1292 MELLOR: THE UNlON OF HYDROGEN AND CHLORINE. CXXX1.-The Union of Hydrogen and Chlorine. VI. Die Period of Induction. By J. W. MELLOR. THE experiments recorded in the preceding paper lead up to the study of the most interesting period of chemical reactions where the velocity gradually increases from zero to a final maximum value. Bunsen and Roscoe call this the u period of induction.” The pheno- menon at first sight appears to contradict the well established law of mass action, because if the reaction takes place directly between, say, two components, the maximum velocity must occur at the beginning of the reaction when the active masses of the reacting components are greatest, as in the action of chiorine on methyl ether in sunlight. ‘6Da,ns les premiers, l’attaque est excessivement vive, mais elle se ralentit imesure que la chloruration fait des progrks ” (Cahours, Compt.rend., 1846, 23, 1070). It is not satisfactory to postulate, with Berthelot and Gilles (Ann. Chim. Phys., 1862, [ iii], 62, 26), ‘‘ une sorte l’inertia de rbsistance b vaincre qui retard la combinaison dans les premiers instants.” Har- court and Esson (Phil. Trans., 1866, 156, 201), Naumann (Annalen,VI. THE PERIOD OF INDUCTION. 1293 1876, 160, 23), and Hell and Urech ( B e y . , 1880, 13, 531) assume that the period of acceleration is due to the fact that '(chemical change consists in the gradual formation of a substance "-the so-called inter- mediate compound-" which at the same time slowly disappears by reason of its reaction with a proportional quantity of a third sub- stance." According t o this view (the intermediate compound theory), the reaction between hydrogen and chlorine in the presence of moisture may be written : CI, + H,O = X + &C.; (X + &c.) + H, = H,O + 2HC1, where X represents the unknown intermediate compound.Although the intermediate compound formed in many reactions has been isolated, with hydrogen and chlorine, unfortunately, indirect evidence is alone available. The reaction is usually studied in Bunsen and Roscoe's actinometer. The mixed hydrogen and chlorine gases are confined in a flak glass bulb (called the '' insolation vessel "), one-third filled with water satur- ated with the two gases. The insolation vessel is connected with an index which is very sensitive t o changes of volume, The lower half of the insolation vessel is painted black in order to screen the water from the influence of light.When the insolation vessel is illuminated, hydrogen chloride is formed and absorbed by the water. The index movements show that the rate of formation of hydrogen chloride gradually increases from zero up to a constant maximum velocity. The rapidity of induction and the rate of formation of hydrogen chloride are each proportional to the intensity of the light. Obviously, a certain amount of hydrogen chloride must be formed before the water begins to absorb this gas a t a constant rate. The rate of absorption of hydrogen chloride by water diminishes as the extremely thin layer of water in immediate contact with the gas approaches saturation.The denser solution of hydrogen chloride formed a t the surface sets up irregular convection currents during its descent to the bottom of the insolation vessel. Such currents are almost absent if ammonia gas is being absorbed instead of hydrogen chloride. The rate of absorption then depends on the rate of diffusion of the dissolved gas from above downwards. The amount of hydrogen chloride which the atmosphere above the water can hold before absorption begins is very small, as will be shown in Part TI1 of this investigation. The period of induction cannot be explained by the delayed solution of hydrogen chloride. Harden and Upon (Proc., 1894, 10,165) have observed a well-defined period of induction with carbon monoxide and chlorine. In this case, the1294 MELLOB: THE UNION OF ~YDROGEN AND cE~LoR~NE].velocity of the reaction is measured by the difference in the pressure of the gas on both sides of the equation C1, + GO = COC1,. Bunsen and Roscoe’s period of induction is sometimes sub-divided into (1) a period of inertness, during which there is supposed t o be no formation of hydrogen chloride, and (2) a period of acceleration. From the results recorded in Part IV of this work, however, it appears that there is no such thing as a period of no Formation of bydrogen chloride (MeIIor and Anderson, Trans., 1902, 81,414).* Assuming the existence of a n intemaediate compound, the period of induction is a necessary consequence o f the bw of mass action. A t the beginning of the reaction, the rate of diminution of chlorine is a maximum, whilst the rate of formation of hydrogen chloride is zero.From that moment, the rate of formation of the intermediate compound X is always equal to the difference in the rate of diminution of chlorine and the rate of formation of hydrogen chloride. During t h e first period of the reaction, the amount of X in the system con- tinually increases, and the system contains the greatest amount of X at the moment when the rate of diminution of chlorine is equal t o the rate of formation of hydrogen chloride. I n symbols, if x, y , x respectively denote the amounts of chlorine, X, and hydrogen chloride in the system a t the time t, then Again, from the law of mass action, the rate of formation of hydrogen chloride will be greatest when the system contains a maximum amount of X.But if y be a maximum, - dY =o, d t , dx dx ” dt dt’ - - = - The rate of formation of hydrogen chloride, therefore, gradually in. creases from zero, a t the beginning of the reaction when y=O, up t o a maximum value when y is a maximum. The so-called period of * The fact that hydrogen chloride is formed during the momentary illumination of a mixture of hydrogen and chlorine was published sin~ultaneously by P. V. Bevan (Proc. Camb. Phil. Xoc., 1902, 11, 264). Mr. Bevan experimented with a platinum wire placed in a mixture of hydrogen and chlorine, and measured the rise of tem- perature which occur8 when the mixed gases are momentarily illuminated by the increase in the resistance of the platinum wire. He believes that the presence of the platinum did not interfere with his results, and concludes that the increase in volume during the Draper effect is equivalent to the heat generated by the combination of hydrogen and chlorine.Mellor and Anderson, however, aFparently obtained much greater expansions than could be explained in this manner.VI. THE PERIOD OF INDUCTION. 1295 acceleration then ceases. Consequently, the duration of the period of induction depends on the relative rates of formation of X and of hydrogen chloride. Pringsheim’s Intermediate Compound. Pringsheim (Ann. Phys. Chem,, 1887, [iii], 32, 384) sought to ex- plain the period of induction by assuming that chlorine monoxide is formed as an intermediate compound according to the following equations : Stage I.-CI, + H,O = C1,O + H, Stage 11.-4H2 + C1,O- = H,O + 2HC1 He appears to have been guided in the selection of chlorine monoxide by the fact that ‘‘ if hydrogen chloride were formed it would be absorbed by the water,” and he observed no change in the volume of the gas during the ‘‘ period of inertness.” First, no reason is offered why the chlorine monoxide formed during this period is not absorbed by the water, the relative solubilities of chlorine monoxide and hydrogen chloride being very nearly as 5 : 1.Second, from experi- ments previously cited, a real ‘‘ period of inertness ” does not exist. The mass law also requires that immediately the smallest trace of inter- mediate compound is formed, hydrogen chloride shall be produced with an infinitely small velocity. Pringsheim also believed that the volume of the intermediate com- pound X must agree with the equation : There are two objections to this conclusion.C3, + H,O = X + &c. n volumes vz volumes. So far as our present knowledge goes, this condition is by no means binding. There is no evidence to show whether the maximum amount of the intermediate compound present in the system when the rate of formation of hydrogen chloride is a maximum is small or great. According t o Recklinghausen (Zeit. physikal. Chem., 1894, 14, 494), when a mixture of hydrogen and chlorine gases, not in direct contact with water, is illuminated, the increase in volume is proportional to the intensity of the light. Bunsen and Roscoe’s actinometer shows that if the gases are in contact with water, the volume decreases.This contraction is the joint effect of a t least four measurable pheno- mena. (1) The contraction due to the absorption of hydrogen chloride ; (2) the evolution of chlorine from the water in the insolation vessel as hydrogen chloride is absorbed (Mellor, Trans., 1901, 79, 216); (3) Recklinghausen’s photo-expansion which, as he says, “ is characteristic1296 MELLOR: THE UNION OF EYDROGEN AND CHLORINE, of a combination (of chlorine with hydrogen, &c.) induced by the light,’’ and which is sometimes as great as 4 per cent. of the whole; (4) the heat of combination of hydrogen and chlorine. If the volume of the intermediate compound does not agree with the above conditions, a fifth factor must be included, namely, ( 5 ) the change in volume due to the formation of the small or great quantity of the intermediate compound.Becquerel (WUT~Z’S Diet. de Chim., 1879, 2, 255), Veley, (Phil. Mag., 1894, [v], 37, 170), and Gautier and Helier (Compt. rend., 1897, 124, 1268), modify Pringsheim’s cycle and assume that the reaction occurs in the following way : STAGE I.-CI, + H20 = HOG1 + HCl CYCLE II. STAGE 11.-H, + HOCl = H20 + HC1 1 Veley gives two reasons for the rejection of Pringsheim’s cycle. (1) ‘‘ It is not probable that the anhydride C1,O would exist as such in the presence of water.” (2) “The investigations of Pedler . . . have shown that chlorine in the presence of water and under the influence of sunlight, the two conditions required, give HOCl as one of the intermediate products of the reaction. . .” In the first place, it is not known whether the vapour of hypochlorous acid is CI,+€€,O (vapour), or HOG1 molecules; nor is it known whether moist chlorine monoxide consists of C1,O + H20 (vapour) or HOCl molecules, yet there is no more reason to suppose that the anhydride C1,O does not exist as such in the presence of water vapour than that SO,, &c,, do not exist as anhydrides under similar conditions. I n the second place, it must be borne in mind that Pedler’s work referred to the action of light on aqueous solutions of chlorine and not particularly to chlorine gas mixed with aqueous vapour.The reaction in the former case is exothermal (Mellor, Trans., 1901, 79, 223, 225 ; see also Richardson, B.A. Reports, 1888, 89), whilst in the latter it is endothermal. There is no satisfactory method for the determination of the amount of hydrogen hypochlorite or of chlorine monoxide in the presence of chlorine, and no experimental evidence has hitherto been published which woald justify the selection of the one intermediate compound in preference to the other.A p i o r i , it appears very probable that either the first or the second of the above cycles represents the actual mechanism of the reaction under consideration. The following considerations have led me to abandon both. I have measured the duration of the period of induction and the rate of formation of hydrogen chloride in the presence of hydrogenVI. THE PERIOD OF INDUCTION, 1297 hypochlorite and of moist chlorine monoxide* in a Bunsen and Roscoe’s actinometer modified as shown in the figure.The top of the insolation vessel A was connected with capillary tubing (about 1 mm. bore) B and C. C was filled with the desired vapour under a pressure of 1-5 mm. of mercury in excess of the prevailing atmospheric pressure. This arrangement was fused at 13 to the exit tube of an ordinary actinometer so that the insolation vessels of both actinometers could be illuminated with the same light. By suitably turning the three-way cock 6, both actinometers can be filled with the same mixture of hydrogen and chlorine. The light FIG. 1. was placed so that both actinometers gave identical readings. The light was then extinguished and a little more chlorine-hydrogen mix- ture was sent through the apparatus. A and C were then placed in communication for a moment so that a little of the gas in C passed into A .As soon as equilibrium was restored (about five minutes are re- quired), the two bulbs mere again illuminated. The following numbers are four sets of readings of the two actinometers standing side by side : * The strongest solutions of hypochlorous acid, free from chlorine, can be pre- pared by passing a current of carbon dioxide through an aqueous solution of bleaching powder and distilling. The chlorine monoxide, contaminated with a little free chlorine, was prepared from dry chlorine and mercuric oxide as described in “ Roscoe and Schorlemmer.” VOL. LXXXI 4 R1298 MELLOR: THE UNION OF HYDROGEN AND CHLORINE. Time in minutes. 1 2 3 4 5 6 3 8 9 10 11 12 13 Efect of adding Hydrogen Hypochlorite.Index movements. Standard . 0 1 3 6 11* 10 - I - - - - HOCl about 1 mm. 0 0 1 2 4 8 9" - - - I - - Time in minutes. 1 2 3 4 5 6 7 8 9 10 11 12 13 Index movements. HOG1 5bou t 5 mm. ) The asterisk u * " indicates that the period of induction is over and that the velocity of the reaction is constant. Efeot cy? Adding Chlorine Monoxide. Time in minutes. 4 5 6 7 8 9 10 11 12 13 14 15 Index movements. Standard. c1,o ibout 1 mm.) - 1 3 10 18 23 21" - - - - - Time in minutes. 7 8 9 10 11 12 13 14 15 16 17 18 Index movements. Standsrd. c1,o rbout 5 mm.) Neither the presence of traces of hydrogen hypochlorite nor of chlor- ine monoxide appears to affect the rate of formation of hydrogen chloride very much, and no efect is produced o n j r s t exposure. It is assumed that if it were necessary for either chlorine monoxideVI.THE PERIOD OF TNDUCTION. 1290 or hydrogen hypochlorite to be formed before chemical action can take place, the period of induction would be abbreviated or annulled altogether when either of these substances is introduced into the system*: hydrogen I water vapour I chlorine. It may be objected (1) (‘nascent ” chlorine monoxide (or hydrogen hypochlorite) is more active, chemically, than when prepared in the usual way ; (2) since chlorine monoxide (or hydrogen hypochlorite) is very difficult to purify, impurities are certain to be present, consequently the accelerating influence of the intermediate compound is balanced, so t o speak, by the retarding influence of the impurity (compare Bunsen and Roscoe, Pld.Trans., 1857, 146, 390, et Sep.). But it must be re- membered that the induced mixture may be kept nearly 30 minutes without losing all its induced activity. This is not a characteristic of ‘‘ nascent ” activity. Again, since hydrogen is not directly concerned with the first stage of either cycle, C1, + H,O = C1,O + H,; or C1, + H20 = HOCl + HCI, the natural inference is that moist chlorine exposed to sunlight will exhibit a greater chemical activity towards hydrogen than chlorine not so treated. This is not the case. Still further, the presence of hydrogen along with the moist chlorine would, according to the mass law, drive back the formation of chlorine monoxide, and consequently this compound should be more easily formed in moist chlorine than in a mixture of moist chlorine and hydrogen.When actively combining hydrogen and chlorine gases, which have passed through the period of induction, are placed in darkness, chemical combination stops. If the gases are re-illuminated, the duration of the second period of induction is less than the first provided that the induced gases have not been more than hnlf-an-hour in darkness. Bunsen and Roscoe (Zoc. cit., p. 395) have also published other experiments which prove undoubtedly that there is a marked difference in the chemical activity of an insolated and of a non- insolated mixture of hydrogen and chlorine. This fact coupled with the experirtients recorded in this communication lead to the con- clusion that the presence of hydrogen as well as of moisture determines the greater chemical activity of the induced mixture of hydrogen arid chlorine gases.No compound formed by the interaction of water and chlorine alone will explain the indifference of insolated moist chlorine towards hydrogen; nor will the formation of any compound of hydrogen and * Jakowkin’s work appears t o show that minute quantities of hydrogen hypo- chlorite are normally present in the atmosphere above a solution of chlorine in water (Zcit. phpikak. Chenz , 1899, 29, 613),1300 MELLOR: THE UNION OF HYDROGEN AND CHLORINE. chlorine explain the part played by water in the reaction; the con- ductivity experiments of J. J. Thomson negative any dissociation or ionic hypothesis, and the law of mass action furnishes a satisfactory explanation of the period of induction provided an intermediate com- pound is postulated ; I therefore infer that ay an intermediate compound is formed a t all, it is a complex containing xC12,yH,0,xH2 (where x, y, x, are positive integers) which acts as the intermediate compound in Bunsen and Roscoe's chlorine-hydrogen actinorneter. Action of Eydrogen on Hydrogen Hypochlorite and om Chlorine Monoxide.According to Balard (Ann. Chim. Phys., 1834, [i], 57, 225), hydro- gen appears to exert no action on chlorine monoxide or on hydrogen hypochorite at the ordinary temperature." IT have confined hydrogen and nitrogen in separate eudiometers over dilute solutions of hydrogen hypochlorite for two days (diffuse daylight) at 15", but no chemical action could be detected. Similar experiments, in which the gases, confined over the aqueous hypo- chlorous acid, were heated at looo by means of a steam-jacket for 1-4 hours, gave negative results.Hydrogen hypochlorite in aqueous solution, however, is a t once attacked by nascent hydrogen from a zinc-copper couple, or sodium amalgam placed in the solution. I n the attempt to find if hydrogen attacks nascent hydrogen hypo- chlorite, sufficient carbon dioxide was passed into an aqueous solution of sodium hypochlorite to liberate all the available acid. The same amount of a mixture of hydrogen and carbon dioxide gases bubbled through a similar solution of hypochlorite made no difference to the amount of '' hypochlorite " obtained in each case. These experiments, therefore, do not support the view that eitber chlorine monoxide or hydrogen hypochlorite reacts with hydrogen during the second stage of cycles I or I1 respectively to form hydro- gen chloride as indicated in the equation C1,O + 2H2 = 2HCl + H,O ; or 2HOC1 + H, = 2HC1+ H20.I have yet t o discuss the explanation offered by the two dependent reactions suggested by the equation : * Compare also Cooke (Chem. News, 1888, 58, 103 ; from Glasgow Phil. Trans.) -k I desire to thank Mr. W. R. Anderson, B.Sc., for valuable assistance with the for the reducing action of hydrogen i n the presence of platinum. experiments on the chlorine oxides,DECOMPOSITION OF WATER VAPODR BY THE ELECTRIC SPARK. 1301 A similar cycle was indicated by Mrs. Fulhame, as far back as 1802, t o explain the essential part played by water on certain chemical reactions. Conclusions. 1. I f the reaction between hydrogen and chlorine in the presence of moisture is assumed to take place with the formation of an intermediate compound, the period of induction is a direct consequence of the law of mass action. 2. Since neither chlorine monoxide nor hydrogen bypochlorite abbreviates the period of induction, neither of these substances can take part, as intermediate compounds, in the reaction between hydrogen and chlorine. 3. Since chlorine acquires no appreciable chemical activity by exposure t o sunlight, the presence of hydrogen as well as of moisture determines the greater chemical activity of an induced mixture of hydrogen and chlorine gases. 4. If a n intermediate compound takes part in the reaction between hydrogen Fand chlorine in the presence of moisture, the most probable ‘( compound ” satisfying the required conditions contains xCl,,yH20,zH2, where x, 9, and x are positive integers. I desire to express my gratitude to Professor Dixon for his kind interest and advice during the course of this work. THE OWENS COLLEGE, M ANCHESTER.