ARJISTRONU ON ELECTROLYTIC CONDUCTION. 125 VIII.-Note on Electrolytic conduction and on Evidence of a Change in the Constitution of W a t e r ; an Addendum to the foregoing Paper. By HENRY E. ARMSTRONG F.R.S. THE determination of the nature of electrolytic conduction is of such supreme importance in any discussion of the nature of chemical change and of the laws which determine it,* and the results contained * Professor Stokes in his Presidential Address to the Royal Society just de-livered uses the striking words " So closely is electricity related to chemica 126 ARNSTRONG ON ELECTROLYTIC CONDUCTION. in Mr. Crompton’s extension of MendelAef’s conception are in my opinion of such a definite and conclusive character that it appears both permissible and desirable to briefly refer in the first instance to current opinions in order that those who are not specially informed on such matters may gain the necessary clue to understand the ques-tions involved.In my address to the Chemical Section a t the British Association meeting a t Aberdeen (Report ISS5) I entered somewhat fully upon the discussion of the subject of chemiraZ action specially directing n ttention to t’he absolute interdependence of electrolytic and chemical action and in order to emphasise the importance of the study of electrolvtic phenomena I quoted the wordq used the previous year by Lord Rayleigh in his Presidential Address to the Association :-“ From the further study of electrolysis we may expect to gain improved views as to the nature of the chemical reactions and of the forces concerned in bringing them about.. . . . . . I cannot help thinking that the next great advance of which we have already some foreshadowing will come on this side. Arid if I might without presumption venture a word of recommendatioii it would be in favour of a more minute study of the simpler chemical phenomena.” My position a t that time is best indicated by the following quota-tions from my address :-“ The questions ‘ What is electrolysis ? What is an electrolyte ?’ are all-importmt to the chemist. . . . . . Helmholtz tells us that electrolytes belong to the class of typical compounds the con-stituents of which are united by ‘ atomic affinities,’ not to the class of ‘molecular aggregates.’ Is this the fact? Before chemists can accept this conclusion many difficulties must be removed which appear to surround the question.. . . . . The current belief among physicists would appear to be that the dissolved electrolyte-the acid or the salt-is almost exclusively primarily decomposed (Wiedemann, Elektricitlit 1883 ii 924) We are commonly told that sulphuric acid is added to water to make it conduct but the chemist desires to know why the solution becomes conducting. It may be that in all cases the ‘ typical compound ’ is the actual electrolyte-i.e. the body decomposed by the electric current-but the action only tukes place when. the t!jpicnl compounds are conjoined nnd .form the molecular agyre-gate for it is an undoubted fact that HC1 and H2SOa dissolve in water forming ‘ hydrates.’ This production of an ‘ electrolyticd scstem ’ from dielectrics is I venture to think the important question for chemists t o consider.I do not believe that we shall be able to action that coitld we only clearly apprehend the nature of electricity it Reems not unlikely that an unexpected flood of light might be shed on chemical combi-nation. ARMSTHOXG ON ELECTROLYTlC CONDUCTION. 127 state the exact conditions under which chemical change will take place until a satisfactory solution has been found. ''F. Kohlrausch (Pogg. Ann. 1876 159 233) has shown that on adding sulphuric acid to water the electric conductivity increases very rapidly until when about 30 per cent. of acid is present a maxi-mum (6914) is attained ; conductivity then diminishes almost as rapidly and a minimum (913) is reached when the concentration corresponds with that of a monohydrate (n[,S04,0H,) ; from this point conductivity increases somewhat (to 1031 at 92.1 per cent.H2SOa), and then again falls and is probably zero for the pure acid ; on adding sulphuric anhydride to the acid conductivity again increases. Solu-tions of other acids and of a number of salts-chiefly deliquescent and very soluble salts-also exhibit maximum conductivity at particular degrees of concentration. In no other case has the existence of two maxima such as are observed in solutions of sulphuric acid been esta-blished ; but probably this is because the experiments either have not been or cannot well be carried out with pure substances or very concentrated solutions. Solutions of less soluble salts increase in con-ductivity as the amount of salt dissolved increases." Kohlrausch has suggested as an explanation of the influence of the ' solvent ' on t,he conductivity of an ' electrolyte,' that in a solu-tion the ions which are being transferred electrolytically come less fre-quently into collision than would be the case in the pure substance. There is therefore less opportunity for the formation of new molecules, and the ions are able to travel farther before entering into combina-tion. " Regarding the question from a chemist's point of view however, I cannot help thinking that this explanation is scarcely satisfactory or sufficient ; and I cannot resist the feeling that the production of electrolytically conducting solutions from dielectrics is in some manner dependent upon the occurrence of chemical change.If the composition of the solutions of maximum conductivity be calculated, it will be seen that they contain but a limited number of water mole-cules ; thus bhe solution of sulphuric acid of maximum conductivity (at 18') contains 30.4 per cent. of acid and therefore has the com-position H,SO 1'2.4 H,O (approximately) ; for nitric acid the ratio is 1 8 ; for acetic acid it is about 1 17. Now it is highly remarkable that the solutions of maximum electric conductivity are also very nearly those in the formation of which nearly the maximum amount of heat is developed ; this will at once be obvious on comparison of the curves given by Thomsen (Thermocheinische Untersuchungen 3 ), and by Kohlrausch.In the chemist's experience the point of maxi-mum heat development is usually near t o the point of maximum chemical change and I think therefore that we are justified in con 12s ARJfSTROKQ ON ELECTROLYTIC COSDUCTION. cluding that even if electrical conductivity be not a maximum a t a particular concentration on account of the presence of a part'iculsr hydrate (belonging to the class of molecular aggregates) in maximum amount a t all events the ' structure' of the system is especially favourable and the ' chemical influence' exerted by the one set o € molecules upon the other is a t a maximum a t the point of maximum conductivity. The fact that the amount of sulphuric acid required to form a solution of maximum conductivity increases with tempera-ture-Temp.0" 10" 20" 30" 40" 50" 60" 70" Per cent . . . 30.2 30.9 31.7 32.5 33.5 34.1 34.5 35.4 and also the fact that the maxima and minims of conductivity tend to become obliterated with rise of temperature (Kohlrausch) are both in accordance with the view that conductivity is in some way depen-dent upon chemical composition as the efTect of rise of temperature would be t o cause the dissociation of hydrates such as 1 have referred to. The increase in conductivity of aqueous solutioiis with rise of temperature would appear to be against the view here put forward ; but it is probable that this may be largely due to diminution in viscosity and increase in the rate of diffusion." Jn March of last year I submitted to the Royal Society a paper '' On Electrolytic Conduction in relation to Molecular Composi-tion," &c.(Proc. Roy. Xoc. 1886 ass) in which I pointed out thst it was perhaps desirable to distinguish between simple eZecfrolytes-such as fused silrer chloride and composife electro Zytes-conducting mixtures of compounds like water hydrogen chloride and sulphuric acid which behave as dielectrics when pure. The difference between simple and composite electrolytes was so great that it was difficult to resist the feeling that the mode in which their electrolysis was effected was different and I ventured to suggest this without how-ever attaching any particular importance to the point and rather with the object of inciting debate ; but I clearly stated my objections to the current dissociation hypothesis of electrolysis and formulated the view that in the case of composite electrolytes bot,h constituents were immediately concerned and that the one the solvent did not merely exercise a screening effect as suggested by Kohlrausch.Considerable discuss on on these points has taken place between Professor Lodge and myself as co-secretaries of the B. A. Elec-trolysis Committee and in a communication to this committee a t the recent Manchester meeting in answer t o his criticisms I stated "t,he chief reasons which cause me to hesitate in accepting the ' atomic dissociation hypothesis ' and which have led me to snggest an alternative ' molecular hjpothesis,' viz. that in the case of composit ARMSTRONG ON ELECTROLYTIC CONDUCTION. 129 electrolytes at all events electrolysis is the outcome of the combined action of the E.M.F.and of some effect which the one set of mole-cules exerts upon the other set while both are under the influence of the E.M.F.” And I added “ I care little at present what the effect is the important question to settle being whether electrolysis is primarily an affair of atoms or of molecules. . . . Ostwald’s remarkable contributions to our knowledge of molecular conductivity appear to me to bear continuous testimony to the existence of such an influence of molecule upon molecule as that I have pictured” (Electrician August 5th and 26th 1887). I venture now to contend that thanks to our illustrious Russian confrdre the “ great advance ” of which Lord Rayleigh spoke is no longer far distant and that it is patent that electrolysis is primarily an affair of molecules ; that electrolysis takes place in consequence of an influence which one set of molecules A exercises upon another set of molecules B.The results which Mr. Crompton has obtained are to my mind conclusive on this point the information afforded by the study of sulphuric acid solutions being alone sufficient ; and they are of special importance as indicating the superiority of electrical values over all others in any discussion of the constitution of complex systems of dissociable compounds.* This was to be expected as on the hypothesis now under discussion the variations in the electrical values would represent variations in the extent to which the one set of molecules affect the other set the electrical values serving in fact t o quantify an injluence; for changes in constitution or structure might well occur which would involve but a slight degradation of energy aiid consequently a slight change in density and many other physical properties and which yet might lead t o a relatively very considerable change in the extent to which the compound could exert an influence on the course of electrolytic or chemical change.In the case of sulphuric acid the evidence is all but complete and it is no exa*ggeration to say that every peculiarityof the acid is faith-fully pictured in the second differential coefficient curve. It is well known that H2S04 readily loses SO3 and that Marignac’s experiments have shown that a stable equilibrium is attained only when the acid has lost a.nhydride to the extent required to form a hydrate H2S04*T1TH20 = 12H2S04.H20.A sudden change in the direction of the curve will be seen to occur at a point where this “ hydrate ’’ is situate ; and if the same interpretation be given of this and of the subsequent inflexions it would appear necessary to admit the exktence of such a hydrate although it is not usually recognised, it being supposed rather that equilibrium is established between * It is to be expected that the discussion especially of optical and magneto-optical data will afford valuable results in the case of non-electrolytes. VOL. LIII. 130 ARIlSTRONG ON ELECTROLYTlC CONDUCTION. €€,SO, H2S04.0H2 &c. when the composition is that ,of Marignac’s hydrate. I shall again refer to this point later on and will only add that MendelBeff’s curve does not show this hydrate.The three hydrates H2SOa*OH2 H2S04*20H2 H2SOa*60H2 are indicated both on the conductivity curve and on Mendel6eff’s density curve and there is also a marked break on both curves approximately a t a point corresponding to a hydrate H2S04*150H20. The conductivity curve however affords an unmistakable indication of a hydrate H,SO4*240H2 which does not occur on the density curve. The discovery of this hydrate is of fundamental importance, owing to the fact that the maximum conductivity of sulphuric acid solutions is manifested between the points where the two hydrates H2SO4*6H2O and H2SO4*24H2O are situate the two sets of molecules which mutually affect each other and induce electrolysis thus become exhibited.Mr. Crompton’s curve shows an inflexion at a point corresponding to the compound 2H2SOa*S03. I am not aware that any such substance is known or that salts of such an acid have been described, but a trichromate of somewhat analogous composition has been obtained and Weber has prepared octosulphates M2SO4*8SO3. On writing to my friend Dr. Messel who is accustomed to observe the behavionr of sulphuric anhydride on the large scale I learnt “that mixtures of the acid and anhydride containing up to 29-30 per cent. of the latter are liquid a t ordinary temperatures but that then crystallisation takes place ; when more than 55 per cent. of anhydride is present liquids are again obtained but solidification takes place when a little more than 70 per cent.of anhydride is present.” Now-2H,S04 * SO, contains 28.9 per cent,. SO,. 2H2S04*3S03 , 55.0 , ,7 2HzSO4.2S03 , 41.6 , 9 9 H,SO4*3SOj , ’71.0 , 9 , There is therefore independent evidence of the existence of the compound indicated on the conductivity curve and there can be little doubt that other similar compounds will be proved to exist when search is made for them. The concurrence of the evidence derived from the mathematical discussion of data deduced by the observation of properties so dif-ferent as density and electric conductivity with what may be termed the chemical evidence afforded by the recognition of a variety of distinct compounds of sulphuric acid with its anhydride or with water is so complete that I feel we need have no hesitation in apply-ing MendelBeff’s method and in accepting the conclusions to whic ARMSTRONG ON ELECTROLYTIC CONDUCTION.131 it logically leads even when these are beyond control by any ordinary chemical method. There is reason to believe that not only water and sulphuric acid: b u t liquids generally consist of complexes of the fundamental mole-cules ; and in any discussion of elect,rolytic conduction it is of primary importance t o recognise that this is probably the case. Mendeleeff is prepared to admit the existence of a hydrate containing 1500H2, but I must confess that I hesitate to do so and I even think that the inflexion in the curve at the point corresponding to this supposed hydrate is evidence of a change in the constitution of water.The occurrence of a similar change in the curve at about the same point in the case of ammonia and acetic acid and at a point somewhat, more distant from its origin in the case of phosphoric acid affords eonfirmation of this view. I t is possible also to put the same inter-pretation upon the sudden inflexion of the curve at the point corre-sponding to Marignac’s hydrate. Any difficulty which might be felt in admitting the existence of hydrates so abnormal in composition as 12H,S04-OH3 and H,S04*150H,0 would then be removed. But it is also conceivable that Marignac’s hydrate has an existence and that it is a hydrate of a polymer of sulphuric acid-not of the funda-mental molecule H,SO4-consisting of a very limited number of complexes of H,S04 molecules associated with a single water mole-cul e .The assumption that sulphuric acid mainly consists of complexes such as may be represented by the formula (H,S04) would serve to explain many features of the conductivity curve near to its origin from the acid thus the low conductivity of concentrated solutions may be attributed to the presence chiefly of (H2S04) molecules and of hydrates derived therefrom; and it may be supposed that as the complexes become more and more resolved into hydrates of the fundamental molecule H,S04 by dilution a greater influence would be exercised by molecule upon molecule. A somewhat similar view to that here indicated has been advanced by Landolt (Optisclzedrehunqsverm~,~e~ p. 59 j in explanation of the influence of “ neutral ” solvents on the rotatory power of opticallr active substances such as t,urpentine.* Rut the chief evidence in favour of such an cxplanation is probably t be found in the influence * Mr.Crompton has discussed among others the values given by Landolt as representing the change in rotatory power on diluting nicotine with water and turpentine with benzene and acetic acid the curves afford the clearest indications of the presence of molecular compounds of the active substance with the solvent., but the determinations are not nearly numerous enough to permit of the discoverj of the cornposition of these compounds. I therefore propose to reinvestigate the behaviour of turpentine. K 132 ARMSTRONQ ON ELECTROLYTIC CONDUCTION. which a small quantity of one metal-for instance lead or bismuth-exercises on another-for example on gold-the most ductile of metals a minute percentage rendering it highly brittle.It is incon-ceivable that the lead can enter into uniform relationship with the gold molecules throughout the mass L e t that it can combine with the whole of them. In the pure gold the molecules are probably throughout uniformly and continuously related and very probably also they are of simple atomic constitution but not truly symmetrical : on the introduction of a small proportion of foreign iricompatibie molecules continuity becomes disturbed ; the atoms become free to re-arrange themselves and in place of the uniform relationship which previously obtained a non-unif orm relationship results in consequence of the formation of complex aggregates which have less power of cohering than the original simpler molecules.It is perhaps an argument in favour of this “ screen ” hypothesis that as already pointed out the marked change in the curve near to the origin from water takes place in the case of phosphoric acid later than in the case of sulphuric it is not improbable that the molecules of the former acid are of greater complexity than those of the latter so that if a certain proportion of foreign molecules be necessary to produce maximum effect a larger absolute proportion of phosphoric acid would be required. Considerations of this character whatever their value in the present instance are certainly of importance in connection with the study of the influence of small quantities of foreign sub-stances on the properties of metals generally and in preventing superheating superfusion &c.It’ the argument advanced in this and the foregoing paper be cor-rect it becomes more than ever necessary to consider the grounds on which the conclusion is based that electrolysis is the outcome of atomic dissociation. It has always appeared to me that Kohlrausch’s curve for sulphuric acid affords the most positive evidence that electrolysis cannot be a simple dissociation phenomenon as in that case it is to be expected that dissociation would take place to a. gradually in-creasing extent until a maximum was reached and not as the curve would indicate that after attaining a maximum it should diminish and then again increase until a much higher maximum is attained.The first differential curve may bs regarded as picturing the manner in which the influence is exercised at various points by the compounds whose existence is indicated on the second differential curve and this curve most clearly indicates that the effect is exerted mainly at two points-the one lying between the acid and the bihydrate and the other between the two hydrates containing 6 and 24OH, the discontinuity of the phenomena at points intermediate between these two principal points is very striking. An explanation of the discon ARMSTRONG ON ELECTROLYTIC CONDUCTION. 133 tinuity has already been given viz. that it is due t.0 the fact that in the one case the polymerised acid in the other the fundamentrat mole-cule is active ; and if it be difficult to understand the changes which occur on passing from acid to water on the dissociation hypothesis if is undoubt,edly still more difficult to understand the sudden rise and subsequent fall in conductivity on passing from sulphuric t o pyro-sulphuric acid and to account for the fact that the latter has a lower conductivity even than the former acid.On my hypothesis however, the fact that H,S,07 is chemically in a certain sense the more stable compound affords a sufficient explanation of the difference in its behnviour and that of sulphuric acid. Any conclusion relating to electrolytic phenomena will doubtless be found to apply equally to allied chemical phenomena so that if it be asserted that too much stress has been laid on the part which atomic dissociation plays in electrolysis the same remark must be made regarding its influence on the course of chemical change.Cer-tainly the more I study the subject experimentally the more the conviction is forced on my mind that an explanation of the majority of chemical changes will ultimately be found in the principle of association and not in that of dissociation. Reference was made in the early part of the paper t o the division of electrolytes into two classes of simple and composite electrolytes. There is no reason to cease making this distinction on the contrary. But it being established as I believe that in the case of solutions electrolysis is the outcome of the influence exerted on each other by molecular aggregates consisting of the same proximate elements but differently constituted there appears no longer to be any necessity to suppose that there is any fundamental difference in the mode in which e1ect)rolysis takes place in solutions and in fused salts Le., simple electrolytes. It is in the highest degree probable that fused salts contain molecules of different orders of complexity and it may fairly be supposed that these would be capable of influencing each other much in the same way that I suppose the different aggregates in aqueous solutions influence each other. With regard t o the nature of the influence thus pictured as exerted between molecule and molecule I do not propose to discuss this now, and would only add Chat my views are entirely based on the assump-tion of the existence of what I have termed “ residual affinity,” and that they necessarily involve the recognition of the potency of this factor. Central Institution, City and Guilds of London I n s t i t u t e