. Clinical trials are extremely time-consuming and consequently expensive. A process improvement must not cause the government regulator to require that your organization repeat a clinical trial! For this reason, the final steps in the manufacture of pharmaceuticals are often purification steps in order to guarantee as much as possible that the final product’s overall purity and impurity profile is unaffected by any of the earlier steps in the process.
If what your chemical process provides is an industrial chemical species that is used as a starting material for other companies’ manufacturing you can be much more flexible towards modifying all the steps, even late-stage steps, of your process. You may even be able to provide different grades of products to different customers with cost advantages that can be shared between your company and the customer organization. For example, suppose your process comprises the isolation of a crude final product (X) which is then purified in two distinct following stages with corresponding products (XI) and (XII). The purification steps that change (X) to (XI) and (XI) to (XII) cost your company money and require you to charge a higher price for these ‘improved’ materials. Perhaps your industrial client operates a process that is insensitive to the presence of the impurities your purification steps remove. In this case, your customer would be advantaged if he could purchase either (XI) or (XII) instead, at a cheaper price, and your company would make more profit selling these less processed materials. No process change would be made at all. Your company catalog would simply now list three different grades of one product at corresponding appropriate prices!
Throughput Changes
Changes in a chemical process that increase the rate at which products can be sent out the door of a manufacturing facility can lower the cost per unit. Sometimes increasing the scale of operations can be accomplished without further outlays for capital and labor. Changes in the efficiency of isolation and purification steps can often be made without any change in the overall purity or impurity profile of any reaction transformation. KiloMentor has written blog articles about throughput improvements. Search the blog using the keyword ‘throughput’.
Process Step Costing as the Basis for Improvement Studies
The entire chemical process from beginning to end needs to be costed. Costing will have been done when the process was initially worked out since no work would ever have been done unless management felt that the product could be manufactured at a cost that would have provided an opportunity to make a profit. This costing may be out of date-some changes make have been made on the fly and input prices change all the time. Nevertheless, this initial costing will be easier to update than to repeat what has already been done.
To reduce overall cost the most efficient strategy is to look at aspects of the process that contribute a larger part to the total. It is the detailed costing that provides this information.
An Expensive Reagent
Often a major element of the overall cost is one particular reagent in one particular step. Before you, as a chemist, start trying to improve the reaction, isolation, and purification aspects of that reaction step, there are two people from whom you should seek help. One is a member of your organization- the sourcing specialist. The second person is the company salesman who sells this particular reagent. Explain to the sales agent that your company is finding the expense of the reagent a problem and you have been tasked with finding some alternative or some improvement.
You need to ask, “Can you purchase in bulk? In solution? In a different quality? To what extent can the actual production process be disclosed to you? Does the company sell other products that do the same job? What methods to prepare this material are taught in the literature/ (could we make the material ‘in situ’ less expensive). Can the salesperson provide access to a scientist or product specialist in his/her company that can answer your questions?”
Don’t think for a minute that you are wasting the time of people in your supplier’s organization. This is the type of interaction they love! They are getting to know your company better. They are building a closer relationship. This is the kind of work that earns them bonuses!
To what extent can you disclose to the salesperson exactly what you are trying to do?
An Expensive Starting Material
As I have already noted. This is a challenge for the person who sources your chemicals. Everything already mentioned with respect to reagents also applies here.
Solvent Expensive
The cost of solvent is very often the largest expense among processing chemicals. The solvent is the chemical most often available in different grades. A lower grade may work just as effectively in your reaction step as the best available grade.
In most fine chemical processes solvent is not recycled. This has two important consequences. First, solvent mixtures can be used as the medium for chemical reaction steps. Second, the cost of disposal must be included in the overall cost. The expense of intermediate drying is related to the volatility of the last solvent from which an intermediate is separated or the last wash solvent.
Solvents can be replaced by less expensive ones within chemical reactions but this involves empirical investigations that will require examination of different reaction times and reaction temperatures. Without actually using a different solvent the throughput of a reaction transformation can be improved by working at higher concentrations of reactants using less solvent diluent. When a reaction step is first developed a higher solvent/solute ratio may be used to ensure that all the reacting components are dissolved as a homogeneous solution. This is usually required for precise reaction rate in-process checks to determine the most appropriate point to stop or quench the reaction and begin the reaction workup. Using a higher dilution than necessary at the outset also improves the ease of adequate temperature control in the reactor.
Increasing the concentrations by decreasing the solvent ratio will simultaneously decrease the cost of solvent, increase the potential throughput by increasing the reaction rate, and increase the throughput by increasing the scale of each individual reaction run.
If the product precipitates in the more concentrated solution the overall yield is unlikely to be affected adversely; however, determining the best stopping point for the reaction will become more complicated since reaction sampling may become more complex.
In a more concentrated reaction solution, any exothermicity of the reaction will be more difficult to control. Care needs to be exercised in this respect.
A Long Processing Time
In the plant at the beginning of the reaction period, the rate may need to be slowed down in order to control the exothermicity of the reaction. This is most often accomplished by controlling the rate of addition of some reaction constituent or by cooling the reactor. Once all the components have been mixed together the reaction is allowed to continue to a prescribed endpoint at a selected reaction temperature. Reaching this stopping point may take a relatively long time when the reaction temperature must be kept lower to provide the required selectivity of reaction or to protect some unstable component of the mixture. Long reactor utilization decreases its availability for increasing throughput or for use in other plant work.
The most common reaction temperature used to complete the reaction is the boiling point of the reaction solvent. If a shorter reaction time is sought the solvent needs to change. This can have other consequences not easily anticipated. Nevertheless, it is the only way to accelerate unimolecular reactions. Reactions that are bimolecular or higher proceed faster at higher concentrations so the rate can be accelerated by decreasing the solvent volume. For reactions conducted under reflux, this just means switching to distilling the solvent and concentrating the reactor charge. Using this approach to reaction time shortening must be done with precautions. If reactants whose stoichiometry must be maintained, are swept out of the reactor with distilling solvent the reacting ratio will be disturbed and the reaction interfered with. To be safe some kind of fractionating column can be used to catch and return these wayward chemicals.
Chemical Yields
The overall yield fraction for a chemical reaction step consists of the product of the reaction yield fraction multiplied by the isolation yield fraction.
Reaction Yield Fraction & Isolation Yield Fraction
The reaction yield fraction is the molar fraction of the moles of product relative to the moles of the limiting starting material. The moles of product are found by analytical assay of a homogeneous sample from the reactor at the stopping point. The isolation yield fraction is the mole fraction of the product obtained as isolated final intermediate compared to the moles of assayed product in the reactor at the reaction’s completion. Thus if the assay of product in the completed reaction indicates that the number of moles is 0.75 of the theoretical based on the limiting reactant (reaction yield fraction), and the moles of isolated product is 0.80 of the theoretical amount based on the titre and volume of the reaction solution (isolated yield fraction) then the overall yield fraction is 0.75 X 0.80 = 0.60 and the overall yield is 60%.
A low reaction yield fraction cannot be improved by working on the workup, isolation, and purification. A low isolation yield fraction cannot usually be improved by working towards improving the product assay in the reactor at the reaction stopping point. The word ‘usually’ is used in the second statement here because most often as the reaction yield is improved the level of impurities declines and this decline can somewhat improve the isolation recovery.
When the High Cost Arises Significantly from a Low Reaction Yield in a Step
I now get down to what the activity that most process chemists think of as process optimization. That is- changing the parameters that control the efficiency of a reaction or even changing some of the steps within the process scheme. As I have clarified above, the overall yield percentage from one intermediate to the next intermediate is the product of the reaction yield fraction times the isolation yield fraction expressed as a percentage. When trying to improve the overall yield in a step it is important to identify in which of these two factors the most substantial yield loss arises and then to experiment to improve that aspect.
When it turns out the reaction itself is not selective or the equilibrium that makes product is not sufficiently favorable, continuous and discontinuous variables affecting the chemistry need to be examined looking for combinations that increase the reaction yield fraction.
One by one variable optimization is inadequate for this purpose. Statistical methods need to be applied used among a group of factors shown to strongly affect the yield outcome. Factors that do not raise costs should generally be preferred over those that do when a choice must be made. For example, increasing the excess of a reagent might increase the yield fraction but it most certainly will increase the overall cost of starting materials. Increasing the solution concentrations of the reacting species might increase the yield fraction but it will definitely increase throughput, decrease reactor utilization time, and decrease solvent expense.
When the High Cost Arises Significantly from a Low Isolation Yield in a Step
More than anything else the focus of the KiloMentor blog has always been to make workup, separation, isolation, and purification simple and rugged; and by so doing increase isolation yields. Improving isolation yields can lower costs by lowering overall yields in-process steps without changing impurity profiles qualitatively, unlike replacing one chemical reaction with a different one to achieve the same outcome. Improving an isolation yield will not change the identity of the coproducts or byproducts that arise in the corresponding chemical reaction. Changing the isolation protocol will have no effect on the conditions of the reaction that precedes it. Usually, more processing time is spent cleaning up the ‘soup’ that is the reaction mixture than is spent mixing together and stewing together all the ingredients. Since more time is spent on the workup etc., there is more potential time to be saved shortening things.