Improving Recovery from Crystallization

KiloMentor consistently advocates a preference for scaling-up processes that are designed, as much as possible, to use intermediates that can be protonated or deprotonated at pH levels accessible in water and which, therefore, can utilize 'phase switches' as part of their purification. Such proposed paper syntheses are rugged to the extent that they do not depend for purification on the crystallization of hypothetical intermediates, whose physical properties are unknown when the route is being planned. Despite this bias against neutral isolated intermediates, once the molecular weight of intermediates in a proposed route exceeds what can be practically distilled, crystallization must remain the predominant isolation and purification method for these neutral, un-ionizable intermediates.

Indeed, crystallization is of such importance that it is taught at the early undergraduate stage in what still remains of laboratory training in universities. The disadvantage is that such early treatment is elementary and actual scientists who will work in the lab during their careers never seem to get around to more sophisticated discussions of it, but learn what more they can through experience - both good and bad.

Quite a bit of what I offer here is taken from a manual that soon will become an authentic antique: Laboratory Technique in Organic Chemistry, Avery Adrian Morton, McGraw-Hill Book Company, Inc. 1938! 

Trying to crystallize without first applying other methods of purification is the most common error. Because the deleterious effect of impurities upon the rate and completeness of crystal formation is so great crystallization of crude products should never be attempted until other methods of purification have been applied. Although immediate crystallization will usually reduce impurities dependably when crystals do form, the extent of crystallization achieved in a reasonable time is almost always far less than it would be with a purer starting sample. 

Thus it might be generally wiser to triturate high molecular weight compounds and distill or co-distill low molecular weight ones before trying to crystallize them.

Consideration should be given to first distilling in vacuo; co-distilling with another solvent (triethanolamine, quinoline, kerosene, glymes, fatty acids ); steam distillation with or without the presence of salt; superheated vacuum steam distillation; exhaustive digestion with a poor solvent or a hydrotrope; trituration with a poor solvent; extraction in a Soxhlet apparatus; passage through a plug of a solid adsorbent; or acid/base extraction. The last of these has utility even when the desired compound is neutral because the extraction has the capacity to remove acidic or basic impurities. Other applicable methods are treatment with insoluble polymeric derivatizing reagents to trap identified (or guessed impurities) or partitioning between two immiscible organic solutions. If the compound is a Lewis base, a cocrystal with another compound such as bis-N,N’-(3-nitrophenyl)urea may be possible. Specific Kilomentor blog articles treat many of these methods and these may be found using the blog's search tool.   

The influence of heat on solute solubility is marked. At higher temperatures, differences in the solubility of a substrate between different solvents are leveled. A purified hydrocarbon fraction is often as good a solvent when hot as, say, nitrobenzene. The only advantage of nitrobenzene in such a situation might be that it could hold back impurities from solidifying when the solvent is cooled while the hydrocarbon would often hold back very little.

More thought needs to be given to the purity of solvents used in crystallizations because solvent impurities, such as other solvent residues, can retard the rate of crystal nucleation and crystal growth just like other higher molecular weight reaction impurities can. Because solvent changes at scale are done by solvent exchange rather than by evaporation to dryness and replacement with the new solvent, the residue from the reaction solvent at scale can be higher than in the laboratory experience.

Besides the physical property difference between homologous alcohol solvents: methanol, ethanol, propanol, etc. process chemists need to be heedful that only ethanol has denaturants added to make the ethanol unsuitable as a beverage. These denaturants are different for different grades of alcohol and can change crystallization kinetics.

Water is the most omnipresent impurity in crystallizations, so much so, that in many cases efforts to work free of it are doomed to failure.  Polyhydroxy compounds such as sugars and glycosides are common materials affected dramatically by the presence of water in the crystallization environment. For example, sucrose of 72% purity has been shown to crystallize twice as fast as a 70% pure sample but only one-fifteenth as fast as the pure sugar. [A. R. Nees and E.H. Hungerford, Ind. Eng. Chem., 28, 893 (1936)] .

Water frequently forms a solvate with a compound of interest when it crystallizes and this can be helpful, or not, depending upon the properties sought.

Digestion

Digestion is hot trituration with the minimum amount of a poor solvent required to cover a crude solid and make it stirrable. Digestion as a purification method at scale requires some means to obtain an initial crude solid without evaporation to dryness.  Digestion is typically done by refluxing the liquid making up the slurry in order to equilibrate the impurities with the solvent solubility.

In addition to water and saturated hydrocarbon solvents, liquids unlikely to dissolve  a large amount of the main product in trituration, the following constant boiling azeotropic mixtures which are either hydrocarbon or water-rich might serve:

97.0% water 3.0% acetic acid azeotrope bp 76.6 C

91.0% water 9.0% Benzyl alcohol azeotrope bp 99.9ºC

87.1%heptane 12.9% water azeotrope bp 79.2 ºC

94.4%hexane   5.6% water azeotrope bp 61.6ºC

95.5% hexane 4.5% allyl alcohol azeotrope bp 65.5ºC

97% hexane 3% 1-butanol azeotrope bp 67.0ºC

91.5% Cyclohexane 8.5% water azeotrope bp 69.8ºC

83.7% acetonitrile 16.3% water azeotrope bp 76.5 C

72.9% Allyl alcohol 27.1% water azeotrope bp 88.2ºC

66% Allyl cyanide 34% water azeotrope bp 89.4ºC

77.5% Formic acid 22.5% water azeotrope bp 107.1ºC

Digestion from an Inert Support

A concept that synthetic organic chemists have not given sufficient consideration to is the evaporation of a reaction mixture onto an inert solid material from which byproducts can be digested away using poor solvents followed by dissolving the main product with a good solvent and filtering away from the inert solid support material.  
One possible reason for this is that synthetic chemists are not familiar with the properties of pharmaceutically acceptable excipients that could be used as the inert material for such evaporations. Inert solids such as microcrystalline cellulose, crospovidone, cross-linked polystyrene, calcium sulfate, calcium carbonate, calcium phosphate, etc. could be heated and stirred to just above the temperature of the solvent in which the reaction mixture is dissolved and the solution of the reaction mixture added onto it.

The solvent would be expected to flash distill out of the reactor and could be collected for destruction or reuse.  The method would have a particular advantage in that it could be used for recovering dipolar aprotic solvents which are up until now difficult to remove.
Evaporation of a reaction mixture onto a solid polymer is one means to evaporate to dryness; something that cannot be done on-scale in a large stirred reactor. Trapping of an intermediate on a polymeric support for isolation has been examined by Frans Muller and Brian Whitlock, An Alternative Method to Isolate Pharmaceutical Intermediates, Organic Process Res. & Dev., 2011 , 75, 84-90.

Evaporation of a solution of a reaction mixture onto chromatography media such as silica or alumina has been done to prepare a concentrated band of material that can be spread on the top of a chromatographic column to be eluted with a series of increasingly polar solvents.

Treatment of a solution of crude reaction mixture with charcoal has often been done to remove small amounts of non-polar high molecular weight impurities.

A paper has been published that illustrates the work-up of a chromic acid oxidation by pouring the crude reaction mixture onto a column of cross-linked, unfunctionalized polystyrene and eluting the column with methanol-water mixtures to remove first the inorganic salts and then the organic products. In this process, the solvent of the reaction mixture is simply diluted with the methanol-water eluate.

In the slightly different process I am contemplating, the reaction mixture would be added to stirred warm cross-linked polystyrene so that the reaction solvent evaporates and leaves the crude product in the solid resin. Then the inorganic components would be removed by digestion or trituration with methanol/water mixtures. Thereafter, the organic compounds would be eluted with a less polar solvent mixture.

Since charcoal has been used as an additive to consume excess oxidant, the first step after reaction completion could be treatment with a small amount of charcoal and filtration followed by evaporation onto free-flowing non-functionalized cross-linked polystyrene.

If we can obtain the crude solid solvent-free, we can also adopt the other digestion technologies using poor solvents for the principal product.