The vapour density of phosphorus trichloride and oxychloride corresponds with their formulæ (Cahours, Würtz)—namely, is equal to half the molecular weight referred to hydrogen. But it is not so with phosphorus pentachloride. Cahours showed that the vapour density of phosphorus pentachloride referred to air = 3·65, to hydrogen = 52·6, whilst according to the formula PCl5 it should be = 104·2. Hence this formula corresponds with four, and not with two, molecules. This shows that the vapour of phosphoric chloride contains two and not one molecule, that in a state of vapour it splits up, like sal-ammoniac, sulphuric acid, &c. The products of disruption must here be phosphorous chloride, PCl3, and chlorine, Cl2, bodies which easily re-form phosphoric chloride, PCl5, at a lower temperature. This decomposition of phosphoric chloride in its conversion into vapour is confirmed by the fact that the vapour of this almost colourless substance shows the greenish-yellow colour proper to chlorine. This dissociation of phosphoric chloride has been considered by some chemists as a sign that phosphorus, like nitrogen, does not give volatile compounds of the type PX5, and that such substances are only obtained as unstable molecular compounds which break up when distilled; for example, PH3,HI, PCl3,Cl2, NH3,HCl, &c. To prove that the molecule PCl5 actually exists, Würtz in 1870 observed that when mixed with the vapour of phosphorous chloride the vapour of phosphoric chloride distils over (from 160° to 190°) perfectly colourless, and has a density which is really near to the formula—namely, to 104—and the same density was determined for the pentachloride in an atmosphere of chlorine. Hence at low temperatures and in admixture with one of the products of dissociation, there is no longer that decomposition which occurs at higher temperatures—that is, we have here a case of dissociation proceeding at moderate temperatures.
An important proof in favour of the type PX5 is exhibited by phosphorus pentafluoride PF5, obtained by Thorpe as a colourless gas which only corrodes glass after the lapse of time; it may be kept over mercury, and has a normal density. It is formed when liquid arsenic trifluoride, AsF3, is added to phosphoric chloride surrounded by a freezing mixture: 3PCl5 + 5AsF3 = 3PF5 + 5AsCl3.
In general, fluorine and phosphorus give stable compounds: PF3, POF3, and PF5, as would be expected from the fact that in passing from Cl to I (i.e. as the atomic weight of the halogen increases) the stability of the compounds with P and the tendency to give PX5 (Note [24]) decreases. Phosphorus trifluoride is obtained by heating a mixture of ZnF2 and PBr3, by the action of AsF3 upon PCl3, by heating phosphide of copper with PbF2, &c. It is a strong-smelling gas, which liquefies at -10° under a pressure of 40 atmospheres, giving a colourless liquid. It dissolves easily in (is absorbed by, reacts with) water, and acts upon glass; when mixed with Cl2 it combines with it (Poulenc, 1891), forming PCl2F3, a colourless gas of normal density, which is transformed into a liquid at 8°, decomposes into PF3 + Cl2 at 250°, and, with a small amount of water, gives oxy-fluoride of phosphorus, POF3 (with a large amount of water it gives PH3O4), which Moissan (1891) obtained by the action of dry HF upon P2O5, and Thorpe and Tutton (1890) by heating a mixture of cryolite and P2O5. It is a gas of normal density, like PF3, and was obtained by Moissan by the action of fluorine upon PF3 (PSF3, see Chapter XX., Note [20]). Thus the forms PX3 and PX5 not only exist in many solid and non-volatile substances, but also as vapours.
[26] Phosphorus oxychloride is obtained by the action of phosphoric chloride on hydrates of acids (because alkalis decompose phosphorus oxychloride), according to the equation PCl5 + RHO = POCl3 + RCl + HCl, where RHO is an acid. The reaction only proceeds according to this equation with monobasic acids, but then RCl is volatile, and therefore a mixture is obtained of two volatile substances, the acid chloride and phosphorus oxychloride, which are sometimes difficult to separate; whilst if the hydrate be polybasic the reaction frequently proceeds so that an anhydride is formed: RH2O2 + PCl5 = RO + POCl3 + 2HCl. If the anhydride be non-volatile (like boric), or easily decomposed (like oxalic), it is easy to obtain pure oxychloride. Thus phosphorus oxychloride is often prepared by acting on boric or oxalic acid with phosphoric chloride. It is also formed when the vapour of phosphoric chloride is passed over phosphoric anhydride, P2O5 + 3PCl5 = 5POCl3. This forms an excellent example in proof of the fact that the formation of one substance from two does not necessarily show that the resultant compound contains the molecules of these substances in its molecule. But other oxychlorides of phosphorus are also formed by the interaction of phosphoric anhydride and chloride; thus at 200° the chloranhydride, PO2Cl, or chloranhydride of metaphosphoric acid, is formed (Gustavson). The chloranhydride of pyrophosphoric acid, P2O3Cl4, was obtained (Hayter and Michaelis), together with NOCl, &c., by the action of NO upon cold PCl3, as a fuming liquid boiling at 210°.
[27] The direct action of the sun's rays, or of magnesium light, is necessary to start the reaction between carbonic oxide and chlorine, but when once started it will proceed rapidly in diffused light. An excess of chlorine (which gives its coloration to the colourless phosgene) aids the completion of the reaction, and may afterwards be removed by metallic antimony. Porous substances, like charcoal, aid the reaction. Phosgene may be prepared by passing a mixture of carbonic anhydride and chlorine over incandescent charcoal. Lead or silver chloride, when heated in a current of carbonic oxide, also partially form phosgene gas. Carbon tetrachloride, CCl4, also forms it when heated with carbonic anhydride (at 400°), with phosphoric anhydride (200°), and most easily of all with sulphuric anhydride (2SO3 + CCl4 = COCl2 + S2O5Cl2, this is pyrosulphuryl chloride). Chloroform, CHCl3, is converted into carbonyl chloride when heated with SO2(OH)Cl (the first chloranhydride of sulphuric acid); CHCl3 + SO3HCl = COCl2 + SO2 + 2HCl (Dewar), and when oxidised by chromic acid.
Among the reactions of phosgene we may mention the formation of urea with ammonia, and of carbonic oxide when heated with metals.
[28] We are already acquainted with some of the chloranhydrides of the inorganic acids—for instance, BCl3, and SiCl4—and here we shall describe those which correspond with sulphuric acid in the following chapter. It may be mentioned here that when hydrochloric acts on nitric acid (aqua regia, Vol. I. p. [467]) there is formed, besides chlorine, the oxychlorides NOCl and NO2Cl, which may be regarded as chloranhydrides of nitric and nitrous acids (nitrogen chloride, Vol. I. p. [476]). The former boils at -5°, the latter at +5°, the specific gravity of the first at -12° = 1·416, and at -18° = 1·433 (Geuther), and of the second = 1·3; the first is obtained from nitric oxide and chlorine, the second from nitric peroxide and chlorine, and also by the action of phosphoric chloride on nitric acid. If the gases evolved by aqua regia be passed into cold and strong sulphuric acid, they form crystals of the composition NHSO3 (like chamber crystals), which melt at 86°, and with sodium chloride form acid sodium sulphate and the oxychloride NOCl. This chloranhydride of nitric acid is termed nitrosyl chloride.
Cyanogen chloride, CNCl, is the gaseous chloranhydride of cyanic acid; it is formed by the action of chlorine on aqueous mercury cyanide, Hg(CN)2 + 2Cl2 = HgCl2 + 2CNCl. When chlorine acts on cyanic acid, it forms not only this cyanogen chloride, but also polymerides of it—a liquid, boiling at 18°, and a solid, boiling at 190°. The latter corresponds with cyanuric acid, and consequently contains C3N3Cl3. Details concerning these substances must be looked for in works on organic chemistry.
[28 bis] This reaction indeed proceeds very easily and completely with a number of hydroxides, if they do not react on hydrochloric acid and phosphorus oxychloride, which is the case when they have alkaline properties. When the hydroxide is bibasic and is present in excess, it not unfrequently happens that the elements of water are taken up: R(OH)2 + PCl5 = RO + 2HCl + POCl3. The anhydride RO may then be converted into chloranhydride, RO + PCl5 = RCl2 + POCl3—that is, phosphorus pentachloride brings about the substitution of O by Cl2. Thus carbonyl chloride, COCl2, boron chloride, 2BCl3, and succinic chloride, C4H4O2Cl2, &c., are respectively obtained by the action of phosphoric chloride on carbonic, boric, and succinic anhydrides. Phosphorus pentachloride reacts in a similar manner on the aldehydes, RCHO, forming RCHCl2, and on the chloranhydrides themselves—for example, with acetic chloride, CH3.COCl (when heated in a closed tube), it forms a substance having the composition CH3CCl3.
Phosphorus trichloride and oxychloride act in a similar manner to phosphoric chloride. When phosphorus trichloride acts on an acid, 3RHO + PCl3 = 3RCl + P(HO)3. If a salt is taken, then by the action of phosphorus oxychloride a corresponding chloranhydride and salt of orthophosphoric acid are easily formed: 3R(KO) + POCl3 = 3RCl + PO(KO)3. The chloranhydride RCl is always more volatile than its corresponding acid, and distils over before the hydrate RHO. Thus acetic acid boils at 117°, and its chloranhydride at 50°. Phosphoric and phosphorous acids are very slightly volatile, whilst their chloranhydrides are comparatively easily converted into vapour. The faculty of the chloranhydrides to react at the expense of their own chlorine determines their great importance in chemistry. For instance, suppose we require to know the molecular formula of some hydrate which does not pass into a state of vapour and does not give a chloranhydride with hydrochloric acid—that is, which has not any basic or alkaline properties; we must then endeavour to obtain this chloranhydride by means of phosphoric chloride, and it frequently happens that the corresponding chloranhydride is volatile. The resultant chloranhydride is then converted into vapour, and its composition is determined; and if we know its composition we are able to decide that of its corresponding hydrate. Thus, for example, from the formula of silicon chloride, SiCl4, or of boron chloride, BCl3, we can judge the composition of their corresponding hydrates, Si(HO)4, B(HO)3. Having obtained the chloranhydride RCl or RCln, it is possible by its means to obtain many other compounds of the same radicle R according to the equation MX + RCl = MCl + RX. M may be = H, K, Ag, or other metal. The reaction proceeds thus if M forms a stable compound with chlorine—for example, silver chloride, hydrochloric acid, and R, an unstable substance. Hence, a chloranhydride is frequently employed for the formation of other compounds of a given radicle; for instance, with ammonia they form amides RNH2, and with salts ROK, with anhydrides R2O, &c.