HOMOGENEOUS DECOMPOSITION OF OZONE. 2463CCLIV.-The Homogeneous Decosnposition of Ozone inthe Presence of Oxygen and Other Gases.By DAVID LEONARD CHAPMAN and HERBERT EDWIN JONES.IT has been shown by H. E. Clarke and one of us (Trans., 1908,93, 1638) that the rate of decomposition of ozone on the surfaceof glass is so slow that even in moderately small globes the amountof ozone destroyed on t.he internal surface of the vessel may beneglected in comparison with that decomposed in the interior o2464 CHAPMAN AND JONES : HOMOGENEOUS DECOMPOSITION OFthe gas. I n other words, it has been demonstrated that the con-version of ozone into oxygen under suitable conditions may beassumed to be a homogeneous change without any appreciable errorbeing made. It is the only slow chemical change in the gaseousstate which has, as yet, been shown to satisfy the condition ofhomogeneity under re alis ab le conditions .Since the quantitative investigation of a chemical change entirelyconfined to matter in its least complex state might be expected tofurnish results of exceptional theoretical significance, an attemptwas made by Mr.H. E. Clarke and one of us to construct anapparatus with which the velocity of decomposition of ozone in thepresence of oxygen and other gases might be measured; but beforethe apparatus had been sufficiently perfected to furnish satisfactoryresults, Mr. Clarke was unfortunately compelled to relinquish thework. The investigation has been continued by the authors ofthis communication with the aid of a slightly modified and improvedform of the apparatus originally designed by Clarke and one of us.Before giving a detailed account of this apparatus, and the modeof conducting an experiment, it will be convenient to state thegeneral conclusions that have been drawn from the results, and toindicate what we believe to be the theoretical significance of theseconclusions.The results demonstrate that :(a) Oxygen, nitrogen, carbon dioxide, and possibly water vapourhave no effect on the rate of decomposition of ozone, that is, therate of decomposition of ozone in the presence of these gases is afunction of the concentration of the ozone only.( 6 ) Nitrogen peroxide (Andrews) and chlorine accelerate in amarked degree the decomposition of the gas.( c ) I f the order of the change can be represented exactly by anintegral ordinal number, that number is the second.I n respect of their influence on the rate of decomposition of ozone,gases may therefore be separated into two classes-those which arewithout effect, and those which act as powerful catalysts.That aclassification based on such a striking distinction should be possiblelends strong support to the view that the catalytic action of thesecond class is chemical rather than physical in its nature, sincea physical property is generally shared, in a greater or less degree,by all gases. Moreover, nitrogen peroxide and chlorine are sub-stances of which the first is known to react with ozone, and thesecond is closely related to an element, namely, iodine, which hasbeen shown to be oxidised by ozone.The facts, so far as they have been made out, indicate that themechanism of the decomposition of ozone in the absence of catalystOZONE IN THE PRESENCE OF OXYGEN AND OTHER GASES.2465is a sinikle process, consisting of the conversion of two molecuies ofozone during a favourable collision into three molecules of oxygen.Such a view is in harmony with the fact that gases having nochemical action on ozone are without influence on its rate of decom-position (for the number of collisions between pairs of molecules ofozone is almost independent of the diluting gas), and also with thefact that the reaction is of the second order.A result of exceptional interest is that which relates to theFIG. 1.influence of moisture.to the experiment,al section of the paper.The discussion of this will be relegatedEXPERIMENTAL.The apparatus used for the preparation and collection of theozonised oxygen is depicted in Fig.1. The oxygen was preparedby heating potassium permanganate. Dust and carbon dioxidewere removed from it by its being passed through a tube packedwith glass wool and soda-lime. It was stored in a small gas-holderA , which contained concentrated sulphuric acid. The gas-holderwas connected by narrow capillary tubing with a Brodie ozonegenerator B, made of thin glass, as recommended by Shenstone(Trans., 1893, 63, 938). The generator was immersed in dilutesulphuric acid, and its inner tube contained metallic mercury.Itwas connected by capillary tubing with a, vessel C, containing con-centrated sulphuric acid saturated with ozone. As the ozonisedoxygen passed into this vessel, the displaced acid entered th2466 CHAPMAN AND JONES : HOMOGENEOUS DECOMPOSlTION OFreservoir 11. The receptacles 1) and C were connected by a widetube a, in which a tap Tawas inserted, and also by a tube b ofvery fine bore.When T, was closed, the acid entered B very slowly, owing tothe resistance offered to its motion by the capillary tube, and therate at which the ozonised oxygen entered the receiver was cormspondingly-slow. The upper end of D was connected with a deviceby means of which the current of gas could be further regulated.The wide tube E , containing powdered potassium hydroxide, wasground into the mouth d of the receptacle B.E was in turn joinedby rubber tubing to a flask F , containing water, which could besiphoned out drop by drop through the fine capillary t-ube G. Byraising and lowering G, the rate at which the water siphoned overcould be regulated, and the flow of gas through the ozone generatorthereby controlled. The potassium hydroxide in E served todestroy traces of ozone which would otherwise have attacked therubber tube. Before use, the apparatus was cleansed with a hotmixture of potassium chromate and concentrated sulphurie acid,and then with hot distilled water, and thoroughly dried.All air was displaced from the apparatus before starting anexperiment by a, current of oxygen.Oxygen was ,collected in thegas-holder, and a volume of ozonised oxygen sufficient for oneexperiment was prepared from it. The tap Tb was then closed,and the oxygen remaining in the gas-holder A was allowed toescape. The gas in the receiver C was next transferred to theholder A , the taps Ta and Tb being left open. It was then againdrawn slowly through the ozone generator into the receiver C, the.coil being in action. The percentage of ozone was appreciablyincreased by the gas being submitted for a, second time to theaction of the silent discharge.The section of the apparatus used to measure the rate of decom-position of the ozone (at looo) is shown in Fig. 2.The glass tubes, A , and A,, of about 100 C.C. capacity, in whichthe ozone was heated, communicated by capillary tubes, on the oneside with the previously described section of the apparatus, andon the other with the left-hand limbs of the manometers m, and m2,which contained concentrated sulphuric acid.The right-hand limbsof the manometers were connected by capillary tubes and groundglass joints with two bottles, B, and B,, of about 2 litres capacityeach. The corresponding parts of the apparatus were made asnearly alike as possible. The apparatus was connected with aninjector pump at P, and was provided with a mercury manometerat M , as shown in the diagram. Taps TI, T,, and T3, and tapsti, t,, t,, t 4 , t,, $6, t,, t,, and t,, provided with mercury seals, werOZONE I N THE PRESENCE OF OXYGEN AND OTHER GASES. 2467inserted in the positions indicated in the figure.The pressure onthe rightrhand side of the manometers was kept constant by thebottles being immersed in a bath of water at a fixed temperature.The water in the bath was stirred by a current of air, and thetemperature was controlled by a delicate electric thermoregulator.During an experiment the tubes A , and A , were kept at loooby means of a current of steam, which entered the jackets sur-rounding them from above. I n it preliminary experiment, it wasshown that the temperature of both of the tubes could be raisedto that of the steam in the same time. Before the experimentsJ twere started, ozonised oxygen was passed through the heated tubesfor several hours.Znfluence of Oxygen on the Rate of Decomposition of Ozone.‘The object of the first series of experiments was to ascertain theeffect, if any, of varying concentrations of oxygen on the rate ofdecomposition of the ozone.The method of conducting an experi-ment was as follows.A quantity of ozonised oxygen sufficient for one experiment wascollected in C, the ozone remaining in the generator and capillarytubes being subsequently driven out through the taps T, and T, bya stream of oxygen from the holder A .With the taps t,, t,, and T, closed, and t,, t,, t,, t,, t,, t,, and t,open, the apparatus was exhausted as completely as possible by theinjector pump. The taps t,, t,, and t , were then closed, and oxyge2488 CHAPMAN AND JONES : HOMOGENEOUS DECOMPOSITION OFadmitted from the holder A through t, and t , until the volumebetween the taps t , and t , and between the taps t , and t, had beenfilled.The pump was again set in action, and the taps t,, t,, andt , cautiously opened. When the limit of exhaustion attainablewith the injector pump had been reached, the process describedabove was repeated, the removal of traces of air from the tubesA , and A , being thereby ensured.Bycautiously opening t , and t,, ozonised oxygen was admitted to thetubes A , and A , from the receiver C, the pressure on both sidesof the manometer being maintained the sanie by the simultaneousadmission of air into the bottles B, and B, through the taps t , andT,. When the manometer M indicated a pressure of a little lessthan half an atmosphere, t , was closed.The ozone left in thecapillary tubes on the left-hand side of the taps t, and t , was dis-placed by oxygen, and oxygen was then introduced into the tubeA , by carefully opening the tap t,, air being at the same timeadmitted through t,. The taps t, and t , were closed when themanometer m indicated a pressure slightly less than an atmosphere.The tube A , was thus filled with ozonised oxygen at the pressureof half an atmosphere, whilst A , contained the same amount ofozone, but approximately twice as much oxygen.The tubes having been filled, the taps t, and $8 were opened,and a rapid current of steam wm passed through the steam jackets.After one an& &half minutes (when the contents of the tubes hadattained the temperature of the steam), the taps t , and t 8 wereclosed.The differences of pressure registered by the manometerswere noted a t regular intervals. Curves were plotted, showing therelation between the increase of pressure in the tubes A , and A ,and the time. It sometimes happened that the total amounts ofozone contained in A , and A , respectively (as indicated by the totalchange of pressure) were not exactly equal. In such cases a simplecorrection was applied in order that the results might be strictlyconip ar ab le.The changes of pressure in cm. of sulphuric acid are plottedagainst the times, and the four pairs of curves thus obtained areshown in Fig. 3. The circles correspond to the changes ofpressure in the tube which contained ozonised oxygen at apressure of half an atmosphere, and the crosses to changes ofpressure in the other tube which contained the same amount ofozone per unit volume, but twice as much oxygen.The numbersattached to the curves indicate the order in which the experimentswere performed.These results point decisively to the conclusion that variation inAfter the final exhaustion the taps t,, t,, and t , were closedOZONE IN THE PRESENCE OF OXYGEN AND OTHER GASES. 2469the pressure of the oxygen mixed with ozone is unattended byappreciable alteration in the velocity of decomposition of the lattergas at looo. This conclusion does not appear to agree with theobservations of previous investigators.S. Jahn (Zeitsch. amrg. Chem., 1906, 48, 260) and Perman andFro.3.7 I -I I___I____ t-- -- 1 1-ripTime : 1 division=+ hour.Greaves (Proc. Roy. Soc., 1908, A , 80, 353) assert that the rate ofdecomposition of ozone varies approximately inversely as the oxygen-pressure. If the reaction were reversible, this result might beexpected; but Perman and Greaves and others have conclusivelydemonstrated that at looo it may be regarded as irreversible.".Ic Fisher and Braehmer (Ber., 1906, 39, 940), have recently shown that ozone isformed in srnall quantities from oxygen a t teinperatures above 1300".VOL. XCVII. 7 2470 CHBPMAN AND JONES : HOMOGENEOUS DECOMPOSITION OFI n order to explain his results, Jahn has suggested that the decom-position of ozone occurs in two stages, expressible by the equations :o,=o,+o .. s . . . . . ( a )0,+0=20, . . . 4 . . . . ( 6 )the first stage being rapid and reversible, whilst the second is slow.These assumptions would require that the rate of decomposition ofthe ozone should be directly proportional to the square of theconcentration of the ozone, and inversely proportional to the con-centration of the oxygen. Jahn’s experiments were carried out at127O, at which temperature the reaction is bimolecular, accordingto Waxburg (Sitzungsber. R. Akad. Wiss. B e r l i n , 1901, 48, 1126).Perman and Greaves, on the other hand, consider that Jahn’shypothesis is not justifiable, and that the alleged effect of theoxygen is due to variations in the gas film on the glass surfaceresulting from the changes in the oxygen-pressure.If, however,the decomposition occurs mainly in the body of the gas, as appearsfrom the work of Clarke and one of us (Zoc. cit.), the suggest.ionmust for that reason alone be discarded,Imfluence of Aqueous Vapour om the Rate of Decomposition ofOzone.A second series of experiments was performed in order toinvestigate the influence of water vapour on the rate of decom-position of the gm-a subject of considerable interest, both onaccount of the diversity of the .results obtained by previous workersand of its bearing on the general problem of the catalytic effect ofmoisture on most simple chemical changes.The experiments were carried out as described above, except tha’tboth tubes were filIed with ozonised oxygen at a pressure of a littleless than an atmosphere, and that the gas which entered the tubeA , was saturated with water vapour by its being passed througha small wash-bottle containing distilled water.To prevent anymoisture being carried over into A,, a tap was inserted on the left-hand side of the wash-bottle, which was shut off while A , was beingfilled.The eight curves obtained by plotting the results of four experi-ments are shown in Fig. 4. Ir? experiments IIIa and IVa, themoist gas was contained in the tube which held the dry gas in thetwo previous experiments. The crosses correspond to the changesof pressure of the moist gas, and the circles to the changes ofpressure of the gas dried by concentrated sulphuric acid.Andrews and Tait, and also Brodie, in their classical memoirs oOZONE IN THE PRESENCE OF OXYGEN AND OTHER GASES.2471ozone, recommend the use of carefully dried oxygen for the pre-paration of Ozone by the silent discharge.Shenstone and Cundall (Trans., 1887, 51, 610) showed thatcarefully dried oxygen can be easily converted into ozone-a factwhich was shortly afterwards confirmed by Baker (Trans., 1894,65, 611), who stated that ‘‘ ozone was formed as rapidly in oxygendried with phosphorus pentoxide as it was in the same tube whenthe oxygen had been dried only by sulphuric acid.”As a result of further investigations, Shenstone (Trans., 1897,71, 471) drew the remarkable conclusion that all previous stateFIG. 4.ments on the subject were wrong. He observed that a high per-centage of ozone is formed by the action of the silent discharge onoxygen saturated with water vapour, aad that the ozone thusproduced is remarkably stable.On partly drying the gas, thepercentage of ozone produced was considerably reduced, and thegas was found to be singularly unstable. Oxygen which had beenthoroughly dried was found to become ozonised exceedingly badly.Subsequent investigations have failed to confirm Shenstone’s work,and Armstrong has suggested that his anomalous results may be dueto the presence of oxides of nitrogen (formed by continuous action of7 u 2472 CHAPMAN AND JONES : HOMOQENEOUS DECOMPOSlTION OFthe discharge from traces of nitrogen contained in the oxygen),which, as Andrews has shown, immediately destroy ozone.Thomson and Threlfall (Proc.Roy. SOC., 1885, 40, 340) assertthat ozone is produced when an electric spark is passed throughvery carefully dried oxygen.Warburg (loc. cit.) maintains that at looo the dry gas is justas stable as the moist.Warburg and Leithauser ( A n n . Physik, 1906, [iv], 20, 751)have, moreover, shown that the formation of ozone both in oxygenand in air is retarded by the presence of moisture, the retardationbeing greater in oxygen than in air.Fischer and Marx (Ber., 1906, 29, 3631), working d t h a Nernstfilament, find that the first traces of moisture lessen the yield ofozone by catalytic action, whereas larger quantities of water vapourincrease the yield of hydrogen peroxide at the expense of the ozone.Perman and Greaves (Zoc.cit.) claim to have shown that watervapour accelerates the decomposition of ozone, and that the effectis roughly proportional to the amount of water vapour present.They consider that the effect is due to the deposition of moisture onthe surface of the glass, which causes the ozone to be more rapidlycondensed. They point out that their results do not agree withthose of Shenstone.Although in our experiments the Oaone mixed with a considerableproportion of water vapour appears to decompose at a slightlygreater rate than that which has been dried with sulphuric acid,the difference is so small that we are disposed to think that it oughtto be attributed to the gradual removal of water vapour, adsorbedon the inner surface of the glass, at the higher temperature, or tosome similar cause.We are, at least, justified in concluding thatat 1000 a large difference in tho quantity of water vapour presentwith the ozone is not accompanied by any marked change in thevelocity of decomposition. Our results agree with those ofWarburg, whose experiments were also conducted at looo.Influence of Nitrogen, Carbon Dioxide, Carbon Monoxide, andChlorine o n the Rate of Decompaitwn of Ozone.The negative character of the results obtained with the first threegases is sufficient proof that no appreciable quantity of impuritycapable of destroying the ozone was contained in them. I n eachcase concentrated sulphuric acid was used to dry the gas. Theexperiments were conducted as follows.A tube containing soda-lime was introduced between the gas-holder and the ozone generator in order to remove any carbondioxide from the oxygen.The tubes A , and A , were filled witOZONE IN TRE PRESENCE OF OXYGEN AND OTHER GASES. 2473ozonised oxygen at a pressure of half an atmosphere in the mannerpreviously described. Oxygen was then admitted to one tube untila pressure slightly less than an atmosphere was registered by themanometer. The gas of which the effect was being investigatedwas then introduced into the other tube through T, or T, until thepressure in both tubes was the same. The rates of decompositionof the ozone in the two tubes were compared. Several experimentswere performed, the gas under investigation being introduced intoA , and A , alternately.r-r-I-r * FIG.5.T h e : 1 division = 0.5 hour.I n the case of nitrogen, no effect on the velocity of decompositionwas observed, while the experiments with carbon dioxide and carbonmonoxide demonstrate that the influence, if any, of these gases issmall. Chlorine, on the other hand, was found to decompose ozoneso rapidly that it was quite impossible to make any trustworthymeasurement of the velocity of decomposition.The six curves obtained by plotting the resulbs of three experi-ments with nitrogen are shown in Fig. 5. I n the experiments I1and 111, the nitrogen was contained in the tabe which in theprevious experiment held ozonised oxygen only. The crosse2474 CHAPMAN AND JONES ; HOMOGENEOUS DECOMPOSITION OFcorrespond t o changes of pressure in the gas which containednitrogen, and the circles t.o the changes of pressure in the othertube.The readings taken in the experiments with carbon dioxide I andcarbon monoxide I1 are plotted in the curves shown in Fig. 6.Thecircles correspond to changes of pressure in the tube whichcontained the ozonised oxygen only.FIG. 6.Time : 1 dkision=O*5 hour.Shenstone and Evans (Trans., 1898, 73, 246), while investigatingthe influence of the silent discharge on atmospheric air, weresurprised to find that much as 98 per cent. of the oxygencontained in the air could be converted into ozone, the maximiuliyield of ozone obtained from pure oxygeii under the sa.me con-ditions being only 13.6 per cent. A similar observation was madeby Brodie (Phil. Trans., 1874, 164, 101) when he submitted carboOZONE IN "HE PRESENCE OF OXYGEN AND OTHER GASES.2475dioxide to the action of the same agency; the carbon dioxide wasdecomposed into carbon monoxide and oxygen, 85 per cent. of theoxygen being in the form of ozone. It was conceivable that theseinteresting and peculiar phenomena might arise from inhibitiveeffects on the thermal decomposition of ozone of nitrogen and carbondioxide respectively. The above experiments demonstrate that suchan explanation is untenable, and that the phenomena in questionmust be due to some obscure influence of the nitrogen on the onehand, and the carbon dioxide or carbonmonoxide on the other, onthe energy of the discharge itself.Observations on the Order of the Change.In examining our results with a view to det'ermining the orderof the change, we have adopted a novel method.Instead ofcalculating the value of the constant for each observation on theassumption that the change is of a given order, we compare theamounts of ozone decomposed at given intervals of time with thosecalculated for changes of a specified order. This method ofexamining the results has the obvious advantage of enabling us t odecide at a glance whether the departure of the experimentalnumbers from those calculated from any given set of assumptionsas to the nature of the change lie within the limits of experimentalerror.A curve of a given order is drawn through three points on theexperimental curves, the points selected being the origin, and thepoints corresponding with the last observation, and am observationmade when about half of the ozone was decomposed.The differencebetween the magnitudes of ordinates (amounts of ozone) of theexperimental and calculated curve at different times are thentabulated. In the present case it was only necessary to comparethe experimental curve with calculated curves drawn through itof the first and second order. It will be seen from the numbersgiven that the reaction is very nearly of the second order. Onlythe numbers obtained in those experiments in which large quantitiesof dry ozone were decomposed, and in which readings were takenfor a considerable period of time are submitted to examinationbelow2476 HOMOGENEOUS DECOMPOSITION OF OZONE.Examination of the Measurements Made in Experiment I V a .(Al.)Time.2510165897178238298397508CalculatedObserved change ofchange of pressure, Difference :pressure. 1st order. k, = 0'00648.0.60 0 -36 -I- 0 '241 -30 0'91 +0-391 *go 1 -68 +0'222.95 2.57 4- 0.387.00 7-00 08 *50 9 '26 + 0.7110.30 11-26 +0'9610.95 11.76 t o 3 111-40 11-97 + 0.5711.85 12.08 + 0-2312'10 12.10 0Sum of difTerences= + 4.51Obs-rvedchange ofpressure.0.601-301 -902 -957-008.5010.3010.9511-4011 -8512'10SumCalculatedchange ofpressure, Differcnce :2nd order. k2=0'00142.0 -48 + 0'121-16 + 0.142-13 - 0'233-11 - 0.167.00 08-65 + 0.10l O - O i - 0.2310.93 - 0'0211'34 - 0.0611.78 - 0.0712.10 0of differences = - 0.51The calculated curves are drawn through the origin, the point(58, 7*00), and the point (508, 12.10).If the difference betweenthe ordinates of the calculated and experimental curves indicatesthat the order of the reaction is of a higher value than thatcorresponding with the calculated curve, a plus sign is attached toit, a minus sign having the reverse significance.It will be evident on inspection of the above numbers that thereaction differs but slightly from one of t,he second order.Examination of the Measurements made in Experiment IZZa.closeIy to a reaction of the second order.(Al.)I n this experiment the rate of change approximates still moreTime.136912182430516075105129156276381489531CalculatedObserved change ofchange of pressure, Difference :pressure. 1st order. k, = 0-00624.0.25 0.17 + 0.080-75 0-50 + 0'251 -55 0 -95 f 0.601.95 1'44 + 0.512-40 1 '88 t-0'523'15 2.69 +0'463-85 3-26 -i- 0.594 *50 4.15 -t 0.356-10 6-16 - 0'066-85 6 3 5 0 .oo7-50 7-82 f - 0 - 3 28.60 9 '24 + O 649-15 10.00 + 0.859-75 10*60 +0%511'40 11.81 + 0'4111.80 11 -85 + 0'0511.85 11.85 0 .ooSum of differences = +7.0611 .oo 11 -63 -t 0'63CalculatedObserved change ofchange of pressure, Difference :pressure. 2nd order. k2= 0~00141.0-25 0 '23 +0'020.75 0-69 +0-061.55 1 -30 + 0'251.95 1.86 -k 0.092'40 2.36 -t 0.043'15 3 '25 - 0-103.85 4'00 - 0.1 54'50 4 '64 - 0'146'10 6.35 - 0.256.85 6 -85 0 .oo7.50 7'60 t-o.108 60 8'61 + 0.019-15 9'19 + 0'049.75 9 '69 - 0.0611 .oo 10.92 - 0'0811'40 11.43 + 0'0311 -80 11'76 - 0 '0411.85 11 -55 0.00Sum of differences = - 0-1METHYL-ORANGE AND METHYL-RED. 2477I n the other experiments, in which the initial percentage of ozonewas less, or the period during which observations were madeshorter, the order of the change falls to a greater extent below thesecond. All the experiments are conclusive in demonstrating,however, that if the order of the change can be represented by anintegral ordinal number, that number is the second.SIR LFNLIXE JENKINS LABORATORIES,JESUS COLLEGE, OSFORD