PLATINUM METALS AS CHLORIDES. (Reactions of grm.-litre Solutions). minute 3. Saturation .. .. cipitate. ation. 2. H2S at 18° în one Immediate brown pre- No precipitate. Decoloration without No precipitate. precipitate. with Blackish brown separ- Blackish brown sepa- Blackish brown pre- Slight brownish turH2S at 80° in ration. one minute 4. Ethyl mercaptan Immediate yellow pre- Yellow solution, Immediate dark brown Unchanged. I : 1000 6. NH3 solution, on warming Platinum, PtC14. Pale yellow. No precipitate. Iridium, IrCl4. Dark brown. Ruthenium, RuC14. Dark brown. Osmium, OsC14. Golden yellow. No precipitate. Light yellow precipitate, slowly. Slow decoloration without precipitate. Brown precipitate, slowly. No decoloration nor precipitate. Red-brown precipitate, slowly. Slow decoloration. 7. Saturation with ammonium chlor No precipitate. No precipitate. Yellow precipitate. Black precipitate. tion. Brown precipitate. cipitate. Red precipitate. Blue precipitate of di- Yellow solution; smell Light solution; smell oxide. of ozone. of per-osmium acid. Unchanged. Unchanged. Unchanged. Unchanged. Blackening owing to Yellow coloration. separation of metal. Blackening owing to Yellow coloration. separation of metal. Yellow coloration. Unchanged. Unchanged. Unchanged. Unchanged. Unchanged. Reddish precipitate. Unchanged. Brownish precipitate. Unchanged. Unchanged. Immediate yellow pre- Unchanged. cipitate. 14. Luteo-cobalt chlor- Unchanged. ide in HCl solu tion 89 behind. It is_then_filtered into a second flask in which water is placed to the amount of about 15 per cent of the alcohol in the filtrate. The liquid as it flows in solidifies at once to a thick crystalline paste from which the yellow needles of the hydrated pure compound Na2PtCl6+6H2O can easily be isolated by means of a suction apparatus. The alcohol still remaining in them rapidly volatilises in the air. (NOTE.-While cold alcohol has no action on pure sodium platinum chloride it energetically decomposes the salt if foreign platinum metals are present, aldehyde and acetic acid being formed. Black insoluble substances are formed, especially on warming. A rule must therefore always be made of acidifying with hydrochloric acid alcoholic platinum chloride solutions before evaporation). The small quantities of double salt contained in the mother-liquor can easily be obtained by evaporation (after acidification with hydrochloric acid). The reduction to metal is usually performed by heating the double salt in a current of hydrogen, and extracting the sodium chloride. It can also be effected in the wet way by electrolysis, or by hydrazine hydrochloride, if the latter is free from impurities. The reaction also takes place (this is at variance with the statements in literature, "Gmelin-Kraut,” 1914, [7], V., 42) even in the presence of a great excess of free hydrogen chloride. Special attention may here be called to the fact that the difficultly soluble precipitates of platinum ammonium chloride, made by igniting on platinum, are well suited for this reduction in the wet way. The yellow powder is suspended in ten times its weight of water made slightly acid with hydrochloric acid, and the mixture is warmed in a dish with gradual addition of a solution of hydrazine hydrochloride. If the liquid is stirred from time to time the reaction begins with the complete solution of the double salt to a clear brownish liquid (with transitory separation of a flocculent precipitate of lower oxide). In a later phase of the reaction | gas is evolved, and the reduced platinum separates from the liquid in a more or less compact state on the side of the dish, while the supernatant liquid gets gradually paler. By regulating the addition of the hydrazine solution and the rise of temperature the velocity of the reaction can be arbitrarily altered. In circumstances which will be described later the reduced platinum may contain appreciable quantities of chloride or lower chloride, which are decomposed on ignition. Small amounts of iridium which might be present in the platinum ammonium chloride remain more completely dissolved with any platinum in the liquid, the more acid the liquid was in this reduction process. II. THE ANALYTICAL Detection of tHE PLATINUM 1. Qualitative Reactions of the Grm.-litre Solutions. From what has been said above it will be seen that there is no prospect of determining the very small amounts of other constituents in platinum unless the large excess of the metal itself is first removed. The next task is then to reduce these constituents to a small volume, and to separate them from one another after suitable concentration. In the case of the platinum metals, even more than with the other elements, the quantitative separation depends upon an accurate knowledge of the qualitative reactions of their solutions. The scientific treatment of the platinum metals in literature only occasionally gives an opportunity of obtaining a comparative view of the differences in question. For this reason it seemed to us appropriate to observe the most important reactions of the platinum metals in solutions containing equal quantities by weight, and to give the following summary of them. Comparable grm.-litre solutions were chosen, which contain a grm. of the metal as chloride in a litre. Their low concentration is favourable for the accurate observation of the differences in the reactions. Such solutions | CHEMICAL NEWS, Aug. 20, 1915 have also been found very useful for comparative observations in the analysis of other heavy metals. The liquids used contain the metals in the state of chlorination (or oxidation) which mostly occurs in analysis. Palladium is dissolved as dichloride, rhodium as trichloride, platinum, iridium, ruthenium, and osmium as tetrachlorides with hydrochloric acid. The grm.-litre solutions were prepared as follows:(1). Platinum.-One grm. of the pure metal was converted in the usual way into H2PtCl6. The residue, free from nitric acid, obtained by evaporation was dissolved in water to make 1 litre. (2). Palladium.-The quantity of crystallised palladous chloride (2.002 grms. PdCl2+2H2O) corresponding to I grm. of metal was dissolved in water to make 1 litre. (3). Rhodium.-The quantity of ammonium rhodium chloride (3.8561 grms. (NH4)3RhCl6+1*5H2O) corresponding to I grm. of metal was repeatedly evaporated with dilute aqua regia, and finally with hydrochloric acid; the residue, free from ammonium chloride, was made up to I litre with water, after the addition of 50 cc. of concentrated hydrochloric acid. (NOTE. The ammonium rhodium chloride was prepared by saturating a solution of sodium rhodium chloride with ammonium chloride and slowly crystallising). (4). Iridium.-The quantity of ammonium iridium chloride (2·2882 grms. (NH4)2IrCl6) corresponding to I grm. of metal was repeatedly evaporated with dilute aqua regia, and finally with hydrochloric acid: after the addition of 50 cc. of concentrated hydrochloric acid to the residue, which was free from ammonium chloride, it was made up to 1 litre with water. (5). Ruthenium.-The quantity of per-ruthenic acid (16293 grms. RuO4) corresponding to 1 grm. of metal was dissolved in 50 cc. of concentrated hydrochloric acid and the solution made up to 1 litre with water. (NOTE. In the evaporation of RuO, described by Gutbier and Rauschoff Zeit. Anorg. Chem., 1905, xlv., 260), the weighing of any desired quantity can be performed in stoppered glass vessels, they can be opened under concentrated hydrochloric acid, and diluted to a suitable volume according to the weight). (6). Osmium.-The quantity of ammonium osmium chloride (2.3031 grms. (NH4)2OsC16) corresponding to I grm. of metal was treated with 100 cc. of concentrated hydrochloric acid and made up to 1 litre with water. 2. Colorimetric. According to this summary each platinum metal gives some reactions which are different from those of the other metals, and which can be used in their presence for qualitative detection. The coloration of the chloride solutions is characteristic. The solutions of platinum and osmium are yellow, of palladium brownish yellow, of rhodium red, of iridium and ruthenium dark brown. The colorations of the grm.. litre solutions are, however, not quite constant, but alter somewhat with the age of the solutions owing to increasing hydrolysis. A colorimetric comparison with the solutions of the chlorides of other heavy metals is interesting. It is, however, of value only if the hydrogen chloride concentrations in the different solutions as well as the concentrations of the metallic chloride in question are nearly the same and are also high, as otherwise hydrolysis causes great discrepancies. Decigrm.-litre solutions are then prepared from the grm.-litre solutions by diluting with ten times the amount of 37 per cent (fuming) hydrochloric acid; these solutions contain the metallic chlorides in concentrated hydrochloric acid. Such solutions can be kept for colorimetric purposes. If solutions of the following are placed in similar test-tubes-Copper (as CuCl2), iron (as FeCl3), nickel (as NiCl2), cobalt (as CoCl2), and gold (as AuC13), and if they are arranged (independently of the shade of colour), according to the obvious intensity of the coloration the following series is obtained :Platinum, pale yellow; nickel, pale yellow; gold, pale CHEMICAL NEWS, } Vapour Pressure of Concentrated Sugar Solutions. Aug. 20, 1915 The depth of colour of platinum chloride is the least, and hardly half that of gold chloride. 91 yellow; palladium, pale yellow; rhodium, rose-red; | between the vapour pressures of the solution and the cobalt, sky-blue; copper, golden yellow; iron, golden solvent to be obtained without relatively large errors. yellow; osmium, golden yellow; ruthenium, brown; In 1912 Dr. Perman and Mr. T. W. Price read a paper iridium, brown. There can be no doubt about the position before this Society, in which they gave the results of a of the last members, but there may be a difference of large number of experiments made by an air-current opinion about the places of rhodium and cobalt, according method. Their data showed variations in the relative to the subjective impression. lowering of vapour pressure which were unexpected from theoretical considerations, and, in consequence, Prof. A. W. Porter suggested to me that it would be useful to attempt measurements by a different method in order to confirm or modify their conclusions. Preliminary results obtained with a very concentrated sugar solution of unknown strength did not agree with Dr. Perman's values, and the following paper contains an account of further observations made on solutions of saccharose. The measurements are by no means as accurate as one could wish, but they are probably the best which have been obtained so far with highly concentrated solutions. The almost equal colour intensity of the copper, iron, and osmium solutions is more than ten times as great as that of the platinum and nickel solutions; this may be seen by comparing them with centigrm.-litre solutions of iron, copper, and osmium, specially prepared by diluting with concentrated hydrochloric acid. It will be seen that they are rather more strongly coloured, though the shade is essentially the same. According to this superficial discussion the colorimetric determination of the individual platinum metals as chlorides is not hopeless, if concentrated hydrochloric acid is used as solvent, which Hüttner (Zeit. Anorg. Chem., 1914, lxxxvi., 341) found convenient in his experiments on cobalt, iron, and copper. The rose-red coloration of rhodium chloride solution is specially characteristic, but it is very much influenced by hydrolysis. The grm.-litre solution containing more water used in the summary is more yellowish red, owing to the admixture of a yellow modification of the chloride. With iridium and ruthenium chlorides the effect of hydrolysis upon the colour of the solutions is also appreciable. In time aqueous solutions of iridium chloride deposit blue dioxide. Osmium chloride is also similarly decomposed bydrolytically, giving a dark hydroxide. Owing to the great colouring power of the brown iridium tetrachloride it is easy to estimate the actual amount of iridium in a technical platinum colorimetrically from the coloration of its hydrochloric acid chloride solution (just as the cobalt in a technical cobalt can be determined colorimetrically by Hüttner's method). It can easily be done with a centigrm. of metal, which is dissolved in aqua regia, and then the residue after evaporation is freed from nitric acid and diluted to 100 cc. with concentrated hydrochloric acid. The decigrm.-litre solutions of pure platinum and pure iridium, also prepared with concentrated hydrochloric acid, are then put together until the colour of the mixture is the same as that of the solution under examination. The proportions by volume of the solutions gives approximately the proportion by weight of iridium to platinum in the technical metal. The coloration is only slightly affected by very small quantities of other impurities. The colorimetric determination of rhodium by means of the rose-red chloride solution is, on the other hand, very dependent upon the impurities; it can be employed only when the rhodium solutions are comparatively pure. (To be continued). Method. After a critical examination of the air-current method it seemed difficult to point to any definite source of error other than that possibly introduced by the use of the perfect gas equation in the calculation of the vapour pressure from the experimental data (for example, see a letter in Nature, March 11, 1915, from Lord Berkeley). This being the case, it seemed desirable to use a direct method which should give the pressures without the necessity of any such calculations. Moreover, the point of immediate interest was the variation of the Babo constant (i.e., the lowering of the vapour pressure relative to that of the solvent) with temperature, and the direct static method lends itself very easily to the determination of the vapour pressure at different temperatures. The more serious objections usually put forward against the method are (see, for example, Lincoln and Klein, Journ. Phys. Chem., 1907, p. 11):— (a) That with so small a quantity of vapour any impurity would produce a large error. (b) That, since evaporation takes place from the upper layers only, they become too concentrated and the measurevapour pressure is too small. With regard to (a) the most serious impurity is dissolved air. The removal of this air under the condition that the concentration of the solution must be known constitutes the main difficulty of the experiment. The method used for removing the air would, however, also remove any traces of other volatile substances, should they ever be present. As for (b), even at the highest temperatures only a small quantity of vapour is evolved, and by stirring the solution electromagnetically the objection is sufficiently overcome. Apart from these possible sources of error there is the disadvantage that the preparation of the apparatus for each solution is somewhat laborious, so that the method does not lend itself easily to making experiments with many different solutions. It is necessary, then, to introduce an air-free solution of known concentration into a vessel in which its vapour pressure can be measured barometrically. After many trials the arrangements shown in Figs. 1 and 2 were THE VAPOUR PRESSURE OF CONCENTRATED adopted. The solution was prepared in the vessel depicted SUGAR SOLUTIONS.* in Fig. 1, and from this vessel it was transferred to the main apparatus (Fig. 2), where its vapour pressure was measured. The part of this latter to the left of the constriction A is required only during the filling process, and when that is completed it is sealed off. The Preparation of the Solution.-The filler consists of a glass bulb blown at the end of a piece of wide glass tubing, about 10 inches long, which serves as a condenser. The tubes C and D are fused to the end of this condenserc having at its upper end the inner portion of the groundglass joint through which the filler is ultimately joined to the main apparatus at c' (Fig. 2). In the preparation of a solution the condenser tube was first cut in two at E, and the two parts thorougly cleaned with chromic acid. The acid was rinsed out with water, then ammonia, and finally 9: Vapour Pressure of Concentrated Sugar Solutions. this was removed with distilled water. The two parts | The bulb was next connected, at D, through suitable drying tubes, to the air-pump, and was carefully warmed. The taps C and D were shielded as far as possible from the direct heat with wet blotting-paper. The whole apparatus being exhausted of air, evaporation was continued until the desired concentration was almost reached. Finally, to remove all traces of air the solution was heated to about 100° C. with the tap D closed. After cooling a little, D was opened, and the air which had been evolved was pumped out. This process was repeated several times. The bulb was then allowed to cool, and was weighed. The mass of the bulb, together with a known mass of crystals, having been determined previously, the mass of the water, and therefore the strength of the solution, was known. The operations just described needed the utmost care in the manipulation both of the flame and the air-pump, in order that the tendency of the solution to boil by bumping might be kept within reasonable limits. The Filling Process.-Referring to Fig. 2, the apparatus consists of a bulb F, into which the solution is to be placed, and the arrangement to the right of F, by which the pressure is measured. The wide glass tube H contains the mercury whose meniscus is in contact with the vapour; close to it is a similar tube K, which is just long enough to project over the top of the water-bath in which F, H, and K are immersed. This tube is connected, through an air trap, by flexible rubber tubing to the barometer M. The level of the mercury through the whole system is controlled by a mercury reservoir attached to N. The vapour pressure is sufficient to make the mercury column continuous, and the difference in the level of the mercury at H and м gives the required vapour pressure (subject to a small temperature correction to be discussed later). During the measure. ments the solution in F is stirred by means of a sealed glass tube containing short rods of soft iron embedded in glasswool. These iron rods are dragged up and down by a circular electro-magnet just outside the bulb. In the process of cleaning and filling, F, H, K, and the other rigid parts of the apparatus were firmly fixed to a wooden frame to which the filler tube, sealed on to the rest by the ground joint cc', was also attached. The apparatus was completely exhausted of air through the tube L, and the mercury was allowed to flow into HK from N, thus preventing access of vapour to the more remote portions of the apparatus. The tap in c was then opened, and the solution, driven by gravity and its own vapour pressure, flowed down slowly into F. When sufficient solution had entered the bulb it was sealed off at the constriction A. This was always one of the critical stages in the experiment; but success was rendered quite certain by softening the constriction with a luminous flame before the apparatus was exhausted. This removed any stress that might have been imposed by the attachment of the comparatively heavy filler bulb, and by the subsequent clamping down on to the supporting frame. In the earlier experiments, as a final precaution, a tube was sealed on at B in a plane perpendicular to that of the rest of the apparatus. This tube was connected to the CHEMICAL NEWS, pump, and the solution was warmed and stirred in F to remove any residual traces of air, the water vapour evolved being condensed in a bulb surrounded by a freezing mixture. This water was afterwards distilled back into F; but a little was inevitably drawn into the drying tubes, and the uncertainty thereby introduced into the concentration was considered to overbalance any advantage gained by the possible removal of last traces of air, of whose presence, moreover, there was never any certain evidence. The Measurements. fully, the apparatus was removed from its supporting When the tube F had been filled and sealed up successframe, and was held by an iron clamp so that the tubes A, H, and X were immersed in a water-bath. This con E FIG. 1. sisted of a large copper tank with a plate-glass window, through which the level of the mercury in the tube H could be observed. The bath was heated by two independent gas-burners, one arranged so that its flame could be adjusted very finely in order to keep the temperature constant. The stirring was carried out most efficiently by a single propellor fixed in a corner of the tank and driven at a high speed by an electric motor. A metre scale was set up in a vertical plane midway between H and M, and the divisions on this scale corresponding to the two mercury levels were read off with two cathetometers used as reading telescopes. The reservoir attached to N was manipulated each time so as to bring the mercury meniscus to about the same level in H. The temperatures were determined with a limited-range thermometer (50° C. to 105° C.), divided into tenths and read with a small magnifying lens. The thermometer was afterwards standardised by direct comparison with another limited-range thermometer whose errors were well known. Three other thermometers were necessary-(i.) to deter mine the average temperature of the emergent stem, (ii.) to measure the temperature in the region between к and м (the thermometer was suspended between the limbs of the U-tube of the air trap), (iii.) to measure the temperature of the barometer tube M. A screen was set up to protect the mercury columns outside the bath from direct exposure to heat. The readings were, at first, corrected for the curvature of the meniscus in H and in м, but as the correction is at the very extreme limit of the possible accuracy and is made very uncertain by the high temperature of the mercury in the tube H, it was afterwards omitted altogether. A small correction is, however, necessary to allow for the difference of temperature of the mercury columns connecting F and |