110 ELECTROLYSIS the colours are arranged in circles long known as Nobili s rings. Similar phenomena are exhibited by salts of bismuth, nickel, cobalt, and manganese, all of which are precipitated as peroxides, usually hydrated(Wernicke, Pogg. Ann., cxli. 109), upon the anode by the action of the oxygen liberated by the passage of electricity. Silver is also thrown down as a black peroxide, together with some oxygen from a solution of sulphate and nitrate, and iron behaves somewhat similarly in an ammoniacal solution of the protoxide in vacuo. i Such secondary actions vary very conspicuously with the density of the current and the temperature. Bunsen (Pogg. Ann., xci.) electrolysed solution of chromic chloride, and by increasing the current density obtained in succession H, Cr.,0 3 , Cr0 3 , and metallic Cr at the cathode ; the reason for this is evidently that with high current densities the supply of ions in any time is greater than can take part in secondary action, and hence some of the original ion is deposited. A rise of temperature favours chemical action, and promotes rapid mixture of the ions with the solution at the same time ; so the higher the temperature the greater is the current density required to isolate the ions. From concentrated sulphuric acid, for instance, below 80 only H and are obtained ; between 80 and 90 oxygen is given off at the anode, while at the cathode H and S, due to reduction of H 2 S0 4 by hydrogen, appear ; above 90 sulphur alone is deposited at the cathode (Warburg, Pogg. Ann., cxxxv. 114). Mixed Instructive and important cases of secondary action occur when electro- the electric current is made to traverse a mixture of several solu- lytes. tions. Magnus (Pogg. Ann., cii. 23) determined by experiments on dilute CuS0 4 solution, in an apparatus with a porous diaphragm of clay, colloid paper, or animal membrane, specially arranged that the lines of flow should be parallel, and the current density therefore uniform, that there was a limiting value of the density above which both copper and hydrogen appeared at the cathode, but below only copper. His results show that this density is inde pendent of length of the electrolyte and material of the electrodes, but varies directly as the size of the electrodes. The specific resistance of the constituents, as well as the relative position of the two ions in the "electro-chemical series" (vid. inf.), are of great importance, the electro-negative metal always appearing first. In order to determine whether the current traversed both elec trolytes or only one, Hittorf (Pogg. Ann., ciii. 48), with the apparatus above described (p. 108), electrolysed mixed solutions of potassium chloride and iodide in different proportions, and arrived at the important conclusion that for all densities the current traversed both electrolytes, as it were in multiple arc (though the resistance of the mixture apparently bears no definite relation to the resistances of its constituents except for some of the haloid salts); but the products liberated depend on the secondary action at the electrode, and hence on the current density. The formation of an envelope of liquid of altered composition would also intro duce complications (Smee, Phil. Mag., xxv. 437). BufF, by experi menting on solution of HG1, with a small amount of H 2 S0 4 , sub stantially confirms Hittorf s results (Ann. d. Chcm. u. Pharm., cv. 156). These considerations are, of course, especially useful in effecting the deposition of alloys by electrolysis. The possibility of so doing appears to depend upon the composition of the solution employed. An acid solution of Cu and Zu deposits only copper, but the addition of potassium cyanide determines the deposition of brass. Gore (Electro-metallurgy, p. 51) points out that, in order to deposit an alloy of two metals, there must be no electric separation when the two metals are in contact with the liquid ; if indeed such were the case, a deposit of the two metals, say of Cu and Zn, would im mediately act as a CuZn couple (see p. 114), and the electro-nega tive metal alone would be deposited at the expense of the electro positive. Migra- Although the amount of a salt decomposed by the passage of a tion of given quantity of electricity is the same whether the salt be fused tlie ions, or dissolved in alcohol, water, or other solvent, yet the presence of the solvent produces an important effect upon the electrolyte, which should not be lost sight of in quantitative experiments. The phe nomenon is known as the "migration of the ions" (Hittorf), or the "unequal transfer of the ions" (Miller). Suppose, for example, we electrolyse a solution of CuS0 4 containing - 16 gramme of salt per cubic centimetre, in a vessel separated by a porous diaphragm into two portions A and B. Let electricity be passed through the solution between platinum electrodes from B to A, until 1 59 grammes of CuS0 4 have been decomposed. Then (1.) 1 59 g. of CuS0 4 has been removed from the solution ; (2.) 63 g. of Cu has been deposited on the platinum cathode ; (3.) "16 g. of has been evolved at the anode, and 80 g. of S0 3 absorbed there by the water of the solution. Now, had the electrolyte been a single fused compound, no com plication could have arisen ; the liquid remaining must still have been homogeneous (except for the presence of the ions near one or tther electrode). But when the salt is dissolved, it is important to In A. In B. Before Electrolysis After ) x g. CuS0 4 . (x- ^l-M g. CuS0 4 -(- 63 g. Cu, including Cu deposited. y g. CuS0 4 . /2/-_.l-59 g. CuS0 4 -f 96 g. S0 4 , including oxygen collected. If the volumes of equal, since the flui Hence A will g the two vessels are equal, x and y are of cours d is originally homogeneous. 1 71-1 consider from what part of the solution the salt has been removed. Suppose that of the CuS0 4 decomposed th was taken from the 71 vessel B, and therefore ^ ths from A. The result of electrolysis n may then be exhibited thus (assuming that no diffusion takes place through the diaphragm) : 1 Ti-1 B will lose 63 g. Cu, and gain 96 g. S0 4 . 7c- 71- We may therefore state the result thus : For every equivalent of copper deposited upon the cathode the entire gain of copper ic the vessel A is th equivalent, and the entire gain of S0 4 in B is n- equiv. The experiment shows that the entire gain of copper in A is "276 x 63 g., and the gain of S0 4 in B is 724 x -96 g. ; and hence, for solutions of CuS0 4 of that strength, = 276, and con- 71-1 sequently ^- = 724, so that, of the CuS0 4 decomposed, 72 4 per cent, is taken from A, and 27 6 per cent, from B, and the solution round the cathode is weakened much faster than that round the anode. This will be observable by the depth of the blue colour of the solution. If the anode be of copper and be vertically above the cathode, the effect is well seen ; for although the total amount of CuS0 4 in solution remains constant, the difference of colour at the two electrodes is very apparent, and, if the action be continued, strong dark-blue solution drops down in thin streams from the anode through the more dilute (Magnus). The value of n differs for different salts, and usually for solutions of the same salt of different strengths, though in some cases, as K 2 S0 4 , KN0 3 , NaCl, and KC1, the variations for great difference of concentration are very slight. The following table shows a few of the results obtained by Hittorf, with the apparatus described above, by which errors due to diffusion were avoided. The numbers 71 1 in the third column indicate what is called above , i.e., the total 7t excess in equivalents of the anion in the vessel containing the anodes corresponding to a decomposition of one equivalent of salt ; or, except in the last few cases, that part of the salt decomposed which is taken from the vessel containing the cathode. Salt. No. of cc. of solvent containing one gramme of salt. n-1 n HC1 2-9 319 HC1 362 168 HC1 1409 171 IIC1 2125-9 210 HBr 8-6 178 HI0 3 13 3 102 K 2 S0 4 118 500 KoS0 4 4128 498 NaCl 207 634 Fe 2 Cl s 25-25 600 CdI 2 -j 42 114 CdI 2 in alcohol .... -, ZnI 2 in alcohol 11 372 05 2-102 1 313 2 16 The iodides of zinc and cadmium are anomalous, but it may be supposed that they are decomposed as double salts thus : 2CdI a = Cd + (CdI 2 +I a ) , or (2CdI s + I 2 ). The total increase in the amount of an ion in one part of a vessel E divided by a porous partition is also affected by a mechanical v < transference of the electrolyte through the pores of the diaphragm, 01 generally in the positive direction of the current, which is very D noticeable in cases of electrolytes of high resistance. This was dis
covered by Keuss in 1807, and observed by Ferret soon after-