HAROLD S . JOHNSTON 21 A THERMAL METHOD OF INVESTIGATING FAST GAS-PHASE REACTIONS PART I.-THE MERCURY PHOTO-SENSITIZED DECOMPOSITION OF ETHYLENE BY A. B. CALLEAR AND JAMES C. ROBB Dept. of Chemistry, The University, Birmingham, 15 Received 22nd December, 1953 This paper describes a method of measuring rates of chemical reactions in the gas phase which occur so quickly as to render inadequate any conventional experimental method based, for example, on pressure mcasuremcnts. The mercury photo-sensitized decomposition of ethylene has been investigated by this method and shown to occur by a mechanism somewhat simpler than that which has been previously proposed. INTRODUCTION The stationary temperature which a reacting gaseous system attains is obviously some measure of the rate of reaction.In photochemical reactions, this temperature is usually less than about 0.1" C above ambient, but it has proved possible to measure rapidly the steady temperature with a reasonably high precision. In a study of the mercury photo-sensitized decomposition of ethylene, investigation is complicated by the products of reaction, hydrogen and acetylene. The hydrogen produced quenches the excited mercury atoms, so forming hydrogen atoms which then hydrogenate the ethylene to ethane and butane. As will be seen later, pressure measurements, which can be made only by allowing a significant fraction of re- action to occur, may be ambiguous, especially at low pressures of ethylene. This thermal method allows measurements of rate of reaction to be made accurately when only an insignificant amount of reaction has taken place, thus removing the necessity for considering the secondary reactions of the products.The major features of the mercury photo-sensitized decomposition of ethylene are well established. (1) The overall process can be written C2& + Hg3P1 --f C2H2 + H2 + Hgl&22 THERMAL METHOD The rate of decomposition of ethylene decreases with increasing pressure (for a constant quantum input of 2537A radiation) and this has been attributed to the formation of an excited molecule C2H4* in the initial step, which can either decom- pose or be deactivated by suitable gas-phase collision. The complete reaction sequence would then be represented by (2) C2H4* -+ C2H2 + H2 (3) C2H4* + M +- C2H4 + M (4) C2H4 + Hg3P1 --f c2&* + HgWo Leroy and Steacie 1 concluded that for a constant quantum input, the dependence of rate of decomposition on total ethylene pressure is rate = A/(l + BP) where A and B are constants and P is the total pressure of ethylene. This leads to a bimolecular deactivation process.Darwent2 showed that the results of Leroy and Steacie were in better agreement with the relationship rate = A(l + BP2) (which would imply a three-body deactivation process) although in a subsequent publication3 he put forward a more complex mechanism involving a partial heterogeneous decomposition of the excited ethylene molecules. The main evidence for this suggestion is the variation of rate with mercury-vapour pressure. He found that lowering of the mercury-vapour pressure caused a decrease in the rate of reaction although the absorption of 2537A radiation was complete, and con- cluded that at the low mercury-vapour pressures, the excited ethylene molecules, produced farther from the vessel walls, had less chance of undergoing hetero- geneous decomposition.Further, Leroy and Steacie 1 found a maximum rate of reaction at about 10 mm pressure of ethylene and suggested that the fall-off in rate below this pressure was due to lack of quenching of Hg* by ethylene. Darwent 3 points out that the measured quenching cross-section of ethylene for Hg* is such that no fall-off in quenching should occur at 10 mm pressure and suggests that the fall-off in rate is accounted for by the deactivation of metastable mercury atoms (Hg3Po) on the walls of the vessel.He assumes that ( 5 ) is slow compared with reaction (2). C2H4 + Hg3Po + c2H4* + HglSo THEORETICAL CONSIDERATIONS The thermal conductivity of most simple gases is independent of pressure if the mean free path is small compared with the distance the heat is conducted. If a thin wire in a gas is maintained initially at constant temperature, the heat loss at low pressures falls off due to the appearance of a temperature discontinuity at the gas-solid interface. At any fixed pressure, the heat loss is proportional to the temperature of the wire, until at high temperatures convection occurs. The temperature differences attained in the gas phase in the investigation of this photo- chemical reaction are about 3 x 10-1" C , with a quantum input of about 1014 quantalcrnz sec.The platinum filament used to measure the temperature changes is maintained at about 0.5" C above the temperature of the thermostatted reaction vessel. In the reaction system and conditions considered here, heat is trans- ferred by conduction solely up to pressures of the order of 200 mm. Fig. 1 shows the absorption of light in a semi-infinite reaction vessel, bounded by a plane transparent window. The vessel contains mercury vapour and a gas which can quench excited mercury atoms. The absorption of light will be ex- ponential and I = I0 exp (- kx),A . B. CALLEAR AND JAMES C. ROBB 23 where k is the extinction coefficient, I is the light intensity per unit area at a point distant x from the window, and I0 is the incident light intensity per unit area.Between x and co the amount of energy appearing in the gas as heat is Iohv exp (- kx). When thermal equilibrium is attained, all the heat must be conducted back to the plane window, maintained at constant temperature. The thermal conductivity K FIG. 1.-Variation of light intensity and temperature in a semi-infinite reaction vessel. is related to the temperature gradient d0/dx and the amount of heat Q conducted per unit area by the relation The amount of energy conducted through unit area distant x from the window is Iohv exp (- kx) per unit area. It follows that Q = - Kdekdx. Kd0ldx = Iohv exp (- kx), whence where A0 is the difference in temperature between the window and a point x. Thus A0 is proportional to the energy input to the gas and inversely proportional to the thermal conductivity of the gas.When 2537A radiation is being absorbed by mercury vapour at its vapour pressure at 30" C, A0 becomes almost constant for x greater than 5 mm. In the practical case, this idealized state of affairs will not be realized since the finite geometry of the reaction vessel will impose disturbances on such a tem- perature distribution. In another paper 4 it has been demonstrated experimentally that for a system of finite size, the temperature difference between any point in the vessel and the walls is proportional to the heat input. A0 = Iohv[l - exp (- kx)]/Kk, EXPERIMENTAL APPARATUS REACTION vmEL.-The Pyrex reaction vessel A (fig. 2) has an upper ground flange to which is fixed a quartz window B. The resistance thermometer D, a platinum wire 10-2 mm diameter, is held in position parallel to B and 7 mm below it, by two thin platinum supports C.Electrical contact is made through the base of the reaction vessel by tungsten seals and finally by insulated connections, through the thermostatting liquid. The vessel is connected t o the high-vacuum line by a side tube E. The flange on the thermostat lid limits the light input and serves as a 4-cm diameter stop. The resistance of D is about 60 ohms a t room temperature. THERMosTAT.-The thermostat consists of a lagged rectangular tank, 15 in. X 11 in. and 14 in. deep, fitted with a close-fitting lid with a circular hole in the centre. The thermistor F forms one arm of an a.c. bridge network and controls the temperature to 10-3" C.Insulated heating wire is stretched near the inside of the bottom of the tank. The thermostatting liquid, a solution of sodium chloride in water, is stirred vigorously through the bottom of the tank. The level of the liquid is maintained constant and above the level of the flange on the lid to prevent ripples forming on the liquid surface through which the light passes.24 THERMAL METHOD ULTRA-VIOLET LAm.-The lamp G is of the low-pressure mercury discharge type in the form of a U cooled by circulating water at 25 5 0.5" C. The lamp is supplied by a Tinsley a.c. stabilizer and a transformer. The u.-v. radiation (2537A) is admitted to the reaction vessel by opening a Pyrex shutter H. MEASUREMENT OF TEMPERATURE.-The platinum wire in the reaction vessel forms onc arm of a d.c.bridge network. The off-balance signal is amplified and displayed on an oscilloscope screen.4 This signal is then reduced to zero by varying the e.m.f. across the bridge. If the u.-v. is admitted to the reaction vessel, the oscilloscope trace is de- E flectcd. With about 30 mm of C02 in the system, the half-life time of the temperature rise in the gas is about 1 sec. This enables one to discriminate the slight rise in temperature of the vessel and supports of half-life time of 1 min. MANOMETER.-A modified Pear- son differential manometer was used, having a sensitivity factor 5 of 200. The best performance was obtained using a platinum electrode with a sharp point, to detect theposi- tion of the mercury surface. Repro- ducibility was improved by sparking the electrode with a Tesla coil dis- charge before use.Electrical con- tact was observed using a triode valve circuit, the mercury contact applying a negative bias to the grid. PREPARATION OF GASES.-The inert gases were supplied spectroscopically pure by the British Oxygen Company. The carbon dioxide was obtained pure by distillation and fractionation of the solid. It was dried over phosphorus pentoxide. The ethylene was obtained from a cylinder supplied by the British Oxygen Company. It was purified by repeated distillation and fractionation. r'"-*->q --L FIG. 2.-Reaction system. RESULTS The data presented herc have been measured at 32" C unless otherwise stated. ARGON.-cUrVC A, fig. 3, demonstrates how the measured temperature varies with argon pressure, when a constant intensity of 2537A radiation is admitted to the reaction vessel.Argon is a very inefficient quencher of Hg* and the temperature measured by the wire arises largely from radiation falling directly on to it. The sharp rise in tem- perature at low pressures is due to thermal conduction from the wire decreasing rapidly due to the temperature discontinuity, while the slight rise at higher pressures is attributed to quenching of the Hg* by small traces of impurity in the argon. CARBON ~ ~ o x r ~ ~ . - C u r v e B, fig. 3, shows the variation in temperature with pressure for constant light input. Carbon dioxide quenches Hg* fairly efficiently and the fall-off in temperature bclow about 50 mm pressure is attributcd to decreased quenching of Hg* while above 50 mm, quenching is almost complete.That this fall-off is not due to thermal conductivity changes of the system can be demonstrated by using a mixture of argon and C02. The high argon content of the mixtures maintains the thermal conductivity of the system constant while the temperatures can be measured down to low C02 partial pressures. Curves C and D, fig. 4, are obtained for such mixtures containing respectively 14 % and 5.2 % C02. To compare the temperatures measured in C and D with those in curve B, it is necessary to correct those temperatures obtained in the C02 + argon mixtures, to the equivalent in CO2. This is easily done since the thermal conductivities of argon and C02 are known and have been measured in this apparatus.4 This cor- rection puts thc curves B, C and D on the same temperature scale so that the hcat pro- duced in the system is proportional to the measured temperature.Plotting the adjusted curves C and D, together with curve B against the partial pressure of C02, producesA . B . CALLEAR A N D JAMES C . ROBB 25 one common smooth curve. This shows that the fall-off in temperature at low C02 pressure is not a function of the thermal conductivity, but solely due to decreased quench- ing of Hg" by C02. Further, the coincidence of these curves leads to the conclusion that the wire is not affected by diffusion of excited atoms and molecules since the variation 14 __o_ T +--- 6 A FIG. 3.-Variation of temperature in reaction vessel as a function of pressure : (A) argon; (B) carbon dioxide.in diffusion constant would in this case show temperature changes with alteration of pressure and composition of the gas mixture. The form of the curve for the fall-off in quenching a t low pressures is the same for a number of gases investigated. For C02, quenching in the reaction system falls off to 50 % at a C02 pressure of 0.73 mm. ETHYLENE.-when ethylene in the presence of mercury vapour is irradiated with light of wavelength 2537 A, the curves obtained for the temperature of the gas as a function of pressure are more complex since in this case it is necessary to take into account the energetics of the decomposition reaction. Fig. 5 shows four curves, E, F, G and H, obtained in this way. The gas compositions are respectively 100 % ethylene, 10-4 % ethylene, 1.07 % ethylene, and 0.09 % ethylene respectively, argon again being the diluent to maintain the thermal conductivities constant for each run over the pressure range Eta/ Pressure (nrm Uy ) /uo -2 5u I .___-----I--- FIG.4.Variation of heat appearing in the reaction vessel as a function of total pressure : (C) argon 86 %; carbon dioxide 14 %; (D) argon 94.8 %; carbon dioxide 5.2 %. involved, The curves are plotted as heat liberated in the gas in arbitrary units as a function of pressure. The heat liberated is obtained by multiplying the temperature observed by the thermal conductivity of the gas mixture. For the curves G and H, where the ethylene concentration is low, the Hg* is still quenched solely by ethylene. Since the excited ethylene molecules so produced are deactivated only by collision with argon, then on either curve, for a constant total pressure,26 THERMAL METHOD the ratio of the number of ethylene molecules decomposing to the number being de- activated is constant.It is possible to estimate the fall-off in quenching as follows. Suppose a point P is chosen on curve H and at that total pressure the corresponding point on curve G is Q. Assuming that at P, quenching is 50 %, then from the form of the curve obtained in fig. 3 for C02, which is generally applicable to the fall-off in quenching by any gas, and knowing the ethylene pressures at P and Q, it is possible to calculate the per cent. quenching which should be occurring at point Q. It is then easy to calculate the heats which would be measured if quenching were 100 % at P and at Q.These shoiild both give the same value. Failure to obtain the same value means that the assumption of 50 % quenching at P is incorrect and another point is chosen and the process is repeated. FIG. 5.-Variation of heat appearing in the reaction vessel as a function of total pressure : (E) pure ethylene ; (F) argon 89.6 %; ethylene 10.4 %; (G) argon 98-93 % ; ethylene 1.07 "/o ; (H) argon 99.91 % ; ethylene 0-090 % By trial, or by a graphical method, the point P can be found which corresponds to 50 % quenching. The method can be repeated for various quenching pcrcentages and table 1 gives the information obtained. It is seen that quenching of Hg* by ethylene is 50 % at an ethylene pressure of 0.050 mm. Leroy and Steacie concluded that quenching in their experiments dropped to 50 % at about 1 mm pressure compared with 89 % quenching at 1 mm pressure in thesc experiments.If the Hg* reradiates its cnergy, and imprisonment of this radiation does not occur, the fraction of the total radiation which is quenched is ~ / ( t + r), where r is the average lifetime of Hg* and t is the average time between collisions for Hg* and ethylene molecules. At 1 mm pressure of ethylene both t and are about 10-7 sec, leading to a value of 50 % quenching. An approximate pressure of C2H4 % radiation graphical treatment of the imprisonment cffect which occurs in practice, leads to this formula being modified to 67/(f i- 67), for a system of the dimensions used hcre. This gives a value of about 86 % quenching at 1 mm.The value of 89 % quenching found experimentally a t 1 mm is in good agreement with the lifetime of Hg*. The curve E obtained for pure ethylene can be corrected for fall-off in quenching and assumes the form shown in fig. 6, curve J. From 1-5 mm pressure the heat liberated in the gas is practically constant. The shape of this curve is described qualitatively as follows. At suficicntly high pressures, practically all the CzH4* formed is deactivated by collision, so that nearly all the radiation entering the system will appear as heat in the gas. As thc pressure is reduced, decomposition of C2H4* will become more significant and a fraction of the energy is absorbed as a result of the chemistry which occurs. Thus the heat liberated in the system will decrease.At low enough pressures, nearly all the CzH4* will decompose. An einstein of 2537 8, is equivalent to an energy of 112.2 kcal. The heat of the reaction TABLE 1 rnrn Hg quenched 4 3 2 1 0.50 0.050 g8*0 96*7 93*5 88.5 81.0 50.O c2H4 -> C2H2 + H2A . B . CALLEAR AND JAMES C . ROBB is calculated to be - 41.71 kcal/mole from the following data : 6 heat of combustion of H2 f; 68-32 kcal/mole at 25^ C to liquid water, 9, 3, C2H4 = 337.23 ,, 9 , ,, Y > 9 , C2H2 = 310.62 ,, Y, 9 , 27 In curve J, the heat liberated is still rising slightly at 150 mm pressure with pressure increase. The curve is asymptotic to a value obtained when all the energy entering the system is appearing as heat. This value was obtained by replacing the ethylene by COz, measuring the heat appearing in the system, and correcting with the thermal conductivities of ethylene and C02, since the quantum input is the same in both cases.This value is marked at Y. The heat measured at low pressure is marked at Z. If the zero on the heat axis is V, then YV is a measure of the quantum input and ZV should correspond to the energy entering the system less the energy of reaction, if each quantum of radiation at low pressure decomposes one molecule of ethylene. Thus for 100 % primary quantum efficiency, YZ corresponds to the heat of reaction, - 41-71 kcal/mole and YV to the energy in an einstein of radiation, 112.2 kcal so that YZ/YV should be 0.37. The value measured is actually 0.36, in agreement with a primary quantum efficiency of 0.97. FIG. 6.-Variation of heat appearing in the reaction vessel for pure ethylene, after correction for fall-off in quenching.EFFECT OF INERT cAsm.-Curve F, fig. 5, can be corrected for fall-off in quenching and gives the dotted line. It will be seen that the heat measured at low pressures in this case is about 10.4 arbitrary units compared with 10.2 measured in pure ethylene, curve J, fig. 6. This close agreement supports the validity of the method. Curves such as F, with added inert gas can give information on the deactivation of C2H4* by collision with the inert gas molecules. For any point X on a curve such as J, it can be seen that XZ is a measure of the number of C2H4* being deactivated. One can deduce the percentage deactiva- tion at any pressure. Table 2 shows the percentage of C2H4* being deactivated at 50 pressure of inert gas.TABLE 2 He 10 Kr 23 A 29 c2H4 63.4 deactivation gas yo deactivation deactivation gas yo deactivation CONCLUSIONS DRAWN FROM THE ABOVE EXPERIMENTS mm (i) The observations are in agreement with a primary quantum yield close to unity. This is not in agreement with the observations of Leroy and Steacie 1 and Darwent 3 who find a yield of 0.37 at about 10 mm pressure of ethylene. (ii) The information presented here demonstrates that at pressures above 1 mm, quenching of Hg* by ethylene is virtually complete, while Leroy and Steacie 1 found about 50 % quenching at 1 mm.28 THERMAL METHOD (iii) Darwent has put forward a complicated mechanism whereby C2H4* partly decomposes on the vessel wails. If this was the case, at low pressures of ethylene, curve J would show a marked decrease in the heat appearing in the gas, due to energy being transported to the walls of the vessel.His main experimental justification for this mechan- ism lies in his observations that the rate of photolysis decreased with decreasing mercury vapour pressure. Points (ii) and (iii) have been further investigated by carrying out the photolysis of ethylene and measuring the rate of reaction by manometric methods. MANOMETRIC RESULTS RATE OF REACTION AS A FUNCTION OF ETHYLENE PREssuRE.-In fig. 7 are shown measure- ments at 20" C of pressure as a function of time as the photodecomposition proceeds. The initial increase in pressure arises from the decomposition of ethylene to acetylene and hydrogen and the subsequent decrease in pressure arises when the hydrogen itself reacts to give atoms and hydrogenation of the ethylene occurs.A cursory inspection of these curves would lead one to the belief that the initial rates of decomposition are quite strongly pressure dependent, but this is not the case. The complication of subsequent hydrogenation becomes important quite early on in the reaction. Fig. 8 shows the initial rates of pressure rise measured with the sensitive manometer over the first minute of reaction, when the extent of reaction is so small that the amount of hydrogen produced is sufficiently FIG. 7.-The pressure increase during the photolysis of ethylene, at different total pressures, carried out at 20" C. (1) 6.5 mm Hg pressure (3) 3.0 mm Hg pressure (2) 4.7 mm Hg pressure (4) 1.7 mm Hg pressure.small not to quench a significant amount of Hg*. Now it is clearly seen that the rate of decomposition is practically constant, falling off by only a few per cent. at 1 mm pressure. The rate of decomposition as a function of pressure for constant light input is shown in fig. 9a. This is in close agreement with the deductions made from the temperature measure- ments above and demonstrates a possible explanation of the results of Leroy and Steacie. EFFECT OF MERCURY VAPOUR PRESSURE ON RATE.-Darwent's 3 observations of variation of rate of photosensitized decomposition with change in mercury vapour pressure is inconsistent with curve J at low pressure. An alternative explanation of his results would be that his radiation was not sufficiently collimated so that radiation which would be absorbed at high mercury vapour pressures would at lower pressures be able to pass through the sides of the vessel although his conditions were such that collimated radiation would all be absorbed before reaching the bottom of the vessel even at the lowest pressure of mercury used. Using an uncollimated beam from the mercury lamp, the rate of decomposition varies as reported by Darwent (table 3), when the mercury vapour is in equilibrium with liquid mercury at 30" C and at 0" C.By collimating the beam with two circular stops such that the most oblique radiation could not fall on the walls of the vessel, the rates obtainedA . B . CALLEAR A N D JAMES C. ROBB 29 were independent of mercury vapour pressure (table 3).These experiments were carried out using 5 mm pressure of ethylene in each case. f i h e (min) (7.5 I FIG. &-Initial pressure rise during the photolysis of ethylene at different total pressures carried out at 20" C. The curves have been vertically displaced for the sake of clarity. (1) 6.1 mm Hg pressure (3) 6 mm Hg pressure (2) 1.0 mm Hg pressure (4) 1.5 mm Hg pressure. TABLE 3 rate (collirnatcd) rate (uncollimated) prcssure riselmin pressure riselmin expressed in mm absolute expressed in mrn absolute temp. O C 0 3-85 x 10-3 74.4 x 1 0 - 3 10 3-93 x 10-3 - 30 3-85 x 10-3 114.0 x 10-3 DISCUSSION The overall rate as a function of pressure measured manometrically is in good agreement with that obtained by temperature measurements. The rate of reaction obtained from curve J, fig.6, is plotted on fig. 9~ for comparison with that obtained by direct pressure measurements. These curves are fitted at 0 and 150 mm pressure. They are both in excellent agreement with the expression where P is the pressure of ethylene and A and B are constants. The plot of R against P2 is shown in fig. 9 ~ . The information obtained here is consistent with the following reaction sequence, excluding decrease in quenching below 1 mm. rate of decomposition = A/(1 + B P ) , > I HglSo 4- h~ -+ Hg3P1 Hg3P1 4- C2H4 -+ HglSo + C2&* c2H4* -+ c2H2 + H2 k2 C2H4* + 2C2H4 -+ 3GHi k3 This leads to a kinetic expression30 THERMAL METHOD The deactivation process involving an apparently three-body collision is rather difficult to interpret.Two possible explanations may be put forward. It may be that C2H4* + C2H4 --f C4H8*, perhaps in the form of a cyclobutane or butene molecule which has a relatively long existence, considerably longer than the dura- tion of the average bimolecular collision. Subsequent collision with an ethylene molecule would then lead to the stabilization of the three ethylene molecules. Since there are no C4 hydrocarbons formed, this mechanism would seem unlikely. An alternative explanation might be that the excited ethylene has energy consider- ably in excess of that required for decomposition, and that two successive fruitful collisions may be necessary before the molecule is rendered incapable of decom- position. One estimate7 puts the energy of activation at about 85 kcal/mole, while the maximum energy available is 112 kcal and the reaction is only 42 kcal endothermic.That the degradation of the energy of excited ethylene may occur in steps, is supported by Laidler's 8 suggestion that since excited ethylene may be FIG. 9. (A) Variation of the initial rate of pressure increase as a function of pressure carried out (B) Variation of reciprocal quantum yield with (pressure)*. at 20" C. The dotted curve is obtained from curve J, fig. 6. in a triplet state, the degradation of the electronic energy will be slower than the degradation of the vibrational energy of the molecule. These two alternative deactivation processes lead to the expression rate = A/((1 + BP)(1 + CP)). introducing a term in P in the lower line as well as the P2 term.This expression interprets the results more satisfactorily than the simple bimolecular expression rate = A/(1 + BP). Assuming that each collision leads to energy transfer, the cunsecufive collision theory leads to an average lifetime of C2H4* of 5 x 10-9 sec, since about half the excited ethylene inolecules are deactivated by a pressure of 40 mm of normal ethylene. If the deactivation is a true three-body process, the average lifetime can be estimated from Bodenstein's formula,g o}h = ratio of ternary collisions to binary collisions. o is the molecular radius and A is the mean free path. Assuming that every ternary collision leads to deactivation, the half-life time of C2H4* is about 2 x 10-6 sec.A . B . CALLEAR AND JAMES C . ROBB 31 The experiments conducted in presence of argon and krypton also indicate a three-body mechanism, but the lower deactivating efficiency compared with ethylene makes it necessary to carry out the reaction over a higher pressure range than could be employed here owing to the dificulties introduced by convection currents in a reaction vessel at pressures much in excess of 150 mm.With helium, it might appear that the form of the curve relating rate of decomposition to pressure is in better agreement with rate = A/((1 + BP)(l -1- CP)}, but the high thermal conductivity of helium makes the tcmperature changes in the system very small so that accuracy is poor. From the data presented on the fall off in quenching of Hg3P1 by C02 and by ethylene, a comparison of the quenching cross-sections can be made directly, by finding the pressure at which 50 % quenching occurs. Taking the quenching cross-section 10 of CO2 at 3.54 A2, the quenching cross-section of ethylene would be 30 A2.A value obtained by Steacie 11 of 48 f 5 A2 is in qualitative agreement with this estimate. No evidence has been found in these experiments to suggest that Hg3P1 is significantly quenched by ethylene to the metastable Hg3Po state, as has been suggested elsewhere,3 and it has also been shown that there is no reason to suppose that C2H4* decomposes heterogeneously on the walls of the reaction vessel. The fresh information obtained in the low pressure region showing that the rate is maintained to about 1 mm pressure is in accord with the known quenching cross- section of ethylene for Hg3P1.The quantum yield of close to unity is considerably higher than previously proposed. The value of the method described above can be said to lie in the speed and accuracy with which measurements of reaction rate can be made. These two advantages arise from the fact that the rate of reaction is obtained from the total quantity measured and not by a difference of two large quantities such as in pressure measurements or in measurements of concentration of reactant by some physical method. It would not be possible, for this reaction, to estimate the amount of products formed since the whole character of reaction changes if significant amounts of hydrogen are produced. This thermal method requires a reaction time of only about 1 sec to enable a rate measurement to be made and in that time no effective change in reactant concentration has occurred and only an insignificant amount of products is formed. I t is of interest to note that it has proved possible to measure a primary quantum efficiency without measuring the absolute quantum input to the reaction system, such measurement being often subject to some uncertainty. The authors wish to thank The Anglo-Iranian Oil Company and Imperial Chemical Industries Ltd., for financial support in obtaining some of the apparatus, and one of them (A. B. C.) wishes to thank the Department of Scientific and Industrial Research for the award of a maintenance grant. They are both apprecia- tive of the interest in this work which has been shown by Prof. H. W. Melville, F.R.S., and for the many valuable suggestions made by him. 1 Leroy and Steacie, J. Chem. Physics, 1941 , 9, 829. 2 Darwent, J. Chem. Physics, 1951, 19, 258. 3 Darwent, J. Chem. Physics, 1952, 20, 1673. 4 Callear and Robb, to be published. 5 Pearson, 2. physik. Chem., A, 1931, 156, 56. 6 A.P.Z. research project 44 (Carnegie Inst. Tech.). 7 Steacie and Leroy, J. Chem. Physics, 1942, 10, 22. 8 Laidler, J. Chem. Physics, 1942, 10, 43. 9 Bodenstein, 2. physik. Chem., 1922, 100, 1 18. 10 Mitchell and Zemansky, Resonance Radiation and Excited Atom (Cambridge Uni- 11 Steacie, Can. J. Res. B, 1940, 18, 44. Selected values of properties ofhydrocarbons, heats of combustion (1945-46), table 0 n, 8 n, 25 n. versity Press, 1934).