The Extraction of Opium Alkaloids


I. Classical Processes for Opium Extraction
II. Study of a New Process
III. Industrial Application of the Process


Pages: 22 to 29
Creation Date: 1950/01/01


The Extraction of Opium Alkaloids

Chemical Engineer

I. Classical Processes for Opium Extraction

There are three Classical processes for the extraction of morphine. They are all very old: the Merck process dates from 1830; in 1835, Pelletier described a lime treatment of 5 kilograms of opium by Thiboumery, the manager of his factory. As regards the Robertson-Gregory process, the exact date of its discovery is unknown, but it was described by Wurtz in his Dictionnaire de chimie pure et appliquée published in 1868. Since that time nothing has been published regarding the extraction of morphine. In his book, Die Fabrikation der Alkaloide, published in 1927, Schwyzer merely describes the Robertson-Gregory-process-with treatment of the morphine of the mother-liquors by acetylation to diacetylmorphine and saponification of the latter after purification. In their Traité de pharmacie chimique (vol. II, section II, 2nd edition, published in 1938), Lebeau and Courtois merely describe the Robertson-Gregory and Thiboumery processes.

The following is a description of the three old processes, followed, in each case, by a statement of their drawbacks.

Merck process. The opium is exhausted with cold water and the resultant liquor concentrated to a syrupy consistency. It is then precipitated hot with powdered sodium carbonate and heated as long as ammonia is given off; the solution should remain alkaline to phenolphthalein. After standing for twenty-four hours, the resultant precipitate is filtered and washed with cold water; when the wash water is colourless, the precipitate is dissolved in alcohol at 85°. The alcoholic solution is evaporated, to dryness and the residue is exhausted by means of dilute acetic acid, which is added in small quantities as neutralization proceeds. The acetic acid solution is treated with decolorizing charcoal, then filtered and precipitated with ammonia, care being taken to avoid an excess. The precipitate is filtered, washed, and purified by crystallization from alcohol; concentration of the alcoholic mother-liquor yields a further quantity of morphine.

I have not tried this process, which is not mentioned by any modern author. The method of purifying the morphine does not appear to be a good one, and the considerable quantities of alcohol required, owing to the slight solubility of morphine in this solvent, made it impossible to consider it for industrial use.

Thiboumery and Mohr process. The opium is cut into thin slices and treated with three times its weight of hot water until a homogeneous paste is obtained. The liquid is filtered off, the residue pressed and again treated with three times its weight of water. The solutions obtained are evaporated to half their volume and poured into boiling milk of lime: one part of lime in ten parts of water should be used for four parts of opium. The precipitate is filtered off and re-treated with three parts of water to one part of opium; it is then filtered off again. The lime solutions are united and concentrated to a quantity twice the weight of the opium used. The solution is filtered, heated to boiling, and the morphine is precipitated by the addition of ammonium chloride. After cooling, it is filtered: the precipitate is washed, then purified by solution in hydrochloric acid and crystallization of the morphine hydrochloride.

This process is an attractive one: there are no technical difficulties and the morphine is well separated from the secondary alkaloids: the morphine solutions are relatively clean. I made certain changes in this process; for instance, instead of precipitating the lime solution with ammonium chloride, I acidified it slightly with hydrochloric acid and salted out the morphine hydrochloride with common salt. In this way I obtained almost all the morphine from the lime solution in the form of very pure hydrochloride. The morphine remaining in the salted-out liquid was precipitated and returned for purification.

I found, however, that this process had one important drawback: the yield was definitely bad. I subsequently discovered that I had not been the only one to encounter this difficulty. It is possible that the bad yield is due to an oxidation of the morphine in alkaline solution, but the fact that the lime always retains morphine was also a contributory factor.

Robertson-Gregory process. The opium is completely exhausted by five to ten times its weight of cold water; the solution obtained is evaporated to the consistency of a soft extract, and then the process is repeated with cold distilled water. This aqueous re-extraction causes impurities to precipitate, they are filtered off and the solution obtained is evaporated until its density is 10° Baumé. One hundred and twenty grammes of calcium chloride for each kilogram of opium are added to the boiling liquor which is then diluted with a quantity of cold water equal to its own volume. A precipitate of meconate and sulphate of calcium is thus formed and is filtered off. After filtering, the liquid is again concentrated, and this produces a new deposit which consists almost exclusively of calcium meconate. This deposit is filtered off and the solution is left standing. After a few days it becomes a crystalline mass composed of morphine hydrochloride and codeine hydrochloride: this is known as "Gregory's salt". The crystals obtained are drained and then placed in a cloth and squeezed out in the press. They are purified by successive crystallizations, the solutions being: decolorized each time with animal charcoal. When the crystals are sufficiently pure, they are dissolved in water and the morphine is precipitated with ammonia; the codeine remains in solution.

The first drawback of this process is that 20 to 25 per cent of the morphine is left with the secondary alkaloids in the brown and viscous mother-liquids after filtration of "Gregory's salt". Schwyzer separates and purifies this morphine by precipitating all the alkaloids, by dissolving the secondary alkaloids in acetone, by acetylating the impure morphine in order to purify it in the form of diacetylmorphine, and then by saponifying the latter to re-convert it to morphine. This process is complicated and cannot be as good as the author claims, owing to the solubility of morphine in acetone which, although slight, is not negligible:

The second drawback is that the hydrochloride of morphine and codeine crystallize in furry needles which retain the mother-liquids in which the crystallization occurred. Several successive crystallizations and subsequent recoveries are required for purification, and this is a time-consuming process.

Schwyzer purifies the mixture of hydrochlorides by malaxation, with a small quantity of ice-water and subsequent pressing; he repeats the process several times. This no doubt enables him to leave less morphine in the mother-liquids, but that fact does not make the process much more attractive.

II. Study of a New Process

Having observed the drawbacks of the various known processes, I thought that it should be possible to evolve a new process, more suitable for industrial use, on the scale which I considered that the first phase of this process was the extraction of opium with water, the concentration of the aqueous extract and the precipitation of all the alkaloids, which is the first part of the Merck process. It seemed to me that it should then be possible to separate the morphine, which is practically insoluble in certain organic solvents, from the secondary alkaloids which are soluble. In this way I hoped to obtain all of the morphine, obviously in a very impure form, but nevertheless completely separated from the other alkaloids. The tests which I made enabled me to arrive at the anticipated result by using benzene. I considered that I had already gained a definite advantage over the Gregory process, but one problem remained: the purification of the morphine obtained. There is no known solvent which allows of the easy crystallization of this base, and the only solution was to use a salt as intermediary and purify the latter. When using hydrochloric acid I encountered the same difficulties in purification as those I have already mentioned; I had to find a morphine salt of small solubility in water and of crystalline form such that the mother-liquor could be easily eliminated. As no salt with these properties had been described, I tried preparing organic salts of morphine, and found that morphine tartrate was the ideal salt for purifying this alkaloid. The use of this salt successfully completed the first part of my work and enabled me to find an opium treatment fully suitable for industry.

I should mention, however, that during my bibliographical research two years ago, I found that Tanner (C. 1905.I, p. 525) had obtained a precipitate of morphine acid tartrate in preparing the neutral salt. This author indicates a solubility of one part of morphine acid tartrate in 100 parts of water. I found that 100 parts of cold water dissolved 1.5 parts of morphine acid tartrate; this means that this salt is only a third as soluble, as the hydrochloride. It has the advantage of crystallizing from very impure solutions in large crystals, which can be filtered off and washed with the greatest ease. It is possible, by using morphine acid tartrate, to obtain morphine from residues which contain only 4 per cent. The purification of morphine by passage through this salt had never previously been described; but it is simpler and less costly than any of the usual processes.

The yields obtained by my process were 95 per cent of the morphine in the opium used (applying the Harrison method of assay).

The League of Nations established specific roles for the sampling of opium, and the sellers mix it in order to provide lots of which all pieces have the same morphine content. Nothing of that kind existed in 1921: the morphine content varied from one cake to another, and the cakes were of all sizes. Sampling was carried out, at the factory, by a representative of the seller and a representative of the purchaser. It was very difficult to obtain an average sample of the lot, as the seller tended to take copious samples from the cakes the appearance of which revealed a high morphine content, while the purchaser did the opposite. Thus, in one lot having the same morphine content, half of which was purchased by us and the other half by a competitor, the sampling showed 12 per cent of morphine in one case and 13 per cent in the other.

Lastly, the various methods of assay gave different results: the Codex, Harrison, and the Dr. Gilbert methods did not show the same morphine content, and the price of opium assayed by the Harrison method was not the same as the price of the same opium assayed by the Dr. Gilbert method.

The international method of assay now used gives almost the same results as the Harrison method.

It is my experience that when assaying opium for morphine by a lime process, it is essential, first, to carry out the trituration of the opium with water with the greatest care until a very fine and homogeneous paste is obtained. Only then can the lime be added, as otherwise it will produce a precipitate of very hard meconate of lime around the particles of opium not reduced to thin paste. This will prevent the morphine covered over by the meconate of lime from dissolving, and thus it will not be included in the assay.

III. Industrial Application of the Process

On the basis of 300 kilograms of opium per operation

The process consists in:

  1. A. Extraction of the opium;

  2. B. Concentration of the liquids;

  3. C. Precipitation of the total alkaloids;

  4. D. Extraction of the secondary alkaloids;

  5. E. Preparation of morphine acid tartrate;

  6. F. Precipitation of the morphine base;

  7. G. Treatment of the secondary alkaloids;

  8. H. Extraction of the alkaline liquids;

  9. I. Natural codeine.


As the aqueous extract of opium must be concentrated before further treatment, methodical extractions should be carried out so that not too much water will have to be evaporated.

Five parts of water are used to one part of opium: the opium is cut into thin slices and placed in the extractor which contains the water for extraction. In order to avoid blocking, the stirrer is set in motion as soon as any opium is placed in the extractor. The water is heated to 45 degrees for the first extraction in order to facilitate the disintegration of the opium; this heating is unnecessary-when opium marc is being extracted.

The extractions are made as follows:

First extraction: Fresh opium
Extracted with the liquid from previous second extraction
Second extraction: Marc of first extraction
Extracted with the liquid from a third extraction
Third extraction: Marc of second extraction
Extracted with the liquid from a previous fourth extraction
Fourth extraction: Marc of third extraction
Extracted with pure water
Marc of fourth extraction
Thrown away after making certain that it does not contain any morphine

The aqueous extracts are always acid to litmus and the marc of the fourth extraction hardly ever contains any morphine. In 1930, however, we had to treat opium, saved from the fire at Smyrna, which contained 10.5 per cent of morphine. This opium, which had been subjected to a fairly high temperature, had lost part of its volatile acids and the morphine was in a form insoluble in water (probably that of the base), but soluble in lime, and this enabled it to be detected by the assay. This particular opium was treated with disastrous results during a period when I happened to be absent for several months. Those in charge at the factory did not heed the warnings given by the accountant who had noted that the entries for morphine yield were abnormally low: they failed to check the acidity of the extraction liquids or to make the indispensable examinations of the extracted marcs. They burned these marcs in the boiler with the morphine which they still contained. When I returned and took stock I found that 125 kilograms of morphine base had disappeared in this way. The lesson to be drawn from this incident is that, even when manufacture is carried out with perfect regularity, the chemical engineer should never relax his vigilance, and that the results obtained in the analytical laboratory are more important than any hypothesis. There was 10.5 per cent of morphine in the opium treated, and that morphine should have been extracted or else an attempt should have been made to find the missing morphine. The assertion that, despite the assay, the failure to extract further morphine was due to the fact that the opium did not contain any more was a fanciful hypothesis which was not justified by the fact that the process had been strictly followed.

I subsequently added to the last extraction, a quantity of pure concentrated sulphuric acid, in the proportion of one cubic centimetre of acid per kilogram of opium; this addition of acid permits of a more complete extraction of bases-narcotine and papaverine, certain quantities of which always remain in the marc.

The apparatus consists of four 3,000 litre extraction vats with stirrers, the whole being made of lead-coated iron. Each extractor has two decanting pipes with bronze cocks, at levels which enable 750 and 500 litres respectively to be left in the vat. In the lower part there is a bronze drop-valve 100 millimetres in diameter, which allows the thick sludge, consisting of the marc and the water remaining after the several decantations, to be drained off. There is one 2,500 litre cylindrical vacuum receiving tank for each extractor, that is, four receiving tanks in all.

At first I installed vacuum filters with a filtration surface of 5 square metres, but the results were not satisfactory. Owing to the fragments of capsules which it contains, Turkish opium gives extracts which can be filtered, although with difficulty. It is almost impossible to filter opiums from other sources. In any event, the period of filtration varies considerably and does not allow the work to proceed regularly.

After making tests with a decanter with clay filter and suction apparatus, I replaced the filters by centrifugal decanters 1.50 metres in diameter with 8-h.p. motors operating directly on the shaft. Two decanters are required, each of them being placed between the two extractors which it serves.

The liquid is first decanted by means of the upper cock; if it is clear, it is drawn directly into the receiving tank; otherwise the decanter is set revolving and the decanted liquid is transferred to the receiving tank by means of the suction apparatus. The liquid flowing through the second decantation cock is then passed into the centrifugal apparatus, the stirrer is set in motion and the marc is centrifuged three times.

The densities of the solutions are as follows:

From the first extraction, 1.030 to 1.035;

From the second extraction, 1.012 to 1.014;

From the third extraction, 1.006 to 1.007;

From the fourth extraction. 1.002 to 1.003.


The opium liquids resulting from a first extraction, which amount to some 1,600 litres, must be concentrated and reduced to 300 litres. The apparatus used has a capacity of 3,000 litres; it is made of lead-coated iron and a coil, also lead-coated, is used to heat the liquid. A copper still could also be used; when making a laboratory test I also used an aluminum apparatus which served the purpose well.

The concentration is made by means of a vacuum; the condensation of the distilled water is effected by means of a nest of copper tubes. The condensed liquids are collected in two receiving tanks which are used alternately; the distillation is a delicate matter, particularly at the start, as the opium liquids froth a great deal. Before being thrown into the drain, the condensed liquids are examined in order to make certain that they do not contain morphine, which might be the case if the froth was carried over. Generally speaking, only at the start does the distillation present difficulties, thereafter there is no need for constant supervision of the apparatus. This would not be the case if one wished-to use a small apparatus with continuous feed.


When the opium liquids are reduced to one litre per kilogram of opium, their volume is 300 litres. They are placed in a 1,000 litre apparatus made of lead-coated iron and equipped with a lead-coated stirrer and a double jacket of sheet iron which permits of steam-heating and water-cooling, The liquids are heated to 85°-90° and precipitated at that temperature by powdered Solvay-process sodium carbonate. When alkalinity to phenolphthalein is reached, no further Solvay powder is added; ammonia is given off. After heating for one hour, the alkalinity to phenolphthalein is again tested and, if it has disappeared, more Solvay powder is added. When persistent alkalinity is achieved, the mixture is cooled; there is a precipitate of small grains which are filtered off and washed. They are then dried without heat in order to remove as much of the water as possible, as the total alkaloids are not dried in the oven before the treatment is continued.


As we have seen, the secondary alkaloids are separated from the morphine by means of a solvent. I gave the preference to ,benzene, but other hydro-carbons and chlorinated derivatives such as carbon tetrachloride, trichloroethylene, etc., may also be used.

The apparatus consists of a 606 litre extractor without stirrer, which is joined at the top to a condenser and at the bottom is equipped with a drain-cock leading to a 400 litre distilling apparatus. There is a perforated plate covered with filter cloth a few centimetres from the bottom of the extractor. On this cloth is placed a tubular crown of perforated steel having a vertical branch which goes out through a tube in the cover and can be connected to the steam. The total alkaloids are then placed in the extractor, benzene is added to cover them} the lid is put on and the mixture is left to soak. After half an hour the lower cock is opened and the benzene containing the secondary alkaloids is allowed to flow into the distilling apparatus. The extraction is repeated with fresh benzene until a sample of 10 cubic centimetres of benzenic solution leaves, after evaporation, a residue of less than 0.1 gr. The cover is then put on and steam is directed through the perforated crown so as to carry off the benzene retained by the raw morphine remaining on the filter cloth.

The benzenic solution which contains the secondary alkaloids-narcotine, papaverine, thebaine, and distilled after each extraction; this makes it possible to use a smaller still and benzene receiving tank. A thermometer dips into the liquid, and distillation is stopped when the temperature begins to rise. After the benzene has been distilled from the last extraction, the mixture is allowed to cool gradually and 20 litres of water are added. It is then re-heated in order to expel the remaining benzene. The residue, which is in the form of a thick tar, is run off hot into a vat; it will be treated subsequently in order to separate the secondary alkaloids.


At the time of my first tests, I wished to convert raw morphine into acid tartrate in aqueous solution, to purify it with charcoal, and to filter and crystallize it. Two difficulties arose: the solution of acid tartrate crystallized in the pipes during vacuum filtration, despite the fact that the filter and the intermediate receiving tank were heated; and the solution, which was transferred to a vat cooled by a herse and having a stirrer, produced an insulating crust of morphine acid tartrate on the vat lifting and this prevented cooling. I had already met with this difficulty in the crystallization of other products and knew that there was no simple remedy. I therefore decided to make all the treatments on the neutral salt and to make the acid salt by adding tartaric acid to the cooled solution of the neutral salt; this gives a real precipitation of morphine acid tartrate in fairly fine crystals from which the mother-liquor can be easily removed by filtration and washing.

The raw morphine separated from the minor alkaloids is placed, with 600 litres of water, in a 1000 litre enamelled cast-iron vat, with a wooden stirrer and water bath for heating: the mixture is heated to 80°C. and tartaric acid is gradually added until there is slight persistent acidity to methyl red. In order to facilitate the filtration and to decolorize the solution, charcoal from the treatment of the acid tartrate is added to precipitate the base morphine (see F. Precipitation of the base morphine).

The acid solution is vacuum filtered on an enamelled cast-iron filter with a wooden filter plate and a double steam-jacket to keep the solution hot. A vacuum is created in an enamelled cast-iron evaporator with a cover and a double steam-jacket; this evaporator, which is linked up with the filter, receives the solution of morphine neutral tartrate. The tubes are made of lead. After the charcoal and the insoluble matter have been carefully washed with hot water on the filter, the filtered solution is discharged into an enamelled cast-iron vat with a wooden stirrer and double jacket for cooling. When the solution is in the vat, the stirrer is set in motion, and the solution is cooled by passing water into the double jacket. When the solution is cold, the same quantity of tartaric acid as had been used to make the neutral tartrate is added, plus an excess of about 1 kilogram. The solution must be acid to methyl orange; otherwise, a further quantity of tartaric acid must be added.

After being stirred for about another two hours, the mixture is left standing for one night and then the morphine acid tartrate is filtered on an enamelled filter 80 centimetres in diameter with a wooden filter plate. When the liquid is cold, a stone jar (in which a vacuum can be created), with a pipe at the bottom, will be used as a receiver. This jar should be washed periodically, as in the course of time a small quantity of morphine bitartrate is deposited there.

The filtered product is washed on the filter with cold water and dried as far as possible. The mother-liquor is concentrated in enamelled vats with a wooden stirrer and double jacket for heating and cooling. After two further precipitates of morphine bitartrate have been obtained, there remains mother-liquor which, contrary to all expectations, contains a small quantity of codeine. The treatment of this mother-liquor will be described later: I would, however, point out at this stage that, whilst it is stored, this liquid releases a further small quantity of morphine acid tartrate, which should, of course, be put aside for extraction of the morphine.


The moist acid tartrates of the first, second, and third precipitates are combined and placed in an enamelled vat provided with a lower pipe with a built-in ceramic cock and a wooden stirrer. Water four times its weight is added to the acid tartrate and the mixture is heated by passing live steam through a movable lead pipe. When the temperature reaches 85 degrees, the lead pipe is removed and the stirrer set in motion. To ensure smooth working and to avoid crystallization of the acid tartrate during treatment, a return to the neutral tartrate is made by adding ammonia to the solution, which should, however, remain slightly acid to methyl red. Decoloration is effected by using charcoal (5 to 10 per cent of the weight of the moist tartrate). I noted that the decoloration was improved by adding 300 gr. of sodium acetate "snow", 200 gr. of ammonium oxalate, and 100 gr. of sodium bisulfite per operation. The decoloration of the solution is observed by filtering a sample and precipitating the morphine content with ammonia. This should give a white morphine; if the precipitate is not satisfactory, a further quantity of charcoal and a little bisulfite are added; if a solution of ammonium oxalate gives a further precipitate, ammonium oxalate should also be added.

When the solution of morphine tartrate, treated as above, gives a white morphine upon precipitation with ammonia, the morphine is run onto an enamelled cast-26 iron filter with a wooden filter plate and is collected in an enamelled cast-iron intermediate receiving tank in which a vacuum is created. The charcoal is washed on the filter and used again when the raw morphine is dissolved with tartaric acid (see E. Preparation of morphine acid tartrate).

The filtered solution is run into an enamelled cast-iron vat having a wooden stirrer, and the morphine is precipitated hot with ammonia (at approximately 60°). The solution should be barely alkaline to phenolphthalein and an excess of ammonia should be avoided. The solution is left to cool for one night and after the alkalinity to phenolphthalein has again been verified, the base is vacuum-filtered on an enamelled cast-iron filter. It is washed with water, and the mother-liquors are preserved for subsequent extraction with butyl alcohol (see H. Extraction of the alkaline liquors).

The morphine obtained is dried on trays in a drying oven at 80°; it is hydrated and contains approximately 2 per cent of impurities. It can be used for the manufacture of codeine, dionine, or heroin.

It is better to prepare a purer morphine for the manufacture of morphine hydrochloride; for that reason, the solution of morphine tartrate decolorized with charcoal is not precipitated with ammonia after filtration. It is run into an enamelled vat and, after cooling, pure hydrochloric acid is added to restore the acidity to methyl orange. This gives ammonium chloride and a precipitate of morphine acid tartrate. This tartrate is filtered and washed; it is treated in the same way as the tartrates of the first, second, and third precipitations in order to precipitate pure morphine. The acid mother-liquors are precipitated with ammonia in order to extract the morphine remaining in solution in the form of acid tartrate; the morphine recovered in this way is filtered and returned for purification by tartaric acid; the alkaline liquids are extracted with butyl alcohol.


The secondary alkaloids, with the exception of codeine, were formerly of little interest; I extracted them nevertheless but did not trouble about the yield. Subsequently, papaverine as such, and thebaine owing to its derivatives, became very important.

In the process which I have described, the secondary alkaloids pass into the benzene when the total alkaloids are treated with this solvent. There is, however, a certain quantity of codeine in the alkaline precipitation liquids of the total alkaloids, and these liquids also contain a certain quantity of thebaine. Moreover, a certain quantity of codeine does not dissolve in the benzene, although the volume of benzene used should be capable of dissolving several times the weight of the codeine contained in the opium. Hence, some codeine remains with the raw morphine, and this codeine is finally left in the mother-liquors of the morphine tartrate process in the form of acid tartrate.

I have always disregarded the narceine in opium, as it can be produced from the narcotine, large quantities of which are unused.

I separated the secondary alkaloids, by using the difference in basicity between thebaine and codeine, which are strong bases, and narcotine and papaverine, which are weak bases.

The secondary alkaloids extracted from the benzenic solution take the form of a more or less hard tar; this tar is placed in an enamelled vat with a wooden stirrer and a double sheet-metal jacket used for heating and cooling. Three parts of water to one part of secondary alkaloids are placed in the vat and heated to 80°, the mixture being stirred; the product melts and tartaric acid is gradually added until neutrality to methyl red is achieved. The solution: with tartaric acid takes it very long time, as the thick and viscous mass mixes badly with water; during the treatment a certain quantity of water evaporates and is replaced in order to preserve a constant level in the vat. When the solution remains neutral, it is cooled; the narcotine-papaverine mixture solidifies into more or less coarse grains. It is filtered and washed with a little water; the solution containing the codeine and thebaine is concentrated until a sample gives an abundant precipitation of thebaine acid tartrate upon addition of tartaric acid. When this degree of concentration is reached, tartaric acid is added to the solution to obtain acidity to methyl orange, and the solution is cooled. The thebaine acid tartrate is filtered and washed; it is purified by crystallization from water. The thebaine base is obtained from pure tartrate by precipitation with ammonia or a filtered solution of sodium carbonate. The precipitation of a hot solution of thebaine tartrate in a mixture consisting of equal parts of water and ethyl alcohol enables thebaine to be obtained in crystalline form.

The mother-liquors of thebaine tartrate contain codeine; their treatment will be described in the special chapter on codeine.

After filtering off the tartaric solution, the narcotine-papaverine mixture is placed moist into an iron cohobator equipped with a stirrer; three parts of methylethylketone and 0.10 part of decolorizing charcoal are added for each part of the alkaloid mixture and the whole is cohobated and stirred. After dissolving, the mixture is passed, under pressure, through a filter and the filtrate is collected in a cast-iron vat cooled by a herse and equipped with an iron stirrer. The narcotine crystallizes; it is filtered and washed. It is purified by recrystallization in the methylethylketone, in the presence of charcoal.

The papaverine is contained in the mother-liquors of methylethylketone; papaverine acid oxalate is precipitated by adding oxalic acid until acidity to methyl orange is reached. The papaverine acid oxalate is purified by several crystallizations in water; the pure oxalate is redissolved in water and precipitated with ammonia in order to obtain pure base papaverine.


Despite the low solubility of morphine in water, it is essential to extract all the alkaline liquids obtained during manufacture. This is all the more essential as, owing to the impurities which they contain, the residual liquids dissolve a quantity of morphine appreciably in excess of the quantity soluble in pure water. I have extracted 0.75 gr. of morphine per litre from the alkaline precipitation liquids of morphine.

Morphine is not very soluble in most immiscible organic solvents; moreover, the alkaline liquids to be extracted readily give emulsions with the solvents. For my purposes, the most suitable solvents seemed to be butyl alcohol and amyl alcohol. I chose butyl alcohol.

As the quantities of water to be extracted were considerable, I made continuous extraction tests in the laboratory. These gave good results and I applied this method in the workshop.

The alkaline liquids deposit a sludge during storage; this is separated by filtration or by centrifuging, and as it contains morphine (approximately 7 per cent), it is again treated with tartaric acid by the method described above.

The apparatus for extraction consists of two 500 litre receiving tanks with an outlet pipe at the bottom. One is used for the liquids to be extracted, and the other for the butyl alcohol. These receiving tanks are connected with a small constant-level receiving tank which enables a regular outlet speed to be maintained. The hourly quantifies of liquids to be extracted and of butyl alcohol are regulated by means of interchangeable pipes with calibrated orifices. The extraction is carried out in a column 25 centimetres in diameter and 2 metres in height filled with Raschig rings and with a coil for steam-heating in the lower half. The lower part of the column is fixed on a flat cylinder 40 centimetres in diameter and 20 centimetres in height, and the upper part carries a cylinder 40 centimetres in diameter and 40 centimetres in height. The extraction takes place at a temperature of 60 degrees C; the butylic alcohol arrives through a pipe at the bottom of the column, and the mother-liquors through a pipe at the top of the column; the force of gravity causes the liquids to descend whilst the alcohol rises. The fresh butyl alcohol is in contact, at the bottom of the column, with liquids which contain practically no more morphine and, as it rises, meets liquid which are increasingly rich in morphine. The butyl alcohol containing alkaloids flows out through a pipe at the top of the upper cylinder resting on the column; the aqueous liquids flow out through a pipe at the bottom of the cylinder at the base of the column and re-ascend through a vertical tube the outlet of which will be some 25 centimetres lower down than the outlet pipe for the butylic alcohol.

The liquids extracted contain butyl alcohol in solution and they must therefore be heated to 100° in order to distil and recover it. This part of the process might be made continuous, but I carried it on intermittently.

On the other hand, the butyl alcohol containing alkaloids is treated continuously. It comes into a 120 litre distilling apparatus with a powerful coil for steam-heating. The mixture of water and butyl alcohol boils at 96° although the boiling point of butyl alcohol alone is 116°. The butyl alcohol from the extraction contains 10 per cent of water and distilling is therefore carried on continuously and in the presence of water. A decanter separates the water, which returns into the still, from the butyl alcohol which passes back again for extraction. If the boiling temperature rises, a further 10 litres of water should be placed in the distilling apparatus.

Every two or three days the contents should be fully distilled and the apparatus drained. The tar obtained is treated with tartaric acid by the method described for the treatment of raw morphine.

I attribute the good results of continuous extraction by means of the column described above to the fact that the butyl alcohol dissolves 10 per cent of water, and that water also dissolves 10 per cent of the butyl alcohol. This facilitates contact between the solvent and the solution to be extracted. I attempted to make continuous extractions with the same apparatus by using the benzene as a solvent, but without success. I attribute this failure to the insolubility of benzene in water; this decreases the contact between the solvent and the solution to be extracted. Apparatus with powerful stirrers should be used for extractions with solvents which are not miscible with water.


In the process described, codeine is contained:

  1. In the mother-liquors of thebaine tartrate;

  2. In the mother-liquors of morphine acid tartrate produced by the treatment of raw morphine and by the treatment of the tar obtained from extraction with butyl alcohol.

Liquids of different origin should not be mixed but the treatment is always the same. A 400 litre sheet-metal apparatus with a powerful stirrer is filled with the liquids up to the level of the decantation pipe which is placed at the 240 litre level; flakes of caustic soda are added until alkalinity to phenolphthalein is reached, and the extraction is made with benzene. After the mixture has been allowed to stand, the benzenic solution is run through a Florentine decanter into a distilling apparatus. The extractions are continued until a sample of benzene no longer contains codeine. The benzene is distilled after each extraction, but not fully, and when extraction is complete, the benzene is distilled without going over 85°; the remaining mixture is left to cool off slightly, 10 litres of water are added and the mixture is heated to 100° in order to expel all the ben zene. The raw codeine, which is in the form of an oil is run off. It is purified by means of the hydrochloride.

After extraction with benzene, the alkaline liquids are run into a vat and iron-free hydrochloric acid is added until acidity to methyl red is reached. The slightly acid solution is heated and precipitated with powdered Solvay's process sodium carbonate until it is alkaline to phenolphthalein. After cooling, the precipitate which contains a little morphine is filtered off; the liquids are extracted with butyl alcohol. The impure morphine filtered off is treated with tartaric acid by the ordinary method.