THE POLYSULPHIDES OB THE ALKALI METALS. PART 111. 1063 XCIV.-The Pol ysdphides of the Alkali Metals. Part III. The Solidifying Points of' the Systems Sodium Monosulp hide-Sulp 12 ur and Potassium i l l 0 7 1 o-sulph id e-Sulplzur. By JOHN SMEATH THOMAS and ALEXANDER RULE. THE formation of polysulphides by the action of sulphur on alcoholic solutions of the arthydrous hydrosulphides of sodium and potassium has previously been investigated and described by the authors (T. 1914 105 177 2819). By means of this reaction it was found possible t o prepare sodium tetrasulphide and potassium pentasulphide in the pure and anhydrous form but it was observed that when attempts were made t o utilise i t for the preparation of other polysulphides of lower sulphur content the above-mentioned compounds contaminated t o a greater or less extent with the corresponding hydrosulphides were always obtained.This result is in general agreement with the work of Bloxam (T. 1900 77, By measuring the quantities of hydrogen sulphide evolved when known weights of the hydrosulphides of sodium and potassium were treated in boiling alcoholic solution with varying amounts of sulphur the present authors (Zoc. cit.) have shown that the reaction between sulphur and the hydrosulphides of the alkali metals in solution tends t o the formation of one polysulphide only; in the case of sodium the tetrasulphide and in the case of potassium the pentasulphide. The only polysulphide of sodium that Bloxam succeeded in pre-paring was the hydrated enneasulphide to which he assigned the formula Na,S,,14H20.He described however a regular series of the potassium compounds including an enneasulphide which he claimed t o have isolated in the anhydrous state. The question of the existence of enneasulphides of the alkali metals will be dis-cussed later. It may however be pointed out here that the present authors have failed to obtain any such compound either of sodium o r potassium. Further in view of the ease with which the polysulphides of these metals undergo hydrolysis and taking into consideration the fact that Bloxam's preparations were carried out in aqueous solutions the 1,urity of the anhydrous polysulphides described by him seems open to question. There is however considerable evidence of the existence in aqueous solution of a series of sodium polysulphides lrom the disulphide t o the hexasulphide and possibly beyond the work of 753) 1064 THOMAS AND RULE THE POLYSULPHIDES Euster and Heberlein on this point being specially important (Zeitsch.anorg. Chem. 1905 43 53; 44 431). These authors found that their results could only be explained satisfactorily by assuming that the solutions dealt with contained a number of anions of the type W SN2 Sllg etc. in equilibrium with each other. The present investigation was undertaken with the view of resolving the present uncertainty as to the number and the relative stabilities of the solid polysulphides of sodium and potassium. Two main objects were kept in view. I n the first,place it seemed desirable to decide whether higher polysulphides of these metals than those which have been described can exist in the solid state, and secondly the nature of the lower polysulphides and the ques-tion of the existence of intermediate compounds required to be settled.The last-mentioned point is of especial interest in view of Bloxam’s conclusion (Zoc. cit. and T. 1895 67 277) that the enneasulphides of sodium potassium and ammoniuin can exist as stable compounds. I n order to account for these intermediate compounds Bloxam suggested that the polysulphides of the alkali metals really consist of solid solutions of sulphur in fundamental compounds of the type R,S, and that therefore the general formula assigned t o them should be R,S, where ~ 1 may be any whole number greater than 5. I n the case of rubidium and cmium Biltz and Wilke-Dorfurt (Zeitsch.aitorg. Chem. 1906 40 297; 50 67; see also Rer. 1905 43 53) who thoroughly investigated the polysulphides of these metals by a thermo-analytical method found no indica-tion of the existence of the intermediate compounds required by this theory and the present authors similarly have failed t’o con-firm the existence of such compounds of sodium and pot.assium. On the other hand positive evidence has been obt>ained in support of the simpler general formula R,S,. E X P E R I M E N T A L . The main portion of the experimental work described in this paper consisted of the determination of the freezing-point curves for the systems sodium monosulphide-sulphur and potassium mono-sulphide-sulphur. These curves were constructed in each case in two sections the first coiisisting of that portion relating to mix-tures in which the percentage of sulphur exceeded that required for the formation of potassium pentasulphide or sodium tetra-sulphide as the case might be the second dealing with mixtures intermediate in composition between potassium pentasulphide or sodium tetrasulphide and the corresponding disulphide OF THE ALEALI METALS.PART 111. 1066 In a second series of experiments the higher polysulphides were heated in a steady stream of hydrogen alld the rate of desulphur-isation was determined by weighing at regular intervals. During each period of heating the temperature was kept as nearly constant as possible but i t was regularly increased for successive periods.Finally in order to decide between the simpler formula R,S, and the double formula advocated by Bloxam molecular weight determinations were carried out in the cases of those compounds which could be obtained in the pure state. For this the ebullio-scopic method was employed alcohol being the solvent. The hygroscopic nature of the substances dealt with made it necessary carefully to exclude moisture a t every stage whilst, owing to the ease with which they oxidise ali solidifying-point measurements had to be made in a stream of pure dry nitrogen. Frequently mixtures were ob t ainecl especially when the composi-tioh was not that of a compound which supercooled to a remark-able extent and it was sometimes possible t o keep these mixtures in the supercooled condition for several days a t the ordinary temperature.The crystallisation of these supercooled mixtures was usually accompanied by considerable increase in volume and it often took place so rapidly that the crucible containing the mixture was shattered. Preliminary experiments showed that a t temperatures above 400° the fused substances attacked glass to an appreciable extent, the disulphides being more active in this respect than the tetra-or the penta-sulphides. Except at much higher temperatures, they are however all without marked action on glazed porcelain, and vessels of this material were therefore used t o contain them. For the determination of solidifying points below 300° the apparatus used was very simple. It consisted of a weighed crucible of glazed porcelain in which the weighed substance from 15 to 20 grams in amount was placed.The crucible was then lowered to the bottom of a hard glass tube about 30 cm. long the lower end of which was sealed. This tube was clamped in a vertical position and its upper end was fitted with a rubber stopper through which passed a thermometer a porcelain stirrer, the exact weight of each having previously been determined and two gas-delivery tubes. One of these tubes just passed through the stopper whilst the other reached almost to the bottom of the apparatus. By their means the air was displaced by dry nitrogen. The mixtures were fused by heating the apparatus in an oil-bath, and a similar bath was used to maintain a constant external temperature during solidification.u u 1066 THOMAS AND RULE THE POLYSULPHIDES In consequence of the high melting points observed in the region of the disulphides it became necessary to employ a form of apparatus more suited to their measurement than the one just described. An electrically heated arrangement was finally adopted. This consisted of an internally glazed porcelain tub0 of approximately 3.5 cm. in diameter and 50 cm. long clamped in an upright position. A portion of this tube was wound with nickel-chromium wire; current from the lighting circuit was used and loss of heat by radiation and oxidation of the wire were prevented by surrounding the wound section of the tube with a thick layer of kieselguhr. It was found that temperatures up to 800° could easily be attained and that when equilibrium conditions had been estab-lished any desired temperature could be maintained with very slight variation over considerable periods of time.The pyrometer consisted of a platinum-platinum-iridium thermoelement in con-junction with a millivoltmeter which was carefully calibrated. The hot junction was protected from the action of the fused sub-stances by enclosing it in a very thin-walled porcelain tube. For the reasons previously mentioned silica could not be used for this purpose and it was found inadvisable to employ platinum.* The lower portion of the furnace tube was packed with kiesel-guhr through which a silica tube of small bore passed. On the kieselguhr the crucible was supported so as to rest in that region of the furnace where the temperature was most uniform.It was placed in position and removed by means of a stout platinum wire looped round it. A rubber stopper pierced with four holes closed the upper end of the tube. Through these holes passed the pyro-meter tube the stirrer the platinum wire referred to and lastly, a delivery tube which used in conjunction with the silica tube a t the lower end allowed of the displacement of the air by dry hydrogen or nitrogen. The pure potassium pentasulphide and the di- and tetra-sulphides of sodium used in these experiments were prepared by the methods described in earlier papers of this series (Zoc. cit.). After preparation they were kept iii a vacuum desiccator over phosphoric oxide. It has previously been pointed out that potassium disulphide 1 In some early experiments the hot junction was protected by a platinum tube but a t the close of the series the metal was found t o be considerably lighter in weight.On dissolving the fusion a brown insoluble powder was obtained which from a partial examination appeared to be platinum mono-sulphide. The quantity obtained however was insufficient to allow of a complete examination OF THE ALKALI METALS. PART fff. 1067 cannot be Qbtained by a method analogous to that employed for t.he preparation of the corresponding sodium compound since the reduction of the pentasulphide by metallic potassium yields a poly-sulphide mixture consisting mainly of the trisulphide but contain-ing a considerable amount of other substances the removal of which is very difficult.In their work on the polysulphides of rubidium and msium Biltz and Wilke-Dorfurt ( l o c . c i t . ) obtained mixtures of low sulphur content by strongly heating the penta-sulphides in a stream of hydrogen. It was decided to use this method for the preparation of potassiuin disulphide. F o r purposes of comparison sodium disulphide was also prepared in this way, and a series of solidifying-point determinations was made with material so prepared as the starting point. The results obtained in this series were found to agree very closely with those recorded in experiments in which sodium disulphide prepared by reduction formed the starting materisl. The sulphur used was twice recrystallised from carbon disulphide and then heated f o r several hours a t looo in order to remove the last traces of solvent.This substance was also preserved in a desiccator over phosphoric oxide. The Desulphurisalion of Sodium Tetrasulphide aud Potassium Pent a s d p hid e . The primary object of these experiments as has already been mentioned was the preparation of the materials necessary for the construction of the complete solidifying-point diagrams of the systems under consideration. The removal of sulphur from the easily prepared higher polysulphides was effected by heating them in a steady stream of hydrogen. At half-hourly intervals the heating was interrupted and after cooling the amount of sulphur lost by the substance was determined by weighing. During each period of heating the temperature was kept as nearly constant as possible and its value was 2 5 O higher for each period than for the one immediately preceding.The variations in the rate of removal of sulphur yielded valuable information with regard to the exist-ence of lower compounds and especially as to their relative stabilities. The heatings were carried out in a slightly modified form of the electrically heated furnace already described. A porcelain crucible of known weight and containing the weighed substance was intro-duced into the furnace tube and the stream of pure dry hydrogen started. The gas was prepared by the electrolysis of a solution of barium hydroxide using a current of constant strength thus u u* 1U08 THOMAS AND RULE THE POLYSULPHIDES avoiding variations in its rate of flow. During the time that the temperature of the furnace was being adjusted to the constant, value desired the crucible was drawn into the cool upper pottion of the tube and on the expiration of each heating period the sub-stance was cooled quickly in the same way.Frc. 1 . IO(700) 9( 650) S(600) 7(55fl) G(500) 4(400! j.( 350) 2( 300) Z(350) O( 200) 0 5 1 0 1 Na2M, Loss in weight per 100 grams of sodium tetrasulphide. The weight of sodiuiii tetrasulphide taken was 15.673 gram and 0.5657 gave 0.4624 Na,SO,. 0.2464 , 1.3178 BaSO,. S=73*47. Na,S4 reqclires Na = 26.44 ; S = 73.56 per cent. The results obtained are represented graphically in Fig. 1 in which the loss in weight per 100 grams of sodium tetrasulphide is an analysis of the substance gave the following figures : Na = 26.47 0s THE ALKALI METALS.PART 111. 1060 pIotted against the total time of heating. The temperature main-tained during the previous half-hour is shown and points are given a t which from the loss of weight. the fused material should have the compositjion of the inone di- and t ri-sulphides. As will be seen from the curve the tetrasulphide undergoes considerable decomposition below 300° aiid from this point the velocity with which the sulphur is reniovetl increases regularly as the temperature rises until the composition of the fusion approaches t h a t of the disulphide the temperature by this timc having risen t o between 625O and 650O. Before the loss reaches t h a t calculated for the disulphide however the direction of the curve changes markedly and the removal of sulphur proceeds much more slowly.This retardation is indeed so marked as to render the preparation of the monosulphide by this method a t the temperatures attain-able in the apparatus described practically impossible. The values given on the curve are of course only relative since the rate of loss of sulphur must obviously depend on the surface area exposed by the fused substance and also on the rate of flow of the hydrogen stream. They do however indicate the existence of a disulphide the stability of which towards heat far exceeds t h a t of the higher compounds. The total loss of weight in the course of the experiment was 6.575 grams and the composition of the product should therefore be Na =45*43 ; S = 54.57 per cent.Analysis of the product gave the following results which agree well with the composition calculated from the loss in weight : 0.4846 gave 0.6795 Na,SO,. 0.3725 , 1.4810 BaSO,. S=54.62. NazSe requires Na=41*76; S=58*24 per cent. From the excellent agreement between the composition of the product as determined analytically and that calculated from the loss in weight it is clear t h a t the sodium polysulphides do not volatilise to any appreciable extent when they are heated a t temperatures within the limits mentioned and this was found to be equally true in the case of the potassium compounds. Biltz and WilkeDorfurt (loc. cit.) found however t h a t when the disulphides of rubidium and czsium were prepared by this method volatilisa-tion invariably occurred and to a greater extent in tho case of the czsium compounds than in the case of those of rubidium.It would seem therefore that the polysulphides of the alkali metals, like the halogen salts show an increasing tendency t o volatilise with rise in the atomic weights of the metals from which they are derived. Na = 45.40 1070 THOMAS AND RULE THE POLYSULPHIDES Turning to the preparation of potassium disulphide the weight of the pentasulphide taken was 17.853 grams. The purity of the salts was established by analysis in which the following results were obtained : 0.3587 gave 0.2816 K,SO,. K=32-76. 0.2752 , 1.3429 BaSO,. S=67*09. K2S requires K=32.78; S=67*22 per cent. FIU. 2. K2SS K,S* K 8 K,S, Loss i n weight per 100 grams of potassium pentasulphate.The removal of sulphur was effect,ed and controlled in the manner already described and the results obtained are shown graphically in Fig. 2 OF THE ALKALI METALS. PART 111. 1071 Whereas sodium tetrasulphide decomposes slightly a t its melt-ing point 274O and loss of sulphur is appreciable a t 300°, potassium pentasulphide appears to be more stable at its melting point although in this case also appreciable loss of sulphur occurs a t 300O. As in the case of the sodium compound the rate of removal of sulphur increases regularly as the temperature is raised but the curve appears t o indicate a slight retardation when the composition of the fusion approximates to that of the trisulphide thus furnish-ing evidence of the existence of that compound as a substance the sulphur pressure of which remains constant betwen definite temperature limits.I n the case of potassium also the great stability of the disulphide may be deduced from the marked change in the direction of the curve when the composition of the fusion approaches that of that compound. Even a t 750° the further removal of sulphur takes place very slowly indeed. The final product of the experiment after cooling was analysed, and gave the following figures: 0.3935 gave 0.4838 K,SO,. K = 55.17. 0.4017 , 1-3008 BaSO,. S=44*49. K,S requires K=54*94; S=45.06 per cent. The total loss in weight was 7.306 grams and from this assuming that no volatilisation of the sulphides had taken place the com-position of the product should have been K-55-46; S=44*54 per cent.It is evident that sulphur had been removed in excess of the amount required for the formation of potassium disulphide. The agreement between the values obtained and those calculated from the loss in weight is not so good as in the previous experiment, but the divergence lies within the limits of reasonable experimental error. Since on examination of the furnace tube no trace of sub-limed sulphide such as Biltz and WilkeDiirfurt describe in the cases of the rubidium and casium compounds was found it was concluded that as in the case of the sodium compounds no appreci-able volatilisation had occurred. Both potassium pentasulphide and sodium tetrasulphide fuse to deep red almost black liquids and this is also true of the various products obtained from them by the removal of sulphur.On solidification the penta- and tetra-sulphides crystallise quickly with very little supercooling and the disulphides also crystallise quite readily although supercooling is more marked in their case. Between the di- and the tetra-sulphides however and especially in the neighbourhood of the trisulphides red glass-like substances were always obtained. I n these cases inoculation often failed t 1072 THOMAS AND RULE THE POLYSTJLPEIDES promote crystallisation and the amorphous substances thus obtained could be kept for considerable periods a t the temperature of the laboratory. Crystallisation usually took place with great rapidity when the temperature was raised and owing t o its being accompanied by considerable increase of volume shattering of the containing vessel commonly resulted.After crystallisation the mass usually remained transparent and red but presented the appearance of fracture in all directions. When the substances after crystallisation were allowed t o remain the crystalline form apparently changed and the mass became opaque and pale yellow, the depth of the colour depending on the percentage of sulphur in the substance. I n all cases the products were soluble in water, forming clear yellow solutions which however rapidly became cloudy owing to oxidation. The solutions of the disulphides were only faintly coloured but increase in the sulphur content was in each case accompanied by a progressive increase in the intensity of the coloration. The System Sodium Monosulphide-Sulphur.The apparatus employed in the investigation of this system has already been described. Every experiment was carried out in a &ream of nitrogen which had been passed over red-hot copper and then over phosphoric oxide, During the cooling of the mixtures the temperatare was read every fifteen seconds. The accuracy of the measurements depended not only on the exactness of the temperature readings which in the case of the thermometer could be trusted to 0.2O and in that of the pyrometer to 1-1.5O but also on the general tendency of the curve and the ease with which crystallisation could be induced. Thus the error in the neighbourhood of the disulphide where the curve is steep is probably far greater than in the regions where the curve is comparatively flat.Similarly higher accuracy is attainable in the case of those mixtures which solidify sharply without supercooling than in thel case of those in which the tendency to supercooling is very great and crystallisation even in the most f avourable circumstances only proceeds slowly. I n these cases the error in single determinations may be relatively great. The results obtained in the analysis of the original material used for the construction of the section of the diagram lying between the disulphide and the tetrasulphide have already been given. On analysis the tetrasulphide employed as starting material for the mapping out of the remaining portion of the curve gave the following figures OF THE ALKALI METALS. PART 111. 1073 0.4931 gave 0.4019 Na,SO,.0.2785 , 1.4904 BaSO,. S=73*52. Na,S requires Na = 26-44 ; S = 73.56 per cent. At the commencement of each series of experiments the sub-stance the crucible containing it and also the thermometer and stirrer employed were weighed. After each solidifying-point determination the fusion was allowed to cool in an atmosphere of nitrogen and when cold after removing the volatilised sulphur, the substance etc. were re-weighed. The increase in weight gave Na =26.39. the amount of sulphur actually added and from this the per-centage composition of each fusion was calculated. As a check on the compositions determined in this way the final mixture obtained in each series of measurements was analysed. I n every case the analytical results agreed well with the percentage compositions calculated from the total increase in weight.I n table I the experimental results obtained f o r the different mixtures investigated are given and they are also graphically represented in Fig. 3. In this diagram the solidifying points o 1074 Ji 0 - r 9; -5.g 2s -3 % &PI M 54.57 55-25 56.39 67.9 1 59.04 60.18 6 1.08 61-91 62.38 62-92 63-63 64.25 64.95 65-74 66.10 66-32 66.74 67-14 67-69 68.05 68-61 69.29 69.67 70-13 70.42 71.25 72.35 73-56 74- 15 74.45 74-62 74-96 75.38 75-72 76.29 76-37 76.97 77.43 77-71 78-03 78.31 78.41 78-63 79.15 80.15 75-1.7 THOMAS AND RULE THE POLYSULPHIDES 8 1 W E %& 3 p 1 j iL P i o d 3% 0 0-723 0.771 0.855 0-970 1.067 1.168 1.252 1.330 1-380 1.437 1.513 1,580 1.662 1.752 1.800 1.829 1.883 1-935 2.010 2.060 2.140 2-241 2.301 2-364 2.42 1 2.550 2.760 3.000 3.114 3-180 3.217 3.294 3.342 3.392 3.522 3.615 3.636 3.794 3.921 4-00 1 4.094 4.179 4.209 4.278 4.445 4.792 81.18 5-200 3 c PI .d .- in $ 3 4 M 504 455 439 445 44 0 432.5 421.5 413 403-5 300 374 356.5 322 290 206 214.5 218 220 223 223.5 235.5 246 252 256 263 2 73 2 75 272.5 269 266 267.5 260.5 257.4 253.5 248.7 250 250 251 251.8 251-7 250 249.5 249.7 249.7 249-5 249.6 249.4 TABLE I.Remarks. Eutectic point doubtful Approximate m.p., Naps,. These points were doter-mined pyro-metrically. Probably m. p. eut,ectic This point is taken from he curve. The eutectic solidified amorphous and this accounts for the wide difference in the values observed. = 206. .m.p. Na,S, taken as 223.5". Mean solidifying point = 222.6". of eutectic Solidifying point of Na,S,. Mean solidifying point of eutectic = 247.6". Solidifying point of Na,S,. Saturated solutions of sulphur in The other layer was sulphur with Na,S with rn. p. 249.6". solidifying point 11 8" OF THE ALKALI METALS. PART 111. 107 5; the mixtures used are plotted against their compositions the latter being expressed as the number of gram-atoms of sulphur added per gram-molecule of sodium monosulphide.As will be seen by reference to the curve the system contains two well-defined eutectics the first of which lies between the tri-sulphide and the tetrasulphide contains 68-15 per cent. of sulphur, and melts a t 222'6O whilst the second lies between the tetra- and penta-sulphide. The melting point of this second eutectic is 247*6O and it contains 75.48 per cent. of sulphur. I n addition t o these there can be no doubt from the general tendency of the curve that eutectics also exist between the di- and tri-sulphides, and also between the monosulphide and the disulphide. Owing to undercooling the melting point of the di-tri eutectic could not be determined accurately hut it is in the neighbourhood of 206O, and the mixture contains approximately 65 per cent.of sulphur. For the definition of the monosulphide-disulphide eutectic only one observation was obtained and this was made pyrometrically . The halt which was fairly distinct occurred a t a temperature of 432O. Although the existence and especially the exact solidify-ing point of this eutectic mixture cannot be considered to be determined by such a single measurement it should be pointed out that the observation is in agreement with the general direc-tion of the curve so that, on the whole the conclusion is justified that the mono-di eutectic does exist and that its solidifying point is probably in the neighbourhood of 430O. It contains approxim-ately 55.8 per cent. of sulphur. The solidifying point of the monosulphide has not been deter-mined since a sufficiently high temperature could not be obtained in the apparatus employed.It would seem probable however, from the direction of that portion of the curve which lies between the monosulphide and the mono-di eutectic as defined by the two points actually observed that the melting point of the mono-sulphide must exceed 800O. Turning to the maxima exhibited by the diagram the first maximum occurs when the composition of the mixture corresponds with the disulphide the temperature being 445O. From t.he fact that on either side of this point the curves fall away very steeply, it follows that the disulphide is quite stable towards heat a t its melting point. From the disulphide the mlidifying points of the mixtures fall nearly 250° until the di-tri eutectic is reached from which point they again rise but much less rapidly.I n the case of the tri-sulphide a definite maximum is not shown but the existence of this compound is established first by the pronounced break whic 1076 THOMAS AND RULE THE POLYSULPHIDES is observed in the curve when the mixture has the composition of the trisulphide and secondly by the presence of both di-tri and tri-tetra eutectics. The absence of a definite maximum is evidence of the unstable character of the trisulphide which must decompose at a tempera-ture below its melting point. From the break indicating the existence of the trisulphide the curve again rises t o a maximurn at 2'75O at which point the fusion has the composition of the tetrasulphide. Thus the existence of this compouiid as a substance stable a t its meltisg point is established.The behaviour of mixtures containing a higher percentage of sulphur than the tetrasulphide is interesting. As the sulphur content is increased the solidifying points become lower until the tetra-penta eutectic melting at 247'6O is reached. From this point the curve appears to be almost horizontal but' on closer inspection a very slight rise can be seen which reaches a maximum a t 251-8O the rise therefore being only 4 - 2 O . A t this maximum point the substance has a composition practically identical with that of the pentasulphide. Kremapn (Moizats?b. 1904 25 1311) has shown that in the case of additive compounds dissociation must theoretically result in the flattening of the maximum and that this flattening must be the greater the greater the extent to which dissociation takes place.Ile has investigated cases for example trinitrobenzene-naphthalene in which the solidifying points of both eutectics and that of the compound only differ by one o r two degrees. It would appear that in the case of sodium pentasulphide the fused sub-stance is so strosgly dissociated that the dissociation products are able to lower the maximum t o a point very little above the tetra-penta eutectic on the one hand and the solidifying point of the saturated solution of sulphur in sodium inonosulphide on the other. Abegg and Hamburger (Zeitach. anorg. Chem. 1906 50, 435) have recorded a similar observation in the case of the poly-iodides of potassium and their results have been confirmed by Kremsnn (Monatsh.1912 33 1081). I n this case the differ-epce between the maximum and the eutectic temperatures was found to be 3.7O. The results obtained by the authors in their present investiga-tion lead to the conclysion that the pentasulphide is the highest polysulphide obtainable by fusiog together sodium monosulphide and sulphur for immediately beyond the pentasulphide maximum the curve becomes horizontal indicating the formation of a saturated solution of sulphur in the latter compound. The melt OF THE ALKALI METALS. PART 111. 1077 ing point of the saturated solution is 249.6") and several points were obtained differing from this value by only 0*1-0-2* although the sulphur content was varied considerably. In order to make quite certain that this horizontal line really represented the formation of a saturated solution the solidifying point of a mixture of equal sinounts of sodium tetrasnlphide and sulphur was determined.After the fused mixture had been thoroughly stirred it separated into two layers the lower of which, the saturated solution solidified a t 249*4O whilst the upper layer of sulphur remained liquid until the temperature had fallen to 11S0 when it solidified to a pale yellow mass. From the appear-ance of the sulphur so obtained and also from its solidifying point, it would seem that the sodium polysulphicles do not dissolve in molten sulphur to any appreciable extent. No evidence whatever is furnished by the results obtained in this investmigation of the existence of intermediate compouiids Such as the enneasulphide.If intermediate compounds exist at all they can only do so a t low temperatures. The System Potassium i~onosulphide-Sulph.2cr. The methods and apparatus employed in the investigation of this system were exactly similar t o those already described when dealing with the sodium series. As starting materials the pure pentasulphide was used in one case and mixtures obtained by expelling sulphur from the penta-sulphide by heating it in a stream of hydrogen were employed in other series. Analyses of the peiitasulphide and of a mixture con-taining less sulphur than the disulphide have already been given. I n addition to these a third material was employed which on analysis gave the following results : 0.4971 gave 0.4639 Ii,SO,.K = 41.88. 0.2333 , 0.9856 BaSO,. S=58*04. These figures agree well with the composition calculated from the loss in weight and the substance therefore approximates to the t r isulphide . A t the end of each series of determinations the final product was analysed. I n every case the figures obtained were found to be in good agreement with the composition calculated from the weight of sulphur added in the course of the series. The results obtained are given in table 11 and in Fig. 4 they are represented graphically 1.078 \ 1.i-t.i sutectic. J Mean m. p. = 250.9. 1 THOMAS AND RULE THE POLYSULPHIDES These points were deter-' mined pyro-metrically. TABLE 11. 44.55 45.56 46.89 48.09 49.18 50.63 51-97 53-02 54.32 55.51 57.06 57-78 68.93 69.62 60.15 60-55 60.95 62.02 62.38 62.72 63.70 64.58 65.74 66.79 67.22 67.99 68.73 69.2 1 69.59 69.81 70.26 70.57 70.86 71.21 71.56 72.26 74.11 76.42 0.959 1-041 1.145 1.259 1-360 1-501 1.638 1-752 1.900 2.042 2-240 2.342 2.499 2.600 2.691 2.743 2.807 2.983 3-043 3.102 3-279 3.445 3.680 3.905 4-00 1 4.179 4-360 4-480 4.550 4.637 4.709 4.847 4-930 6.03 1 5.135 5.352 6-98 1 6.902 467.5 464-5 452.5 433 390.5 3GO 311 25 1-5 251.6 238 217.4 189 160.5 101 ? 130.7 141.6 145-5 152 166 178.2 192 20 1.9 204.3 206.5 204.2 200 196.7 192.3 190 187 185.2 188.6 1ss.9 188 188.1 187.9 185.3 468 -264 249 249 251.5 --101 ? --144 142.5 145 142 --183 183.7 183.1 182.4 182.2 -I I Remarks.Aniorphous. Probably tri-tetra eutectic. Tetra-pen ta eutec tic. Mean m. p. = 143.4. Pcnta-hexa eutectic. Mean m. p. = 182.9". Saturated solution of sulphur in I<,S6 m. p. = 188.1. I The sulphur layer solidified a t 117.5". An examinat,ion of the diagram renders quite clear the existence of three well-defined eutectic horizontal lines which lie between the di- and tri-sulphides the tetra- and penta-sulphides and the penta- and hexa-sulphides respectively. The eutectic mixtures corresponding with these horizontal lines have the following com-positions and melting points OF THE ALKALI METALS.PART 111. 1079 Di-tri eutectic S=54.32 per cent. m. p. 250.9'. Tetra-penta eutectic S = 62-31 per cent. m. p. 143.4O. Penta-hexa eutectic S=70*39 per cent. m. p. 182.9'. Owing t o the difficulty of obtaining the necessary material and the previously mentioned limitation of the apparatus employed the region between the mono- and di-sulphides has not been investi-gated. a region was encountered in which the tendency of the fusions to form highly supercooled amorphous masses was very marked and here no trust-Between the trisulphide and the tetrasulphide worthy transition points could be obtained. It seems probable that the amorphous substance referred t o in the table really consists of the tri-tetra eutectic mixture. The curve exhibits five maxima oE which those corresponding with the di- tri- and penta-sulphides are quite distinct.They occur a t temperatures of 470° 252O and 206.5O respectively. The compounds mentioned therefore can exist as stable substances a t their melting points. Of the two remaining maxima that which indicates the exist-ence of the hexasulphide and is found a t 1 8 9 O is flatbned in th 1080 THOMAS AND RULE THE POLYSULPHIDES same way as the previously discussed sodium pentasulphide maximum being but lo higher than the saturated solution line on the one hand and 6 O above the penta-hexa eutectic on the other. Potassium hexasulphide is therefore very considerably dis-sociated a t its melting point. I n the case of the tetrasulphide the curve shows no definite maximum but the existence of this compound the least stable towards heat of the polysulphides of potassium is indicated by the decided break observed in the curve at the point where the com-position of the fusion corresponds with that of this compound.The temperature a t which this break occurs is 145'5O. A t 18B0 when the mixture contains 71.5 per cent. of sulphur, the curve becomes horizontal and the addition of comparatively large amounts of sulphur does not affect the solidifying point. When a sufficiently larg0 amount of sulphur is added separation into two layers can be observed the one consisting of a saturated solution of sulphur in the hexasulphide solidifying a t 1 8 8 O whilst the other solidifies a t 1 1 7 * 5 O and is practically pure sulphur. As in the case of the sodium series this investigation furnished nos evidence whatever of the existence of intermediate compounds.A comparison of the results of this investigation with the work of Biltz and Wilke-Dorfurt (Zoc. cit.) on the polysulphides of rubidium and cmium is interesting and brings out more clearly the observation made by the authors in an earlier paper that, whilst the polysulphides of potassium rubidium and caesium resemble one another very closely the sodium compounds exhibit notable differences. Thus whilst the three elements first named all form hexasulphides and stable pentasulphides in the case of sodium no trace of the existence of a hexasulphide could be detected and the pentasulphide is relatively unstable. On the other hand sodium tetrasulphide is stable and the trisulphide unstable this order being reversed for the potassium rubidium, and caesium compounds.As regards their appearance stability, and eolidifying points all the disulphides appear to be similar in character. Towards heat these compounds show greater stability than any other of the polysulphides of the alkali metals. For purposes of comparison the solidifying points of the com-pounds and eutectic mixtures investigated by the authors together with the values obtained by Biltz and Wilke-Dorfurt in the rubidium and czsium series are given in table 111 OF THE ALKALI METALS. PSRT 111. 1081 Compound or eutectic. Disulphido ..................... Di-tri eutectic ............... Trisulphide ..................... Tri-tetra eutectic ............Tetrasulphide .................. Tetra-penta eutectic ......... Pent asulphide ............... Penta-hexa eutectic ......... Hexnsulp hide .................. Saturated solution ............ TABLE 111. Sodium. Potassium. 445.0' 471.0" 206.0 250.9 223.5 252.0 222.6 ?(amorphous) 275.0 > 145.0 24i.6 143.4 251.8 206.0 - 182.9 - 189.0 249.6 1SS.l R,ubidinm. Czsi urn. 420.0" 480.0" 200.0 205.5 213.0 217.0 148.5 151-0 > 160.0 > 160.0 159.5 159.6 232.0 210.0 189.8 178.0 201.0 186.0 184.6 172.5 The Molecular Weights of t h e Polysulphides of Sodium and Potassium in Alcoholic Solution. The view expressed by Rloxani (loc. cit.) that the character of the polysulphides of the alkali metals is better denoted by the general formula R,S than by the usually accepted formula R,S,, and that they are to be regarded as derivatives of a compound of the type R4S5 has already been referred to.The existence of compounds of the last-mentioned type has not been established by substantial experimental evidence but the theory was put forward as an attempt to bring the various intermediate conipounds which Rloxam clainied to have isolated of which the socalled ennea-sulphides are the best examples into line with those compounds, such as the tetra- and peiit?a-sulphides the existence of which was generally recognised. As regards the rubidium and ca%uiii compounds Biltz and Wilke-Dorfurt (loc. cit.) found no trace of any such intermediate compounds and the present authors conclude similarly in the cases of the sodium and potassium compounds.Intermediate com-pounds if they exist at all must be of an extremely unstable character. In order to obtain further evidence as to whether the single or the double formula is to be preferred for these compounds the molecular weights in alcoholic solution have been determined for such of them as can readily be obtained in the pure state. For this purpose the ebullioscopic method was employed. Carefully dried alcohol was used and the polysulphides employed had also remained for several days over phosphoric oxide. I n every case their purity was established by analysis. Several determinations were carried out with each substance, solutions of different concentrations being used. It' should be pointed out however that owiiig to the sparing solubilities of sodium sulphide and potassium pentasulphide it was not possibl 1082 THOMAS AND RULE THE POLYSULPHIDES in the cases of these substances to vary the concentration of the solutions used to any considerable extent.The following results were obtained : Sodium Disvlphide. Weight of substance Weightl of Number of taken. alcohol. experiment. Gram. Grams. E. M. W. 3 0,3684 3 1.67 0.184 73.97J 70.57 1 1 0.2396 32-20 0.098" 2 0.3441 31.24 0-173 74-49 Mean = 73.01. Na,S requires M.W. = 220.28. Na$ , , = 110.14. Sodium Te t ?.as u 1p12 ide. Weight of substance Number of t,aken. experiment. Grams. 1 0.2210 2 0.4278 3 0-6194 4 0.8988 5 1.1836 Weight of alcohol. Grams. E. M. w. 31.46 0*071° 115.76 30-91 30.13 30.65 0-290 118.05 31.07 0.358 123-07 Na,S8 requires M.W.= 348.56. Na,S 9 , = 174.28. Potassium Pentaszdphide. Weight of substanca Weight of Number of taken. alcohol. experiment. Gram. Grams. E. M.W. 30.69 E:;: ig;:::} Mean = 157.25 1 0,2761 2 0.3478 31.41 K,S:, requires M.W. = 467.1. K,S Y , = 238.55. I n the case of each polysulphide investigated the mean observed value for the molecular weight was found to be considerably lower than that required by even the simple formula R,S, and there can therefore be little doubt that under these conditions the double formula is quite out of the question. The uniformly low results obtained require however some further explanation. They may be accounted for in two ways. I n the first place alcoholysis of the solution may occur in accord-ance with the equation The result of such a reaction would be to reduce the weight of alcohol actually acting as solvent whilst a t the same time the R,S + 2C2H,*OH 2C2H,*OR + H,S + S - 1 OF THE AT;RAtI METALS.PART 111. 1083 number of molecules in solution would be increased. Both these results would tend to lower the observed molecular weight. That alcoholysis does occur to some extent is certain for in every determination the solution obtained was bright green. This phenomenon has been noted in an earlier paper (T, 1914 105, 188) and was there shown t o be due to the separation of minute particles of sulphur formed by the decomposition of the poly-sulphide by alcohol. Since however the amounts of hydrogen sulphide given off from such solutions even during prolonged boil-ing were so small as to be practically negligible the alcoholysis cannot be very considerable.In the case of sodium tetrasulphide if one supposes the above equation to represent the reaction and the sulphur produced to be insoluble the substance would have to be alcoholysed to the extent of more than 45 per cent. in order that the molecular weight might have the observed value. Taking all the known facts can-cerning the polysulphides of the alkali metals into consideration, this cannot be deemed probable. This opinion is strengthened by the behaviour of sodium tetrasulphide in aqueous solution. One would naturally expect this compound to be decomposed by water to a far greater extent than by alcohol yet Kuster and Heberlein (Zoc.cit.) have shown that in O.1N-solution at 25O the extent of the hydrolysis is only 11.8 per cent. On the other hand the low values obtained by the ebullioscopic method for the molecular weights may be due to the ionisation of the polysulphides in alcoholic solution. Supposing this to be the case and assuming that the alcoholysis is so slight as t o be negligible in order t o account for the difference between the observed and the theoretical values the degree of ionisation must be of the order of 25 per cent. Turner (Amer. Chem. J . 1908 40 588) has studied the ionisa-tion of alcoholic solutions of various halogen salts of the alkali metah in particular of potassium iodide. He employed the con-ductivity method and the results of his work lead to the con-clusion that a 0-1N-solution of this salt is ionised to the extent of 35 per cent.a t its boiling point'. Whilst therefore the evidence is insufficient t o permit of a definite conclusion being drawn the authors are of the opinion that although alcoholysis may in some measure give rise to the observed abnormalities its effect must be comparatively slight and the main cause must be looked f o r in the direction of ionisation of the polysulphide solutions. With the view of gaining further information on this point the conductivities of alcoholic solutions of the polysulphides will be investigated a t a later date 1084 THOMAS AND RULE THE POLYSULPHIDES Surn?tzury and Con el us i o n s. The systems sodium monosulphide-sulphur anci potassium monosulphide-sulphur have been investigated by a thermo-analytical method and in each case the existence has been estab-lished of a complete series of compounds to which the general formula R,S may be assigned where x is a whole number having the maximum value 5 in the sodium series and 6 in the case of the potassium series.No evidence whatever has been obtained of the existence of intermediate compounds such as the enneasulphide described by Bloxam. The thermal diagram for the potassium series closely resembles those obtained by Biltz and Wilke-DSrfurt for rubidium anci czsium. In the sodium series however the resemblance is much less striking. The sodium compounds differ in appearance and properties the maximum combining power of the metal is less anti strict parallelism can no longer be observed between this series and the potassium rubidium and czsium series as regards the com-parative stabilities of compounds of corresponding compositions.This is in accordance with the behaviour of the alkali metals in their compounds with other elements notably with iodine. I n connexion with the preparation of the materials necessary for the main research the rate at which sodium tetrasulphide and potassium pentasulphide lose sulphur when they are heated in a steady stream of hydrogen a t regularly increasing temperatures has been investigated. The results obtained lead to the coiiclusioii that the disulphides in each case are extremely stable compounds from which sulphur can only be removed with difficulty.The bearing of this on the constitution of the polysulphides is of some interest. This question has been discussed by many authors but the views put forward by Spring and Demarteau (Bull. SOC. chim. 1889 [iii] 1 11) and Kuster and Heberlein (Zeitsch. anorg. Chent. 1905 43 72) appear to be the most important. The first-mentioned authors point out that whilst towards inorganic substances the polysulphides behave in accord-ance with the general formula R,SS, their reactions with alkyl haloids are best explained by assuming that they possess the formula R,S,S, in which the first two sulphur atoms in the mole-cule play quite a different part from the remainder. Spring and Demarteau favour the latter formula considering the higher poly-sulphides merely to be solutions of sulphur in the disulphides.Kiister and Heberlein criticise this view on the ground that if it were true all polysulphides being essentially salts of hydroge OF THE ALKALI METALS. PART 111. 1085 disulphide should be hydrolysed in solution t o the same extent, f o r the hydrolysis could not be influenced by the mere physical solution of sulphur. They consider that the constitutions of the polysulphides and the polyiodides are analogous and since it is generally recognised that the latter compounds are best repre-sented by the formula RI,Iz they conclude that the constitution of the polysulphides is best expressed by the formula R,S,S,. With the criticism of Kiister and Heberlein the present authors are in agreement. Nevertheless their results certainly favour the view that in the polysulphide molecule two atoms of sulphur are in a different state of combination from the remainder.They suggest that the disulphides should be regarded as being derived from the form of hydrogen disulphide analogous to the tautomeric form of hydrogen peroxide. They would thus possess the formula R*S*S*R. From this substance the higher polysulphides are obtained not by solution of Thus the trisulphide would R.q:S R*S:S’ the tetrasulphide by I n this way the criticism sulphur but by further combination. be represented by the formula E?:S , R*S etc. of Kiister and Heberlein is obviated, whilst still retaining an explanation of the difference in behaviour of two sulphur atoms in the molecule from that of the remainder. Taken in conjunction with the work of Biltz and Wilke-Dorfurt, the results of the present investigation establish that the volatili-ties of the polysulphides of the alkali metals increase with rise in the atomic weight of the metal. This is in agreement with the well-known behaviour of the halogen salts of these metals. With the object of deciding between the simple formula R,S, and the doubled formula R,S, the molecular weights of several polysulphides in alcoholic solution have ,been determined by the ebullioscopic method. Under these conditions there can be no doubt that the simpler forinula is the correct one. The results obtained were however very considerably lower than even the simple formulz require. The authors attribute this to the ionisa-tion of the solutions and i t is calculated that assuming alcoholysis t o be negligible such ionisation takes place to the extent of about 25 per cent. In conclusioii the authors wish to express their thanks to Mr. F. Hughes who has rendered valuable assistance in the prepara-tion of some of the materials used in the work. INO~WANIC LABORATORIES, UNIVECSITY OF LIVERPOOL. Llteceiued S e p t e t i d m 7th 1t)l’i.