mixtures of solution of salts (such as K,SO, and Al(SO4)3), which together give alums, that although in dilute solution the experimental and calculated numbers agree, yet in more concentrated solution the observed conductivity is less than the calculated. Similar results have been obtained for mixtures of solutions of the alkali sulphates and the vitriols, which together give salts of the type K2SOMgSO6H2O.1 It thus appears that complex ions exist in solutions of double salts, except at great dilution. It has also been shown that simple ions exist to a very small extent in solutions of complex salts. Thus Morgan found that a normal solution of potassium ferrocyanide contains o'000533 gram-equivalents of the cyanogen ion (CN) per litre. 2 The scheme of ionisation for complex salts is therefore the same as that for double salts. (1) K,Mg(SO4)22K + Mg(SO.)," (a) For double salts, the ionisation corresponding to equation (b) is practically complete, and for complex salts the reaction corresponding to equation (¿') takes place only to a very small extent so small, in the case quoted, that our chemical tests for the ion Fe" give no reaction. Of course, the concentration of the Fe" is much smaller than the small concentration of cyanogen given above. The influence of successive ionisation such as that described above, on the transport numbers, may be conveniently examined here. In the case of potassium magnesium sulphate, the ionisation (a) produces ions 2K and (Mg(SO4)2)", and when a current is carried by these ions, for every (Mg(SO4)2)" ion 1 Archibald, Trans. Nov. Scot. Inst. Sci., 9. 307 (1891). travelling to the anode, there are 2K ions travelling to the cathode. But now, if the solution be diluted, the (Mg(SO))" ion will be replaced by ions Mg" and 2SO", and thus for every 2SO" travelling to the anode, there are 2K and Mg travelling to the cathode. It is easy to see from this that the changes of concentration at the anode and cathode caused by electrolysis (from which the transport numbers are calculated) must be different for the two solutions, and the difference will be greater the greater the splitting up of complex ions caused by dilution. Thus a variation of the transport numbers with dilution indicates the presence of complex ions. If, however, the variation is small, it may be due to other influences not yet fully investigated. It has been found that in many cases the transport numbers determined from solutions of simple salts show variations with dilution. Thus the transport number for cadmium determined from a solution of cadmium iodide of medium concentration was found by Hittorf to be negative, while that for iodine determined from the same solution was greater than unityi.e. it appeared that the current carried by the anion was greater than the total amount of current passed-which is absurd. In dilute solution, however, the transport numbers both for cadmium and iodine were positive and less than unity. Similar large variations were found with cadmium chloride, zinc chloride, and zinc iodide,1 Smaller variations of the same kind were found with barium chloride, strontium chloride, calcium chloride, and magnesium chloride." The explanation is that all these solutions contain complex ions; and it has been determined that they are complex anions in the following way. Compounds of this type can either ionise in stages, giving in the first stage complex cations, e.g. BaCl, BaCl + Cl'; or can give complex anions of the type BaClg' or BaCl,". Now, the transport number for Ba" determined from a solution containing ions BaCl will appear to be greater than 1 Hittorf, loc. cit. 2 Bein, Zeitschr. phys. Chem., 27. 1 (1898); 28. 439 (1899). Noyes, Zeitschr. phys. Chem., 36. 61 (1901). when determined from a solution containing only Ba" and Cl. And since BaCl splits up into Ba" and Cl' in dilute solution, the transport number for Ba" determined from a solution containing BaCl will appear to diminish with dilution. That is, the transport number for the cation, determined from solutions containing complex cations, will appear to diminish with dilution. And similar reasoning shows that the transport number for the anion, determined from solution containing a complex anion, will appear to diminish with dilution. Since, in all the cases mentioned above the transport number for the anion diminished on dilution of the solution, the complex ions present are anions, eg. in solution of CdI, we have probably CdI", and similarly for solutions of CdCl2, ZnI2, and ZnCl2 in solutions of the chlorides of barium, strontium, calcium, and magnesium, the complex ions are probably of type BaCl or BaCl,". No account has been taken here of the effect of the differences in the velocities of the complex ions and the simple ions into which they dissociate. For a full discussion, see a paper on this subject by B. D. Steele.1 2 Compounds of Metallic Salts with Ammonia.-The compounds of ammonia with the salts of chromium, platinum, and cobalt, have been examined by Werner and others. The following is a slight sketch of their work on the ionisation of these compounds :— Luteocobaltic chloride, Co(NH)¿Cl3, contains six molecules of ammonia, which, according to Werner's theory of these compounds, are attached to the metal. For it has been shown that four ions are present in solution, and these are [Co(NH)] and 3Cl'. The ammonia may be partly replaced by water, as in the compound [Co. Cl, without any *(H2O)3. change in the mode of ionisation. (NH3)37 But if the number of molecules of ammonia and water together is made equal to five, it is found that one of the chlorine atoms cannot split off as an ion. Thus, chloropur 1 Phil. Trans. A. (1902), p. 105. 2 Werner and Miolati, Zeitschr. phys. Chem., 12. 34; 14. 506. pureocobaltic chloride [ [CONCI, gives ions [Co)] Co (NH3)5 (NH3) and 2C1'. And, as is expressed by the formula written for the compound, the explanation is that a chlorine atom has taken the place of the missing ammonia molecule, and now forms part of the radical, which, owing to the presence of the chlorine atom, has become divalent. On again reducing the number of ammonia molecules, a precisely similar effect is observed; [Co(NH)]Cl has two negative groups in the radical-which has become monovalent (NH).7 and Cl'. --and gives ions [Co(NU)] (NO2)2 NH3 The next compound in the series is CI which by the rules given should give no ions at all, i.e. should be a non-electrolyte. Experiment shows that under conditions unfavourable to the chemical alteration of the substance, which is not at all stable, the conductivity is extremely small. Exactly similar remarks apply to the compounds derived from platinic chloride, which have the formulæ : The second compound of the series, [Pt(NH)]Cl, has not yet been isolated. The further withdrawal of ammonia from the neutral compound [Pt(NH)2] can be accomplished if the conditions are such that an anion is present which can go into the radical, e.g. in presence of potassium chloride, and the comand K. NH pound obtained is [PINK, giving ions [PH] Pt The next compound in the series is [PtCl]K, the ordinary potassium platinichloride. These remarkable relations, which show the gradual transition from a tetravalent cation [Pt(NH)]**** to a divalent anion [PtCl]", through the intermediate stage of a non-electrolyte, have been observed in various other series of similar compounds. 1 Electro-affinity.-We have seen (Chap. I.) that a cation is an atom minus an electron, and an anion an atom plus an electron. The theory of electro-affinity 1 deals with the forces by which the charges of electricity are bound to the ions, i.e. with the force by which an electron is bound to an atom to form an anion (affinity for negative electricity), and with the force between a cation and an electron opposing the formation of an atom (affinity for positive electricity). It has not been found possible to measure these forces, which, however, remain constant for each ion. But the values of the electrode-potential (see p. 156) for the various ions follow the same order as these forces, and will be referred to as the electro-affinities of the ions. According to Abegg and Bodländer, the ionisation of a simple salt is the effect of these affinities for electricity, which act in opposition to the chemical force between the atoms. And since the chemical forces in a molecule are different for each different compound, we must, in comparing the ionisation of different compounds, take the chemical forces, as well as the electro-affinity, into account. We now have an explanation of the fact that practically no connection has been found between the extent of ionisation of a salt and that of the acid and base from which it is formed. Thus, if the base is sodium hydrate and the acid acetic acid, although the affinity for positive electricity (or force producing ionisation) of the sodium is the same in caustic soda as in sodium acetate, and similarly, the affinity for negative electricity of the CH,COO group the same in acetic acid as in sodium acetate, yet the chemical forces with which the sodium is bound to hydroxyl, hydrogen to the CH,COO group, and sodium to the CH,COO group (which, 1 Abegg and Bodländer, Zeitschr. f. Anorg Chemie, 20. 453 (1899). |