The manufacture of alkaloids from opium
Assistant Professor of Chemistry, University of Montreal, Montreal, Canada
Author: Walter R Heumann
Pages: 34 to 40
Creation Date: 1957/01/01
Assistant Professor of Chemistry, University of Montreal, Montreal, Canada
The classical processes for the extraction of alkaloids from opium of Merck, Thiboumery-Mohr and Robertson-Gregory have been critically reviewed a few years ago by Barbier (1), (2). Characteristic features of all these processes are the tendency to obtain opium extracts as concentrated as possible and to precipitate the alkaloids from such highly concentrated extracts. Any one of these processes has to use heat in some of its phases, either for concentration by evaporation or for precipitation from alkaline solutions. The aim of these principles may be to obtain high yields of morphine or to reduce the quantities of liquors for filtration. It must, however, not be overlooked that the alkaloids are the more soluble, the higher the concentration of impurities is. Prolonged action of heat can have a deteriorating effect on morphine, especially in alkaline solutions and in contact with air. Concentrated solutions of high-molecular impurities are often difficult to filter or to handle properly without loss, as they are sticky and viscous. The raw morphine obtained by these methods is necessarily of low purity and yields of secondary alkaloids tend to be low or even null.
In 1924, Kanewskaja (16) described a process, which appears interesting by its relative simplicity and the tendency to use organic solvents and extractions instead of the rather crude and inefficient ancient methods for the separation and isolation of the alkaloids. The aqueous opium extract is first precipitated in presence of alcohol, whereby only morphine and narcotine separate, the other alkaloids remaining in the solution together with the resinous impurities. Narcotine is removed from the morphine-containing precipitate by extraction with chloroform. The other secondary alkaloids are extracted with benzene from the alkaline mother-liquors of the raw morphine. Upon extraction with dilute acetic acid this benzene solution releases only codeine and thebaine into the acid solution, whereas papaverine remains in the benzene. Thebaine is separated from codeine by precipitation with ammonia, which does not precipitate codeine.
In 1927, Schwyzer (18) published a detailed commercial-scale process, which was based on the old Robertson-Gregory method. In 1931 the same author (19) published a modified method, which was aimed at improving the recovery of secondary alkaloids. In this procedure Schwyzer made use of his interesting finding that concentrated opium extracts, when diluted with several times their volume of water, release considerable amounts of resinous impurities, and the author purified in this way the concentrated opium extract before precipitating the alkaloids.
The process of Barbier (1), (2), first described in 1947, did not abandon the classical principles of the ancient processes, but this author made an incontestable contribution to the improvement of existing methods by introducing the slightly soluble acide morphine tartrate as a very efficient and convenient means of purification of raw morphine. He also put more emphasis on the recovery of the secondary alkaloids and he was the first to describe the recovery of morphine from mother-liquors by counter-current solvent extraction.
Characteristics of the New Process
Although the process to be described is to a certain extent derived from principles used in the above mentioned more or less conventional procedures, it deviates from them in the following essential points:
Precipitations of raw alkaloids from aqueous opium extracts are made from rather dilute solutions. Experience showed that the alkaloids are less soluble in solutions less concentrated in impurities, and the alkaloids thus precipitated are not only of higher purity, but also more easily filtered and washed.
Close control of pH is used to obtain fractionated precipitations of alkaloids. Thus, meconic acid and narcotine are first precipitated from the initial aqueous opium extract by adjusting a lower pH than that required for the precipitation of the other alkaloids. The raw total alkaloids obtained by that method contain considerably less impurities, their weight being approximately only one-half that of the product obtained by the usual bulk precipitation. They are more conveniently separated and purified, and the equipment needed for these operations can be of much smaller size.
Concentration by evaporation of alkaloid-containing aqueous liquids is avoided. The vacuum evaporation of aqueous opium extracts is a time-consuming and delicate operation complicated by frothing and scale formation. Certain alkaloids, morphine in particular, are affected by heat and furthermore, the quantitative removal of the alkaloids from the syrupy concentrated opium extract is always doubtful.
Concentration by evaporation and subsequent precipitation of the alkaloids are replaced by liquid-liquid extraction. The alkaloids are first extracted from dilute aqueous solutions such as mother-liquors with organic solvents. Then they are extracted from the solvent with a small amount of dilute aqueous acid, giving thus a concentrated solution. There are now far more organic solvents commercially available and at lower prices than a few decades ago, so that the continuous counter-current liquid-liquid extraction has become, in certain cases, a more economic unit operation than the vacuum evaporation. Liquid-liquid extraction can be used not only to separate and isolate the alkaloids rapidly, but it may also contribute considerably to their purification, since impurities may be insoluble in the properly chosen solvent.
Extreme conditions of concentration, temperature and pH are avoided as far as possible. As already mentioned, the handling of highly concentrated opium extracts is not very convenient, and such extracts release the alkaloids less easily and in a considerably less pure state than a dilute solution does. Higher temperature, particularly when combined with an alkaline reaction or when the solution is in contact with the air, can be deleterious to morphine. The same applies to very high or very low pH. Addition of mineral acids to the extraction of opium causes the extracts to be much darker and higher in content of resinous impurities. Barbier (2) considered it necessary to add sulfuric acid to the last extraction step to insure the extraction of all secondary alkaloids of importance. I found, however, that 1% to 2% of acetic acid added to the extracting liquid gives the same result, with the advantage that the extract thus obtained is much lighter in colour. Minerals acids, when added even in only slight excess, can also provoke decomposition and discolouration at other points of the separating and purification procedures.
This principle still has another advantage - namely, the possibility of using in many phases of the process rather simple and inexpensive equipment such as wooden vats, precipitation vessels or even vacuum filters. The use of expensive corrosion-resistant apparatus is limited to a minimum.
Instead of purifying the alkaloids by repeating the same operation such as crystallization or precipitation, a sequence of several treatments, different in principle from each other, is used where possible. This principle is an efficient means of removing complex mixtures of chemically different impurities. The raw morphine, for example, is first precipitated as the free base from the alkaline solution of its calcium salt; then the base in powder form is extracted with alcohol, and, finally, it is crystallized as the acid tartrate from acid solution.
Emphasis is given to the recovery of the secondary alkaloids papaverine, codeine and thebaine. Elimination of impurities in the earliest possible stage of the process not only simplifies the purification of morphine, but also helps in obtaining higher yields of secondary alkaloids. The above-mentioned fractionated precipitation from the initial opium extract is an example of this principle.
The recovery of the secondary alkaloids is interesting from a commercial point of view with raw materials high in codeine and thebaine content, such as Iranian opium, and it becomes particularly attractive when the price of the opium is based on its morphine content only.
The process described, when applied commercially to Turkish or Iranian opium, gave morphine yields of at least 96%. The average percentage yields of secondary alkaloids were as follows:
0.4 - 0.5
0.4 - 0.5
0.3 - 0.4
1.5 - 1.8
0.5 - 0.6
1.3 - 1.5
Description of the Process
It is not so much the aim of this article to provide a detailed working formula as to discuss the principles of the process. More detailed technical data are therefore given only in a few instances, where the author found their mention to be of particular interest.
Extraction of the Opium
The dissolution of the opium is done in the usual manner. The opium is cut into slices and stirred with five times its weight of water. Heating to 45-50°C is applied to speed up disintegration. The one-pound bricks of Iranian opium can be used without cutting them.
The filtration and washing of the completely disintegrated opium follows to some extent the known pattern of the counter-current method as described by Schwyzer (18) and Barbier (1), (2), but does not aim at obtaining a highly concentrated extract. The marc which remains after the separation of the aqueous extract by filtration is washed five times with 1 to 2 parts of 2% acetic acid for one part of opium. The main extract, together with the first two washings-i.e., about eight parts of liquid for one part of opium, is used for precipitating the alkaloids, whereas the following three washings are used instead of water to extract the following batch of opium. Once the extraction process is started, pure water will be used for washings only, and acetic acid will be added, starting with the second washing.
The filtration is done by vacuum, which has already been recommended by Schwyzer (18). As contrasted with Barbier's observation (1), (2), even the highly resinous Iranian opium can be satisfactorily filtered by vacuum, under certain conditions. The opium extract, together with the marc, is placed on the filter and allowed to drain without vacuum over night. The following morning the vacuum is applied and the filtration completed within approximately six hours. The remaining marc, which now forms a solid pasty mass, is thoroughly mixed with the washing water and allowed again to drain without vacuum over night, and so forth. As the washings progress the filtration becomes easier, so that the last two washings can be done within one day with vacuum being applied right from the beginning. On a cylindrical vacuum filter of the Buchner type of 1,400 mm in diameter at least 200 lb of Iranian or 250 lb of Turkish opium can be filtered as described. Wooden filters are very convenient, as they are inexpensive, durable and easy to maintain. One batch will occupy a filter for six days, which means that even a medium-size manufacture will require quite a number of filters and much space. This may appear to be a disadvantage, but there are definite merits in this method. All operations with the opium are done on the filter itself, which means clean work without losses. Furthermore, the washings are very efficient, as the marc is macerated thoroughly when the washings are allowed to drain freely during the night. In other methods, such as centrifuging or decanting, work is either dirty or separation inefficient and loss of alkaloids therefore inevitable.
Precipitation of the Alkaloids
The opium extract is neutralized by the addition of sodium hydroxide solution and the pH is adjusted to 7.0. The resulting crystalline precipitation of narcotine is easily filtered off by centrifuging and washed with water until free from other alkaloids. Sometimes this raw narcotine may contain small amounts of papaverine. In such cases the precipitation should be tried at a slightly lower pH. To the filtrate and the washings of the raw narcotine, a concentrated solution of calcium chloride is added to precipitate meconic acid as its insoluble calcium salt. The latter is filtered by centrifuging and washed until free from alkaloids as easily as the narcotine. If narcotine is not to be used, both narcotine and meconic acid can be precipitated together by replacing sodium hydroxide with calcium hydroxide.
The combined filtrates and washings from the meconate are further precipitated by adjusting the pH to 9.0-9.2 by the addition of powdered sodium carbonate. The total alkaloids, which precipitate now, are centrifuged and washed the following day. After drying at a temperature not exceeding 60°C, the raw total alkaloids are obtained as a brown, granular powder. The mother-liquors from these raw alkaloids still contain a small amount of alkaloids, which are recovered by solvent extraction as described later.
Separation of the Raw Alkaloids
The separation is based on the known fact that morphine is practically insoluble in non-polar solvents such as hydrocarbons or their halogenated derivatives, whereas the other alkaloids and most impurities are very soluble in these solvents. Barbier (2) recommended benzene, but the author prefers trichlorethylene for several reasons. Its vapours are less disagreeable, less harmful from a medical point of view, and non-explosive. This solvent is easier to drain off from the undissolved morphine, and the fraction which remains in the latter is easily removed by steam distillation.
The apparatus used somewhat resembles the one described by Barbier (2). It consists chiefly of a cylindrical extractor, which is at the same time a pressure filter and distilling apparatus. Filtration is obtained by a cloth-covered false bottom placed as close as possible to the bottom of the vessel. The latter has a central outlet, by which the filtrate can be drawn off to another distilling apparatus. The extractor has a steam jacket for indirect heating and a perforated steam-pipe right under the filter bottom for distillation with direct steam. A strong and rather slow stirrer rotates close to the filter bottom in order to keep the solids, which have a considerable resistance, in continuous slow agitation. In the cover of the extractor there is a connexion leading to a combined condenser for both reflux and distillation. Two other inlets are connected to compressed air and vacuum respectively. The apparatus may be used for both vacuum and ordinary pressure distillation. The raw total alkaloids from 100 kg opium weigh between 15 and 20 kg, depending on the opium used. For this quantity a capacity of the extractor of 100 l above the filter bottom is required.
Fifty litres of trichlorethylene for every 100 kg of opium are first introduced into the extractor. While the stirrer is rotating, the dry alkaloids are added through a manhole. Then the solvent is brought to the boil, and stirring is continued for one hour, whereupon compressed air is introduced and the solvent drawn off through the bottom outlet. This treatment is repeated twice with half the volume of solvent each time, but without the use of heat. The raw morphine remaining in the extractor which should be free of secondary alkaloids, contains a certain amount of solvent, which is driven off by introducing direct steam from beneath the filter bottom. As soon as all the solvent has passed over, heating is continued only by the steam jacket, and the major part of the water, now present in the morphine, is removed by vacuum distillation. The almost dry raw morphine is removed as a light brown powder. When completely dry, it is difficult to handle without loss due to dusting. The yield from 100 kg of opium is from 18 to 20 kg of this product, containing about 20% water. The solvent extract contains the secondary alkaloids papaverine, codeine and thebaine, the recovery of which is described later.
Purification of the raw morphine
The raw morphine contains 10% to 20% impurities, and it can be purified in several ways. Depending on the purity of this product and on the desired purity of the refined product, either one or two or even all three of the following treatments will give the wanted result.
Purification by lime. - The slightly humid raw morphine is mixed in an open stirring vessel with four times its weight of water, and slaked lime is added until the solution remains strongly alkaline to phenolphthaleine. The solution now contains all the morphine as calcium morphinate and is filtered off by vacuum on a Buchner filter. The residue on the filter, a mixture of lime, impurities and may be some calcium meconate, is washed free of morphine and then discarded. The combined filtrates are transferred to an open stirring vessel, where a 20% ammonium chloride solution is added to bring about complete precipitation of the morphine base. The latter is filtered either by vacuum or centrifuging after two days and washed with water. It contains then 20% to 25% of water. The combined mother-liquors, which contain a small amount of morphine, are extracted with solvent, as will be described later.
Purification by alcohol. - The raw alkaloids contain certain resinous impurities which accompany the morphine tenaciously, so that they are not removed either by the solvent extraction of the total alkaloids or by the lime treatment of the raw morphine. These impurities are probably of phenolic nature, like the morphine itself, and they are therefore easily soluble in ethanol and can be removed by treating the raw morphine with a mixture of about 4 volumes of alcohol with 6 of water. When a higher percentage of alcohol is used, too much morphine will dissolve together with the impurities and will be difficult to recover. The apparatus used is a cylindrical extractor equipped with a jacket for heating and cooling, a bottom outlet, medium speed stirrer and condenser for both reflux and distillation. Its capacity is of 90 to 100 l for each 100 kg of opium. The vessel is charged with 45 to 50 l of 40% alcohol for each 100 kg of opium and the morphine is added under stirring. The content is brought to boiling, which is maintained for one hour. Then the mixture is cooled and drawn off into a crystallization vessel, where it is allowed to stay over night. The next day the mixture is filtered on a Buchner filter and washed with small portions of 40% alcohol. The morphine thus treated forms a light brownish or greyish crystalline mass, containing about 25% humidity.
From the dark-brown alcoholic mother-liquors a small amount of morphine is recovered as follows. The alcohol is distilled off, whereupon the resins separate from the residual water. They form a liquid viscous mass when hot and are drawn off by the bottom outlet. These resins are insoluble in water or in pure acetone, but they are soluble in acetone of 50% to 70%, whereas morphine is insoluble in aqueous acetone. The resins are dissolved in up to twice their volume of warm aqueous acetone and filtered after two days on a Buchner filter. The residue is washed with 60% acetone and dried. It consists of raw morphine, representing about 1% of the total morphine content of the opium and it is purified in the same way as the main batch of raw morphine.
Purification by tartaric acid.-This operation is based largely on that described by Barbier (2). Following the treatment by either lime or alcohol, or both, the humid morphine is mixed in an open heated stainless steel vessel with five times its weight of water. Tartaric acid is added until a solution is obtained, which will be slightly acid to methyl red (pH approximately 5). An excess of acid has to be avoided, otherwise morphine bitartrate could crystallize. The solution is brought to 80°C and charcoal is added to de-colourize it. The type of charcoal must be chosen carefully by laboratory tests, as certain commercial grades absorb considerable amounts of morphine, which would be rather difficult to recover. After several minutes the charcoal is separated by vacuum filtration and washed with hot water. The combined filtrates are transferred to an open stainless steel crystallizing vessel, equipped with a water cooling jacket. Somewhat more than the quantity of tartaric acid which has been used to dissolve the morphine is now added, while stirring, so that the solution will be acid to methyl orange. The morphine bitartrate begins to crystallize immediately and in order to prevent the formation of crusts or hard lumps, stirring must continue at least until the major part of the morphine has crystallized. The bitartrate is centrifuged the following day, and after washing with water it forms a slightly brownish crystalline mass. It is dissolved in three times its weight of water by addition of sodium hydroxide until the solution is only slightly acid to litmus. If this solution were still dark-coloured, it could be decolourized once more with charcoal to which some sodium bisulfite can be added. If decolourizing is not necessary, the morphine can be precipitated immediately. This is done by adding, first, ammonia until the reaction is slightly alkaline and then sodium carbonate powder until the pH reaches 9.0 to 9.2. If carbonate were added right from the beginning, instead of ammonia, the solution could froth over. The morphine is thus precipitated as a fine, light powder, which is centrifuged and washed with water the following day. After drying at 60 to 70°C an almost white powder of at least 98% purity is obtained, which purity is sufficient for the conversion into either morphine hydrochloride or codeine or ethylmorphine.
The mother-liquors of the bitartrate, which retain about 10% of the morphine, are made strongly alkaline with slaked lime, whereupon a massive precipitate of calcium tartrate forms. This precipitation is necessary only when these mother-liquors are mixed afterwards with calcium-containing liquors, before entering the solvent extraction, where precipitation of calcium tartrate must be prevented for obvious reasons. The calcium tartrate is centrifuged and washed with water until free of morphine, and then discarded. The combined filtrates are treated with a 20% solution of ammonium chloride in order to precipitate the morphine, which is centrifuged and then purified again, either by alcohol or tartaric acid or both. The alkaline mother-liquors of this recovery procedure are extracted by solvent as described in the following paragraphs in order to recover the small amount of morphine remaining in solution. The same refers to the mother-liquors of the main batch of purified morphine base.
Recovery of Morphine from Mother-liquors
All alkaline mother-liquors from morphine precipitations contain varying amounts of morphine from one-half to several parts per thousand, depending on the purity of the liquors. It is best recovered by extraction with organic solvents at pH 9.1, the isoelectric point of the morphine. From every 100 kg of opium about 1,000 to 1,200 l of such liquors are obtained, which contain up to one-tenth of the morphine originally present in the opium. The mother-liquors are first collected in an open wooden vat or iron storage tank, and after adjustment of the pH to 9.1 to 9.3, allowed to stay for several days. It is advantageous to mix liquors of different morphine content. The more concentrated liquors may, after mixing, release some morphine together with resinous impurities, which they could release otherwise during the extraction in the packed column, thus causing the formation of emulsions, the principal enemy of liquid-liquid extractions. Moreover, the mixed liquors, varying less in morphine content, allow an easier control of the column operation. If, after several days, a deposit has formed in the storage tank, the clear liquid is decanted into the feeding tank of the extraction column. The remaining slurry is filtered off, the filtrate transferred into the feeding tank and the residue treated in the same way as the raw total alkaloids.
The proper choice of the solvent is important for several reasons. Morphine is only slightly soluble in all water-immiscible non-polar solvents, so that a large quantity of these would be needed. It is more soluble in hydrophylic polar solvents, such as aliphatic alcohols and phenols, but these solvents are more or less miscible with water. Barbier (2), who described for the first time the counter-current solvent extraction of mother-liquors from the opium extraction, used butanol, working at 60°C. This author ascribed the satisfactory results he obtained to the partial solubility of butanol in the aqueous liquor, which seems to be in contradiction to current concepts on partition phenomena as applied to liquid-liquid extraction. It appears rather probable that the higher temperature used by Barbier may be credited with the good extracting effect, as the solubility of morphine in polar solvents increases considerably with the temperature. It has been known for some time that certain mixtures of solvents show a markedly increased dissolving power for morphine. In 1936 F. Hoffmann-LaRoche and Co. (12) developed a procedure for the manufacture of alkaloids from poppy straw, where a mixture of equal volumes of butanol and benzene was used for the extraction of morphine from the aqueous poppy straw extract. Indeed, mixtures of aliphatic alcohols with cyclic hydrocarbons were found to be some of the best solvents for this purpose. One gramme of morphine dissolves in 360 ml of a benzene-butanol mixture 1 : 1, whereas the solubilities in the individual pure solvents are 1 part in about 5,000 for benzene and 1 part in about 140 for butanol. Morphine is also quite soluble in phenols such as the cresols, and a small amount of the latter added to the above-mentioned mixture has a striking effect on the solubility of morphine. As little as 5% to 10% of cresols added to the mixture of equal volumes of benzene and butanol will double the solubility of morphine. Moreover, the solubility of the solvent in the aqueous phase, which is about 7% for the binary mixture, decreases markedly by the presence of cresols. In the concentration range of about 1 g of morphine per litre of aqueous liquors the partition coefficient of morphine between equal volumes of the cresol-benzene-butanol phase (10 : 45 : 45) and water at pH 9.1 is 95/5, whereas for the same system without cresol it is of the order of 90/10. Benzene is a rather volatile and hazardous solvent. The addition of cresols somewhat decreases its volatility, but toluene used instead of benzene may still better serve this purpose. The solubility of morphine is only slightly less in a toluene-butanol mixture than in a benzene-butanol one. It can be figured out that for the complete extraction of liquors containing 1 g of morphine per litre, at least half their volume of benzene-butanol 1:1 is needed. For higher concentrations of morphine, either a larger amount of this solvent or the addition of some cresol to the latter is necessary.
The extraction is carried out in a packed-column counter-current apparatus. Both liquids, the alkaloid-containing mother-liquors and the solvent, are introduced into the column from feeding tanks, which are placed above the column, their flow by gravitation being controlled by flow meters. The solvent, being less dense than the aqueous phase, is introduced at the bottom of the column, while the aqueous liquors enter at the top. The height of the packed column is at least eight times its diameter. When Raschig rings are used as packing material, the amount of mother-liquors that can be safely extracted per hour will be roughly equal to the total volume of the column. For example, a column of 300 ? 2,700 mm filled with rings of 15 to 20 mm in diameter, will have a capacity of 200 l of aqueous liquors per hour, or a 500 ? 4,000 mm column will have about four times this capacity. This throughput will somewhat vary with the quantity of solvent needed. The latter depends on the composition of the solvent mixture, as well as on the concentration of morphine in the aqueous phase, and will vary between 0.5 to 1.0 volume per volume of aqueous phase. The column extends at both ends beyond the packed zone to provide space, where the two liquids can separate before leaving the column. The height of each of these two extensions is about twice the diameter of the column.
The aqueous liquors leaving the column by the bottom are free of alkaloids, but they contain a small amount of dissolved solvent which can be recovered by distilling off one-tenth of the volume of the liquors. The solvent leaves the column by the top charged with the alkaloids. It is further introduced into the bottom of another packed column, about half the size of the preceding one, where it meets an aqueous solution of sulfuric or phosphoric acid, which easily extracts the alkaloids. The first column can be made of sheet iron, whereas the second one has to be of stainless steel. Ten per cent sulfuric acid enters the latter column by the top, flowing from a feeding tank at such a rate that a concentrated, slightly acid solution of alkaloids can be drawn off from the bottom. If the mother-liquors are so rich in codeine that the only slightly soluble codeine sulfate precipitates from the acid extract, then phosphoric acid should be used instead of sulfuric acid. The acid alkaloid solution is adjusted to pH 9.0 to 9.2 by the addition of ammonia, and the precipitated morphine is filtered off after two days. It contains 70% to 75% of pure alkaloid and is purified the same way as the raw morphine. The mother-liquors of this morphine precipitation contain varying amounts of codeine and thebaine, which are recovered by extraction with trichlorethylene as described later.
The solvent, which leaves the acid column by the top, contains small amounts of the acid alkaloid solution, which are extracted by washing with water in a third packed column similar to the preceding acid column. This washing water is used to prepare the dilute acid for the second column. The solvent is then collected in a storage tank, and its composition is determined and readjusted to the desired formula by the addition of fresh components. It is then ready to re-enter the extraction in the first column. As it retains only small amounts of impurities after each cycle, redistillation may be necessary first after some 10 to 15 cycles. With a properly designed extraction plant the losses of solvent per cycle will not exceed 1% to 2%.
The Secondary Alkaloids
The separation of narcotine from the opium extract prior to the precipitation of the total alkaloids is described in one of the preceding chapters. The other secondary alkaloids, papaverine, codeine and thebaine, appear at two points of the process. The mother-liquors from the morphine recovered in the extraction column contain codeine and thebaine, whereas the trichlorethylene extract of the raw total alkaloids contains thebaine and papaverine and sometimes also codeine. The latter forms a soluble complex with ammonia. Since ammonia is always present in any aqueous opium extract, codeine is usually not precipitated by sodium carbonate from opium extracts. In the case of opium of very high codeine content, such as Iranian, some codeine may, however, appear in the precipitated raw total alkaloids. It may also be mentioned here that Schwyzer (16) has already pointed out that, contrarily to the general opinion, codeine can easily be extracted by solvents from only slightly alkaline solutions of its ammonia complex.
The secondary alkaloids are extracted from aqueous solutions with trichlorethylene rather than with hydrocarbons such as benzene. The latter emulsifies more readily and separates less easily after mixing with the aqueous liquid in batch-wise extractions of the mix-and-settle type. This may be due, in part at least, to the greater difference in density between trichlorethylene and water than between benzene and water. The main advantage of trichlorethylene, however, is that it extracts much less impurities than does benzene. The recovered secondary alkaloids are therefore of much higher purity and in certain cases the solvent can be used again many times without being redistilled.
The extraction apparatus consists of a stainless steel closed cylindrical extractor with an efficient high-speed agitator, conical bottom with outlet and reflux condenser. The alkaline mother-liquors are stirred for several minutes with one-fourth their volume of trichlorethylene, and then allowed to separate. The solvent is drawn off by the bottom outlet and the operation is repeated twice with other portions of fresh solvent. The remaining aqueous liquor does not contain any more secondary alkaloids, but still retains a small amount of morphine. It is therefore brought back to the storage tank, where the morphine-containing mother-liquors are conditioned for the extraction column. The combined solvent extracts, now containing codeine and thebaine, are transferred back into the extractor and thoroughly mixed with sulfuric acid of 10%. The quantity of acid must be such that the aqueous phase remains acid to congo red even after thorough contact with the solvent. The aqueous solution almost immediately becomes a soft crystalline mass of codeine sulfate and separates easily on the surface. It can be skimmed off by hand with a stainless steel sieve or strainer. After crystallization in a cooling chamber over night, it is filtered off on a Buchner filter and washed with 70% alcohol. It forms a yellowish, almost white, crystalline mass, which is purified by recrystallization from water. The mother-liquor from the codeine contains the thebaine together with a small amount of codeine. The former is precipitated by the addition of ammonia and separates easily as a resinous dark-brown mass. This raw thebaine is purified by crystallization from water as the bitartrate as described later. Its mother-liquor is recycled into another batch of codeine-containing liquors. The trichlorethylene from which the alkaloids have been extracted, can be used again for another extraction of mother-liquors. In the course of a longer series of such cycles some accumulation of impurities can take place and may necessitate an occasional redistillation.
The trichlorethylene extract from the separation of the raw total alkaloids is evaporated almost to dryness. Some water is then added and distillation continued until all the solvent is driven off. The residue is a dark viscous mass containing the alkaloids, which is drawn off while still hot before it solidifies by cooling off. It is digested by heating with twice its weight of water, to which tartaric acid is added, so that after thorough stirring a solution slightly acid to methyl red (pH 5.5 to 6.0) is obtained. After being allowed to settle for a short time, the clear solution is decanted and the remaining resins are digested once more with a smaller amount of water and then discarded. The combined tartaric solutions are brought to 80°C and an amount of tartaric acid equal to that used for the digestion is added, whereupon thebaine bitartrate begins to crystallize rather quickly. In order to prevent the formation of hard lumps the mass has to be stirred until it is cooled off. It is filtered the following day, either on a Buchner filter or by centrifuging, and washed with cold water. This raw thebaine bitartrate is an almost white, sometimes slightly violet crystal mass. It is dissolved in 3 to 4 times its weight of hot water by addition of sodium hydroxide until the resulting solution is only slightly acid to methyl red. This solution is decolourized with charcoal and bisulfite; then tartaric acid is added to produce another crystallization of bitartrate which will be pure enough to be converted into pure thebaine base.
The mother-liquors of the raw thebaine bitartrate are made alkaline with ammonia in order to precipitate papaverine together with a small amount of thebaine. The resinous precipitate is digested as described for thebaine, using oxalic acid instead of tartaric acid. An excess of oxalic acid causes papaverine bioxalate to crystallize. After filtration, the almost white product is recrystallized several times from ten times its weight of boiling water to free it from cryptopine. The ammoniacal mother-liquor from the raw papaverine base is extracted with thrichlorethylene as described before to recover the codeine. The combined mother-liquors from the papaverine bioxalate recrystallization are precipitated with ammonia, and the precipitate, which contains papaverine with some thebaine, is added to the mixture of raw secondary alkaloids before they are digested with tartaric acid.
As morphine and some of its derivatives are highly toxic and addict-forming drugs, many attempts have been made to replace them by less harmful products. The first partially successful achievement along these lines was the synthesis of Dolantine by I. G. Farben in Germany in 1939 (8), (9). Although this drug, now better known internationally as pethidine, is today generally accepted together with some other products of minor importance as a substitute for morphine in certain instances, it is far from being the right answer to the morphine problem, since it is also apt to induce addiction. Furthermore, even a perfect substitute for morphine cannot be the solution, as only about 5% of the manufactured morphine is medically used as such, the remaining 95% being converted into derivatives such as codeine and ethylmorphine. Efforts have been made for some years to synthesize substitutes for codeine, and some of these preparations are already on the market; the consumption of codeine, however, does not appear to have been affected by them.
The following figures are all taken from official publications of the United Nations Division of Narcotic Drugs (20). The total legal world production of morphine reached the record figure of 84 tons in 1954, whereas the pre-war record was the 1929 production of 55 tons. Codeine consumption, which in 1935 amounted to 18 tons, increased to a record of 68 tons by 1954. The consumption of pethidine, on the other hand, seems in 1954 to have reached a fairly stabilized level of about 11 to 12 tons yearly, whereas all the other substitutes together did not at that time exceed a yearly consumption of one ton.
New interesting developments seem to be foreshadowed by the complete synthesis of morphine by Gates and Tschudi in 1952 (11), followed by a different synthesis of the same product by Ginsburg and Elad in 1954 (12). Although far from being commercially practicable, because of their expensive multi-step complexity, these syntheses are certainly more than a mere laboratory curiosity and in a more distant future may possibly become an important concern to manufacturers of natural morphine.
Another development of interest is the extraction of morphine from poppy capsules. Although considered with much scepticism by most manufacturers, this process has become technically and economically feasible since it was initiated by Kabay in Hungary in 1931 (15). The raw material, dry, ripe poppy capsules, is very bulky and its alkaloid content is very low. Therefore its transportation over long distances is not economical and its utilization is economical only when combined with the commercial use of the poppy seeds. Furthermore, successful cultivation of the poppy is possible only under special climatic conditions and in regions where cheap labour is available. That is why the manufacture of morphine from poppy capsules will most probably always remain restricted to certain countries and will hardly be able to supersede the manufacture of morphine from opium as a whole; in spite of that, this process supplies at present about one-fourth of the world's production.
Reviewing all these facts, it appears fairly certain that the manufacture of alkaloids from opium will not decline in importance in the near future, and therefore it may be worth while to discuss here some possibilities of improving today's manufacturing methods.
The search for a better solvent for morphine may still be an interesting task despite the numerous data accumulated on this subject in the past. The number of versatile commercially available solvents has been increasing steadily for some years, in particular of oxygen-containing solvents such as esters, ethers and glycols, which should be promising solvents for morphine.
The surface active agents, which were developed into an impressive technical field during the last two decades, have not aroused too much interest in the alkaloid industry, except for a few minor investigations. There seem, however, to be several possibilities of using surfactants in order to improve processes of manufacturing opium alkaloids. The addition of surfactants to the extraction of opium with water would probably not only help the maceration of the marc, but also speed up the filtration of the extract. The solvent extraction of aqueous alkaloid solutions, an over-all important unit process in the field, is often hampered by the formation of emulsions, which can sometimes be avoided by the addition of small amounts of certain surfactants. The author applied this principle successfully in several instances among others to the solvent extraction of aqueous poppy capsule extracts, which behave quite similarly to opium extracts. Finally, precipitation of alkaloids from solutions containing large amounts of high-molecular impurities may be improved to give higher yields of alkaloids when surfactants are added. Systematic investigations of these possibilities with the different types of commercial surface active agents therefore seem worth consideration.
The use of inorganic coagulants for high-molecular organic impurities of opium extracts has been studied to some extent. Busse and Busse (4) proposed, as early as 1933, the precipitation of resinous impurities by repeated addition of sodium chloride to aqueous opium extracts. Their method, however, appeared to be cumbersome and of limited effectiveness. More interesting are the attempts that have been made to prevent these impurities from getting into the aqueous extract at all. Eder & Waeckerlin (6) found in 1940 that when adding manganese or ferrous salts to the opium extraction, the alkaline extracts were much less coloured than usual, and did not emulsify as easily as ordinary extracts. Feinstein & Hannan (10) suggested in 1950 the addition of aluminium sulfate to the aqueous extraction of anabasine or nicotine from the corresponding plants. Decker (5) received in 1951 a patent which describes the extraction of opium alkaloids from poppy straw by a 1% aqueous solution of copper sulfate, which is claimed to give much purer extracts than water. Considering the numerous troubles which impurities are able to cause in the course of the involved manufacturing process, the possibility of getting rid of them right at the start may be found rather appealing.
Apart from the high-molecular impurities, one of the main technological difficulties in alkaloid manufacture is the liquid-liquid extraction. Emulsions may form with ease, particularly in batch-wise extractions of the mix-and-settle type. They can sometimes seriously hamper the extraction or even make it impracticable. Continuous counter-current extraction in gravitational packed columns is a more satisfactory solution but its operation needs much skilful super- vision and the through-put capacity is limited. More recently continuous centrifugal liquid-liquid extractors have appeared on the market, and they probably offer the best answer to the problem so far. The Podbielniak extractor (3), (17) has been an established process unit in the chemical and related industries for a number of years. Its European counterpart, the German Luwesta-extractor (7), which is based on a principle somewhat similar to that of the known milk separators, appeared on the European market in 1950. Both these extractors offer several interesting advantages as compared with other extracting equipment. They occupy very little space, have an unusually high through-put capacity, need only a few minutes to reach continuous operation conditions and are easy to operate and maintain. Owing to the use of centrifugal forces several thousand times the force of gravitation, they are able to mix and separate efficiently without formation of emulsions. These extractors have already proved their usefulness in the alkaloid industry and according to the manufacturer of the Luwesta-extractor (21), the latter has been used successfully in the solvent extraction of morphine from aqueous poppy capsule extracts, which are quite similar in nature to opium extracts. If in some way or other the solvent extraction could be improved, it could occupy a still larger field in the opium alkaloid manufacture, replacing less efficient operations such as evaporation, precipitation or crystallization.
The ion-exchangers, another new tool for chemical processing, have not yet been investigated thoroughly, as far as the alkaloid industry is concerned. Their application to the analysis of alkaloids was reviewed recently by Jindra (14). To the author's knowledge their use for purification of opium extracts for manufacturing purposes has not been investigated so far. The fact that such extracts contain beside the alkaloids rather large amounts of both ionic and non-ionic impurities, some of them in colloidal suspension rather than in solution, does not allow us to foresee whether this unit process could be economically applied to the separation of the alkaloids from the impurities. On the other hand ion-exchange has already found many successful uses in industrial large scale purification procedures, so that a systematic experimental study within the opium alkaloidal manufacture is certainly indicated.
BARBIER, A., Annales Pharm. Franc ., 5, 121 - 140 (1947).002
BARBIER, A., Bull. Narcotics , 2, No. 3, 22 - 29 (1950).003
BARSON, N. and BEYER, G.H., Chem. Engin. Progress , 49, 243 - 252 (1953).004
BUSSE, S. and BUSSE, V., Khim. Farm. Prom ., 1933, 127 - 129.005
DECKER, W., USP 2,565,067 (Aug. 21, 1951).006
EDER, R. and WAECKERLIN, E., Pharm. Acta Helv ., 15, 227 - 256 (1940).007
EISENLOHR, H., Chemie-Ingenieur-Technik, 23, 12 - 14 (1951). EISENLOHR, H ., Industrial Chemist and Manufacturer , 27, 271 - 73 (1951).008
EISLEB, O., German Patent 679,281 and US Patent 2,167,351 (1939).009
EISLEB, O. and SCHAUMANN, O ., Deutsche Med. Wochenschr ., 65, 967 - 68 (1939).010
FEINSTEIN, L. and HANNAN, P. J., US Patent 2,525,784 (1950).011
GATES, H. and TSCHUDI, G., J. Am. Chem. Soc ., 74, 1109 - 10 (1952).012
GINSBURG, D. and ELAD, D., J. Chem. Soc . (London) 1954, 3052 - 56.013
HOFFMANN-LAROCHE and Co., Swiss Patent 186,666 (1936).014
JINDRA, A., Bull. Narcotics , 7, No. 2, 20 - 27 (1955).015
KABAY, J., Hungarian Pat. 109,788 (1931).016
KANEVSKAJA, S. I., Journal f. prakt. Chemie , 108, 247 - 56 (1924).017
PODBIELNIAK, N. J., US Patent 2,044,996, 2,093,646, 2,153,640, 2,209,577.018
SCHWYZER, J., Die Fabrikation der Alkaloide , Julius Springer, Berlin, 1927.019
SCHWYZER, J., Die Fabrikation pharmazeutischer und chemischtechnischer Produkte , Julius Springer, Berlin, 1931.020
UNITED NATIONS, Report of the Permanent Central Opium Board for 1954 (Document E/OB/11, November 1955). See also Bull. Narcotics , 8, No. 1, 33 - 35 (1956).021
WESTFALIA SEPARATOR A.G., Oelde, Germany; private communication.