Good Reasons for Solvent Mixtures for Chemical Reactions?

 

Why do process chemists regularly optimize a chemical step using only a single reaction solvent? Neal G. Anderson in his valuable text, Practical Process Research & Development, dedicates an entire chapter to solvent selection; however, all that he says about using mixtures of solvents is “Sometimes a mixture of solvents will dissolve a compound much better than any one solvent .....” Rather obliquely, regarding the goal of solvent selection, he writes, “[Other] important considerations are to decrease waste and allow for efficient solvent recovery and reuse.”

Although purifying and recycling solvent is almost certainly easier if it is a single substance; generally, in the fine chemical industries, including making pharmaceuticals and pharmaceutical intermediates, solvents are not reused. The exception would be some product that achieves a massive volume. Thus, it might make sense for the owner of a composition of matter patent for a major pharmaceutical to use a single-component solvent so that solvent recovery would be simpler if that product became a blockbuster drug. 

The reason recycling is rare is economic. In a multipurpose plant, batch sizes are too small and the number of different solvents used is too many to make it worthwhile accumulating and storing used solvents for delivery to a solvent recycling specialist. As for purifying solvents themselves, the reactors in a multipurpose plant are too costly to be used for solvent recycling. Finally, recycling solvent within a fine chemical facility would only be acceptable so long as the same strict specifications could be met for recycled solvent as for other input reactants and this testing brings its own costly analytical burden.

My conclusion: most of the time using mixtures of solvents as the reaction medium is just as practical as using a single solvent. That is despite the almost universal traditional practice of using a single component solvent without asking the reason why it must be so. 

Let us examine some of the reasons solvent mixtures could be advantageous.

Throughput

A solvent mixture may well dissolve more substrate per liter than any single component medium can. Getting more substrate dissolved homogeneously in a reactor can improve the economics by increasing throughput, especially in early process steps which need to be run multiple times. (Solubility is the single advantage that Neal Anderson did mention.)

Cost

One particular solvent may possess a specially advantageous property while, at the same time, being prohibitively expensive. Using a mixture of this expensive solvent and a cheaper cosolvent may adequately preserve the special property while reducing the overall cost.

 

Increasing the Heat Capacity

The preferred solvent for yield optimization may be one that boils substantially above the best reaction temperature. Adding a co-solvent that boils at the desired reaction temperature can increase the heat capacity of the medium at the reaction temperature. The lower boiling solvent’s vaporization into the condenser and the returning condensate will stabilize the reactor temperature. Consequently, the addition rates of reactants can be higher without overshooting the optimum reaction temperature.

Changing a Phase’s Density

Some solvents are more dense, and some less dense than water.  In work-ups with water, sometimes having the product-containing liquid phase more or less dense than water can have an advantage. There can be fewer transfers between vessels. The number of large vessels needed to execute a process step may depend upon it. Processing times can be reduced and throughput increased. Fewer vessels mean less cleaning and a smaller burden on plant facilities.

Reducing the Solubility of a Product or Co-product

Decreasing the solubility of a product or co-product can cause it to precipitate as the reaction proceeds. This can drive an equilibrium towards completion, simplify isolation or raise the overall yield.

Making Telescoping Reactions Easier

Sometimes it is not useful to isolate a process intermediate but the solvents appropriate for the present and subsequent process steps are not the same. A solvent switch is required. Evaporation to dryness is not possible at scale. It would be advantageous if the second step in the telescoped pair was optimized in a solvent mixture consisting of a minor amount of the first solvent and a majority of the second solvent. If this were done it would not be required to substantially remove the first solvent. This might save substantial time and substantially reduce waste.

Because the Solvent Mixture Selected is a Constant Boiling Azeotrope

A constant boiling azeotrope has a fixed composition and it boils at a constant boiling point. In these respects, it is the same as a pure single molecular species. It can usually be purified by simple distillation. However, many azeotropes have the advantage that by changing the pressure-usually by reducing the pressure- the azeotrope can be split into its component substances for distillation. This distillation at a different pressure can potentially remove the better solvent and lead to the precipitation or crystallization of a solute.  

To Reduce Solvent Viscosity

Viscous solvents are often usefully high boiling but their viscosity is a problem for stirring and for heat conduction. Mixing with another solvent can reduce the viscosity of the reaction medium.

To Provide a Distillation Chaser

Adding a higher boiling solvent into a reaction solvent mixture can provide a chaser for reaction mixtures that are subsequently worked up by distillation. In ordinary distillation, sometimes a substantial amount of product is lost in the still pot and the distillation column. A solvent component that can act as a chaser can eliminate this loss. Of course, such a chaser could also be added after the reaction is over but before the distillation step.

Drying Simplicity 

Drying solvents at scale with inorganic salts followed by filtration of the inorganic salt hydrates uses labor, equipment, and time inefficiently.  It is greatly disfavoured for work at scale. The preferred method for solvent drying selects a solvent that forms an azeotrope with water and distills a portion of the solvent as the azeotrope. Such a solvent may usefully be part of the original reaction solvent liquid.

Raising the Freezing Point 

At what temperature does the solvent that is being considered solidify or become highly viscous? The freezing point can limit the range of temperatures that can be used in the optimization.  Lowering the temperature is often the best option for increasing the selectivity of the desired reaction versus competing reactions that produce by-products. If low temperatures create vicious reaction mixtures, these can result in hot spots during reagent additions. Inadequate mixing leads to incorrect stoichiometry, creating in turn by-products, and poor crystallization control. For example, DMSO when diluted with a small amount of toluene is more resistant to freezing and so can be cooled to a lower reaction temperature.

Pyridine-Water Selective Precipitation with Pyridine Recovery

 

Dissolution of a solute in a water-miscible solvent followed by crystallization or precipitation of the solute by gradual or portion-wise addition of water is an established method of separation and purification.  It is frequently applied to the separation of mixtures of different polymers.

 Solvents commonly used are methanol or ethanol. When lower alcohols are used with small molecule substrates it amounts to the same thing or at least strongly resembles crystallization from mixed alcohol-water solvent. When more expensive organic liquids are used as solvents to be practical at scale there must exist a cheap straightforward method to recover that solvent.

Pyridine is miscible with water in all proportions. It can be used to purify solutes or separate mixtures of solutes by the gradual addition of water so as to cause fractional precipitation. Typically one starts with something like a mixture of 5 parts pyridine and 1 part solute which can be warmed to dissolve what may be a solid or oily mixture; then, one gradually adds water with vigorous stirring until faint turbidity persists. At this point, optionally, a small amount of pyridine (a drop or two at the laboratory scale) can be added to just clear the haziness. Crystallization may begin after some time. In Aleksandra Smoczkiewiczowa and Jan Bielawny's paper in  P. Zakresu Towarozn. Chem.,Wysza Szk. Ekon. Poznaniu, Zesz. Nauk., Ser. 1 1970 No. 36, 149-62, it says that their cholesterol oxidation mixture was dissolved in a 5-fold amount of pyridine and by addition of water fractionally precipitated about 15% androstenolone acetate.

 Pyridine is somewhat expensive as solvents go. No obvious simple means to recover the pyridine when the precipitation is complete makes this an infrequently used methodology  Pure pyridine cannot be recovered by distillation because pyridine/water forms an azeotrope. Fortunately, there is a technical trick that does achieve this separation. Pyridine is not particularly soluble when sodium hydroxide is dissolved into the aqueous pyridine so the addition of enough caustic causes pyridine-water to separate into two phases. The pyridine layer can be separated and the mostly- layer discarded.

Propylene Carbonate as a Useful Solvent for Organic Synthesis Processes

Propylene carbonate has a boiling point of 242 °C. The density of propylene carbonate is 1.189 g/cc, and the dielectric constant and dipole moment, respectively, are 64.94 and 16.5.  It is a solvent that can be expected to provide good solubility for a wide range of organic reaction substrates.

Propylene carbonate may be the only solvent that is (a) usefully immiscible with water, (b) does not contain a halogen in its formula, yet (c) has a density greater than water. At 25°C the solubility of propylene carbonate in water is 8.3% and the solubility of water in propylene carbonate is 17.5%. Excess water forms a second phase on top of the water-saturated propylene carbonate. When the mixture of liquids is cooled to near 0° C the separation of phases is even greater. Thus, propylene carbonate can provide two phases that can be used for liquid-liquid extractions. Moreover, the more predominantly organic layer is the lower phase, and in a reactor can be simply cut through the bottom valve. Thus, when the product can be taken into the aqueous phase by acid or base, the organic phase can be removed, leaving the product in the reactor. This can save a vessel in a chemical process work-up.

 In fact, propylene carbonate is thermotropic with appropriate water mixtures. As the temperature is varied between 0 and 61°C, the two phases that derive from a particular weight fraction of propylene carbonate and water change compositions. It would therefore be expected that the partitioning of a mixture of substrates, such as might be the products from a reaction step,  could be optimized between the two phases both by varying the propylene carbonate/water weight fraction and by changing the temperature of the two-phase mixture. The UCST for propylene carbonate and water is about 72 C. At this temperature, only a single distinct, clear phase remains. 

Propylene carbonate can be hydrolyzed by both aqueous acid and aqueous base. There are both good and bad aspects to this. The bad news is that the stability of the solvent in contact with water is 

somewhat limited. This, however, is also true of ethyl acetate, where it is not regarded as a severe limitation. The good part is that small amounts of the solvent mixed with a hydrolytically stable cosolvent can be removed by hydrolysis since the products, carbon dioxide and propylene glycol, are both water-soluble.

Reaction Solvents that could be Worked-Up with Acetic Anhydride

Solvents that could be distilled away from acetic anhydride (bp. 140 C) and taken up into any solvent immiscible with acetic acid, after the hydrolysis of the acetic anhydride chaser and admixture with a little additional water to enhance immiscibility of the two layers:

Chlorobenzene

Nitro propane

Methyl chloroacetate

Cyclopentanone

Diethyl carbonate

Dimethyl sulphite

Tetrachloroethylene

2-nitropropane

Methylisobutyl ketone

N-methylmorphiline

Nitromethane

Toluene

1,1,2-trichloroethane

Trifluorotoluene

1,4-dioxane

Nitromethane

Methylcyclohexane

Heptane 

Propionitrile

Dibromomethane

Dimethylcarbonate 

Trichloroethylene

Isopropyl acetate

1,2-dimethoxyethane

Fluorobenzene

1,2-diethoxyethane

1,2-dichloroethane

Cyclohexane

Acetonitrile

Cyclohexane

Benzene

methylethylketone 

2-methyltetrahydrofuran

Ethyl acetate 

Butyl chloride

Carbon tetrachloride

Petroleum ether

Hexane

Diisopropyl ether

Chloroform

Perfluorohexane

1,1-dichloroethane

Methyl acetate

Carbon disulphide

Dimethoxymethane

Pentane 

Diethyl ether

Methyl t-butyl ether

Organic Solvents and Various Means for their Removal

 

Co-distillation

If two liquids are essentially immiscible, distillate comes over when the sum of their vapor pressures equals the pressure inside the distillation apparatus. The effect is that a small amount of a high boiling material will co-distill along with a larger amount of a lower boiling material. So long as the low boiling material is inexpensive and the higher boiling material is easily separable from this large amount of low boiler, co-distillation can physically separate the high boiler from less volatile or non volatile  components mixed with it. Aside from using water (which is classified below as steam distillation), the inexpensive material most frequently used in a co-distillation is kerosene/paraffin/lamp oil/coal oil. Much less frequently silicone oil (dimethicone KF-96L-2cs) has been used as the high boiling component.

Steam Distillation

Steam distillation represents the particular case of co-distillation using water. Many higher boiling solvents can be chased by steam distillation. Nitrobenzene and 1,1,2,2-tetrachloroethane are frequently removed this way after Friedel-Craft reaction. The steam can be preheated to temperatures above 100 C thereby co-distilling a larger portion of the lower boiling material while minimizing the volume of water. Codistillation including steam distillation can also be done under vacuum so that the material of interest is not exposed to as much heat.

Reversible Degradation

Some special compounds can be used as solvents that are reversibly cracked into smaller more volatile fragments by heating. Piperylene sulfone has solvent characteristics very similar to tetramethylene sulfone. The essential difference for our consideration here is that the former is fragmented by heat into  1,3-pentadiene and sulfur dioxide. These can be trapped together as a distillate whereupon they reform piperylene sulfone.

  Likewise, dicyclopentadiene upon heating can disaggregate into the monomeric cyclopentadiene and be distilled out of a reactor and away from less-volatile products.

Hydroxymethanesulfonic acid is a strong acid solvent that upon heating breaks apart into formaldehyde, water, and sulfur dioxide before reforming when they are recondensed together. 

Degradation

Some solvents can be hydrolyzed into water-soluble fragments.

 

Acetic anhydride can be hydrolyzed into acetic acid.

 

Dimethylformamide can be hydrolyzed by acid into dimethylamine and formic acid.

 

Dimethylacetamide can be hydrolyzed into acetic acid and dimethylamine.

 

Propylene carbonate can be hydrolyzed into 1,2-propanediol and carbon dioxide.

Is Glycerol the Best Green Solvent for Process Chemistry?

In a paper published online in 2006, Glycerol as a Green Solvent for high product yields and selectivities,  A.Wolfson, C. Dlugy, and Y. Shotland provide ineluctable evidence that increased utilization of glycerol should be forthcoming in organic chemistry processing.  Their reasoning included environmental, economic, safety, handling, and product isolation considerations. The only difficulty using glycerol appears to be a viscosity much higher than with standard organic solvents.

Yet irrefragable as their teaching was there are still things unsaid that further boost this processing chemical! Glycerol can be used to conveniently perform all kinds of solvent switches which are otherwise a weakness with the standard reactors in multipurpose plants. 

The best solvent for conducting a chemical reaction may not be the best solvent to purify the product thereof. Furthermore, when a first and a second reaction in a process scheme are telescoped (avoiding the isolation and purification of the product of that first reaction) the best reaction solvents for the first and second reactions are likely to be different. But, as I have explained in another blog article, switching solvents in the plant cannot be done in the simple fashion of evaporating a first solvent to dryness on a rotary evaporator and pouring in the second solvent. Because of the non-zero minimum stirrable volume problem, the complete removal of a first solvent becomes complex and time-consuming.

 

This problem can be solved if a volume of glycerol sufficient to completely occupy the minimum stirrable volume is placed into the reactor.  Then, all of a first solvent (for example a first reaction solvent) can be removed without distilling any of the glycerol, which is so high-boiling, yet the reactor remains stirrable throughout because glycerol occupies the minimum stirrable volume and provides continuous mass and heat transport.   All of the product, coproducts, byproducts, and other processing chemicals (everything that can’t distill with the first solvent) remain with the glycerol. Now, all that is necessary to complete a solvent change is to add a second solvent that is immiscible with glycerol but will solubilize and extract at least the desired product and potentially the entire non-solvent reaction contents.

Here, another benefit of glycerol becomes apparent. According to the (1974-1975) CRC Press, Handbook of Chemistry and Physics, (the old one I have) glycerol is immiscible with at least:

acetone, benzene, butyl acetate, carbon tetrachloride, chloroform, dibutyl ether, diethyl ether, ethyl acetate, isoamyl alcohol, methyl isobutyl ketone (MIBK), nitromethane, petroleum ether, tributyl phosphine.

Furthermore, by analogy, glycerol can be expected to be immiscible with the additional common solvents:

 hexane, cyclohexane, methylcyclohexane, heptane, toluene, methylene chloride, trichloroethylene, tetrachloroethylene, isopropyl acetate, nitroethane, 2-nitropropane, and t-butyl methyl ether.

So a variety of extractions are possible. 

In a further extension, glycerol can be used as a cosolvent with fluids with which it is miscible, such as the lower alcohols methanol, ethanol, isopropanol, etc. It can also form gas-expanded liquid phases with such amines as ammonia, methylamine, and ethylamine. A mixed solvent will generally have a more moderate viscosity than glycerol alone and so will at the same time make processing easier. All the above materials, immiscible or miscible with glycerol, can be completely expelled from the reactor so long as the amount of glycerol equals, at least, the minimum stirrable volume.

When a reaction is conducted wherein the desired product is itself usefully volatile, glycerol can act as an effective chaser in a distillation that purifies the desired product.

Glycerol is cheap, biodegradable, and has a bp of 182 ºC @ 10 mm Hg. It would be expected to remain behind in a standard distillation when combined with any of the common organic reaction solvents. Even DMSO (bp 189℃), DMF (bp153℃), and NMP (bp 81-82℃ 10 mm Hg) would be expected to be chased by glycerol.

The distilled first solvent contaminated with traces of glycerin upon simple treatment might be ready for reuse. Thus the first solvent is no longer waste and there are no mixed fractions of solvents to dispose of. The waste glycerin is a biodegradable material and the quantity used is no more than the minimum stirrable volume of the reactor.

Whether used just to drive the removal of a first solvent during reaction work-up or as cosolvent for the first reaction and then a chaser for the first solvent, glycerol it seems has the properties that enable simpler processing.

What can You Trust about Chemical Patents?

Advice About Reading Patents

In my career, I have been the inventing scientist and I have been the person drafting a patent before being sent to a patent agent to incorporate the legalese.  I have even been both for the same patent! What can you learn from a patent? What part is ‘fake news’?

The conscientious scientist writes down the experimental description of the significant experiments carried out. The scientist explains what his/her interpretation of these results means generally. He explains what application of the present theoretical understanding of his subject leads him to predict generally based on this work. He states why he believes it is useful and may prove more widely useful.

Then the conscientious patent agent takes over. The patent agent's professional responsibility is different. You see, you have to pay to hold a patent: there are annual fees that must be paid to maintain this monopoly and it isn’t worth paying money unless the barrier to use is strong enough to prevent competitors from learning from your insight, slightly tweaking your protocol and getting essentially the same benefit. But it isn’t worth the time filling out your precious scientific expertise to perform all the possible permutations and combinations of every aspect of the methodology so everything can be presented in its own experimental and claimed in its own claim. Instead, the patent agent asks you to imagine, using your experience, imagination, knowledge of the literature, and knowledge of chemical theory to guess all the other conditions that would give, to some degree, the claimed outcome. 

The patent agent then writes a description of your invention including the full breadth of your considered opinion of what could work. Claims are also constructed encompassing both your educated guesses and what you have shown experimentally works. These are called the broad claims. They are there to prevent those who follow what you have taught from escaping your legal protection by making inconsequential or obvious changes or substitutions.

So what can you trust? According to US patent law at least one narrow claim must cover the best protocol for practicing the invention that the scientist has found out at the time that the invention patent is filed. Normally what is described in the experimental and talked about most completely in the body of the patent is the part that has been proven in the lab.