HEYROVSK~ ELECTROAFFINITY OF ALUMINIUM. PART n. 27 1V.-The Electroafinity of Aluminium. Part 11. The Aluminium Electi*ode. By JAROS LAV HEYROVSK G. Passiu e A luminizcm . Potentials.-The position of aluminium in the electro-potential series has long remained uncertain. Thus Streinz (Ber. lVien. ilkad. 1878 77 410)) from measurements in aluminium nitrate solutions placed aluminium as follows : . . . . - A1 Mg Zn Cd Sn Pb Fe . . . . +, and from measurements in aluminium chloride solutions thus : - . . . . Mg Zn Al Cd . . . . +. Wright and Thompson (Phil. Mag. 1885 [v] 19 102 197) found aluminium to be more positive than zinc in solutions of chlorides by 0.280 volt of bromides by 0.295 volt and of sulphates by 0.537 volt although from thermo-chemical determinations they expected a potential one volt? more negative than for zinc.Neumann (Zeitsch. physikal. Chem. 1894 14 193) using amalgamated aluminium placed it' as follows : . . . Mg Al Mn Zn . . . Burgess and Hambuechen (Electrochem. Ind. 1903 1 165) found that the potential of aluminium wires varied from -0.3 to -1-3 volt (referred to the normal hydrogen electrode). Van Deventer (Chem. T'lieekblad 1907 4 625 771) found that amalgamated aluminium had a potential similar to magnesium, whilst the inactive metal was more noble than zinc. Obviously in all cases where the metal is not amalgamated, aluminium remains in a passive condition the passivity being caused by a skin of oxide or hydroxide as is evident on dissolving aluminium in dilute alkalis or during amalgamation when the coherent skin peels off.The potential of metallic aluminium like that of any passive electrode is not influenced by the presence of aluminium ions in solution. The metal behaves rather as a gas electrode being sensitive t o oxidising and reducing agents besides being influenced by anions (compare Jory and Barnes Trans. Anzer. Electrochem. SOC. 1903 3 95). Great sensitivity to shocks was also observed. The following table of experimental results shows the passivity of aluminium wire 28 HEYROVSK~: Cell (at room temperature), first electrode is +. calomel H g I N-KCIsoln. j *l calomel ‘-Ig ~ N/lO-KCl s ~ l n . ! ~ ~ ’ ELECTROAFFINITY OF ALUMINIUM. PART 11. 0 bserved E.M.F. Behaviour. in volts. 0.77-1.67 Showing sudden changes and great fluc-twations of E.M.P.When not moved 0.7 6-0.80 Immedat-ly after the aluminium surface has been rubbed on glass fragments. Allowed to remain. < 1.0 Again rubbed. 1.48 1.40-1.47 Current of oxygen passed round the 0.6-0.7 electrode. (least observed E.M.F.) Current of hydrogen passed. 1-38 The potential is most negative in solutions of chlorides; it is less negative in bromides iodides sulphates and most passive in nitrate solutions. Evidently solutions of compounds of a more oxidising character passivify aluminium more intensively. This is in accord-ance with the known fact (compare Miiller Zeitsch. physill-d. Chem. 1909 69 481) that the passivifying influence of an anion is inversely proportional to the solubility of the product formed a t the electrode (that is the oxide).Dissolution of 2lXetaZZic A lzcminium .-The combined action of the active cation H’ and the anions Cl’ Br’ and I’ activates aluminium so that it decomposes water evolving hydrogen even in the dilute acids. Centnerszwer and Sachs (Zeitsch. physilcd. Chem. 1914 87 692) found that the rate of evolution of hydrogen in N-hydrochloric acid was 0.066 C.C. per sq. cni. per minute in N-hydrobromic acid 0-002 c.c. and in hydriodic acid still less. The dissolution in N-sulphuric acid was much slower whereas in nitric acid no hydrogen was evolved. The following experiments were made at the ordinary temperature : Evolution Acid used. of hydrogen. Notes Snlphuric acid N/10. None. But solutions in contact with metal for some days were found to contain aluminium salt.Slow. - 3 N . , concentrated. Slow. Yellowish or orange coating of sulphur Nitric acid diluts or None. Solutions were found to contain alu-concentzrated. minium salt. If some chloride is Hydrochloric acid di- Strong. I f potassium chlorate is added evolu-Sulphurous acid. Moderate. Odour of hydrogen sulphide. deposits on metal. added bubbles are evolved at once. lute or concentrated. tion increases HEYROVSK$ ELECTROAFFINITY ox ALUMINIUM. PART 11. 29 Evolution Acid used. of hydrogen. Nohes. Organic acids. None. Insoluble. Potassium iodide and Slow. -iodine solution. Potassium sodium or Strong. Concentrated ammonia and potassium barium hydroxides carbonate solutions also dissolve the dilute or concen- metal.trated. The dissolution in pure Concentrated sulphuric acid is interest-ing. If we suppose that electrochemical processes cause corrosion, then the precipitation of sulphur on aluminium must be regarded as cathodic deposition the few cations S""" which might exist in minute quantities in concentrated sulphuric acid being thrown down as a more noble element with less solution tension than aluminium. Electro-deposition of Metals by *4 luminium.-Similarly the deposition of any other more noble metal on aluminium is influenced by anions present in solution as they determine the potential of the metallic aluminium. Solution into which Deposition of I n this connexion the following results were obtained: Evolution aluminium is dipped. of metal. hydrogen.Notes. Gold chloride dilute. Gold deposits at once. Strong. Mercuric chloride Instantly amalgamated. Strong. dilute. Mercurous chloride. Slow amalgamation. Slow. Solution of mercuric No action. oxide in nitric acid. ' -Silver nitrate dilute Fine crystals of silver. Cupric chloride cu- Copper deposits readily. None. Slow. or concentrated. pric bromide cupric chloride and potass-ium iodide. (jopper sulphate. Very slow deposition oi None. copper. Fehling's solution. Y 9 s Slow. Cupric nitrate. Scarcely any action. -Ammoniacal copper No action. -Ferrous sulphate. No action. -Zinc sulphate. -solution. Ferric chloride. Dark powder deposits. Slow. Zinc chloride. No action. -Alkaline solution of Grey crystalline powder. Slow. 9 , zincate.Solution turns violet. Mercurous chloride must be in contact with metal. If trace of chloride is added amalga-mation occurs. Solutions decolorise 30 HEYROVSKg ELECTROAFFINITY OF ALUMINIUM. PART 11. Active Aluminium. Aluminium when active that is when dipping into a solutioii of hydrochloric acid or alkali hydroxide decomposes water vigor -ously and cannot be used for precise E.M.P. measurements. Neumann (Zoc. cit.) therefore used amalgamated aluminium, which causes less rapid evolution of hydrogen. He obtained as the mean value of rather variable potentials in N-aluminium sulphate solution - 1.317 volt8 in N-aluminium chloride solution - 1.292 volts in N-aluminium nitrate solution - 1.052 volts, referred to the normal hydrogen electrode as zero.Even such an electrode is far from being a reversible one since the hydrogen ions discharging on the electrode make the potential more positive just as silver ions do in silver concentration cells. I f the evolution of hydrogen could be prevented that is if the potentJal of hydrogen could be lowered below that of aluminium, a reversible aluminium electrode would be obtained. The potential of - amalgamated aluminium wires in 0-0213A'-aluminium chloride solution saturated with hydrogen under atmo-spheric pressure was indeed found to be less variable and more negative namely - 1*330+0.003 volts but slow evolution of hydrogen could not be prevented. The high overvoltage of hydrogen on a mercury surface makes it possible for a dilute amalgam of a very negative metal to behave as a reversible electrode because the evolution of hydrogen is almost entirely prevented.Lewis ( J . Amer. C'hem. Joc. 1910 32 1458; 1912 34 119; 1913 35 340; 1915 37 1893) has been able to determine the electrolytic potentials of alkali metals using dilute amalgams and the same method has been adopted here. for aluminium. Preparation of A lunainium A ma1gam.-About 0.4 gram of aluminium (99.6 per cent.) was dissolved in 200 grams of pure dry mercury by boiling for two to three hours in an atmosphere of dry carbon dioxide. On cooling some solid amalgam separated o u t on the surface showing that this very dilute (about 0.1 per cent.) amalgam is saturated. This anialgain is extremely easily decom-posed in moist air instantly losing its lustre and becoming covered by hydroxide this being no doubt due to the great affinity of aluminium for oxygen and its small affinity for mercury.Whether this saturated amalgam shows any difference of poten-tial from pure active aluminium or not could not be ascertained, since aluminium was found to be passive in dry acetone ether or piperidine. However since the liquid amalgam is in contact with solid amalgam their solution tensions must be identical and a HEYROVSKg ELECTROAFFINITY OF ALUMINIUM. PART II. 31 the aluminium seems to be very loosely bound to mercury the heat of oxidation of the solid amalgam has been found to be the same as the heat of oxidation of aluminium (Baille and FQry Ann. Ghim. Phys. 1889 [vi] 17 246). The electrolytic potential of the liquid amalgam must be very near to that of ideal active aluminium.Measurement of E.M.F.-The glass apparatus in which the amalgam had been prepared was inverted and the amalgam allowed to pass through a side tap by which the flow could be regulated into a sealed-on capillary tube (of 1 mm. bore) with a platinum contact. The lower end of the capillary tube was bent up and opened out so as t o provide a larger surface of amalgam. The electrode dipped into a solution of aluminium chloride which was stirred by means of a stream of hydrogen bubbles. The space above the solution was thus kept filled with hydrogen under atmo-spheric pressure. The second electrode was a hydrogen electrode consisting of platinum coated on glass as used by Loomis and Acree (Smer.Chem. J. 1911 46 585 621 638); there was also a calomel electrode attached to the vessel filled with the same aluminium chloride solution to check the hydrogen electrode from time to time. A t the beginning of each experiment hydrogen was passed through the cell until the potential difference between the calomel and hydrogen electrode became constant. Then the tap on the capillary tube was opened and the amalgam allowed to drop out slowly. It was found better to allow the amalgam to flow slowly, as on fresh surfaces after a few seconds bubbles of hydrogen appeared. Since however during readings the potential increased by several millivolts this being no doubt due to electrical adsorp-tion of ions on drops of mercury the solution round the electrode was stirred by bubbling hydrogen through it and simultaneously the aluminium chloride solution through which hydrogen was first passed was allowed to flow into the space round the electrode.A second series of measurements was made with new amalgams and capillary tubes and the results obtained did n o t differ by more than 10 millivolts from the first series even in the most dilute solutions whilst in the stronger solutions the agreement was within 3 millivolts. I n this way the following readings were obtained (using accumulator Weston cell Lippmann electrometer and potentio-meter giving readings to 0.1 millivolt): The whole apparatus was kept in a thermostat a t 25.0° 32 HEYROVSE< ELECTROAFFINITY OF ALUMINIUM. PART 11. Concentration of Mean E.M.P. of gram-equivalents AlCI Al-aluminium chlorides in per litre.the cell : + - ~ 2 1 solution 1 amdgam. (2.88 1.128 volt 0.1845 1-164 0-0675 1.160 0 0337 1-157 0.02 13 1.145 0.0107 1-161 (0.006Ti5 1.135 (0.00213 1.136 AM referred to the normal hydrogen electrode. - 1.284 volt) - 1.382 - 1-381 - 1-397 - 1.370 - 1.383 - 1.381) - 1.377) The determinations of single hydrogen electrode potentials are described in the preceding paper (p. 15) from which the potentials of aluminium amalgam electrodes rA (third column) could be calculated. The solutions were prepared from aluminium chloride purified by precipitation with hydrogen chloride and to each solution excess of freshly precipitated aluminium hydroxide was added. In such solutions the solubility product [Al"'] .[OH133 should be constant and equal to k . [Al(OH),]. Then the electrolytic potential of aluminium becomes where K is a constant. Such an electrode therefore behaves as a hydrogen electrode or an oxygen electrode of the type Hg [ HgO and consequently its potential qAl should always differ from the platinised electrode by a constant. The difference between the two electrodes (column 2) is very nearly constant except in the case of the first solution which was very viscous and opaque. I n the two most dilute solutions the J?.M.F.'s were rather variable sometimes approaching 1- 16 volts. The value of zA1 on the whole falls with the increase of acidity and concentration of aluminium chloride. In order t o determine how the potential of the aluminium electrode is influenced by different solutions a simpler form of apparatus was used having two capillary amalgam electrodes dipping into solutions saturated with hydrogen and covered by a layer of liquid paraffin.In this case the amalgam was not allowed to drop continuously but the surface was renewed every few minutes. The values obtained in %his way (at 2 5 O ) are given in table I. Single potentials are referred to the normal hydrogen electrode Solution used in the Al-electrode. NIlO-HCl 2.90 N-AICl, 0.368 ,, 0-0675 ,, 0.0409 ,, 0.0184 ,, 0.0046 ,, 0.00306 ,, 0.0028 ,, 0.30 N-Al,(SO,), 0-030 ,, NIlO-KCI N-KCl N/lOO-KOH+Al( OH), N/10- 9 9 ) NI10-KOH Second electrode. NIlO-KCl calomel. N-KCl calomel.Calomel of the same solu-tion. Y, 9 9 99 ?? N-KCl calo$el. 9 , Y, 9 , Y Y Hg+HgO of the same solution. 7 9 N-KOH saturated with $ 9 -4I(OH)3 TABLE I. Observed E.M.F. in volts. 1.5790 1.5A1-55 1.674 1.699 1.702 1-724 1.755 1.770 1.630 1.700 1.742 1.785 1.750 1.910 1.774 1.752 1.713 Potential of Al- electrode, AAI. - 1,209 1 2 6- 1.27 1.374 1.359 1.342 1.343 1.349 1.345 1.347 1.417 1.459 1.502 1.467 1.622 1.576 1.613 1.596 rH a dipping th 34 HEYROVSKP ELECTROAFFINITY OF ALUMINIUM. PART 11. Some determinations of single electrode potentials rA1 and rH, a.re calculated from the results described in the preceding paper; the electrode potentials in potassium hydroxide solutions were obtained by elimination of the diffusion potential with concentrated potassium chloride solutions.With fresh surfacw the E.M.F. rose a few millivolts to a maximum and after three minutes began to fall slowly. If the solution round the electrode was stirred or the electrode shaken, the E.M.F. fell about 20 millivolts but reverted to the original value on keeping. Oxidising agents such as dilute solutions of ferric chloride or hydrogen peroxide or even a current of air, caused a considerable decrease of negative potential amounting to several decivolts ; reducing agents had no influence. The more acidio the solution the more stable was the E.M.F. and fewer bubbles of hydrogen appeared on the surface of the amalgam. In N/10-hydrochloric acid no bubbles were formed and the E.M.F.was constant for t7en minutes a t 1.5788 volts,. whereas in alkaline solutions there was a visible elvolution of minute bubbles. Discussion of Results. From the most trustworthy measurements of rA1 in 0.1845N-AlC1 = - 1.370 volts where [A41*'*] = 0*0130* 0.0675 , = - 1.382 , , =0.0080 0.0337 , = - 1.383 , , =0*0044 the theoretical value for t.he electrolytic potential of aluminium in a normal solution of aluminium ions E.P. has been calculated by the formula E.P. = rA - ~ ~ l o g J A l " ' ] = rA - 0.0591 log, [ Al' 'I. 3 F 3 The values -1.333 -1.341 -1.336 calculated in this way give E.P. = - 1.337 volts as the most probable value. The approximate value of the electrode potential can be obtained from the heat of the electro-chemical reaction if we neglect the term T .__ in the equation T = Q + T. - and assume that the heat equivalent & of the reaction is equal t o the total change of energy. d?r drr d I' d T The heat of the reaction +A1 + HCl (in 200 as.) + iA1C1 + iH2 * See Part I HEYROVSKP ELECTROAJYFINITY OF ALUMINIUM. PART 11. 35 is 41.033 calories (Thomsen). of heats of ionisation of the processes: It can be regarded its the difference $A1 -f &Al**' +@ H + H + +(-) --giving +Al+H' -f QAl++++H, from which the calculated potential of the process A1 + Al"' t 3 0 is 1.76 volts whereas the potential corresponding with 31.000 calories of the reaction &A1 + H,O + &Al(OH) + &HZ leads to 1-34 volts. This would apply to the electrochemical process A1 + 30H' -+ Al(OH) + 3 0 .The coinoidence of the latter value with the observed one was pointed out by Kistiakovsky (Zeitsch. physikal. Chem. 1910 70, 260) who suggested similar electrode reactions for magnesium, iron and chromium. The reason why only the process A1 + 30H' -+ Al(OH),+ 3@ is the source of electrical energy must be sought in the extremely small solubility product of aluminium hydroxide (= 10-33; see Part I) effecting considerable hydrolysis. Owing to this the layer of solution close to the surface of the amalgam is saturated with aluminium hydroxide so that the potential is determined by hydroxyl ion concentration directly by aluminium ion concentra-tion only through the solubility product equilibrium. This is evident from table I where with decreasing acidity rA1 increases.Owing however to increasing oxidation of the amalgam in less acidic solutions the values are shifted towards more positive potentials so that in these solutions the potential density from the hydrogen electrode (column 4 table I) decreases instead of being constant. The reason why even in acidic solutions (like 0-1N-hydrochloric acid) the ordinary ionisation potential of A l e Al'" + 3 0 is not attained but the value remains roughly that of A1 + 30H' -+ Al(0H) + 3 0 , must be sought in the extremely slight dissociation of aluminium hydroxide which does not react sufficiently quickly with the acid to form aluminium ions and remains far behind the reaction c 36 MASON AND WHEELER THE PROPAGATION OF determining the electrode pot eyatial. tion might be a molecular one not ionic thus: Al(OH) + 3HCl+ AlCl,aq., Moreover the second reac-and need not take place at the electrode. Summary. (1) Various potentials of passive aluminium have been measured, and the influence of the anion on potential dissolution and deposition of metals is discussed. (2) The potential of active aluminium in the form of saturated liquid amalgam has been measured in different solutions. (3) The theoretical value of E.P.= -1.337 volts a t 2 5 O the normal hydrogen electrode being taken as zero. The correspond-ing electrochemical reaction being A1 + 30H' -+ Al(OH) + 3 0 . I n conclusion I desire to express my thanks to Prof. F. G. Donnan a t whose suggestion this work was undertaken for his kind interest and advice. 7 also wish to express my indebtedness to Dr. R. E. Slade for his constant help throughout this investigation. CHEMICAL DEPARTMENT, UNIVERSITY COLLEGE LONDON. [Received October 20th 1919.