HADOW ON THE XI.-On the Composition of the Ptatinidcyanides. BY EDWARD ESQ, A. HADOW As the platinidcyanides are derived from the platinocyanides it will be well briefly to mention in the first place the rnde of forma-tion aiid properties of the latter. The platiiiocyanides are a remark-able set of salts belonging to the stable class of double cyanides exceeding the ferrocyanides in the force with which they retain the electro-negative metal and disguise it to ordinary tests. Coucen-tratecl and boiling nitric or hydrochloric-acids alone or mixed extract no platiuum from them; they are unaffected by digestion with peroxide of mercury -and coacentrated boiling sulphuric acid liberates cyanide of platinum with difficulty only. They have COMPOYlTlON OF THE PLATZNIDCYANIDES.107 the general formula MCy.PtCy or MPtCy2 the platinum in them existing in the state corresponding to protochloride of platinum so that before preparing these salts from a solution of bichloride of platinurn it must be reduced to the condition of protochloride either by sulphurous acid or by evaporation to dryness and expul- sion of half the chlorine by heat. The product in either case is treated with excess of cyanide of potassium until a clear solution is obtained which if previously warm and sufficiently concen- trated will deposit on cooling long prisms of the potassium-salt from which the other piatinocyanides may easily be derived by precipitating its solution with that of a salt of copper washing the precipitate suspending it in water and decomposiGg it by a current of sulphuretted hydrogen ;the solution which contains hydroplatinocyanic acid and after filtration from the sulphide of copper may be saturated with any required base.A more couve- nient way I find in many cases is to treat the protochloride of platinum directly with the required base and sufficient hydro- cyariic acid to form the double cyanide; the combination takes place easily if assisted by a gentle heat and it is only uecessary to crystallise a few times to get rid of the chloride present in the solution. These salts are remarkable for the great beauty and variety of the colours their crystals exhibit while their solutions are trans- parent and colourless. The platinocyanide of magnesium is perhaps the most beautiful of these salts; it forms by slow evspo- ration large and regular prisms of a deep red by transmitted light but viewed by reflected light the sides of the prisms exhibit a brilliant beetle-green and the extremities a deep blue or purple colour.The red salt gently warmed even under water becomes bright yellow wliich is also the colour of crystals deposited from a solution at a temperature of 160 OF.; heated to 212O F. the salt becomes quite white and again at a higher temperature bright yellow. These changes of colour correspond to successive losses of water the ordinary red salt contaiuing 7 equiv.; the yellow according to Wesclsky 6 equiv. At 212" F. the salt still retains 2 equiv. which are only expelled by a heat of between 300" and 400",when it becomes anhydrous and again yellow.If a portion of this yellow anhydrous salt be laid on the red salt in powder it will soon abstract water (from the latter) and a white layer will be formed between two yellow borders one of these yellow compouuds being anhydrous and the other contriining 6 equiv. of water. In 108 HADOW ON THE endeavouring to prepare this salt I on one occasion made use of a method given by Quadrat which consists in evaporating to dryness mixed solutions of sulphate of magnesia and platino-cyanide of potassium and digesting the dried mass in a mixture of alcohol aid ether; instead of which however I used only alcohol and that somewhat weak and by so doing obtained a solution which gave in addition to the ordinary dark red salt another set of crystals of a paler red and exhibiting a steel-blue lustre by reflected light instead of the usual emerald-green of the ordinayy salt.Analysis proved it to be a double platinocyanide of magnesium and potassium having the formula MgPtCy2.KPtCy + 2H0 + 5aq = 416.5. Calculated. Found. 2Pt = 47-54! 47.2 -4'7.1 4Cy = 24.97 - MgK = = 3-00 9.36 3.02 9.41 2HO = 5H0 = 4.32 10.81 4.40 expelled above 212' F. 10.97 expelled at 212. In this as in many other platinocyanides a portion of water is retained at 212'. This sdt csnnot be recrystallised except from a considerable excess of platiiiocyanide of magnesium and even when this is present sudden cooling or agitation of the saturated solution causes the separation of the potassium salt free from magnesium.These platinocyanides represented by the general formula MPtCy, may be transformed into thi salts termed platinidcya- nidep by the action of chlorine bromine nitric acid &c.-a set of compounds characterised in the crystalline form by a peculiar coppery lustre and to which the general formula M,Pt,Cy has hitherto been assigned. They are thus represented a3 differing from the platinocyanides by the addition of an atom of cyariogen to every two atoms of the latter salts. Mg,Pt,Cy = 2(MgPtCy2) .Cy. This somewhat improbable formula seemed yet to be strongly confirmed by the following facts. Gerhardt at first doubted it COMPOSITION OF THE PLATINJDCYANXDES.109 but afterwards confirmed it by an analysis of the potassium-salt. A determination of the platinum and potassium made by myself agreed so closely with the requirements of the formula as also strongly to confirm it. These salts act like the ferridcyanides in presence of a free alkali,-exerting a powerful bleaching action on cochineal ; they also liberate iodine from iodide of potassium corifirming the analogy between the two sets of salts. They evolve cyanogen when heated sufficiently leaving a residue of platinocyanide. They arc produced equally by the action of bromine ziitric acid and other oxidizing agents on the platiuo- cyanides showirig that no single one of these is ebsentially neces- sary for their formation. But then cn the other hand the following difficulties pre- sented themselves The platinum and basic metal remain in the same proportion to each other in the platinidcyanides as in the platitlocyanides from which they are derived the difference being merely an excess of cyanogen in the latter over that in the former.Whence is this cyanogen derived? In the case of the ferrid- cyanides the proportion between the iron and the basic metal in the ferrocyauide has been altered by the chlorine which abstracts a portion of the basic metal leaving the cyanogen behind to explain the excess found in the ferridcyanide But to account for that in the platinidcyanides it is necessary to assume the simultaneous removal by the action of chlorine on the platinocyanides of equal proportions of platinum and basic metal an explanation which appears very improbable to any one who has observed the great stability of the platinocyarzides and the remarkable ease with which the transformation by chlorine occurs.These salts it is true have an oxidizing effect in presence of potash in consequence of which by the addition of an oxidizable body they are reduced to platinocyanides; but then it ought to occur with simultaneous formation of free cyanide of potassium but no free cyanide can be detected as a product of such a reaction in any case. Moreover there exists another salt t4e HADOW ON THE ultimate product of the action of chlorine on the platinidcyanide of potassium (the salt chiefly examined) usually rcpreseiited by the formula PtCy,.KCl the formation of which from the plati- nidcyvanicle is equally difficult of explanation with that of the latter from the platinocyanide; and further still is it difficult to represent the mode of the reduction of the chlorinated salt to the state of platinidcyanide which however actually occtirs very readily under the action of reducing agents The determination of the proportions between the cpanogen and the platinum at once served to clear up most of these difficulties.The proportion mas found to be identically the same in the platinidcyanides as in the platinocyanides ; and the difference between the two sets of salts i4 composition was found to consist in the addition of a certain amount of the chlorine bromine or other elementary or compound salt-radical used for their forma- tion,-so that there exist not one set but many sets of platinid- cyanides requiring the prefix of cl'doro bromo &c, to distinguish them.Such being the case it appeared most probable that Knop's and Gerhardtjs salts had been true cgano-platinocyenides as the above formula indicates formed under some peculiar condi- tions perhaps in presence of excess of cyanide of potassium so that cyanagen instead of chlorine had really been supertdded to the original platinocyanide and that the salt under examination differed from theirs only in containing chlorine in place of cyanogen I was therefore nat a little surprised to find instead of 1 atom of chlorine to 2 atoms of platinum which a chlorine-compound analogous to this cyanogen- compound should contain not above a third of that quantity; accordingly even if the chlorine ever were truly replaced by cyanogen-which after several attempts I failed to effect-the formula M,Pt,Cp never could represent the composition of the cyano -platinocyanides the super- added cyanogen being too large in amount by two-thirds Considerable difficulty was found in determining the amount of chlorine in the salts on account of their high atomic weight.Repeated corn bustions of the chloro-platinocyanide of potassium with nitrate of potash and carbonate of soda gave a proportion of 1 equiv. of chloi-ine to between 6 and 7 equiv. of platinum. This analytical method proving unsatisfactory from not giving a probable formula a spthetical method was tried with better success; it was found that a solution of pltltina-cyanide acidulated with hydrochloric acid rapidly and perfectly COXPOSITIOX OF THE PLATISIDCYARIDES.decolorises a solution of permanganate of potash until the platino- cyanide has taken up the rnaxiinum amount of chlorine and become converted into the salt PtCy,KC1 tthich has been before mentioned as the ultimate product of the action of chlorine on the cldoro-platinocyanides and which might be termed the percliloro-platinocganide of potassium. As no method could be devised to indicate when the firat stage of the change namely from platinocyanitie into chloro-platinocyanide is complete it was not possible to determine the amoirtit of chlorine necessary for its foi-mation in this direct way; but starting from the chloro-platinocyanide it was easy to determine by means of permanganate of potash how much additiond chlorine was necessary to convert it into the percliloro-platinocyanide the true composition of which had been determined and found to have been correctly represented in its elementary proportions by the formula hitherto given.It was thus found that while 6 equiv. of the platinocyanide containing 6 equiv. of platinum required 6 of chlorine for the formation of the perchloro-platinocyanide,a quantity of the chloro-platinocyanide contaiiiing the same amount of platinurn required only 5 equiv. of chlorine to convert it into the same salt proving that 1 equiv. already existed in this quantity of the salt arid therefore that its true composition in the anhydrous state is 6(KPtCy,) C1.This composition explains its oxidizing power in alkaline solutions and why no free cyanide can be detected when its force has been spent on an oxidisable body an alkaline chloride and not a cyanide being formed The evolution of cyanogen and its reduction to platinocyanide when heated is due to the chlorine partly entering the salt and dis-placing cyanogen; (a certain arnoiint of chloride of ammonium is however likewise formed at the time when moisture is present) fj(KPtCy,)Cl = 5(KPtCy,) + KC1 + PtCy + Cy. The salt after this gives a sornevhat turbid solution from PtCy which has been set free. The formation of the salt from the platinocyanide arid its ready conversion into the perchloro-platinacyarlide and the reduction of the latter into the chloro- platinocymide and platinocyanide successively by reducing I12 HADOW ON THE agents is due to the successive additions of chlorine in one case and successive abstractions of it in the other without any forma- tion or separation of cyanide of platinum or of chloride of potassium in either case- The chloro-platii~ocyanide of potassium as ordinarily dried in a hot-water oven loses the percentage of water given by Gerhardt in his analysis but still retailis 3 equiv.which tQgether with the chlorine overlooked correspond pretty nearly in weight with the excess of cyanogen supposed to exist in the salt and which thus appears to have been determined by difference only.The compo-sition of the crystallised salt is The direct determinations of chlorine as before stated were unsatisfactory inasmuch as they showed on the average a deficiency of about 0.20 per cent. The salt has a slight tendency to lose water of crystallisation in dry weatter. The water was therefore determined in two samples one dried spontaneously in the air the other by strong pressure in bibulous paper immediately after removal from its solution as the two extremes to furuish a oorrect mean. Air-dried. Dried by Aver-Calculated Pressure. age. for 18HO 18H0 Loss per cent at 212" = 11.34 20.9 12-34 11.53 11-87 3H0 1) 400' = 1.94 1-97 -1'96 1.98 Composition of salt dried at 212' :-6Pt = 49.27 Found.50.07 48-7 Average 49*34 Ca+lculated. 49.39 6K 12Cy = = 19.68 25.83 - 19.92 L 19*8Q 25.83 19.47 25.95 C1 = 2-76 2.82 2.79 2.79 2.95 3HO = 2.19 2.23 _I 2.21 2.24 99-97 100~00 The quantities of the salt taken for the various determinations varied from 7 to 13 grs. The platinum and potassium were COIMPOSITION OF TRE PLATINIDCYANXDES. 113 obtained by evaporation with pure concentrated sulphuric acid followed by gentle ignition. The cyanogen was determined by combustion of 10.6 grs. with soda-lime. The above formula represents the empirical composition of the chloro-plstinocy anide of potassium. Its true ration a1 cornposi tioil was cliscovcrecl on precipitating its solution with a solution of zinc when it was found to have resolved itself into a platinocyanide which mas precipitated in combination with zinc and a percliloro- platinocy anide which remained in solution ; and accordingly on mixing a solutioii of the perchloro-platinocyanide a salt which forms large colourless crystals with another of the nearly colour- less platinocyanide the mixture if sufficiently concerritrated will deposit immediately an abundant crop of copper-coloured crystals of chloro-platinocyanide + KPtc*v2c1 It appears in fact to be a double salt of the two.The solution accordingly of the so-called platinidcyanides exhibits the reactions of a mixture of its component salts and it is thus impossible by any direct tests to ascertain whether either is in excess or not; this p0in.t can only be determined by careful crystallisation.By this latter method the formula of the chloro- platinocyanides was further confirmed ;the two component salts were mixed in various proportions and the compound salt was crys- tallised out when one or other of its components would be found in excess unless the right proportions had been taken. The results obtained in this way were not absolutely satisfactory in conse- quence of partial decomposition during repeated crystallisation ; but they served to counterbalance the determinations of chlorine by analysis the errors in the numbers obtained being in the opposite direction and indiccting tho proportion of 1 equiv. of chlorine to between 5 and 6 equiv. of platinum. It was found that from whatever solution the chloro-platinocyanide was crystal-lisecl it always contained the sanie amount of chlorine so that there appears to be no double salt intermediate between tbe chloro-platinocyanide and perchloro-platiriocyanide; and accord- ingly if a pixrely cyanogen-platinocyanide existed it may be concluded from analogy that it mould have the composition 5(PtR@y,) + PtBCy and not (PtKCy2 -+ PtKCy,) -vvhich the hitherto accepted formula requires.The bromo- and VOL. XIII. 1 114 HADOW ON THE COMP9SITION OF THE PLATINIDCYANIDES. nitro-platinidcyanides so closely resemble the corresponding chlorinated salts as hitherto to have been confourided with them under the general term of platinidcpanides and as their appearance and facility of formation are so much alike there is every reason to conclude that they have an analogous composition being in fact double salts of a platitlocyanide and a corresponding perbronio- and pernitro-platinidcyanide which like the highest chlorine com- pounds are colourless.In the caqe of nitric acid NO must represent the chlorine and bromine of the corresponding salts. There appear to be no analogous iodine coinponnds sirice any one of these salts treated with iodide of potassium immediately liberates iodine. But there are analogous compounds containing the salt radical SO in place of chlorine bromine &c. and doubtless similar compounds with the salt-radicals of other acids; the existence of the sulphizric compouncl was inferred in the first instance from finding that a solution of a platinocyanide acidulated with sulphuric acid rapidly decolorises a perrnan-ganate.The crystallised oxysulpho-platinocyanideof potassium is readily obtained by adding a little peroxide of lead to a saturated solution of platinocyanide of potassium acidulated with sulphuric acid ;-each particle of peroxide dropped in becomes iiistantly enveloped in a moss-like growing mass of the copper-coloured salt. From a few experiments recently made it would appear that these remarkable cornpounds may also be regarded as salts in which the platinocyanide MPtCy plays the part of a metal since in the sulphuric compoundt the sulphuric acid readily admits of replacement by another acid by doizble decomposition with a baryta salt; aud accordirigly it Eeems to be not unlikely that there are oxygen-compounds capable of acting the part of bases and combining directly with acids to form the per-compounds which by combination with the platinocyanides form the platinid- cpdnides.It is on account of the feeble affinity that exists between the platinocyanide (as a compound metal) and the salt- radicai thTt reducing agents so readily decompose them into the platinocyanide and the corresponding acid just as they would the corresponding salts of gold the gold being supposed to represent the platinocyanide in its feebly electro-positive character. If sulphurctted hydrogen be passed into their solutions which are neutral they immediately become stronglyv acid. BUCHTON ON THE STIBETHYLS AXD STIBXETHYLS.115 The most distinctive reactions between the platinocyanides and perchloro-platinocyanides are that the former give au abundant small-blue precipitate with nitrate of suboxideof rrmcury and a $f~occuZent blue precipitate with sdts of copper. The latter give a white with nitrate of suboxide of mercury and a fine sandy blue with salts of copper ;they liberate iodine from iodides ; aitd if dropped in the solid form into a concentrated sdution of platiuocpanide of potassium they are immediately coated with copper-coloured needles The platinidcyanides in solution exhibit both sets of reactions aid accordingly cannot be distinguished until crystallised ont when their dark coppery lustre at once characterizes them. The neatest way of forming the chlorine and bromine salts at once in a state of purity is to take a concentrated solution of a platino-cyanide measure off a sixth-part and cotivert it 5y an excess of chlorine or bromine into the perchloro- or perbromo-platinocgznide expelling the excess of the cklorine and bromine subsequently by a gentle heat and then adding the remaining five-sixths of platinocyanide; in a very short time the double salt will form abundautly in a state of perfect purity.