Changes during the Reacting Phase of a Chemical Transformation that Increase Throughput While Minimizing the Risk to the Overall Yield for the Step

Early Reactions in Scheme are the Best Targets

When attempting to improve the early reactions in a reaction sequence, it is more difficult to make changes that have a major effect on the overall cost. The reason for this is that improvements in yield only reduce the cost of inputs for the reaction you are working on and any others that precede it in the scheme. If few reactions precede the one you are working on, then there are few inputs whose costs can be caused to decline.

For throughput, it is a different story. The biggest improvements in throughput occur for reactions that are required to convert the most kilograms. These reactions occur at the beginning of the reaction scheme where the same reaction must be repeated more frequently to amass the required intermediate amounts.  Therefore, for the initial reactions in a reaction scheme, the emphasis most beneficially go towards increasing throughput.

Increases in throughput for reactions closer to the final product do not improve cost by much. Furthermore, they increase risk by committing more expensive intermediates to every single run. They also build up expensive product in inventory increasing the company’s carrying costs.

Safest Targets

The safest way to increase throughput without any risk to yield is to operate at a scale where at the point of maximum volume in the procedure, the plant reactor you are provided is at its maximum volume. That is— use every litre of the reactor’s volume consistent with the procedure. Getting more substrate into a reactor can also be achieved by distilling reaction solvent after the reaction’s quench to reduce the total reactor volume at the point of maximum volume.

Other ways to increase throughput that would need to be tested to see if they might not appreciably reduce overall reaction yield are:

• Decrease the solvent/substrate ratio
• Drive the reaction to completion by distilling the reaction solvent to concentrate reactants
• Operate at a higher reaction temperature to decrease the reaction time
• Find a reaction catalyst to reduce the reaction time

Decreasing the Solvent/Substrate Ratio

First a word of caution. The ratio of solvent to substrate affects more than the environment of the transition state. Increasing the concentrations often increases the reaction rate and this very often brings with it increased exothermicity. The cooling capacity of the reactor must be able to handle the increase.

Decreasing the solvent/substrate ratio to increase throughput needs to be tested in aggressive steps in the laboratory. In the laboratory, the available ratios of flask surface area to flask volume are very much higher than in the pilot plant. Since the effectiveness of an external cooling bath is proportional to this ratio (because the cooling is delivered through the walls of the vessel) increasing the cooling to maintain the previously established internal reaction will be easier in the laboratory when lower solvent/substrate ratios are tested. Since a significant increase in throughput is being sought, a 20% reduction in the solvent seems like a good starting point for a reaction where the original solvent/substrate ratio is about 20/1. That would be a ratio of 16/1.

Driving the Reaction to Completion by Distilling the Reaction Solvent to concentrate Reactants

This is a strategy that I have never seen demonstrated yet it is theoretically sound. A bimolecular reaction proceeds rapidly at the beginning when the concentrations of both reactants are relatively high but then slows as these concentrations are reduced by consumption to give the product(s). The most reaction time is passed during this terminal stage while the experimentalists wait for the final portions of starting materials to react. If the reaction period is significantly longer compared with the overall processing time for the step it presents an opportunity to increase throughput meaningfully by reducing the reaction time.

If the reaction temperature is the reflux temperature of the reaction solvent, it makes the process easiest to manage. When the reaction slows down solvent is distilled out of the reactor at the reflux temperature which is also the desired reaction temperature. As the overall volume inside the reactor is reduced by the distillation, the reaction rate increases because of the effective concentrations in the reactor increase. The overall reaction time to completion is shortened and with it to some degree the overall cycle time.

When the reaction temperature is below the boiling point of the solvent the same thing can be achieved by reducing the pressure in the reactor with a vacuum pump to the point where the distillation point of the reaction solvent becomes the reaction temperature. Then distillation at this reduced pressure drives the reaction to more rapid completion as in the simpler previous situation.

Operating at a Higher Reaction Temperature to decrease the reaction time

Raising a reaction’s temperature by 20C degrees will increase the rate of reaction by a factor of 2 according to a rule of thumb used by chemists. The long drawn-out reaction time during which the last portions of the reactants are converted can be accelerated by ramping up the reactor temperature towards the end to consume the last bits of the reactants. This would improve the throughput by reducing the reaction phase. Raising the final temperature could cause problems, however. Utilizing a higher reaction temperature than necessary can often give rise to byproducts that will contaminate a desired product. If the reaction that the experimenter is trying to conduct is an equilibrium reaction, the higher temperature could even drive the reaction back towards more starting materials and less product. The actual outcome can only be determined by experiments. Nevertheless, increasing the reaction temperature to finish a reaction can work and if it is a long reaction time that is being shortened, it can significantly improve cycle time. 

Find a Reaction Catalyst to reduce the reaction time

This last idea is the most radical and the least specific. Still, it is often overlooked. Using topic searching in a chemical database, a modern chemist can quickly determine whether there is any catalyst known for widely-used name reactions. This literature searching is properly done at the stage of route scouting but it might be overlooked and can be reconsidered here. A catalyst by definition is used non-stoichiometric amounts because it is recycled back into the transition state. It should accelerate a reaction without any further change of conditions. So long as the catalyst is reasonably priced it should not add appreciably to the costs.

KiloMentor has written elsewhere about methods to decrease cycle time by making changes in the non-reaction portions of the reaction process step.  

The Graduate Student’s Academic Organic Synthesis Project

 So you have convinced your thesis advisor that you should pursue a new synthesis of a previously unknown molecule. Probably it was discovered by separation from a natural product fraction or perhaps this is a new approach to an old molecule of some importance. Whatever; congratulations devising and executing a synthetic project is excellent training for the world of work and besides it is good fun!

You need to understand that working in an academic environment brings some particular advantages and disadvantages. The advantages are numerous and well known. You will probably have access to a broad range of analytic tools for both separation and characterization. Also, you will not need to produce very much of the desired target substance; enough to analyze and characterize. It can be kept in a vial rather than a bottle. Literature access and literature searching will be available and relatively inexpensive.

What is not mentioned is that your own labour will be considered free and money will not be spent to make your work faster. Expensive starting materials will not be purchased. You will probably be limited to substances that cost at least less than $1/gram.

When you work out ‘paper syntheses’ to evaluate for practicality pay attention to your starting materials. Although something may be commercially available, it may be beyond your budget.  Plan to make your starting material if at all possible from things that cost less than $1/gram; preferably much less. This is quite the opposite from what you would encounter in industry. There you would be expected to purchase the most advanced intermediates available. Time is money and your time is costing your company big bucks.

You should not be discouraged though because there is a silver lining to this academic cloud. If you make your own starting material, you can make lots of it which means that you can execute all your steps on a larger scale. You won’t need to spend as much time bringing forward small batches of intermediates and you won’t feel so constrained in what experimental conditions you can risk.

The biggest error I made as an undergraduate and a graduate student was failing to improve the early steps of my route so that I could move larger amounts of material forward quickly. Get the biggest reactors you can find and move forward as much material as your confidence in your early steps will allow. 

An added benefit to learning to quickly scale-up is that this is a major skill required in industry and if you eventually apply for work in the industry citing this experience will enhance your application.

The Purpose of Standard Operating Procedures (SOPs) in Chemical Manufacturing

 The Purpose of Standard Operating Procedures (SOPs) in Chemical Manufacturing

Chemists with post-graduate educations including experimental laboratory training are expected to know a variety of safe and effective ways to achieve common chemical goals. As well, they are expected to have document research skills sufficient to find approved methods for unfamiliar operations. They do not need standard operating procedures. In fact standard operating procedures may be fatal to new discoveries.

  Chemical processes are executed by operators, who are trained differently. Operators are most valued because they can follow instructions of a process in a totally reproducible and detailed fashion while watching and reporting anything unusual. Often they may not know the theoretical basis behind the exact operations that they need to perform.

It is desirable to provide every possible means to assist them to do their work methodically so that an operation is performed precisely the same way each time whether it is part of one process or another and no matter which operator is in charge. This is the purpose of Standard Operating Procedures (SOPs) in the manufacturing environment. SOPs can also be written for operations for analyses, for receiving, for shipping, for reporting accidents, for filling out batch sheets etc. SOPs help to achieve faithful and exact repetition of a procedure wherever a specific action is called for. It is not the basis for a decision between alternate actions. What sets an exceptional operator or analyst apart from an ordinary one are two things- detailed exact repetitive execution and the ability to observe small differences in what happens, remember them, and report them.  

Liquid Sulfur Dioxide as a Means to Switch Solvents

Liquid sulfur dioxide is immiscible with saturated hydrocarbon liquids and will form a separate lower phase when combined with them. Thus, if a saturated hydrocarbon solvent is used as a chaser to drive off the reaction solvent from a first reaction, sulfur dioxide can be added to extract the reaction mixture content into a lower phase provided the desired material is soluble in sulfur doxide. The viscosity of a hydrocarbon fraction that is sufficiently high boiling to work as a ‘chaser’ may mix only sluggishly with liquid sulfur dioxide at -10 C. If it proves necessary to decrease the viscosity of such a hydrocarbon solution it can be mixed with a lower molecular weight saturated hydrocarbon.

Because at 1.46 g/ml the density of liquid sulfur dioxide is almost twice that of a typical hydrocarbon and since saturated hydrocarbons have only a limited solubility in liquid sulfur dioxide, they will form a separate liquid phase from which the liquid sulfur dioxide lower layer can be cut away.

Therefore, liquid sulfur dioxide can be used to transfer a solute from a higher-boiling hydrocarbon solvent mixture to a lower-boiling solvent such as ethyl acetate by first extracting the solute into a liquid sulfur dioxide phase and then displacing the sulfur dioxide with the second solvent (such as ethyl acetate).

There is a shortage of usable solvents that have a higher density than water since halogenated solvents have fallen under a regulatory cloud. Liquid SO2 is such a dense solvent.

Unjustified Shortcomings of DMSO as Reaction Solvent

 

 It is true that DMSO has no known azeotropes. This and its high boiling point account for the difficulty of removing it from reaction mixtures. 

It is not true that DMSO is difficult to dry. DMSO distilled under vacuum after taking a forerun is water-free. 

Whatever its disadvantages DMSO is too valuable a solvent to rule it out.

DMSO is reported to be immiscible with cyclohexane, heptane, hexane, pentane, 2,2,4-trimethylpentane, and diethyl ether, so these liquids can be used in solvent/solvent extractions. Silylation of the solutes in a reaction mixture should improve their extraction into these less polar solvents which can provide a second layer with DMSO.

DMSO is reported to be miscible with methyl t-butyl ether (MTBE). Based on the reported immiscibility with diethyl ether, this would not necessarily be expected. Perhaps adding a small amount of hydrocarbon to the MTBE could provide a two-phase mixture. An anhydrous DMSO produces a phase separation with diisopropyl ether (DIPE) according to the “Solvent of the Week” website. This would permit more flexible liquid/liquid extractions to separate a product from DMSO.

Another potential way to work up reactions done in DMSO could be to concentrate the solution as much as possible under vacuum, and then add the minimum stirrable volume of glycerol (enough to still provide some slight agitation even if the DMSO gets completely removed) and displace the remaining DMSO, still operating under vacuum. Then,  any suitable solvent that is immiscible with glycerol (there are many) could be added and the substrates of interest taken into it. Mixtures of solvents both immiscible with glycerol can be used to increase the extraction’s effectiveness.

Paraffin can also be used instead of glycerol as the chaser for DMSO and this could be useful if the substrate you are trying to recover is polar. Then, extraction from paraffin into lower alcohols becomes possible since these alcohols will be immiscible with the saturated hydrocarbon (paraffin) medium.

Reaction Addition Modes: Controlling Chemical Reaction Stoichiometries

 

  A + B  Gives C + D

Most chemical reactions are either unimolecular or bimolecular in their kinetics. They are either addition/condensations where there is only one product, ie C and no D, or in other instances, some transformation also provides a coproduct, D, created along with C.

Reactions proceed when the energy supplied to a system is appropriate for the free energy of the reaction. More energy being supplied to the system can cause the reaction to proceed faster and/or it can promote alternative transformations that compete with the desired one. Applying excess energy for longer than necessary can degrade any of A, B, C, or D.

If any of the reactions

 C + (A or B) gives (E or H)

or 

A + A gives F  

or

B+ B gives G

can occur, then having the correct stoichiometry present throughout the time when the system experiences the required activation energy becomes important. If the stoichiometry is wrong then starting reactant polymerizations or overreaction modes can compete to the detriment of the desired outcome. A procedure that keeps the reactants in the correct stoichiometry throughout that period when adequate activation energy is supplied is likely to give a cleaner reaction.

Standard Chemical Processing

Now in the standard mixing procedure, a solution either of reactant A or B is added slowly to a solution of the complete charge of B or A already stirring in the reactor. If the reaction proceeds so quickly that it overheats, the addition is slowed or stopped allowing the rate of reaction to slow down as the reactant being added gets consumed and no more is being added. That is, the reaction is prevented from running out of control by letting the stoichiometry inside the reactor change radically. To control an exotherm the operators in such instances are forced to let the concentration of the reactant being added into the vessel fall towards zero.

For greater clarity and to emphasize the necessary safety issue let us perform the thought experiment of mixing all of the charge of these same A and B in methylene chloride and starting to heat the reactor. When the reaction between A and B sets in, the methylene chloride will start to reflux but as the temperature rises towards the boiling point of the methylene chloride the reaction between them goes faster and faster. The methylene chloride now boils faster than it can be condensed and gaseous methylene chloride blows liquid methylene chloride and all the rest of the reactor’s content out of the reactor and probably onto the roof of the plant building!

Identifying Reactions that Could be Affected by Stoichiometry

The standard mixing procedure is unlikely to be adversely affected by the ‘in situ’ stoichiometry if the reaction rates of all steps occurring in the reactor are slow compared to the addition time. This is easily determined. When one-half of the addition of the controlling reactant has been added, an in-process check should be done. If either the product or any reaction intermediates are significant compared with unreacted starting materials then the transformation is likely to be sensitive to the reaction’s stoichiometry. If nothing significant besides unreacted starting materials (or substances produced during the quenching of the reaction before analysis) are present, then it is very unlikely that the yield will be affected by stoichiometry. When hardly any change can be detected in this test it indicates that almost the reacting occurs mainly after the reactant being added is completely in the reactor and the stoichiometry is therefore essentially fixed throughout the actual reacting period.

Importance of the Enthalpy of Reaction Calculation

If you are going to perform experiments in which all the reactants are mixed together at once, you must take appropriate precautions. 

You must do a calculation of the enthalpy of reaction based on standard bond energies; however, the enthalpy of reaction (∆H) is only a crude approximation to the free energy of reaction (∆F) since ∆F = ∆H - T∆S. The change in entropy of the reaction (∆S) also contributes to the reaction’s driving force. For example, if the reaction leads to more moles of products than the moles of reactants, the entropy change is most likely going to be positive and the overall free energy decidedly more negative because of (-T∆S). The reaction will be faster than the approximation.  It will be necessary to absorb the net free energy of reaction through the agency of first- the temperature rise of the reactor contents ( mostly solvent),  then the energy absorbed in vaporizing some part of the reaction solvent, and finally by the energy drained off by the refluxing of a part of the solvent.

Absolute Necessity for Small Scale Trials in a Laboratory Fume Hood

Theory is just theory. If you are going to perform a reaction in which all the reactants are to be mixed together and then reaction initiated (either by heat or catalyst), you must try out the method at a scale where even if you have miscalculated there will be no more than a mess to clean up, not a laboratory disaster! The reaction which you have examined might not even be the reaction that occurs. Something much more energetic could surprise you.

Get Emergency Control Means Ready

Of course, your reaction vessel must be equipped with a means to directly in real-time read the internal temperature. Otherwise, you won’t be alerted soon enough that you have a problem.

If despite your calculations the temperature of the reactor accelerates in an unexpected way you want to be prepared to intervene to slow things down. Having a cold -bath ready into which the reactor flask can be plunged often can tame a reaction that is misbehaving. The advantage of this form of moderation is that it can be removed when the danger passes and the experiment continued.

Some reactions need to be quenched by injection. In many cases that quench material can be as simple as water. Of course, such an experiment is wasted but a runaway condition is avoided. 

When a runaway reaction is the least bit possible, the experimentalist should both wear appropriate personal protective equipment and have a fire extinguisher handy. It is important to prepare by imagining what might happen. For example, are there other chemicals in the hood you will be using that could exacerbate your runaway? 

Well then—can we actually mix reactants A and B together safely so that the correct stoichiometric is maintained at all times and a runaway exotherm is prevented?

 Two Simultaneous but Separate Additions

Simple in concept but mechanically difficult would describe the simultaneous addition to one reactor of two separate solutions, one carrying the A reactant and the other carrying the B reactant. The two addition rates of each need to be controlled so that the same required molar amount of each reactant is delivered throughout the addition time. If unwanted exothermicity occurs in the reactor both reactant streams need to be slowed the same amount or stopped simultaneously. Technically this is a big ask and practically it will only be approximately successful but even if it is only achieved approximately it can provide a closer approximation to constant ideal stoichiometry. This method of reactant addition, even if performed somewhat imperfectly, can give the experimentalist an indication of whether getting better stoichiometry using a different but more practical method is likely to deliver a useful benefit.

Storage Solvent and Reaction Solvent Method

Both reactants can be dissolved together completely in a volatile solvent (storage solvent) at a temperature and concentration at which reaction will not proceed and that solution can be added in a controlled fashion to a higher boiling solvent (reaction solvent) preheated to below its own bp. but above the bp. of the addition solvent so that the storage solvent will flash distill as it is added and the reaction between A and B will occur at the set stoichiometry in the reaction solvent solution.

Flow System Method

A continuous flow system may be possible where the reactants are bled separately but in the proper stoichiometry into a heated reaction zone and the product is continuously removed from the reaction zone. The problem with the flow technique is that for many reactions operating at the lowest normal reaction temperature, the heated zone would be quite long. Such specially configured equipment is not typically available. This can be solved by operating at a higher temperature in a higher boiling solvent treated in the reaction zone near the higher boiling point of this alternative solvent. 

For example, suppose it is known that A will react with B in refluxing dichloromethane and the reaction is essentially complete in three hours. The same reaction however might proceed equally cleanly in hot propylene carbonate in one minute. Then mixing two solutions one of A and the other of B; each dissolved in hot propylene carbonate so that a correct stoichiometric ratio and delivering it into the hot reaction zone in a flow reactor so that the transition of material through that zone was one minute could produce C more cleanly than heating much longer in dichloromethane because the correct stoichiometry has been more faithfully observed and the reaction period is very abbreviated.  If the reaction time is lessened, the reaction zone length can be shortened and some standard equipment may suffice. Now, whether creating the advantage of having the correct stoichiometry throughout the reaction period outweighs the disadvantage of applying more energy to the system than is actually required to get reaction can only be answered by experiment.

Reactants- All-In Method

Another means to achieve the same result would be to mix the correct stoichiometric quantities of A and B in the bottom of a reactor and then add a high-boiling solvent to it; in our example above this would be the propylene carbonate. The reactor is then heated to start the reaction. Normally mixing all of the reactants together and then heating to start their reaction is never done at scale because, as already pointed out, the exothermicity of the reaction is likely to cause a runaway wherein the entire contents of the reactor including any solvent would overheat and potentially even explode; however, using a sufficient quantity of a high-boiling solvent with a significant heat capacity and large enough heat of vaporization will allow any exothermicity to be absorbed first by the heating of the high-boiling solvent and subsequently by the heat of vaporization (refluxing) of a small portion of this solvent.

Using High-Boiling Solvents to Control Exothermic Reactions

The advantage to using a high-boiling solvent for such a reaction is not primarily its ability to dissolve the reactants and bring them together homogeneously. Good low-boiling solvents can do that. The high-boiling solvent can use its higher heat-of-vaporization, as well as its heat capacity, as a heat sink to soak up the exothermicity of the reaction.

In fact, the high-boiling liquid in which the reaction is being conducted does not need to be able to dissolve the reactants for it to be effective as a heat sink. So long as the reactants can blend together without solvation help the liquid medium does not have to dissolve them. Reactions ‘on-water’ for example are of this type. If at least one of the reactants is itself liquid there will be a reduced need for a dissolving type solvent. Another purpose for using a high-boiling liquid medium in which the reactants are insoluble is to provide sufficient bulk volume to satisfy the minimum stirrable volume requirement for the reactor.

Phenylboronic Acid: A Functional Tag to Enable Simple Removal of Excess Reagent or Coproduct using Chromotropic Acid

Chromotropic Acid for Extracting Boronates

The KiloMentor Blog articles emphasize ways to make the workup, separation, and purification of the product from organic reactions more cost-effective. Often this is enabled by phase switching methods that quickly take the desired material into one bulk phase and byproducts, coproduces, and the processing chemicals into another.

One way to dot this is to use a reagent or coreactant that has built into its structure some functionality that allows it to be subsequently extracted into an aqueous phase. 1-(3-Dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride is an example of such a reagent. The pendant dimethylamino functional group makes an excess reagent or coproduce or byproduct basic and so soluble in aqueous acid.

A functional fragment that can be used in this way but has rarely been adopted is the p-dihydroxy-boryl benzyl function.  This substructure appears in the Dobz protecting group for peptide synthesis [D.S. Kemp and David C. Roberts, Tet. Lett. 52, 4629-4632, 1975]. The dihydroxyboryl group forms a strong covalent linkage with the sodium salt of chromotropic acid [1,8-dihydroxynaphthalene-3,6-disulfonic acid] which is very soluble in water.

In the absence of complexing species, boronic acids are in reversible equilibrium with their cyclic trimers and water. Other species containing the group may be partially converted to Boronic anhydrides.

Although many reactions can be conducted in the presence of the free Boronic acid function  as a second option the Boronic acid can itself be protected as the N-methyldiethanolamine complex. N-methylethanolamine is of course itself readily taken into water in a workup.