SUEOXIDE OF PHOSPRORUS. PART 11. 1235 CXXXL-Non-existence of the so-called Suboxide of Phosphorus. Parat 11. By CHARLES HUTCIIENS BURGESS and DAVID LEONARD CHAPMAN. IN a previous paper (Chapman and Lidbury, Trans., 1899,75,973), it was claimed : firstly, that the grounds adduced by previous investigators for the existence of a definite suboxide of phosphorus having the formula P40 are insufficient ; secondly, that the so-called suboxide of phosphorus, prepared by their methods, has properties identical with those of red phosphorus ; thirdly, that when the substance is carefully purified and analysed, the percentage of phosphorus is invariably higher than that required by the formula. Amongst the several ‘‘ suboxides ” prepared and examined by us was one recently obtained by Michaelis and ,Pitsch (Bey., 1899, 32, 337) by precipitating the red solution which results from the action of aqueous alcoholic potash on white phosphorus, with dilute hydro- chloric acid.The percentage of phosphorus found by us in the sub- stance was 90.5, which is 2 per cent. higher than that required by the formula P40, and in its properties the “suboxide” resembled the other preparations, A complete account of the experiments of Michaelis and Pitsch did not appear until several months after their first publication on the subject (Annalen, 1899, 310, 45). I n this paper? the authors lay claim to having prepared a substance free from hydrogen and other impurities, and containing the theoretical per- centage of phosphorus required by the formula, P40. It is moreover supposed to differ from amorphous phosphorus in its behaviour towards aqueous alcoholic potash.I n the present communication, we shall explain why Michaelis and1236 BURGESS AND CHAPMAN: NON-EXISTENCE OF THE Pitsch’s results cannot, in our opinion, be regarded as proof of their conclusions, and shall describe experiments which confirm the views already expressed by Chapman and Lidbury.” The Composition of the so-called Suboxide. Our analyses show : firstly, that the percentage of phosphorus in the ‘‘ suboxide ” prepared by Michaelis and Pitsch’s method varies from 86 per cent. to 93 per cent.; secondly, that hydrogen is always present in considerable quantity ; thirdly, that other impurities, amongst which is carbon, are also present. The explanation of the striking difference between these results and those of Michaelis and Pitsch is, we believe, to be found in the method of analysis employed by them to estimate the hydrogen.I n the first place, the substance was gently heated before the analysis for hydrogen was made, in order to drive off the water which they and previous observers had found to be present. The anthors evidently regard this moisture, the presence of which was directly proved by heating with copper oxide, as entirely derived from the atmosphere during manipulation, but if this view as to its source is incorrect, then the process of heating can only result in removing before analysis part of the water which it is their object to find, Some idea may be gained as to the rate at which moisture is absorbed from the air from the following experiment.0.5946 gram of the substance was introduced into an open, wide-mouthed tube, and the rate of increase in weight observed. I n 10 minutes it gained 0*0006 gram, in 35 minutes 0.0014 gram, and in 135 minutes 0.0036 gram. Another point in connection with their experiment deserves atten- tion, The hydrogen in the previously warmed suboxide was estimated by heating it with lead oxide in a combustion tube, and driving the resulting water by means of a current of dry air into a, weighed calcium chloride tube. During the heating of the mixture of suboxide and lead oxide, a cloud was formed, which was so difficult to control that it could not be prevented from entering the calcium chloride tube. If the cloud were due to phosphorus pentoxide, and this does not seem unlikely, it would, unless the utmost precautions were taken t o cause it to combine with the lead oxide, retain a large proportion of the water in the combustion tube.Again, there is no mention of the lead oxide having been heated while the current of dry air was drawn through, a precaution which certainly seems necessary. By carefully washing and drying the substance, prepared with all the precautions given by Michaelis and Pitsch, Chapman and Lidbury * Michaelis and Arend (Annalen, 1901, 314,2d9) have recently published a paper on the same subject, to which we shall also have occasion t o refer later,SO-CBLLED SUBOXIDE OF PHOSPHORUS. PART 11. 1237 found that the phosphorus exceeded by 2-3 per cent.that required by the formula P,O. Michaelis and Arend, in reply, state that the solu- tion should be cooled t9 Oo, and the precipitation of the red solution effected with acetic in the place of hydrochloric acid, and that under these circumstances the substance will not contain an excess of phos- phorus. From our point of view, the acetic acid is only another possible source of impurity, and it is not at all surprising that the percentage of phosphorus should have been by this means reduced. We have in the present research made no serious attempt to free the red phosphorus from this impurity, but we have simply adopted Michaelis and Arend’s method of preparation, and shown by analysis that the substance obtained cannot have the formula P,O. Estimation of the Phosphorus in the so-called Suboxide.-The specimen was weighed out into a Geissler flask, and dissolved in a small quantity of dilute nitric acid by warming on the water-bath.As soon as the solution was complete, strong nitric acid was alIowed to flow down the condensing tube together with a few C.C. of hydrochloric acid, and the mixture heated on the water-bath for 12 hours. The contents of the flask were then washed out into a large porcelain basin, and the excess of acid evaporated off, I n order to be quite certain that no error was introduced from the action of the acid on the glass, the phosphoric acid was first precipitated with ammonium molybdate. The ammonium magnesium phosphate obtained from this in the usual way was dissolved in hydrochloric acid, and reprecipi- tated by the very slow addition of ammonia from a burette with constant stirring, 10 C.C.of magnesia mixture being added before the whole was left to stand, Estimation of the Hydrogen in the so-called Su6oxide.The method employed consists in distilling the specimen, whereby phosphorus vapour, steam, hydrogen, and phosphoretted hydrogen are formed, The phosphorus vapour is made to combine with copper, and the water is decomposed by red hot aluminium, EO that it is thus possible to obtain all the hydrogen as gas. To avoid the absorption of moisture during manipulation, a soft glass tube, B, was closed at one A B end, and another tube, A , was ground into the other end. A was drawn out to a fine capillary which communicated with the air. The specimen can be kept for days in such a bottle without any appreciable alteration in weight.The weighing bottle was dried in an air-bath, the ‘‘ suboxide ” rapidly introduced, and the stopper, A , inserted, the weight of the bottle having been previously obtained. After finding VOL. LXYIX. 4 Q1238 BURCIESS AND CHAPMAN: NON-EXISTENCE OF THE the weight of the specimen, the hydrogen was estimated in the following manner :- A hard, Jena glass tube of 8 mm. internal diameter was closed at one end, and the bottle, A , which contained I' suboxide," dropped to the bottom. A glass rod, B, was then introduced, and about one-half of the remainder of the tube was first tightly packed with clean copper gauze, C, and then with rolls of aluminium foil, D. The surface of the aluminium foil had been previously cleaned by warming with ether for many hours.The Jena tube was fixed in a horizontal position, and communicated through a phosphoric oxide tube with a continuous Sprengel pump. It was thus left, with the pump working, until a good vacuum had been obtained. The part of the tube containing the copper and aluminium was then heated to a temperature at which the aluminium just began to melt. A carefully chosen tube will stand the pressure of an atmosphere at this temperature. During this heating, a little gas was pumped out from the aluminium and copper, but the volume rapidly diminished, and was soon D C B A negligible. On heating that part of the combustion tube which contained the " suboxide " with a Bunsen burner, hydrogen was evolved in considerable quantity, and continued to come off until all the phosphorus had volatilised.The gas was collected at the pump, and retained for analysis. The residue remaining in the bottle, which contained the suboxide, was a perfectly black, non-volatile substance, .and was proved to contain both carbon and phosphorus," I n two experiments, the substance was heated in the same way as described above, with the excoption that no copper or aluminium was introduced into the combustion tube. The amounts of hydrogen were in these cases also far too large to allow of the substance having the formula P,O. Examination of the Gas EztoZued.-The samples of gas which had been collected at the pump were transferred to a eudiometer andexploded with * This residue will not burn in oxygen even when strongly heated.I n one experiment, it was shown to contain carbon by heating it with lead chromate in a Combustion tube and collecting the carbon dioxide in potash.SO-CALLEb SUBOXTDE OF PHOSPHORUS. PAKT 11. 1239 excess of oxygen. I n several cases, the contraction occurring was too large for pure hydrogen. This was apparently due to the presence of some hydrocarbon, since the gas remaining after explosion was in these cases partly absorbed by potash, The improbability that the substance examined is the oxide P,O is just as great whether the gas contained a hydrocarbon or consisted of pure hydrogen. Pwpamtion of the Samples.-SnmpZe I was prepared by precipitating the red solution which results from the action of aqueous alcoholic potash on white phosphorus with very dilute hydrochloric acid.The details given by Michaelis and Pitech (Annalen, 1899, 310, 56) were observed, the only variation being that the hydrochloric acid was kept in constant rotation by means of an automatic stirring apparatus. After allowing one quantity of the red solution to pass through the filter paper, the hydrochloric acid was left to stand for some time, in order to regain the temperature of the laboratory, before another quantity was prepared and filtered. As only a small amount of the substance was prepared and the time taken in its preparation was considerable, we are inclined to think that the low percentage of phosphorus is due to the action of the caustic alkali on the filter paper. The washing and drying were performed as described by Michaelis and Pitsch.The complete analysis shows that the substance cannot be a suboxide, P,O. The clear red solution was allowed to run from a separating funnel, drop by drop, into very dilute hydrochloric acid, which was kept stirred as before. The acid was cooled from time to time with cold water. The precipitate was washed successively with water, alcohol, and ether, and then with carbon disulphide, which has been shown to remove ordinary phosphorus completely from such a preparation (Chapman and Lidbury, Zoc. c&, p. 976), should i t happen to contain any. It will be seen from the analytical numbers that the amount of phosphorus is 4.3 per cent. higher than that required by the formula P,O, also that the black residue left on distillation in a vacuum is much smaller than in sample I.The sample wad completely soluble in aqueous alcoholic potash. Xarnple I11 was prepared exactly according to the new directions given by Michaelis and Arend (Zoc. cit., p. 263), the utmost precautions being taken to keep both the red solution and the precipitating acid at 0'. The percentage of phosphorus is still 1.5 per cent. higher than that required by the formula P,O, A'amyde I T was prepared from ammonium hypophosphite in the following way (Michaelis and Arend, loc, cit,, p. 266). Thirty grams of ammonium hypophosphite were dissolved in 100 grams of glacial acetic acid, and 45 grams of acetic anhydride were gradually added. Xccmple I1 was prepared without employing filter paper. The red solution was prepared from sample IV.4 Q 21240 BURGESS AND CHAPMAN : NON-EXISTENCE OF THE The mixture was warmed on the water-bath until a precipitate began to appear. It was then cooled, another 45 grams of acetic anhydride were added, and tbe mixture was heated again on the wnter-bath. Finally, the whole was poured into excess of water, filtered, washed, and dried. We experienced great difficulty in removing acetic acid from this precipitate by washing. The analysis shows that the sub- stance cannot be the oxide, P,O. Sampls Y was prepared from sample IV by warming for some time on the water-bath with dilute hydrochloric acid. This removed the im- purities to such an extent that the amount of phosphorus rose 1 per cent. Xample YI was prepared from the red solution, obtained by dissolving sample IV in aqueous alcoholic potash, by precipitating with hot hydro- chloric acid (one volume of strong hydrochloric acid to two volumes of water).Analytical Results.-The analytical results obtained were : Sample. 1. 11. { 111. { IV. { v. VI. Per- centage uf P. a6 7 92'87 92.81 90'01 86 5 0 ~ 7 . 5 ~ 89.75 C.C. of hydrogen per 0.1 gram of substance. 11.14 8.87 8-13 7.05 3-63 7 *55 4 '72 7-51 Contraction Found 1 C.C. 17.06 13.35 12.17 contraction 7 -66 12.97 7'43 Calculated for hydrogen, C.C. 16-71 13'31 12-19 not found 5-45 11.33 7-07 Residue per 0.1 gram. Gram. 0.0059 0-0013 0'0029 0'0052 0-0045 Formula, as- suming that the substance is a chemical com- pound, and neglecting residue. Since 0.1 gram was ahout the amount of substance taken, the hydrogen is calcu- In calculating the formula in the last column, no allowance is made for the residue, The second value for the hydrogen with samples I11 and I V was obtained when It lated as c.c., and the residue as grams per 0.1 gram of substance.and therefore the number attached to the oxygen is somewhat too large. the substance was distilled in a tube which contained no copper or aluminium. was smaller becalise the water vapour was not decomposed. In this experiment, the residue was heated with a solution of potash, and then This residue gave, on combustion with lead chromate, 0'0006 gram of carbon. a This does not represent all the hydrogen which can be obtained by heating alone, since during the experiment the stopper came out of the little bottle, and ~ o m e of the " suboxide " was blown through the tube by the escaping hydrogen.* The absorption which occurred on the addition of potash after explosion was 0.77 C.C. The residue gave, on combustion with lead chromate, O*OOQQ gram of carbou. washed with water, proving that a considerable portion is insoluble in potash.60-CALLED SUBOXIDE OF PHOSPHORUS. PART 11. 1241 Michael+ und Pitsch's Pure and Impure Suboxides. Michaelis and Pitsch distinguish between impure suboxides, which are supposed to contain red phosphorus and solid phosphoretted hydro- gen, and p w e suboxides, free from these substances. The following are their analytical numbers for the impure preparations : P 90.35; 90.1 89.80; 89-72 90.48 90.89 91.85 per cent. H 1-73 0.3 1.36 0.95 1.95 ,, The calculated percentage of hydrogen in the compound P,H, is 1-58, and therefore some of the impure suboxide specimens contain more hydrogen than the solid hydride of phosphorus to which the impurity is supposed to be due.Again, there is nothing in the methods of preparation of the impure and pure substances to suggest such a marked difference in composition as the authors claim for them. For example, a pure suboxide containing no hydrogen is obtained by the action of acetic anhydride on ammonium hypophosphite, and an im- pure substance, containing a large percentage of hydrogen, by the action of acetyl chloride on hypophosphorus acid. Our analytical numbers afford a strong oonfirmation of the view already expressed that the '' suboxide of phosphorus " is only impure red phosphorus.The Red Xolution of Phosphovus in Aqueous Alcoholic Potash.--When ordinary phosphorus is dissolved in aqueous alcoholic potash, a red solution is obtained. The solution, when first formed, is apparently quite clear, but on allowing it to remain at the temperature of the laboratory, it rapidly decomposes with evolution of hydrogen and phosphoretted hydrogen, and loss of colour. If a strong solution is allowed to decompose, it becomes muddy, and a brown solid slowly settles to the bottom, When the solution is kept at Oo, the same changes take place, but much more slowly. On warming on the water- bath, the decomposition proceeds very rapidly, and no brown solid is formed. The point to be borne in mind is that the red solution itself, when separated from the phosphorus from which it is obtained, is at the ordinary temperature very unstable, and readily decomposes with evolution of hydrogen and phosphoretted hydrogen.Now, some stress has been laid by Michaelis and Pitsch upon the supposed fact that ordinary phosphorus is soluble in aqueous alcoholic potash with evolution of hydrogen, whereas the '' suboxide " dissolves in it without liberation of this gas, which would indicate, if the observation were correct, that phosphorus must be partially oxidised before solution can take place. On performing comparative experiments, we find that the 4' suboxide '' and ordinary phosphorus dissolve in aqueous alcoholic --1242 BURGESS AND CHAPMAN : SOX-EXISTENCE OF THE potash at the ordinary temperature, in both cases with evolution of considerable quantities of hydrogen.A t Oo, the ‘( suboxide ” dissolves very much more rapidly than ordinary phosphorus, and in neither case is an appreciable quantity of hydrogen given off. The hydrogen can, therefore, in all cases be regarded as derived from the decomposition of the red solution, and certainly affords no evidence of the state of oxidation of the substances dissolved. No conclusions can be drawn from the relative rates of solution of the two substances, since the particles of the one are so small that their shape can scarcely be seen under a microscope (one-sixth inch objective), whereas the granules of the other can readily be seen with the naked eye. If ordinary phos- phorus could be obtained as finely divided as the “suboxide,” there is every reason for thinking that the red solution mould be formed at least as readily in the one case as in the other.The red solution may be precipitated either by neutralisation with an acid, or by the addition of an ammonium salt. The colour of the precipitate obtained with a strong acid is light yellow or green, with a weak acid it is a darker green, and with ammonium chloride it is still darker. The dark precipitates, when warmed with a strong acid, become bright yellow, indicating that the dark colour is probably due to the presence of a base in the precipitate, which can only be com- pletely removed by a strong acid. The substances obtained in the absence of any base are always bright yellow, such, for example, as the ‘‘ suboxide ” formed by the action of acetic anhydride on hypophos- phorus acid.It may here be mentioned that the brown solid, obtained by the decomposition of the dark red solution, becomes bright yellow when warmed with hydrochloric acid. Michaelis and Pitsch assert that certain specimens of impure “ sub- oxide” leave a residue of red phosphorus when dissolved in aqueous alcoholic potash. All our preparations, whether they contain 94 per cent. or only 86 per cent. of phosphorus, are completely soluble in this reagent, and a precipitate is formed only on allowing the resulting clear red solution to stand. Concerning the nature of the red solution, it may be pointed out that the presence of alcohol greatly accelerates the solution of flowers of sulphur in potash, persulphides being formed, By analogy, we are led to suspect that the solution of phosphorus in the same reagent contains a perphosphide.This supposition, if correct, would account for its colour, and for the precipitation of phosphorus on the addition of an acid. The potassium phosphide is soluble in alcohol with forma- tion of a deepred solution, but the reaction is so violent that it is attended with considerable decomposition,SO-CALLED SUBOXIDE OF PHQSPBOBUS. PART 11. 1243 Solubility of Red Phosphorus in, Aqueous Alcoholic Potash. Besides establishing a very close relationship between the so-called suboxide of phosphorus and red phosphorus in other respects, Chapman and Lidbury have shown that Pedler’s phosphorus (obtained by the action of sunlight on a solution of phosphorus in carbon disulphide) is, like the “suboxide,” soluble in aqueous alcoholic potash.This form of phosphorus was used because it can easily be obtained in a fine state of division. Michaelis and Arend find that this phosphorus contains, as might be expected, both carbon and sulphur, and they believe that some oxygen is also prosent, which is derived from the wash-water by the substitution of oxygen for gulphur. By using carbon tetrachloride in the place of carbon disulphide as a solvent for the phosphorus, they obtained, by the action of sunlight, a substance which they were unable to dissolve in aqueous alcoholic potash. It is not, however, necessary to consider the ex- periments of Michaelis and Arend on this point in detail, since the com- plete solubility of red phosphorus in aqueous alcoholic potash is so easily proved that it cannot for a moment be doubted. To establish this fact, the following experiments were per- formed.The apparatus employed to prepare the red phosphorus is shown in the accompanying diagram. The tube B was partially filled with red phos- phorus, which was prevented from entering A by means of a tight plug of glass wool. The red phosphorus, before being introduced into the tube, was washed in turn with water, alcohol, and ether, and dried in a vacuum over phosphoric oxide. The T-piece, B, was fused on to a tube, which communicated through a phosphoric oxide tube with an automatic Sprengel pump in such 4 manner that A and B were horizontal. After exhaustion, the phosphorus was very slowly dis- tilled into A .During the distillation, a little hydrogen was evolved which was pumped out of the apparatus after sufficient phosphorus CLASS W O O L .1244 SUBOXIDE OF PHOSPHORUS. PART 11. had come over. The tube A was finally fused off at E and F, and the phosphorus was distilled into the end remote from the tube D. The ordinary phosphorus may be completely reconverted into red phosphorus by passing the silent discharge through the vapour. This was done by connecting one pole of an induction coil with copper wire wrapped round the middle of the tube A, and the other pole with mercury con- tained in D. The conversion takes place only slowly a t the ordinary temperature on account of the low pressure of the vapour; it is, however, accelerated by heating the tube to 100' in an air-bath.The red phosphorus was deposited over the inner surface of the tube. When it was thought that the conversion was complete, the tube was opened under carbon disulphide, and allowed to remain in contact with this liquid for Bome hours. The carbon disulphide did not leave any residue of phosphorus on evaporation. The red phosphorus was finally removed from the walls of the tube, ground in a mortar, col- lected on a filter paper, and washed with ether. On treatkg the product with aqueous alcoholic potash, a deep red solution passed through the filter paper, and this gave a voluminous precipitate with hydrochloric acid. A residue consisting of much larger grains dis- tinctly visible to the naked eye remained bebind, which was rendered soluble by regrinding.The same experiment may be performed with commercial red phos- phorus. The sample used by us was in the form of a powder, and did not contain ordinary phosphorus. It was reduced to a fine powder by grinding in an ordinary porcelain mortar. An agate mortar cannot be used for this purpose as its surface is so smooth that the pestle slides over the particles. The fine powder, after being washed successively with hydrochloric acid, alcohol, and ether, was readily soluble in aqueous alcoholic potash, giving a considerable quantity of a deep red solution, which was then precipitated by hydrochloric acid. The residue re- maining on the filter paper was easily rendered soluble by regrinding. By repeating the process of grinding several times, all the phosphorus, with the exception of a small quantity which could not be removed from the filter paper, was coaverted into the red solution. Conclusions. From the experiments described in this paper, and from the facts established by Chapman and Lidbury, we conclude : (1) That no suboxide of phosphorus having the formula P,O has ever been prepared, because the percentage of phosphorus is variable, and because the latest substance, described as having the composition P,O, contains other elements in addition to phosphorus and oxygen, hydrogen being present in considerable quantity.ACTION OF AMMONIA ON METALS AT HIGH TEMPERATURES. 1245 (2) That the substance described as a suboxide of phosphorus is impure red phosphorus, because the properties of both are the same, and because direct analyses have shown that the impurities in the alleged suboxide (regarding it as red phosphorus) are such as might be expected from the methods of preparation. THE OWENS COLLEGE, MANOHESTER.