THE CHEMICAL NEWS.

VOL. CXXI., No. 3159.

A MODIFICATION OF SKRAUP'S QUINOLINE SYNTHESIS By EDWARD DE BARRY BARNETT, B.Sc., F.I.C.

THE yield of quinoline obtained when aniline is heated with nitrobenzene, glycerine, and concentrated sulphuric acid (Skraup, Monatshefte, i., 317, ii., 141; Walter, J. pr. Chem. [2] xlix., 549) is about 50 per cent. of that theoretically possible when calculated on the aniline used, but is only about 27 per cent of that theoretically possible with regard to the glycerine. The use of arsenic acid as an oxidising agent in place of nitrobenzene (Kneuppel, B., xxix., 709; D.R.P., lxxxvii., 334) although it leads to somewhat increased yields with reference to the aniline is even more extavagant in glycerine, as only about one-fifth of that used enters into the reaction. More recently, Druce (CHEMICAL NEWS, 1917, cxvii., 346) has described the preparation of quinoline' from aniline stannichloride, glycerine, and sulphuric acid, and has stated that 50 grms. of aniline stannichloride yield 20 grms, of quinoline. These figures correspond to a yield of 80.5 per cent of that theoretically possible with reference to the aniline present in the stannichloride. The formation of a molecule of quinoline, however, requires the oxidising powers of a whole atom of tetravalent tin:C,H,NH,+Sn(SO1)2+C ̧H ̧O,=

C,H,N+3H,O+SnSO,+H2SO, and consequently Druce's result is difficult to understand, as calculated on the tetravalent tin present in the stannichloride his yield is 161 per cent of that theoretically possible. It seemed superfluous to prepare the stannichloride and then to treat it with sulphuric acid, but a few experiments were made in which aniline was heated with glycerine, sulphuric acid, and stannic sulphate. These resulted in yields of quinoline corresponding to about 30 per cent of the theoretical provided only about the calculated amount of glycerine was used. If, on the other hand, a large excess of glycerine was employed, as in the experiments described by Druce, the yields of quinoline were very poor. The experiments with stannic sulphate were abandoned, however, as it was found that much better results were obtained when the cheaper ferric sulphate was employed. Even with this oxidising agent, however, yields exceeding 60 per cent of theory could not be obtained, and it seemed probable that the relatively poor yields were due to the destruction of the glycerine before quinoline formation had had time to take place. Experiments were therefore carried out in which (a) a great excess of glycerine was employed; (b) the glycerine was slowly dropped into a heated mixture of aniline, ferric sulphate, and sulphuric acid; (c) the reaction was carried out at 130°; (d) the sulphuric acid was diluted to 80 per cent strength. None of these modifications, however, resulted in any definite improvement in the yield. Experiments were also made in which an excess of boric acid was added in the hope that a boric

ester would be formed which would not be so readily destroyed by the sulphuric acid. This method has been very successfully used for the protection of hydroxyl groups during the preparation of hydroxyanthraquinones by direct oxidation (Bayer & Co., D.R.P., lxxix., 768; lxxxi., 244; lxxxi., 481; lxxxi., 956; lxxxi., 960-1-2, etc.), but in the case of quinoline formation the boric acid seemed to have no effect at all. It is probable therefore, that the relatively poor yields obtained are due not so much to the destruction of the glycerin as to the destruction of the aniline, and, in fact, it was found that aniline is fairly easily destroyed when heated with ferric sulphate and sulphuric acid at 140°.

The

When stannic sulphate was used as an oxidising agent the best results were obtained when aniline (50 grms.) was added to concentrated sulphuric acid (75 cmm.), the solution cooled and then the glycerine (50 grms.) and stannic sulphate (190 After heating on an oil-bath at grms.) added 180-190° for six hours, the melt was poured into water, unchanged aniline destroyed by the addition of sodium nitrite, the solution made alkaline, and the quinoline distilled off with steam. yields obtained were about 30 per cent of the theoretically possible. In one or two experiments in which a large excess of glycerine (150 grms.) was used, hardly any quinoline was obtained. Much more satisfactory results were obtained when ferric sulphate was used as an oxidising agent, and in fact, it was found that the use of this oxidising agent provides by far the most convenient method of preparing quinoline in the laboratory, as the very violent reaction which often takes place when nitrobenzene or arsenic acid is used is completely absent, and no tar is formed. It is unnecessary to employ dry ferric sulphate, as ordinary calcined ferric oxide, which has been previously treated with sufficient concentrated sulphuric acid to convert it into the sulphate, gives equally good results. The following procedure was found to give yields of quinoline of 45 to 50 per cent of the theoretical.

One hundred grms. of aniline was dissolved in 150 cmm. of concentrated sulphuric acid, and the solution thus obtained added to 450 grms, of dry ferric sulphate, or to 180 grms, of calcined ferric oxide, to which 370 cmm. of concentrated sulphuric acid had been added an hour or two previously. One hundred grms, of glycerin were then added, and the whole heated for six hours on an oil-bath at 180-190°. Without cooling, the melt was then poured into water, made alkaline with caustic soda, and the quinoline and unchanged aniline distilled over with steam. The distillate was then made acid with hydrochloric acid, and aniline destroyed by adding sodium nitrite. After boiling for a few minutes, so as to destroy the benzene diazonium chloride, the solution was again made alkaline and distilled with steam. The quinoline was extracted from the distillate by shaking with benzene, the extract dried over solid, finely powdered caustic soda, and the benzene then removed by distillation.

A very rapid modification of the above process which gives somewhat better yields, about 60 per cent of the theoretically possible, consists in mixing 50 grms. of aniline with 65 grms, of glycerine, and 100 grms. of calcined ferric oxide, and then adding this mixture as quickly as possible to 150

cmm. of concentrated sulphuric acid in an evaporating basin. A brisk reaction sets in almost at once, and while it lasts, the mixture is kept well stirred. When the reaction subsides, the whole is allowed to stand for half-an-hour without heating, and is then poured into water and worked up as before. Several hundred grms. of quinoline have been prepared by this method which has been found to give very satisfactory resuits.

a

Both ferrous sulphate and stannous sulphate are very readily oxidised by atmospheric oxygen, so that it seemed probable that devised catalytic process might be for the preparation of quinoline. Experiments have shown that this assumption was correct, although the reaction is somewhat slow, probably owing to the insolubility of the catalyst. In carrying out the catalytic process, it is necessary to select conditions such that the glycerine is not rapidly destroyed by the sulphuric acid, and it was found that this could be done by diluting the acid to 80 per cent strength, and then working at a temperature of 130-140°. A trace of either ferric sulphate or stannic sulphate can be used as a contact substance, but the latter is the more satisfactory. Owing to the slowness of the reaction, the process is not convenient for laboratory use, but indications were found which pointed to satisfactory results being obtained if the reaction was carried out in a closed vessel, using oxygen under pressure. The process would probably be a convenient one to use on the large scale should any demand for synthetic quinoline derivatives arise.

RUBBER SEED OIL.

FROM time to time in recent years references have been made made to the possibility of extracting the oil from the seeds of the Para rubber tree. Experiments have been made, both in this country and in the Federated Malay States, and the results were such as to encourage those who claimed that the seeds had a commercial value. At first it was believed that rubber seed oil would prove a substitute for linseed oil; but as yet this belief has not been substantiated, though doubtless further experiments by the chemists may bring about developments which may make it worth while for the planting companies to collect the seeds for the sake of the oil they contain.

In his report on the Agricultural Department of the Federated Malay States in 1919, just to hand, the Director of Agriculture (Mr. L. Lewton-Brain) states that the experimental hydraulic oilexpression plant of the Department has been lent to the Malayan Oil Mills, Ltd., a local company formed primarily to manufacture rubber seed oil. It has been ascertained that seed on storage deteriorates and produces an oil containing up to about 25 per cent free fatty acids, and that such oil is not generally suitable as a substitute for linseed oil. Further, this oil is not suitable for many purposes owing to its slower drying power, compared with linseed oil, and it is necessary to prepare a "boiled" oil for commercial purposes.

Considerable progress has been made in the work of refining the oil, but it is probable that the processes required can only be carried out in a factory under the supervision of a trained chemist.

October 29, 1920

THIS method of solution by digesting with aqua regia, adding sulphuric acid and evaporating to fumes, and then boiling with ammonia, brings all the molybdenum into solution. The residue rarely, if ever, contains molybdenum. Nitric acid used in place of aqua regia in combination with sulphuric acid taken to fumes will also give excellent results. Either of these two methods will completely dissolve the molybdenum in molybdenite and wulfenite ores, and also in ferro-molybdenum if ground sufficiently fine. The silica, iron, alumina, lead, and other sulphates and insoluble matter will be found in the precipitate.

Copper and arsenic, if present, can be removed by adding several grms. of zinc to the acid solution before passing it through the reductor, where some of the arsenic is evolved as arsine. The copper when precipitated by the zinc will also precipitate the remaining arsenic. If much arsenic is present, the addition of a few cc. of a 10 per cent solution of copper sulphate may be necessary. Antimony, if present, acts in a manner similar to arsenic.

Procedures for the separation of molybdenum from tungsten and vanadium are given in standard works listed in the bibliography at the end of this paper. When vanadium is present in the orefor example, wulfenite occurring with vanadinite, as it often does-the vanadium can be determined volumetrically by titration with the permanganate solution, as described by Scott (W. W. Scott, "Standard Methods of Chemical Analysis," 2nd ed. rev., 1917, p. 282). Vanadium can also be determined by the rapid method of reduction with HCI, in which arsenic or molybdenum do not interfere. This method is described by Moore and Kithil (R. B. Moore, and K. L. Kithil, “A Preliminary Report or Uranium, Radium, and Vanadium": Bull. 70, Bureau of Mines, 1914, p. 90).

Molybdenum may be precipitated as the sulphide from slightly acid solutions by passing H1S under pressure into the solution, as described fully in standard text-books and in the references given in this paper. The addition of 2 grms. of tartaric acid will prevent precipitation of vanadium and tungsten along with the molybdenum. The same procedure may be applied to ammoniacal solutions, but the solution must be made slightly acid before the molybdenum will precipitate as the sulphide. The molybdenum sulphide may then be redissolved in strong nitric acid, fuming with sulphuric acid as previously described. Any copper sulphide liable to precipitate with the molybdenum sulphide does not interfere with the volumetric method, but does with the gravimetric sulphide method of determination.

29,

Decomposition by Fusion with Sodium Peroxide. Decomposition by fusion with sodium peroxide is practically the same procedure for preparing the solution of molybdenum that is used by one of the largest producers and refiners of molybdenum ores and products in this country. This method has proved fairly accurate for determining molybdenum in low-grade materials, and a complete determination can be made in 30 minutes.

Procedure-Weigh out a o'5 to 5 grm. cample, according to the grade of the ore; the content of Mo should not exceed o15 grm., as explained under "Size of Sample" (prox.). Place the weighed sample in a 60-cc. iron-spun crucible and thoroughly mix with six times as much sodium peroxide. Cover the crucible with an iron lid, and heat in an electric muffle maintained at a temperature of about 600° C., or in the flame of a Meeker burner, until perfect fusion is obtained. Fusion takes place in about ten minutes in the muffle, and in about five minutes over the burner. Allow the crucible to cool for a few minutes; then while still warm place in an 800-cc. beaker containing 400 cc. of water. After solution, which takes place quickly, remove the crucible and the cover with iron tongs, and wash them with water from a wash bottle. If excess silica is present, boil before making up to volume. Make up the volume in a graduated flask to 500 cc. and filter 'the solution through four folded sheets of 8-inch filter paper placed in a 5-inch funnel.

A good grade of bleached paper should be used, because with unbleached paper the organic matter dissolved by the hot caustic solution will require several cc. of permanganate solution for the blank titration. One grade of paper tested was so poor that as much as 10 cc, of one-twentieth normal permanganate was required to colour the filtrate permanaently after it had been made acid. The quantity required for blanks in the peroxide fusion method should never exceed o8 cc.

Filtration takes place quickly. When 250 cc. have filtered through, stop filtering, add enough dilute (11) sulphuric acid to the filtrate so that after the caustic has been neutralised the solution will contain approximately 125 cc. of free sulphuric acid. Heat the 250 cc. of acidified solution on the hot plate for several minutes to drive off any hydrogen peroxide that may have formed. The acid solution is then ready to be passed through a Jones reductor.

Interfering Elements.-Fusion with sodium peroxide will completely and readily convert all the molybdenum into soluble sodium molybdate. When the fusion is dissolved in water the iron will be completely precipitated with other insoluble hydroxides, including copper. With low-grade material the amount of molybdenum retained by the bulky precipitate is of no consequence in the calculations. However, with high-grade material, the precipitate should be redissolved and again precipitated, as only an aliquot part of the filtrate is taken for analysis and the amount of molybdenum retained by the precipitate may materially affect the results. Other interfering elements that may occur in the leached molybdenum solution may be removed or determined as in acid decomposition (ante).

If the alkaline leached solution obtained by fusing an ore rich in silica is neutralised with dilute acid and an insufficient excess, only a few cc., of acid is present, colloidal silicic acid will be precipitated when the solution is heated. On titrating the solution after it has been passed through the reductor, low results will be obtained, and the end point will fade quickly and be urcertain (see para. "Precaution against Fading End Point,' later). The concentration of acid used, 125 cc. in 250 cc, volume, prevents this precipitation if the solution is not boiled too long. If too much silicic acid separates, the better plan is to take a new sample rather than attempt to run the gelatinised solution through the reductor

Jones Reductor. The Jones reductor is illustrated and described by Lord and Demorest (N. W. Lord, and D. J. Demorest, "Metallurgical Analysis," 1916, pp. 29-30), also by Scott (W. W. Scott, "Standard Methods of Chemical Analysis," 2nd ed., 1917, p 281). It consists of a glass tube which is contracted at the bottom, with a 3-inch funnel inserted at the top. The internal bore and length of tube are important considerations. The stem of the funnel enters a small one-hole rubber stopper, which fits tightly into the top of a glass tube. The stem, or contracted end of the tube enters the rubber stopper of a 2-litre side-neck flask that is connected by pressure tubing to a suction pump.

The zinc for the reductor should be of 20-mesh size and contain less than 0.01 per cent iron. It is best amalgamated as follows: With very dilute H2SO, about 3 cc. to 100 cc. of water-moisten in a beaker a sufficient quantity of zinc to fill the reductor, add a small drop of mercury, and stir until uniformly white. Wash the zinc frce from acid and put it into the tube One-half grm. of mercury is sufficient for 150 grms, of zinc. Avoid using more than just enough.

The reductor is filled by placing some glass beads at the neck; on the beads is placed some glass wool mixed with bits of small broken glass rods to prevent the wool from forming a compact plug that might hinder the passage of the solution. Above the wool may be placed a perforated platinum disk or gauze, to assist the clear passage of the solution and to prevent clogging. The tube is then filled to within a few inches of the top with the amalgamated zinc.

The internal bore of the reductor should not be greater than five-eigths of an inch. The reductor should be 20 to 30 inches long. Shorter reductors ranging down to 10 inches, and larger bores than five-eighths of an inch were tried, but the results were unsatisfactory, owing to the following

reasons:

(a) The molybdenum solution was not completely reduced by passing the liquid through once, which necessitated passing the solution through a second time, and required a larger blank correction. This factor must be carefully controlled, as shown later.

(b) The quantity of solution or wash water necessary to wash a reductor completely when using one of larger bore than five-eighths of an inch makes the solution too bulky to titrate In order to obtain a sharp end point the solution must not be too dilute. The titrations obtained on blank runs, with the same quantity of acid and