154 TILDEN AND BARNETT: THE NOLECULAR WEIGHT XVI1.-The iWolecuZw Weight and F o r m d a of Phos- p hokc Ada ychide a l a d of M P t ap hosp hor ic A cid. By W. A. TILDEN, D.Sc., F.R.S., and R. E. BARNETT? B.Sc., Assoc. R.C.S. Plzoyhoric Adayd?-ide. Considerable advances have been made within recent years in the study of compounds of the elements belonging to the phosphorus group, and among the results established may be remarked especially the fact that the oxides generally., and wveral of the sulphides, con- form to the molecular type of the elements themselves, that is, they cont,ain four atoms oE phosphorus, arsenic, or antimony. The formulm, As40G, and Sb40s, have been est.ablished for arsenious and antimonious oxides respectively by the vaponr density deterniinations published by V.and C. Neyer (Ber., 1879, 12, 1117, 1282). The formulae, P406, aiid P406S4, have in like manner been settled by Thorpe and Tutton (Trans., 1890, 57, 551, and 1891, 59, 1022), and the lower sulphide has the corresponding formula, PASs, calculated from the density of its vapour (Isumbei.t, C.R., 1886, 102, 1386). The case of the higher sulphide will be discussed later on. The compound described by Thorpe and Tutton under the name of phosphorus tctroxide (Trans., 1886, 49, 833), is regarded by them as having the molecular formula, P204, and hence as the analogue of nitrogen tetroxide, but no determination of the vapour-density has yet been attempted. Phosphoric anhydride alone remains unexamiued among the com- pounds of phosphorus which are volatile without dissociation, but it is usually still expressed as pentoxide, P,O,, from analogy with the sulphide, the vapour density of which has been determined by V.and C. Meyer (Bey., 1879, 12, 610). According to Shenstone and Beck (Trans., 2893, 63, 475), phos- phoric anhydride usually contains lower oxides of phosphorus ; for the preparation of the pure oxide we have, therefore, adopted their process with slight modifications. The sublimed product, after pass- ing together with dry oxygen through the heated platinum sponge,AND FORJfULX OF PHOSPHORIC ANHYDIIIDE, ETC. 155 was collected in a narrow tube of hard glass united by means of tho blowpipe tlo the end of the combustion tube. When full, this tube was drawn off at its junction with the combustion tube, exhausted by the Sprengel pump and sealed.Separate small portions could then be sealed off for successive experiments. The oxide thus obtained was chiefly in the form of a crystalline crust adherent to the glass, but also partly in a loosely flocculent con- dition. When free from phosphoric acid, i t does zot melt on the application of heat, but sublimes rapidly below a red heat. An attempt was made to melt the oxide in a tube attached to an apparatus in which i t could be exposed to a pressure of nearly 7 atmospheres. On applying the flame of a spirit lamp, however, the substance sublimed without melting. I n order to determine the vapour-density of this substance, it is necessary, on account of its extremely hygroscopic nature, to avoid contact with ordinary undried air.Platinum is the only available substance capable of resisting to a sufficient extent its action at a red heat, though even platinum is liable to be corroded, especially in the presence of oxygen (Barnett, Trans., 1895,67,513). Hence peculiar difficulties stand in the way of exact determination of the density of the vapour of phosphoric anhydride, and the results of our earlier experiments were unsatisfactory owing to various causes which it is unnecessary to discuss here, being attributable in most cases to in- sufficiently high temperature, or to some imperfection in the ap- paratus. It may, however, be stated that the value deduced for the molecular weight was in erery case more than double the nnm- ber, 142, which corresponds f o the himula, P,O,.The method used was Victor Xeyer’s air expulsion process. The apparatus consisted of a c3;- lindrical platinum bottle, 70 cm. high and 6 cm. in diameter, hold- i n g about 300 C.C. and haviug a tubular neck 40 cm. long by 1Q cm. internal diameter. Into the end of t,his platinum tube w as fitted a glass tube with two branches ; one, a narrow bore gas-leading tube, a , the other a short straight tube b, which could be closed by a stopper. The top of the glass tube was closed by a1.56 TILDEN AND BARNETT: THE MOLECULAR WEIGI-IT rubber stopper traversed by a perforation through which passed, gas-tight, a glass rod having a mark upon it so that i t could always be pushed in to the same extent. The small tube containing the substance could be inserted into the rubber stopper from below, and would be held in the perforation until pushed out by the glass rod.The source of heat was in the earlier experiments a gas muffle, but this was afterwards replaced by a gas furnace in which the platinum bottle could stand upright. It was found necessary to protect the platinum from contact with the furnace gases, and ac- cordingly we procured some glazed fireclay cylinders closed at one end, into which the bottle would just slide, and with this covering i t was placed in the furnace. The top of the cylinder projecting into the air was packed with asbestos card 80 as to prevent cooling by convection. The expelled air was at first collected over oil of vitriol, but as this was inconvenient it was replaced by kerosene of known density and vapour-pressure.I n order to fill the bottle with dry air, a long tube passed through the rubber stopper to near the bottom of the bottle. As the end of this tube when wholly of glass was found to collapse when hot, about 30 cm. of the lower end was replaced by a tube of platinum foil, Bulbs filled with oil of vitriol and phosphoric anhydride were connected with the side-branch, and thus a current of dried air entered here, passed down into the bottle, and escaped up the long tube. This operation was performed after the platinum bottle had been heated to the requisite tlemperature and all was ready for the experiment. By making use of the same side-branch and long tube at the end of an experiment, the vapour could be expelled and the apparatus left ready for a second operation.An approximate estimation of the temperature obtained was made by means of the melting-points of salts. A few crystals of the dry salt were wrapped in a little cylinder of platinum gauze and lowered into the bottle when heated. Suspended at a distance of 2 c.m. from the bottom, sodium chloride was always melted, while potassium sulphate only showed signs of incipient fusion. The latest values for the melting-points of these salts are 815' and 1078" respectively (V. Meyer, Riddle, and Lanib, Bcr., 1894, 27, 3128). It may, there- fore, be concluded that the temperature in the later experiments was in the neighbourhood of 1000°. The little tube containing the phosphoric anhydride having first been nicked by a, file near one end, and wrapped in a piece of platinum foil, was weighed, A crack was then started in the tube by the application of a hot rod, the extremity broken off, and the tube in- stantly fixed in its place.The glass was weighed after the experi- ment was over and it was always found that the two pieces fitted together perfectly.AND FORMUL,Q OF PHOSPHORIC ANHYDRIDE, ETC. 157 The following results were obtained in the manner described. Weight of Vol. of air phosphoric reduced to Vapour Moleculay No. anhydride. 0" and 760 mm. density. weight. 1 . . . . .. 0.2533 gram 15.96 C.C. 177.0 354 2 ...... 0-0841 ., 5.06 ,, 185-0 370 3 ...... 0.2055 ,, 14.02 ,, 163.0 326 4 .. .>.. 0.1904 ,, 13.83 ,, 153.6 307 5 .... .. 0.2600 ., 18.22 ,, 163.0 326 6 .. .. .. 0.1859 ,, 12-33 ,, 167.5 335 Calculated molecular weight, for P,Os = 142, for P,O,* = 284.Experiments numbered 3, 4, 5 and 6 were made at a temperatuiae much higher than I and 2. The trustworthiness of 2 is much less than that of the others on account of the small quantity of substance operated on. For the sake of comparison, a determination of the rapour density of mercury was made immediately before the last ex- periment, and as it gave the value 99.3, it is obvious that the opera- tion mas rightly conductcd. The numbers for phosphoric anhydride therefore undoubtedly point to the formula P401, for this substance, though even at the temperature of bright redness it seems to be only imperfectly gasified. It seems remarkable that while the 4 atoms of phosphorus remain associated in this compound and in the oxjsnlphide, P406S4, dis- covered by Thorpe and Tutton, the molecule should divide when the whole of the oxygen is raplaced by sulphur giving the pentasulphide, P,S,.It is noticeable also that the lower sulphide is represented as P,Ss on the evidence of the vapour density (Isambert, Compt. rend., 1886, 102, 1386). Now, supposing the molecular constitution of the higher sulphide to be P4S10, and that on vaporisation it dissociates into P,S, and S4, the vapour density would have the value (7.67) found by V. and C. Meyer. The boiling point of the so-called penta- sulphide is 530°, and Meyer determined the density at a temperature described as a d w k red (" Die Teinperatnr war so gewahlt, dass das Blei dunkle Rothgluth zeigte," Rer., 1879, 12, 611). A very low red heat, as shown by fireclay bricks, corresponds to about 700".The density of sulphur vapour at 606' was found by Biltz, using Dumas' process, to be 4.734 ; and from estimations through a range of lower temperatures it was found steadily t o diminish from 7.937 at 467.9". It may fairly be supposed, therefore, that the density of sulphur vapour at, say, 650°, assuming this as approximately the temperature of the lead bath in Meyer's experiments on phosphorus pentasulphide, would be somewhat less than 4.7, though not very much less, as i t is evident, on plotting Biltz's figures on squared paper, that the curve is very irregular, and that there has been a sudden fall in value from138 TILDEN AND BARNETT: THE MOLECULAR WEIGHT the figure given immediately before.* Assume the density at a '' dark red " heat to have been 4.3, then equal volumes of such rapour and the vapour of P,S, mould give a mixture having the mean density of 45+10-9/2 = 7.7, which is almost exactls the value found by V.and C. Meyer. The vapour of phosphorus sulphide is brownish- yellow, and, though paler, is similar in colour to sulphur vapour. Altogether the proof that the highest sulphide of phosphorus is a. pentasulphide is far from complete, and certainly the densit.$ of its vaponr cannot be accepted as evidence bearing on the case of the corresponding oxide. Metctphospphoric acid. I n the course of preparing successive batches of phosphoric anhy- dride, our attention was repeatedly drawn to the presence of drops of liquid in the front part of the tube when the distillation was COII- ducted a t too high a temperature.On these occasions the fused, glassy residue usually left in the platinum boat WAS not to be seen, and we came to the conclusion that the drops consisted of metaphos- phoric acid, and that this compound is far more readily volatile than is commonly supposed. H. ROSC, i t is true, states (ArtnuZen, 1830, 76, 2, 13, and 1851, 77, 319) that inetaphosphoric acid volatilises at a bright, red heat, bnt i t has never been regarded as a compound that could be easily distilled. With the knowledge of this fact, however, we thought that it might be possible to take the vapour density, and that the results would be interesting. The acid was prepared by dissolving commercial phosphoric anhy- dride in nitric acid, evaporating the solution in a platinum dish anti1 the liquid ceased to evolve bubbles of mpour, and then boiling away about half of it in a platinum crucible heated to redness.Samples of the residue were then quickly taken out by means of a little platinum dipper, dropped into ft small tube of platinum foil, and immediately enclosed in a glass tube, which was sealed up and weighed. Two determinations of density made a t a bright red heat gave the following results. Calculatd for. r-------\ I. [I. HPOp H,P,O,. HAPJOIp 76.8 78.2 40 SO 160 * The three last values gircii in Biltz's table (Be,,., 188S, 21, 2017) :we as fol- lows :- Temperature. ..... 680.9" Drnuitp. ..... 5.607 ? I ...... 5809 ........ 5'412 ,, ...... 606.0 ........ 4.734 from which it is clear that great accuracy cannot be claimed for these result,s.AND FORMULA OF PHOSPHORIC ANHYDRIDE, ETC.159 Experiments showed, however, that the acid prepared in this man- ner was liable to vary in composition as to the percentage of anhy- dride it contained. It was therefore necessary to analyse every specimen of the phosphoric acid of which the vapour density was taken; in all cases it was found that it contained a quantity of anhydride in excess of the amount, 88.75 per cent., which corresponds to the formula (HPO,),,. The two following series of experiments were made on two different preparations. Nuaaple 1. 0.0978 gram gave v. d.. . . . 0.1070 ,, ,, . . . . 69.1 1 71.6 1 Mea,n v. d. = 70.3. 1.4798 ,, 90.18 per cent. of phosphoric anhydride. Sample II.Fg*9}Meiln v. d. = 69.8, 0.0861 gram gave T. d.. . . . 0.0719 ,, ,, . . .. 69.7 = 90.86 per. 0.7880 ,, 90.79 ,, 9 , 1 Mean cent. anhydride. 1-2997 ,, 90.94 per cent. anhydride Sample I contains therefore 8'7.25 per cent. (HPO,),,, and 12.75 per cent. anhydride, corresponding to 12H2P,06,P4010 ; and Sample II has 81.2 per cent. (HPO,),, and 18.8 per cent. of anhydride, corresponding to 8HaPzOs,P4010 approximately. It is obvious from the three series of experiments, that although the composition of metaphosphoric acid varies a little, the vapour of this substance consists chiefly of a dimetaphosphoric acid, HZP206, which is apparently liable to undergo partial dissociation at a high temperature, and even during ebullition to part with 8 small quantity of water.Comtitzi tion. It is, perhaps, idle to speculate as to the possible constitution of compounds such as the oxides and acids of phosphorus, but a, few words may not be out of place. The group o f 4 atoms of phosphorus is a very stable form of struc- ture, inasmuch as it bears very high furnace temperatures without breaking up. It is only at a white beat that there is eridence of even incipient dissociation (Biltz and V. Meyer, Ber., 1889, 22, 725). The molecule of phosphorus being represented as - - =P - P=160 PHOSPHORIC A KHYDRIDE AND METAPHOSPHORIC ACID. i t is evident that there are but 12 units of disposable valency. This can only provide for the accession of 6 atoms of oxygen, unle,, 0s we assume either the linkage of oxygen to oxygen in a chain, or the disruption of the bond between the phosphorus. The former is so improbable a hypothesis in the present instance that we are reduced to the latter, and the formula of the two oxides assume the following shape, in which the dot represents phosphorus. Formulce somewhat in this sense have already been suggested for P406 by Thorpe and Tutton (Trans., 1890, 57, 563). On the intro- duction of water the molecule of phosphoric anhydride is divided into two parts, and metaphosphoric acid results, which, from the formula given above for the anhydride, would have the constitution I. ,OH 0 = P-OR I 11. 0 From this the usual formula for pyrophosphoric acid, 11, is immeiii- ately derived, Royal College of Science, Loledon.