(C) A remarkable example of the complex compounds of platinum was observed by Schützenberger (1868). He showed that finely-divided platinum in the presence of chlorine and carbonic oxide at 250°-300° gives phosgene and a volatile compound containing platinum. The same substance is formed by the action of carbonic oxide on platinous chloride. It decomposes with an explosion in contact with water. Carbon tetrachloride dissolves a portion of this substance, and on evaporation gives crystals of 2PtCl2,3CO, whilst the compound PtCl2,2CO remains undissolved. When fused and sublimed it gives yellow needles of PtCl2,CO, and in the presence of an excess of carbonic oxide PtCl2,2CO is formed. These compounds are fusible (the first at 250°, the second at 142°, and the third at 195°). In this case (as in the double cyanides) combination takes place, because both carbonic oxide and platinous chloride are unsaturated compounds capable of further combination. The carbon tetrachloride solution absorbs NH3 and gives PtCl2,CO,2NH3, and PtCl2,2CO,2NH3, and these substances are analogous (Foerster, Zeisel, Jörgensen) to similar compounds containing complex amines (for instance, pyridine, C5H5N), instead of NH3, and ethylene, &c., instead of CO, so that here we have a whole series of complex platino-compounds. The compound PtCl2CO dissolves in hydrochloric acid without change, and the solution disengages all the carbonic oxide when KCN is added to it, which shows that those forces which bind 2 molecules of KCN to PtCl2 can also bind the molecule CO, or 2 molecules of CO. When the hydrochloric acid solution of PtCl2CO is mixed with a solution of sodium acetate or acetic acid, it gives a precipitate of PtOCO, i.e. the Cl2 is replaced by oxygen (probably because the acetate is decomposed by water). This oxide, PtOCO, splits up into Pt + CO2 at 350°. PtSCO is obtained by the action of sulphuretted hydrogen upon PtCl2CO. All this leads to the conclusion that the group PtCO is able to assimilate X2 = Cl2, S, O, &c. (Mylius, Foerster, 1891). Pullinger (1891), by igniting spongy platinum at 250°, first in a stream of chlorine, and then in a stream of carbonic oxide, obtained (besides volatile products) a non-volatile yellow substance which remained unchanged in air and disengaged chlorine and phosgene gas when ignited; its composition was PtCl6(CO)2, which apparently proves it to be a compound of PtCl2 and 2COCl2, as PtCl2 is able to combine with oxychlorides, and forms somewhat stable compounds.

(D) The faculty of platinous chloride for forming stable compounds with divers substances shows itself in the formation of the compound PtCl2,PCl3 by the action of phosphorus pentachloride at 250° on platinum powder (Pd reacts in a similar manner, according to Fink, 1892). The product contains both phosphorus pentachloride and platinum, whilst the presence of PtCl2 is shown in the fact that the action of water produces chlorplatino-phosphorous acid, PtCl2P(OH)3.

(E) After the cyanides, the double salts of platinum formed by sulphurous acid are most distinguished for their stability and characteristic properties. This is all the more instructive, as sulphurous acid is only feebly energetic, and, moreover, in these, as in all its compounds, it exhibits a dual reaction. The salts of sulphurous acid, R2SO3, either react as salts of a feeble bibasic acid, where the group SO3 presents itself as bivalent, and consequently equal to X2, or else they react after the manner of salts of a monobasic acid containing the same residue, RSO3, as occurs in the salts of sulphuric acid. In sulphurous acid this residue is combined with hydrogen, H(SO3H), whilst in sulphuric acid it is united with the aqueous residue (hydroxyl), OH(SO3H). These two forms of action of the sulphites appear in their reactions with the platinum salts—that is to say, salts of both kinds are formed, and they both correspond with the type PtH2X4. The one series of salts contain PtH2(SO3)2, and their reactions are due to the bivalent residue of sulphurous acid, which replaces X2. The others, which have the composition PtR2(SO3H)4, contain sulphoxyl. The latter salts will evidently react like acids; they are formed simultaneously with the salts of the first kind, and pass into them. These salts are obtained either by directly dissolving platinous oxide in water containing sulphurous acid, or by passing sulphurous anhydride into a solution of platinous chloride in hydrochloric acid. If a solution of platinous chloride or platinous oxide in sulphurous acid be saturated with sodium carbonate, it forms a white, sparingly soluble precipitate containing PtNa2(SO3Na)4,7H2O. If this precipitate be dissolved in a small quantity of hydrochloric acid and left to evaporate at the ordinary temperature, it deposits a salt of the other type, PtNa2(SO3)2,H2O, in the form of a yellow powder, which is sparingly soluble in water. The potassium salt analogous to the first salt, PtK2(SO3K)4,2H2O, is precipitated by passing sulphurous anhydride into a solution of potassium sulphite in which platinous oxide is suspended. A similar salt is known for ammonium, and with hydrochloric acid it gives a salt of the second kind, Pt(NH4)2(SO3)2,H2O. If ammonio-chloride of platinum be added to an aqueous solution of sulphurous anhydride, it is first deoxidised, and chlorine is evolved, forming a salt of the type PtX2; a double decomposition then takes place with the ammonium sulphite, and a salt of the composition Pt(NH4)2Cl3(SO3H) is formed (in a desiccator). The acid character of this substance is explained by the fact that it contains the elements SO3H—sulphoxyl, with the hydrogen not yet displaced by a metal. On saturating a solution of this acid with potassium carbonate it gives orange-coloured crystals of a potassium salt of the composition Pt(NH4)2Cl3(SO3K). Here it is evident that an equivalent of chlorine in Pt(NH4)2Cl4 is replaced by the univalent residue of sulphurous acid. Among these salts, that of the composition Pt(NH4)Cl2(SO3H)2,H2O is very readily formed, and crystallises in well-formed colourless crystals; it is obtained by dissolving ammonium platinosochloride, Pt(NH4)2Cl4, in an aqueous solution of sulphurous acid. The difficulty with which sulphurous anhydride and platinum are separated from these salts indicates the same basic character in these compounds as is seen in the double cyanides of platinum. In their passage into a complex salt, the metal platinum and the group SO2 modify their relations (compared with those of PtX2 or SO2X2), just as the chlorine in the salts KClO, KClO3, and KClO4 is modified in its relations as compared with hydrochloric acid or potassium chloride.

(F) No less characteristic are the platinonitrites formed by platinous oxide. They correspond with nitrous acid, whose salts, RNO2, contain the univalent radicle, NO2, which is capable of replacing chlorine, and therefore the salts of this kind should form a common type PtR2(NO2)4, and such a salt of potassium has actually been obtained by mixing a solution of potassium platinosochloride with a solution of potassium nitrite, when the liquid becomes colourless, especially if it be heated, which indicates the change in the chemical distribution of the elements. As the liquid decolorises it gradually deposits sparingly soluble, colourless prisms of the potassium salt K2Pt(NO2)4, which does not contain any water. With silver nitrate a solution of this salt gives a precipitate of silver platinonitrite, PtAg2(NO2)4. The silver of this salt may be replaced by other metals by means of double decomposition with metallic chlorides. The sparingly soluble barium salt, when treated with an equivalent quantity of sulphuric acid, gives a soluble acid, which separates, under the receiver of an air-pump, in red crystals; this acid has the composition PtH2(NO2)4. To the potassium salt, K2Pt(NO2)4, there correspond (Vèzes, 1892) K2Pt(NO2)4Br2 and K2Pt(NO2)4Cl2 and other compounds of the same type K2PtX6, where X is partly replaced by Cl or Br and partly by (NO2), showing a transition towards the type of the double salts like the platino-ammoniacal salts. (The corresponding double sodium nitrite salt of cobalt is soluble in water, while the K,NH4 and many other salts are insoluble in water, as I was informed by Prof. K. Winkler in 1894).

In all the preceding complex compounds of Pt we see a common type PtX2,2MX (i.e. of double salts corresponding to PtO) or PtM2X4 = Pt(MX2)2, corresponding to Pt(HO)2 with the replacement of O by its equivalent X2. Two other facts must also be noted. In the first place these X's generally correspond to elements (like chlorine) or groups (like CN, NO2, SO3, &c.), which are capable of further combination. In the second place all the compounds of the type PtM2X4 are capable of combining with chlorine or similar elements, and thus passing into compounds of the types PtX3 or PtX4.

[12] The platinum salt and ammonia, when once combined together, are no longer subject to their ordinary reactions but form compounds which are comparatively very stable. The question at once suggests itself to all who are acquainted with these phenomena, as to what is the relation of the elements contained in these compounds. The first explanation is that these compounds are salts of ammonium in which the hydrogen is partially replaced by platinum. This is the view, with certain shades of difference, held by many respecting the platino-ammonium compounds. They were regarded in this light by Gerhardt, Schiff, Kolbe, Weltzien, and many others. If we suppose the hydrogen in 2NH4X to be replaced by bivalent platinum (as in the salts PtX2), we shall obtain NH3 NH3 Pt X X —that is, the compound PtX2,2NH3. The compound with 4NH3 will then be represented by a further substitution of the hydrogen in ammonia by ammonium itself—i.e. as NH2(NH4X)2Pt or PtX2,4NH3. A modification of this view is found in that representation of compounds of this kind which is based on atomicity. As platinum in PtX2 is bivalent, has two affinities, and ammonia, NH3, is also bivalent, because nitrogen is quinquivalent and is here only combined with H3, it is evident what bonds should be represented in PtX2,2NH3 and in PtX2,4NH3. In the former, Pt(NH3Cl)2, the nitrogen of each atom of ammonia is united by three affinities with H3, by one with platinum, and by the fifth with chlorine. The other compound is Pt(NH3.NH3Cl)2—that is, the N is united by one affinity with the other N, whilst the remaining bonds are the same as in the first salt. It is evident that this union or chain of ammonias has no obvious limit, and the most essential fault of such a mode of representation is that it does not indicate at all what number of ammonias are capable of being retained by platinum. Moreover, it is hardly possible to admit the bond between nitrogen and platinum in such stable compounds, for these kinds of affinities are, at all events, feeble, and cannot lead to stability, but would rather indicate explosive and easily-decomposed compounds. Moreover, it is not clear why this platinum, which is capable of giving PtX4, does not act with its remaining affinities when the addition of ammonia to PtX2 takes place. These, and certain other considerations which indicate the imperfection of this representation of the structure of the platino-ammonium salts, cause many chemists to incline more to the representations of Berzelius, Claus, Gibbs, and others, who suppose that NH3 is able to combine with substances, to adjoin itself or pair itself with them (this kind of combination is called ‘Paarung’) without altering the fundamental capacity of a substance for further combinations. Thus, in PtX2,2NH3, the ammonia is the associate of PtX2, as is expressed by the formula N2H6PtX2. Without enlarging on the exposition of the details of this doctrine, we will only mention that it, like the first, does not render it possible to foresee a limit to the compounds with ammonia; it isolates compounds of this kind into a special and artificial class; does not show the connection between compounds of this and of other kinds, and therefore it essentially only expresses the fact of the combination with ammonia and the modification in its ordinary reactions. For these reasons we do not hold to either of these proposed representations of the ammonio-platinum compounds, but regard them from the point of view cited above with reference to double salts and water of crystallisation—that is, we embrace all these compounds under the representation of compounds of complex types. The type of the compound PtX2,2NH3 is far more probably the same as that of PtX2,2Z—i.e. as PtX4, or, still more accurately and truly, it is a compound of the same type as PtX2,2KX or PtX2,2H2O, &c. Although the platinum first entered into PtK2X4 as the type PtX2, yet its character has changed in the same manner as the character of sulphur changes when from SO2 the compound SO2(OH)2 is obtained, or when KClO4, the higher form, is obtained from KCl. For us as yet there is no question as to what affinities hold X2 and what hold 2NH3, because this is a question which arises from the supposition of the existence of different affinities in the atoms, which there is no reason for taking as a common phenomenon. It seems to me that it is most important as a commencement to render clear the analogy in the formation of various complex compounds, and it is this analogy of the ammonia compounds with those of water of crystallisation and double salts that forms the main object of the primary generalisation. We recognise in platinum, at all events, not only the four affinities expressed in the compound PtCl4, but a much larger number of them, if only the summation of affinities is actually possible. Thus, in sulphur we recognise not two but a much greater number of affinities; it is clear that at least six affinities can act. So also among the analogues of platinum: osmic anhydride, OsO4, Ni(CO)4, PtH2Cl6, &c. indicate the existence of at least eight affinities; whilst, in chlorine, judging from the compound KClO4 = ClO3(OK) = ClX7, we must recognise at least seven affinities, instead of the one which is accepted. The latter mode of calculating affinities is a tribute to that period of the development of science when only the simplest hydrogen compounds were considered, and when all complex compounds were entirely neglected (they were placed under the class of molecular compounds). This is insufficient for the present state of knowledge, because we find that, in complex compounds as in the most simple, the same constant types or modes of equilibrium are repeated, and the character of certain elements is greatly modified in the passage from the most simple into very complex compounds.

Judging from the most complex platino-ammonium compounds PtCl4,4NH3, we should admit the possibility of the formation of compounds of the type PtX4Y4, where Y4 = 4X2 = 4NH3, and this shows that those forces which form such a characteristic series of double platinocyanides PtK2(CN)4,3H2O, probably also determine the formation of the higher ammonia derivatives, as is seen on comparing—

PtCl2NH32Cl23NH3
Pt(CN)2KCNKCN3H2O.

Moreover, it is obviously much more natural to ascribe the faculty for combination with nY to the whole of the acting elements—that is, to PtX2 or PtX4, and not to platinum alone. Naturally such compounds are not produced with any Y. With certain X's there only combine certain Y's. The best known and most frequently-formed compounds of this kind are those with water—that is, compounds with water of crystallisation. Compounds with salts are double salts; also we know that similar compounds are also frequently formed by means of ammonia. Salts of zinc, ZnX2, copper, CuX2, silver, AgX, and many others give similar compounds, but these and many other ammonio-metallic saline compounds are unstable, and readily part with their combined ammonia, and it is only in the elements of the platinum group and in the group of the analogues of iron, that we observe the faculty to form stable ammonio-metallic compounds. It must be remembered that the metals of the platinum and iron groups are able to form several high grades of oxidation which have an acid character, and consequently in the lower degrees of combination there yet remain affinities capable of retaining other elements, and they probably retain ammonia, and hold it the more stably, because all the properties of the platinum compounds are rather acid than basic—that is, PtXn recalls rather HX or SnXn or CXn than KX, CaX2, BaX2, &c., and ammonia naturally will rather combine with an acid than with a basic substance. Further, a dependence, or certain connection of the forms of oxidation with the ammonia compounds, is seen on comparing the following compounds:

PdCl2,2NH3,H2OPdCl2,4NH3,H2O
PtCl2,2NH3PtCl4,4NH3
RhCl3,5NH3RuCl2,4NH3,3H2O
IrCl3,5NH3OsCl2,4NH3,2H2O