摘要:

256 GENERAL DISCUSSION AUTHOR INDEX* Adams, G. K., 97, 138. Bamford, C. H., 208, 245. Barb, W. G., 208,242,245. Barrett, K. E. J., 221. Bateman, L., 190. Bawn, C. E. H., 120, 141, 181, 240. Baxendale, J. H., 160, 234, 239, 249 Becker, E. D., 128, 130. Bell, R. P., 118. Bell, W. E., 70, 131. Berlie, M. R., 50. Bickel, A. F., 25 1. Blacet, F. E., 70, 13 1. Blackman, J. G., 234. Bottomley, G. A., 234. Bowen, E. J., 143,146,229, 233. Bywater, S., 125. Chamberlain, G. H. N., 89. Dacey, J. R., 84, 133, 135. Dainton, F. S., 114, 199, 231, 235, 239,244. Darwent, B. de B., 55, 121, 123, 129. Davies, A. G., 140. Dixon, J. K., 122, 255. Dyne, P. J., 112. Eaton, R. S., 104. Elton, G. A. H., 141. Finkelstein, A,, 76. GiguBre, Paul A., 104, 141. Glockler, G., 138. Gray, P., 135, 137. Groh, H.J., Jr., 128, 130. Harris, G. M., 126, 131. Haszeldine, R. N., 134. Herington, E. F. G., 234. Herzberg, G., 11, 11 1, 113, 127. Hey, D. H., 216,249,252. Hoare, D. E., 89. Hughes (Mrs.) Hilda, 190. Ingold, M. U., 34, 115. Ivin, K. J., 199. James, D. 0. L., 244. Jaquiss, M. T., 246. Kooyman, E. C., 251. Le Roy, D. J., 50, 120. Livingston, R., 228. Lossing, F. P., 34, 1 15. Luft, N. W., 114, 139. Macoll, A., 252. McDowell, C. A., 132. Magee, J., 160. Majury, T. G., 45. Mayo, F. R., 231, 250, 254. Melville, H. W., 122, 150, 231, 232. Mesrobian, R. B., 242. Minkoff, G. J., 141. Morris, A. L., 190. Nicholls, R. W., 127. Norrish, R. G. W., 16. Noyes, W. Albert, Jr., 76, 128, 130. Orr, R. J., 170. Parker, W. G., 97, 139. Polanyi, J. C., 115. Porter, George, 23, 111, 114, 115, 123, 133, Ramsay, D.A., 11, 11 1, 112. Rice, 0. K., 117. Robb, J. C., 122, 150, 231, 232. Roberts, R., 55, 121. Robertson, A. J. B., 115. Rohatgi, Miss K. K., 146. Schiff, H. I., 63, 121, 128. Secco, E. A., 104. Sivertz, C., 229. Stannett, V., 242. Steacie, E. W. R., 9,45, 118, 229. Szwarc, M., 125, 236, 246, 248, 251. Tickner, A. W., 34, 115. Trotman-Dickenson, A. F., 124, 230. Tutton, R. C., 150, 231, 232. Twigg, G. H., 240. Uri, N., 127, 234, 236. Volman, D. H., 116, 141, 253. Walsh, A. D., 89, 112, 127, 136, 140. Waters, W. A., 131, 221, 228, 229, 233, 238, 247, 255. Whittle, E., 120. Wijnen, M. H. J., 118. Williams, Gareth H., 216, 252. Williams, H. Leverne, 170, 237. Winkler, C. A., 63, 128. Wolfhard, H. G., 97, 139. Wright, Franklin J., 23, 115.138. * The references in heavy type indicate papers submitted for discussion.256 GENERAL DISCUSSION AUTHOR INDEX* Adams, G. K., 97, 138. Bamford, C. H., 208, 245. Barb, W. G., 208,242,245. Barrett, K. E. J., 221. Bateman, L., 190. Bawn, C. E. H., 120, 141, 181, 240. Baxendale, J. H., 160, 234, 239, 249 Becker, E. D., 128, 130. Bell, R. P., 118. Bell, W. E., 70, 131. Berlie, M. R., 50. Bickel, A. F., 25 1. Blacet, F. E., 70, 13 1. Blackman, J. G., 234. Bottomley, G. A., 234. Bowen, E. J., 143,146,229, 233. Bywater, S., 125. Chamberlain, G. H. N., 89. Dacey, J. R., 84, 133, 135. Dainton, F. S., 114, 199, 231, 235, 239,244. Darwent, B. de B., 55, 121, 123, 129. Davies, A. G., 140. Dixon, J. K., 122, 255. Dyne, P. J., 112. Eaton, R. S., 104. Elton, G. A.H., 141. Finkelstein, A,, 76. GiguBre, Paul A., 104, 141. Glockler, G., 138. Gray, P., 135, 137. Groh, H. J., Jr., 128, 130. Harris, G. M., 126, 131. Haszeldine, R. N., 134. Herington, E. F. G., 234. Herzberg, G., 11, 11 1, 113, 127. Hey, D. H., 216,249,252. Hoare, D. E., 89. Hughes (Mrs.) Hilda, 190. Ingold, M. U., 34, 115. Ivin, K. J., 199. James, D. 0. L., 244. Jaquiss, M. T., 246. Kooyman, E. C., 251. Le Roy, D. J., 50, 120. Livingston, R., 228. Lossing, F. P., 34, 1 15. Luft, N. W., 114, 139. Macoll, A., 252. McDowell, C. A., 132. Magee, J., 160. Majury, T. G., 45. Mayo, F. R., 231, 250, 254. Melville, H. W., 122, 150, 231, 232. Mesrobian, R. B., 242. Minkoff, G. J., 141. Morris, A. L., 190. Nicholls, R. W., 127. Norrish, R. G. W., 16. Noyes, W. Albert, Jr., 76, 128, 130. Orr, R. J., 170. Parker, W. G., 97, 139. Polanyi, J. C., 115. Porter, George, 23, 111, 114, 115, 123, 133, Ramsay, D. A., 11, 11 1, 112. Rice, 0. K., 117. Robb, J. C., 122, 150, 231, 232. Roberts, R., 55, 121. Robertson, A. J. B., 115. Rohatgi, Miss K. K., 146. Schiff, H. I., 63, 121, 128. Secco, E. A., 104. Sivertz, C., 229. Stannett, V., 242. Steacie, E. W. R., 9,45, 118, 229. Szwarc, M., 125, 236, 246, 248, 251. Tickner, A. W., 34, 115. Trotman-Dickenson, A. F., 124, 230. Tutton, R. C., 150, 231, 232. Twigg, G. H., 240. Uri, N., 127, 234, 236. Volman, D. H., 116, 141, 253. Walsh, A. D., 89, 112, 127, 136, 140. Waters, W. A., 131, 221, 228, 229, 233, 238, 247, 255. Whittle, E., 120. Wijnen, M. H. J., 118. Williams, Gareth H., 216, 252. Williams, H. Leverne, 170, 237. Winkler, C. A., 63, 128. Wolfhard, H. G., 97, 139. Wright, Franklin J., 23, 115. 138. * The references in heavy type indicate papers submitted for discussion.

ISSN:0366-9033    

DOI:10.1039/DF95314FX001

出版商:RSC    

年代:1953

数据来源: RSC

摘要:

256 GENERAL DISCUSSION AUTHOR INDEX* Adams, G. K., 97, 138. Bamford, C. H., 208, 245. Barb, W. G., 208,242,245. Barrett, K. E. J., 221. Bateman, L., 190. Bawn, C. E. H., 120, 141, 181, 240. Baxendale, J. H., 160, 234, 239, 249 Becker, E. D., 128, 130. Bell, R. P., 118. Bell, W. E., 70, 131. Berlie, M. R., 50. Bickel, A. F., 25 1. Blacet, F. E., 70, 13 1. Blackman, J. G., 234. Bottomley, G. A., 234. Bowen, E. J., 143,146,229, 233. Bywater, S., 125. Chamberlain, G. H. N., 89. Dacey, J. R., 84, 133, 135. Dainton, F. S., 114, 199, 231, 235, 239,244. Darwent, B. de B., 55, 121, 123, 129. Davies, A. G., 140. Dixon, J. K., 122, 255. Dyne, P. J., 112. Eaton, R. S., 104. Elton, G. A. H., 141. Finkelstein, A,, 76. GiguBre, Paul A., 104, 141. Glockler, G., 138. Gray, P., 135, 137. Groh, H.J., Jr., 128, 130. Harris, G. M., 126, 131. Haszeldine, R. N., 134. Herington, E. F. G., 234. Herzberg, G., 11, 11 1, 113, 127. Hey, D. H., 216,249,252. Hoare, D. E., 89. Hughes (Mrs.) Hilda, 190. Ingold, M. U., 34, 115. Ivin, K. J., 199. James, D. 0. L., 244. Jaquiss, M. T., 246. Kooyman, E. C., 251. Le Roy, D. J., 50, 120. Livingston, R., 228. Lossing, F. P., 34, 1 15. Luft, N. W., 114, 139. Macoll, A., 252. McDowell, C. A., 132. Magee, J., 160. Majury, T. G., 45. Mayo, F. R., 231, 250, 254. Melville, H. W., 122, 150, 231, 232. Mesrobian, R. B., 242. Minkoff, G. J., 141. Morris, A. L., 190. Nicholls, R. W., 127. Norrish, R. G. W., 16. Noyes, W. Albert, Jr., 76, 128, 130. Orr, R. J., 170. Parker, W. G., 97, 139. Polanyi, J. C., 115. Porter, George, 23, 111, 114, 115, 123, 133, Ramsay, D.A., 11, 11 1, 112. Rice, 0. K., 117. Robb, J. C., 122, 150, 231, 232. Roberts, R., 55, 121. Robertson, A. J. B., 115. Rohatgi, Miss K. K., 146. Schiff, H. I., 63, 121, 128. Secco, E. A., 104. Sivertz, C., 229. Stannett, V., 242. Steacie, E. W. R., 9,45, 118, 229. Szwarc, M., 125, 236, 246, 248, 251. Tickner, A. W., 34, 115. Trotman-Dickenson, A. F., 124, 230. Tutton, R. C., 150, 231, 232. Twigg, G. H., 240. Uri, N., 127, 234, 236. Volman, D. H., 116, 141, 253. Walsh, A. D., 89, 112, 127, 136, 140. Waters, W. A., 131, 221, 228, 229, 233, 238, 247, 255. Whittle, E., 120. Wijnen, M. H. J., 118. Williams, Gareth H., 216, 252. Williams, H. Leverne, 170, 237. Winkler, C. A., 63, 128. Wolfhard, H. G., 97, 139. Wright, Franklin J., 23, 115.138. * The references in heavy type indicate papers submitted for discussion.256 GENERAL DISCUSSION AUTHOR INDEX* Adams, G. K., 97, 138. Bamford, C. H., 208, 245. Barb, W. G., 208,242,245. Barrett, K. E. J., 221. Bateman, L., 190. Bawn, C. E. H., 120, 141, 181, 240. Baxendale, J. H., 160, 234, 239, 249 Becker, E. D., 128, 130. Bell, R. P., 118. Bell, W. E., 70, 131. Berlie, M. R., 50. Bickel, A. F., 25 1. Blacet, F. E., 70, 13 1. Blackman, J. G., 234. Bottomley, G. A., 234. Bowen, E. J., 143,146,229, 233. Bywater, S., 125. Chamberlain, G. H. N., 89. Dacey, J. R., 84, 133, 135. Dainton, F. S., 114, 199, 231, 235, 239,244. Darwent, B. de B., 55, 121, 123, 129. Davies, A. G., 140. Dixon, J. K., 122, 255. Dyne, P. J., 112. Eaton, R. S., 104. Elton, G. A.H., 141. Finkelstein, A,, 76. GiguBre, Paul A., 104, 141. Glockler, G., 138. Gray, P., 135, 137. Groh, H. J., Jr., 128, 130. Harris, G. M., 126, 131. Haszeldine, R. N., 134. Herington, E. F. G., 234. Herzberg, G., 11, 11 1, 113, 127. Hey, D. H., 216,249,252. Hoare, D. E., 89. Hughes (Mrs.) Hilda, 190. Ingold, M. U., 34, 115. Ivin, K. J., 199. James, D. 0. L., 244. Jaquiss, M. T., 246. Kooyman, E. C., 251. Le Roy, D. J., 50, 120. Livingston, R., 228. Lossing, F. P., 34, 1 15. Luft, N. W., 114, 139. Macoll, A., 252. McDowell, C. A., 132. Magee, J., 160. Majury, T. G., 45. Mayo, F. R., 231, 250, 254. Melville, H. W., 122, 150, 231, 232. Mesrobian, R. B., 242. Minkoff, G. J., 141. Morris, A. L., 190. Nicholls, R. W., 127. Norrish, R. G. W., 16. Noyes, W. Albert, Jr., 76, 128, 130. Orr, R. J., 170. Parker, W. G., 97, 139. Polanyi, J. C., 115. Porter, George, 23, 111, 114, 115, 123, 133, Ramsay, D. A., 11, 11 1, 112. Rice, 0. K., 117. Robb, J. C., 122, 150, 231, 232. Roberts, R., 55, 121. Robertson, A. J. B., 115. Rohatgi, Miss K. K., 146. Schiff, H. I., 63, 121, 128. Secco, E. A., 104. Sivertz, C., 229. Stannett, V., 242. Steacie, E. W. R., 9,45, 118, 229. Szwarc, M., 125, 236, 246, 248, 251. Tickner, A. W., 34, 115. Trotman-Dickenson, A. F., 124, 230. Tutton, R. C., 150, 231, 232. Twigg, G. H., 240. Uri, N., 127, 234, 236. Volman, D. H., 116, 141, 253. Walsh, A. D., 89, 112, 127, 136, 140. Waters, W. A., 131, 221, 228, 229, 233, 238, 247, 255. Whittle, E., 120. Wijnen, M. H. J., 118. Williams, Gareth H., 216, 252. Williams, H. Leverne, 170, 237. Winkler, C. A., 63, 128. Wolfhard, H. G., 97, 139. Wright, Franklin J., 23, 115. 138. * The references in heavy type indicate papers submitted for discussion.

ISSN:0366-9033    

DOI:10.1039/DF95314BX003

出版商:RSC    

年代:1953

数据来源: RSC

摘要:

THE REACTIVITY OF FREE RADICALS GENERAL INTRODUCTION BY E. W. R. STEACIE It is the custom at these Discussions for the President of the Society to welcome “ the distinguished overseas members and guests ”. I am sure the situation must have been a little confusing for the President to-day since the Society is itself over- seas, and the identification of the guests is somewhat difficult. I should, therefore, like to take this opportunity to welcome the Society on its first visit to Canada. I should also like to say how happy we are to have in Canada such a distinguished group of workers in the field, from Britain, from the United States, and from other countries. Faraday Society Discussions have a unique reputation, and we in Canada are particularly pleased at having the honour of being your hosts on the occasion of your first overseas visit.Free radicals have frequently been a subject of Faraday Discussions. Full- scale discussions were those on Free Radicals in 1934 and on the Labile MolecuIe in 1947. Free radical reactions have also figured largely in other discussions on hydrocarbons, oxidation, photochemistry, polymerization, etc. To-day’s title, “ The Reactivity of Free Radicals ” is, I think, appropriate since it emphasizes a definite change in outlook which has recently made itself felt. From a pessimistic point of view one might perhaps wonder what we have all been doing during the past eighteen years, since many of the questions asked in 1934 have still to be answered in 1952. Thus the role of free radicals in the decomposition of ethane was a burning question in 1934, and is still open to vigorous argument to-day.From a more optimistic point of view, however, I think that we can fairly say that much progress has been made in the meantime. In 1934 the emphasis was on the existence of radicals. Almost any type of mechanism could be, and was, postulated. To-day we know enough at least to predict the types of reaction which are likely to be of importance, and the maker of mechanisms has to exercise considerably more self-restraint. By the time of the 1947 Discussion the general state of affairs had, I think, become reasonably satisfactory from a qualitative standpoint. General agreement seemed to have been reached as to which types of reaction were fast and which were slow, and as to general orders of magnitude of activation cnergies.Also, there was fairly general agreement as to which overall reactions involved free radicals and which did not, even if there was by no means agreement on the quantitative aspects in many cases. Again, by 1947 at least the main features of the mechanisms of many important thermal and photochemical reactions had been worked out. In the last five years there has been a sharp shift in emphasis, with attention being focused on the rates of free radical reactions themselves, rather than on overall reactions with the elementary steps regarded as merely incidental stages. Along with this has come a demand for more quantitative data resulting in the inevitable reinvestigation of many old established reactions. As a result we are again to-day in a period of flux, with many previously accepted results now being regarded with suspicion.It is, therefore, an appropriate time for such a discussion as the present one, and in particular for emphasis on the reactivity of free radicals. At the moment there is much interest in an accurate knowledge of activation energies, and in a knowledge for a series of compounds of the way in which the activation energy varies with chemical structure. It looks at present as if a be- ginning has been made along these lines. This is a matter of great importance 910 GENERAL INTRODUCTION since it seems to the writer that the numerous efforts to make some such correlation in the past, and to compare the results with theory, have usually taken the form of a plot of f (activation energy) against f (bond dissociation energy) in which both sets of data were completely unreliable.There are also signs of a decided improvement in the bond dissociation energy situation. This has long been the scandal of kinetics, and it seems extraordinary that we are still by no means certain of many of the values for the simplest com- pounds. There has, however, been much accomplished in the past five years. I am by no means convinced that the present situation is as satisfactory as some workers assume, but it is certainly a great improvement. Another feature of recent work has been the successful determination of the absolute rates of certain recombination reactions. This has, in the first place, necessitated a change in outlook regarding third-body effects. In the second place, it has given standard reactions for comparison, and has led to a knowledge of the absolute values of frequency or steric factors for many reactions.The results have been surprising and have proved the misleading nature of calculations in which arbitrary values of steric factors have been assumed in the past. Much reinvestigation of simple reactions is needed before the matter can be cleared up. Much progress has been made in certain specific fields involving free radical reactions. In the field of polymerization there have been great advances in the formal kinetics of the process. Information on specific chemical steps is in by no means as satisfactory a state, and it is gratifying to see that a fundamental contribution to this question is part of the present discussion.Concerning oxidation reactions we are still in a position where nearly everyone is free to have his own mechanism. Great strides have been made in the way of formal mechanisms, but the identity of the chain carriers remains an open question in most cases. It may also be pointed out that the recent work on recombination, mentioned above, suggests that the wholesale postulation of third-body effects in oxidation reactions rests on doubtful ground. The question merits re-examin- ation. It is hoped that to-day’s discussion will be a step towards the final solution of some of these problems. In the liquid phase the problems are more complex, although in some ways the subject is further advanced. We have for our discussion a very good variety of problems on oxidation, polymerization, substitution, etc. All in all, I think the discussion is a timely one, and that we are unusually fortunate in the very distinguished group we have with us. We can hardly fail to effect some clarification of some of the issues.

ISSN:0366-9033    

DOI:10.1039/DF9531400009

出版商:RSC    

年代:1953

数据来源: RSC

摘要:

I. REACTIONS IN THE GAS PHASE THE ABSORPTION SPECTRUM OF FREE NH2 RADICALS BY G. HERZBERG AND D. A. RAMSAY National Research Council of Canada, Division of Physics, Ottawa, Canada Received 31st March, 1952 Ammonia was irradiated in a flash photolysis apparatus and the absorption spectrum of the decomposition products was taken within 1 millisecond of the photolysis flash using a 21 ft. grating spectrograph. The so-called tc bands of ammonia, long known froin emission spectra of oxy-ammonia flames and electric discharges through ammonia, were found to occur in absorption in the region 4500 to 7400A. These bands are shown to be due to the free NH2 radical. Corresponding band systems were obtained by the flash photolysis of N15H3 (60 %) and ND3. From the isotopic shifts between the NI4H2 and N15H2 spectra it is concludcd that the origin of the band system lies near 10,400A.As the first bands observed are considerably removed from the origin of the system, the upper and lower states must differ considerably in the geometrical arrangement of the nuclei. For this reason it has not becn possible as yet to obtain an unique vibrational assignment. Free polyatomic radicals like CH2, CH3, NH2 and others have been assumed to exist in many chemical reactions. Definite proof of their existence and in- formation about their structure may be obtained if their spectra can be identified. While several spectra have been tentatively identified as due to certain free poly- atomic radicals, only in one case so far, that of CF2, can the identification be con- sidered as unambiguous.l.2 In this case both the structure of the spectrum and its observation in absorption support the identification.A new case which we believe to be of this type is the subject of the present paper.3 The so-called a bands of ammonia which occur in emission in electric discharges through ammonia, in oxy-ammonia and other flames and in the spectra of comets have long been suspected of being due to the free NH2 radical although they have also been ascribed to NH3, N2H4 and NH. We have observed the a bands in absorption in photodecomposed ammonia and have definitely eliminated the possibility of N H 3 , N2H4 or NH being the carriers. In addition the structure of the spectrum and the isotope shifts observed when H is replaced by D and N14 by Nl5 confirm the conclusion that NH2 is responsible for the a bands.EXPERIMENTAL Photochemical investigations 4 strongly suggest that NH3 by absorption of light in the region of its diffuse ultra-violet absorption bands decomposes according to NH3 +,hv + NH2 -t H. (1) It appeared therefore possible that NH2 might be dctected in absorption in photochemically decomposed ammonia. Ammonia was irradiated in our flash photolysis apparatus which has been described elsewhere,s and the absorption spectrum of the decomposition products was investigated within 1 millisecond of the photolysis flash using the second order of a 21-ft. concave grating spectrograph. A large number of very fine absorption lines were observed extending from 4500 to 7400A. Many of the stronger lines were found to coincide exactly (within & 0.05 cm-1) with the lines of the 01 bands as obtained in emission from an oxy-ammonia flame taken with the same instrument. This leaves no doubt that the carrier of the flame emission spectrum is identical with that of the absorption spectrum.6 1112 SPECTRUM OF FREE NH2 RADICALS The spectra were taken on Eastman-Kodak 111-F, 111-0 and hypersensitized I-N plates.It is quite clear from the spectra obtained that the high resolving power used (- 100,000) was essential for the observation of the very fine and weak lines and that the high contrast of the plates used was a further aid in detecting many of the weaker lines. The exposure varied from 85 to 180 flashes depending on the spectral region. The absorption tube was 1 m long.The pressure of NH3 first used 3 was 10 mm but further investigation showed that at lower pressures (- 1 mni) the spectra were appreciably improved. Variation of the time interval between the photolysis flash and the source flash used for supplying the continuous background showed that the best spectra were obtained when the source flash immediately followed the photolysis flash (within 1 nisec) though weak absorption was still obtained with a time delay of 4 msec. These results indicate that the lifetime of the radicals responsible for the absorption lines is of the order of a few milliseconds. The spectra of the corresponding isotopic molecules were obtained by using ND3 and a sample of ammonia containing 60 % N15H3. The ND3 was prepared by dropping 99.6 % D20 on to Ca3N2 and the N15 enriched sample was prepared by treating NH4N03 containing 60 % N15HqN03 with potassium hydroxide solution.Fig. 1 shows small sections of all three spectra obtained. The isotope shift between the bands of the N15 and the N14 compound are readily seen by comparing fig, la and b, the latter identifying the N14 lines in the former. For each of the isotopic species several thousand lines are present on the spectrograms, most of them rather weak. Until now only partial measure- ments of these lines have been made. NATURE OF THE CARRIER OF THE a BANDS The fact that the a bands occur in absorption suggests very strongly that the lower state is the ground state of the molecule involved. The possibility of ab- sorption from a metastable state seems very remote since as far as we are aware there is no case in which absorption by a metastable diatomic or polyatomic molecule has been observed in the gaseous state.In addition theobservation of the a bands in comets by Swings, McKellar and Minkowski 7 is not compatible with a metastable lower state if the usual fluorescence mechanism for their ex- citation applies. Accepting therefore the conclusion that the a bands have the ground state as their lower state one can immediately eliminate NH3 and N2H4 as carriers since their absorption spectra are well known. Moreover, the NH band a t 3360 8, does not appear in absorption in photodecomposed ammonia when the a bands are observed and hence the a bands cannot be due to NH. This is in agreement with the previously mentioned assumption about the primary process of photo- dissociation of NH3.In experiments with photodecomposed hydrazine 8 (N2H4) the NH band at 3360A was observed in absorption in addition to the a bands. In view of the difficukies that were experienced in assigning the 4050A group to the correct carrier,8a it is perhaps not superfluous to point out that for the a bands the observation of an isotope shift both on substituting D for H and N15 for N14 does prove that both hydrogen and nitrogen are present in themolecule responsible for the a bands. The elimination of NH3, N2H4 and NH and the requirement that both H and N are present leaves as possibilities only NH2, NzH, N2H2, N2H3 for the carrier. The last three can be eliminated for two reasons.(i) The moment of rnertia when two N atoms are present is large and therefore a narrow fine structure is expected while a fairly open fine structure is observed. (ii) The molecules N2H, N2H2, N2H3 would be nearly symmetric tops and therefore the spectrum should con- sist of simple [I, 1, or hybrid bands. Actually the structure is typical of that of a strongly asymmetric top of small moments of inertia agreeing very well with that expected for non-linear NH;! ; and valence considerations certainly predict NH2 to be non-linear. For all these reasons the concIusion seems to be inescapable that the carrier of the a bands is the free NH2 radical.FIG. 1.-The 5708 A band of N14H2 and N15H2 and the 6030 A band of N14Dz. [To face page 12.13 G . HERZBERG A N D D.A . RAMSAY STRUCTURE OF THE a BANDS On the basis of the electron configuration and the assumption that NH2 is non-linear, Mulliken 9 has predicted the three low lying electronic states 2B2., 2A1 and 2B1 as shown in fig. 2 (changing from Mulliken’s choice of axes to that usually used for XY2 molecdes 10). Since the transition from the 2B2 ground state to the 2B1 state is for- bidden by the electronic selection rules, it appears highly probable that the a bands correspond to the allowed transition to the 2A1 state. In this case, the transition moment is perpendicular to the plane of the molecule. When observed in emission from discharges or flames, the a bands show the typical appearance of a many-line spectrum and no individual bands are obvious. In our absorption spectra a grouping of the lines into individual bands is apparent, partly due to the fact that fewer vibrational levels of the ground state are involved and partly due to the smaller number of rotational levels populated in the absorption experiment.The principal bands I I I I FIG. 2.-Predicted low elec- tronic states of NH2. are listed in table i. There are a number of bands with clearly developed central (Q) branches accompanied by somewhat irregular and much more widely spaced P and R branches (e.g. the bands at 6302, 5708 and 5166 A). TABLE 1 .-PRINCIPAL BANDS OF N14H2 AND N14D2 ; ISOTOPE SHIFTS N*4H2 -N15H2 average N14H2 -N15H2 approximate centres a shifts b of corresponding approximate centres a band lines of N14D2 bands of N 14H2 bands 7354 w 6619 m 6302 s 5977 s 5708 s 5385 m 5166 s 4925 m 4718 w 4524 w 13594 15104 15864 16726 17514 18565 19372 20299 21 189 22098 11.4 21.9 28.7 28.3 31.4 28.3 38.4 41.1 47.0 49.7 6498 m 6280 m 6030 s 5809 m 5600 w 5381 s 5200 m 5112 w 5020 w 4865 m 4701 w 4561 w 4412 w 15385 15919 16579 17210 17850 18579 19225 19556 19915 20549 21266 21919 22659 a The band centres are frequently not defined to better than & 20 cm-1.For N14H2 the longward edges of the central (Q) branches have been used ; for Nl4D2 these central branches and therefore the wave numbers listed are much less well defined. bThe isotope shifts do not depend on the correct identification of the band centres, and are much more accurate than these centres, probably to f 0.3 cm-1. The fact that the bands do not form readily recognizable progressions suggests that the origin of the band system is at some distance from the observed part of the system.This is confirmed by the study of the N14-Nl5 isotope shifts. For all observed bands the isotope shift is fairly large and always in the same direction (see fig. 1). For ND2 compared to NH2 the isotope shifts are so large that a one-to-one correlation between ND2 and NH2 bands cannot at present be made.14 For a diatomic molecule the isotope shift is given approximately by the relation (2) where p = dpT, (p and pi being the reduced masses) and where v, is the separ- ation of the band considered from the system origin. In fig. 3 the observed iso- tope shifts are plotted against v. It is seen that with two exceptions they fall fairly closely on a straight line whose slope gives a value of 1 - p N 0.00395. The line intersects the abscissa axis at 9600 cm-1.This point which is well beyond the observable part of the spectrum corresponds to the origin of the band system. It must, however, be realized that the simple equation (2) would hold for a poly- atomic molecule only if only one vibration is excited in the upper and lower state. SPECTRUM OF FREE NH2 RADICALS AV = (1 - p)v,, FIG. 3.-Isotope shifts between corresponding lines of N14H2 and N15H2. For different vibrations different p values apply. For two vibrations the isotope shift is approximately given by (3) where vz; and vv2 are the parts of the wave number of the band due to the two vibrations. It is immediately clear from this that in such a case, as long as both vV1 and v,,, are positive, the isotope displacements plotted against v should lie be- tween two straight lines of slope 1 - pl and 1 - p2 respectively which intersect at the origin of the band system.According to the Franck-Condon principle applied to polyatomic molecules,ll in the present case only the two totally symmetric vibrations v1 and v2 can be excited. Transitions with Av3 =I= 0 must be very weak and those with odd Av3 are forbidden. The 1 - p values depend on the H-N-H angle and for v l and v2 also on the individual frequencies. Table 2 gives the 1 - p values for four different angles assuming a valence force system with an N-H stretching force constant of kl = 6 x lG5 dyne/cm and a H-N-H bending force constant kd/P = 0.6 x 105 dyne/cm.For v3 the 1 - p value is independent of the force constants. The frequencies v1, v2 and v3 of N14H2 corresponding to the adopted force constants at the different angles are also given in table 2. If NH2 were Av = (1 - Pf)VVI + (1 - p2)v,,, TABLE 2.-kOTOPE FACTORS ANGLES ASSUMING apex angles v1 (cm-1) 180" 3179 150" 3194 120" 3236 90" 3292 1 - p AND VIBRATIONAL FREQUENCIES FOR VARIOUS APEX kl = 6.0 x 105 AND kd/P = 0.6 x 105 dynejcm. 1-A vr (crn-1) I-p2 v3 (cm-1) 1 --P3 3400 0.00419 0.0 1075 0.00419 0.00033 1070 0.00386 3386 0.00395 0.001 18 1056 0.00301 3346 0.00325 0.00227 1038 0.00193 3291 0.00223G . HERZBEKG AND D. A . RAMSAY 15 linear the vibration v l would not show any isotope shift. For the adopted force constants the slope of the line in fig.3 (0.00395) fits best the calculated 1 - p value of v2 at an angle of about 160". This angle would of course refer to the upper state. The calculated 1 - p values are not very sensitive to the force constants chosen. It is rather surprising that the observed shifts do almost aU lie on one straight line in spite of the fact that the 1 - p value for v l is very much smaller. This indicates that v l is only slightly, if at all, excited in the transition. The large distance of the bands of maximum intensity from the system origin requires that the internuclear distances and angles in the upper state are ap- preciably different from those in the lower. Since the bending vibration v2 appears to be the one that is mainly excited it will be necessary to assume that the H-N-H angle changes appreciably in going from the lower to the upper state.There- fore 1 - p will be somewhat different in the upper and lower state. This may account for the appreciable scatter of the points in fig. 3 about the straight line. However, such a scatter would arise only if the vibration is excited both in the upper and lower state. This is in conformity with the somewhat complicated appearance of the band system which consists certainly of more than one pro- gression. In this connection it is of interest to point out that for the NH radicals which occur in the photodecomposition of N2H4 the presence of molecules in the v = 1 level of the ground state is demonstrated by the observation of the 1-1 band in addition to the 0-4 band at 3360A.If all possible wave number differences of the stronger NH2 bands are taken it is found that the values cluster near 900, 1820, 2700 and 3520 cm-1 while for ND2 they cluster near 660, 1370, 2010 and 267Ocm-1. It may be significant that these values seem to be approximate multiples of the first one, in each case suggesting that this first one is a vibrational frequency in the upper state. CONCLUSION Further progress in the vibrational analysis of the a bands will have to await the results of the rotational analysis of some of the bands. We are at present engaged in such an analysis. It is interesting to note that since the electronic transition is one in which the transition moment is perpendicular to the plane of the molecule the allowed vibrational transitions will have a type C rotational structure.10 This is a band type that cannot occur in rotation-vibra- tion spectra of XY;! molecules since there are no vibrations with a dipole moment variation in a direction perpendicular to the plane of the molecule (direction of the axis of largest moment of inertia).Type C bands of not too asymmetric molecules are characterized by a strong central branch 10 as is the case for some of the a bands (see fig. 1). In order to obtain more information about the electronic structure of the NH2 molecule experiments are being prepared with a view to obtaining the absorption spectrum of NH2 in the vacuum ultra-violet. A number of electronic transitions of the Rydberg type are expected which should be very much stronger than the a bands and may therefore be observable in spite of the overlapping NH3 absorption. Even without further progress in the interpretation of the NH2 absorption spectrum the a bands can now be used as a relative measure of NH2 concentra- tion in chemical reactions. In order to obtain an indication of the absolute con- centration it will be necessary to investigate the a bands under conditions in which the NH2 concentration is known. 1 Venkateswarlu, Physic. Rev., 1950, 77, 676. 2 Laird, Andrews and Barrow, Trans. Faraday SOC., 1950,46, 803. 3 A preliminary note was published in J . Chem. Physics, 1952, 20, 347. 4 see Noyes, Jr., and Leighton, The Photochemistry of Gases (New York, 1941).16 FLASH PHOTOLYSIS 5 Herzberg and Ramsay, Faraday SOC. Discussions, 1950, 9, 80 ; Rainsay, f. Chem. 6 A spectrogram showing two small spectral regions of the absorption and emission 7 Swings, McKellar and Minkowski, Astrophys. J., 1943, 98, 142. 8 Ramsay, to be published. 9 Mulliken, in ref. (7). 10 see Herzberg, Infra-red and Raman Spectra of Polyatomic Molecules (New 11 see Herzberg and Teller, 2. physik. Chem. B, 1933,21,410. Physics (in press). spectra side by side was given in ref. (3). see Douglas, Astrophys. J., 1951, 114, 466. York, 1945).

ISSN:0366-9033    

DOI:10.1039/DF9531400011

出版商:RSC    

年代:1953

数据来源: RSC

摘要:

16 FLASH PHOTOLYSIS FREE RADICALS IN EXPLOSIONS STUDIED BY FLASH PHOTOLYSIS BY R. G. W. NORRISH Department of Physical Chemistry, The University, Cambridge Received 9th May, 1952 The application of the technique of flash photolysis to the spectroscopic study of explosions is described. The method yields valuable evidence as to the nature of the short lived intermediates taking part, and, in some cases, of their elementary reactions within the flame. This is achieved by observations of the relevant absorption spectra and their change with time. Reaction processes, complete within a few milliseconds can be readily followed, and processes contributory to the explosion mechanism identified. Examples are given of the explosion of hydrogen with oxygen, and acetylene with oxygen. The method of " flash photolysis " 1 s 2 has provided a technique for the study of the absorption spectra of free radicals, and of following their reactions. By its means sufficient concentrations of free radicals in the gaseous phase may be momen- tarily obtained for spectrographic investigation in the visible and ultra-violet parts of the spectrum.It is the purpose of this paper to describe how the method may be applied to the study of explosive reactions, for the identification of the nature of the free radicals involved, and the elucidation, in part at least, of the reactions in which they are concerned. The method of flash photolysis depends upon the use of a powerful flash, of energy up to 10,OOO J as a photochemical source. The flash is obtained by dis- charging a bank of condensers through an inert gas such as argon or krypton.In a recent form of the apparatus 3 the lamp is of quartz, 50 cm long, with tungsten electrodes at each end. A capacity of 35 pF at SO00 V is used, giving a flash of ca. loo0 J, of duration 10-4 sec, whose useful photochemical efficiency is of the order of 6 x.4 The reaction vessel of quartz is of the same length as the lamp and placed close beside it, the two being surrounded by a reflector consisting of an aluminium tube coated on the inside with magnesium oxide. The flash, whose spectrum is very largely continuous, is capable of bringing about a very large measure of decomposition in suitable photochemically reactive species, such as C4, N02, CHzCO, CH3COCH3, at pressures of ca. 10 rnm, and therefore of pro- ducing momentarily a high concentration of free radicals.The photographic record of the absorption spectra of the free radicals so pro- duced is made possible by the use of a spectroscopic source, in principle similar to the photolysis source, capable of giving a flash with an approximately continuous spectrum in the visible and the ultra-violet as far as 2200 A, of duration 0.05 msec, and also by the fact that neither the photo-flash nor the spectro-flash lamps operate spontaneously, but require trigger currents from auxiliary electrodes to precipitateR. G . W. NORRISI-1 17 their discharge. By means of commutator or electronic devices these two trigger currents can be separated by specific time intervals as low as 30psec, and the corresponding flashes similarly timed.The time between flashes can be checked on an oscillograph using a photoelectric cell, and by this means also the duration of the flashes is measured. With a spectro-flash lamp, using a discharge of 35 pF at 5000 V through argon, an approximately continuous spectrum can be photo- graphically recorded by one flash on a Hilger E. Littrow spectrograph, with the lamp at 1 m distance from the slit.2 It can be shown that the absorption in 1-3 msec of sufficient radiant energy to bring about a large measure of photochemical decomposition in a suitable reactant must also release sufficient heat to raise the temperature of the medium momen- tarily from lo00 to 3000" C if the process were adiabatic. Actually the rise is less than this, since conditions are not ideally adiabatic and can be largely reduced and kept within reasonable bounds by the use of a relatively high pressure of inert gas as coolant.For example, the photochemically active species may be at pressures of ca. 0.1 mm, and the inert gas at pressures up to 1 atm, and greater in vessels capable of surviving higher pressures. Thus an approximation to isothermal conditions can be achieved, and this is important in the study of simple reactions of elementary free radicals and atoms such as ClO and I, which have been under way in this laboratory.5~ 6 It will be realized from the above statement, however, that the method provides also the possibility of administering an approximately adiabatic shock to a system, and this aspect has already proved of interest in the study of the photolysis and pyrolysis of ketene7 and the deposition of carbon resulting from the sudden pyrolytic cracking of substances such as ketene, acetone and acetylene.spectroscopic methods, it has been possible in the past to identify many of the free radicals taking part by their characteristic emission spectra, but such observation is limited to those radicals which occur in their excited states. Further, it is im- possible by these means to analyze the ignition process in time, for the rapid changes of concentration of these intermediates in the reactions constituting the flame front cannot be followed. Hitherto, due to the narrowness of the ignition zone and the low concentration of radicals, the methods of absorption spectroscopy have failed to detect most of the intermediates observed in emission, and until means of achieving a long spectroscopic path through the burning zone could be found it was obvious that they must fail.It is just this possibility which is presented by flash photolysis, for by its means ignition can be generated homogeneously throughout the whole volume of the reaction vessel, providing an " instantaneous " path as long as the vessel itself; and by taking absorption photographs at intervals increasing by 30psec after the initiation of explosion, the generation and decay of all radicals showing absorption between 7000 and 2200A can be followed, and the short induction periods and other phenomena dependent on the composition of the reaction mixture observed.In this we are not limited to radicals occurring in the excited state ; the method is applicable to all intermediates showing recog- nizable absorption spectra, and can in principle be extended to other spectral ranges than that at present instrumentally available. In order to obtain a homogeneous explosion it is necessary to achieve simul- taneous initiation throughout the reaction vessel. This may be done by flash photolysis if one of the reactants has a chromophoric group leading, on photolysis, to free radicals capable of initiating reaction chains or, alternatively, if the reactants do not absorb the light from the flash, by the addition of a suitable photo-sensitizer to the system to fulfil the same function. In either case the mechanism of the process is twofold and consists of (i) the formation of initiating radicals by photo- lysis, and (ii) the raising of the temperature very rapidly (within a few psec) above the thermal ignition limit characteristic of the mixture.Once initiated, the explosion will proceed homogeneously throughout the reaction vessel, and its STUDIES OF EXPLOSIVE REACTlONS.-In the study Of flames and eXpl0SiOnS by18 FLASH PHOTOLYSIS progress can be observed by means of absorption spectra photographed at in- creasing time intervals after initiation. The following are examples of explosive reactions initiated by this method. (a) ABSORPTION BY AN EXPLOSIVE REACTANT (6) ABSORPTION BY ONE REACTANT (absorbing substance written first) : methyl nitrate, ethyl nitrate. chlorine + hydrogen, chlorine + methane, bromine + hydrogen, ketene + oxygen, acetone + oxygen, diacetyl + oxygen, carbon disulphide + oxygen.(c) ABSORPTION BY A SMALL AMOUNT OF SENSITIZER : reactants sensitizers C12 or Br2 C12 or Br2 NO2 hydrogen + oxygen methane + oxygen hydrogen, methane acetylene, ethylene ethane, hexane or benzene with oxygen All these explosions may be produced in gases at comparatively low partial pressures of 10-20mm. They may be suppressed by the addition of sufficient inert gas to act as coolant, and studied with varying partial pressures of reactants. The explosive process itself is of 10-3 - lO-4sec duration, but the decay of the free radicals produced in the process may take from 0.1 to 10 msec. Induction periods of 0.1 msec or greater may readily be measured.Only a few of the reactions in the above table have been investigated in any detail, and here I shall give examples from the observations on the hydrogen + oxygen 8 and acetylene + oxygen 3 reactions to illustrate the possibilities of the method. THE HYDROGEN + OXYGEN REACTIoN.-Fig. 1 shows the waxing and waning of the 1,O and 2,O absorption bands of the OH radical during an explosion of 10 mm H2 + 5 mm 0 2 using 0.75 mm NO2 as sensitizer. Each spectrum represents a separate experiment, the interval between the photo-flash and spectro-flash being gradually lengthened throughout the series; they were taken in a quite arbitrary order, and the fact that they fall into a uniform sequence is indicative of their repro- ducibility. The photographic intensity of the absorption bands gives a qualitative measure of the OH concentration. In comparing intensities one series of spectra is arbitrarily taken as a standard and the intensities of all the others measured against it by means of a comparator. This gives a curve of OH concentration against time on an arbitrary scale which is certainly not linear, but which is sufficient to show the growth and disappearance of the radical in a semi-quantitative way.When standards of comparison can later be established, absolute concentration of radicals will be obtainable. Fig. 2 shows curves obtained in this way for the explosion of a mixture of x mm H2 + 2 mm NO2 according to the equation It will be seen that as the pressure of hydrogen increases beyond the stoichiometric value of 2 mm the rate of disappearance of the OH radical increases, and roughly at a rate proportional to the excess hydrogen.Bearing in mind that the photo- lysis of nitrogen peroxide is known to occur by the reaction the immediate appearance of the OH group gives evidence of the elementary re- action H2 + NO2 = NO + H20. NO2 + hv = NO + 0, H2 + 0 = OH + H,m sec 0 0.8 1 -0 1.2 1-6 2.4 3.2 4.8 7-2 1 2850 A [To ,fuw page 18. 1 1 1 11 1 2750 2800 OH I , 0 2700 OH 2, 0 2650 FIG. 1.R. G . W. NORRISH 19 while the increase in the rapidity of disappearance of OH with increasing excess of H2 bears witness to the reaction We thus obtain direct experimental evidence of the existence of two of the ele- mentary reactions which have been postulated in the explosive combination of hydrogen and oxygen? OH + H2 = H20 + H.FIG. 2. In the absence of excess hydrogen the rate of disappearance of OH is compara- tively slow, and is further reduced by the addition of nitrogen, as is shown in table 1. The above figures illustrate the coolant effect of the inert gas for, at its highest pressure, the rate of OH disappearance is not only slowest, but the maximum concentration of (OH) is also lowest, and achieved more slowly as would be the case in a lower temperature explosion Other details of the hydrogen i- oxygen reaction have been elucidated by similar methods, but will not be further discussed here. TABLE I.-DEPENDENCE OF OH INTENSITY ON TIME AND NITROGEN PRESSURE P NO2 = 2 mm, P H2 = 2 mm time (msec) 0.8 1.2 1.6 2.4 3.2 6.4 10.0 p N2 (mm) 0 20 17 12 3 0 0 0 4 30 23 20 15 10 0 0 22 16 18 17 15 13 6 0 THE EXPLOSION OF ACETYLENE AND OXYGEN.-Fig.3, 4 and 5 show results ob- tained for the explosion of acetylene and oxygen sensitized by nitrogen peroxide. Fig. 3 shows the waxing and waning of the OH concentration for mixtures of different composition. Fig. 4 shows curves for four radicals obtained for a par- ticular mixture slightly rich in acetylene. These curves show well a marked in- duction period of 0.5 msec, the rapid generation of the explosion, and the slow decay of the radicals. Since the absorption coefficients of the various radicals are at present unknown, the curves cannot be taken as indicating the relative con- centrations of the various radicals. In fig. 5 are plotted the various maxima in intensity achieved by each radical for varying mixture strengths.20 FLASH PHOTOLYSIS A stoichiometric mixture for the explosion of acetylene is C2H2/02 = 1/1 as represented by the equation Owing to the very high temperature reached in the reaction, the NO2 is com- pletely decomposed, the nitrogen appearing as CN, NH and probably as N2.Thus the oxygen contained in the nitrogen peroxide must be added to the partial pressure of the oxygen in calculating the true stoichiometric mixture, and with C2H2 -+ 0 2 2CO 4- H2. FIG. 3. FIG. 4. this in mind, we see from fig. 5 the profound change in the abundance of free radi- cals as we pass through the equimolecular mixture. On the oxygen-rich side, OH is the predominant radical, while on the fuel rich side its concentration falls to what is comparatively a very low value.On the fuel-rich side too, the free radicals CN, C2, CH, NH, and C3 (not shown in diagram) are in evidence, and with the exception of CN, are completely absent on the oxygen-rich side. Such marked changes in the nature of the intermediates as we pass from oxidizing to reducing conditions cannot fail to have bearing on the elucidation of the explosion process when further evidence has accumulated. There can be little doubt that the excess acetylene is cracked by the heat generated in the explosion, as isR. G. W. NORRISH 21 evidenced by the free carbon deposited, but by what precise mechanism we shall not further speculate here. This question of the mechanism of carbon formation has been discussed by Porter,lo but it must be admitted that at present the evidence is limited by the fact that atomic carbon, oxygen, nitrogen and hydrogen are not detectable by the methods of flash photolysis since their absorption lies outside the spectral range available.The sharp increase in the concentration of hydroxyl, as we pass to the oxygen- rich side of the equimolecular mixture, emphasizes the preferential burning of carbon which takes place in equimolecular mixtures of acetylene and oxygen.11 It is only when the oxygen is in excess that the present observations give evidence of the appreciable burning of the hydrogen. It is temptingJo compare this result FIG. 5. with the slow combustion of acetylene in which formaldehyde is the principle intermediate product isolated, and to postulate for the latter a chain mechanism involving OH radicals and the intermediate formation of formaldehyde as, for example, OH + C2H2 = CH + CH20 (1) CH + 0 2 = CO + OH.(2) With the intervention of ignition, however, the temperature becomes too great for the formation of formaldehyde, and reaction (1) may well be replaced by OH + C2H2 = CH + CO + H2. (1’) In this way we can account for the preferential burning of carbon and the fact that the hydrogen only begins to burn as the combustible mixture passes to the oxygen-rich side of the equimolecular mixture. Besides yielding evidence bearing on the mechanism of the explosion process, it will be seen that flash photolysis applied to explosion processes provides another means of obtaining free radicals in sufficient quantity for the recording of their absorption spectra even when other methods, including ordinary flash photolysis, have failed.In addition to the examples already quoted, this may be achieved by the introduction of a small quantity of foreign substance to the explosive medium22 FLASH PHOTOLYSIS in a similar way to the well-known methods of obtaining emission spectra in flames. The radicals Sic1 and SnCl, for example, are readily obtained in absorption if traces of Sic14 or SnC14 are added to the explosive mixture, and the high tempera- tures reached, which greatly exceed those attainable if the reaction vessel itself has to be heated, make the method applicable to the most refractory substances. In conclusion, experiments designcd to obtain the HO2 radical spectrum should be mentioned.Many different systems have been used, in all of which kinetic requirements would indicate the presence of this radical. It was shown by the present author 12 that hydrogen and oxygen may be exploded by the photosensitive action of chlorine at temperatures of 100-300" C . This reaction is readily initiated from room temperature by flash photolysis, when the dissociation of chlorine is followed by the reaction C1 -1- H2 = HCl -1 H (3) H 4- 0 2 Z- OH + 0. (4) It was thought that, by working just below the ignition limit, the reaction (4) would be replaced by All the spectra obtained, however, both within the explosive region and during the slow reaction, can be attributed to the OH radical, the excited oxygen molecule or the spectrum discovered in chlorine and oxygen alone and attributed to ClO.2, 13 Although it is possible that H02 has no spectrum in the region investigated, this seems rather unlikely, and further attempts to obtain this radical are in progress. The detailed study of the radicals so far detected, and others which we may hope will result from similar investigations, provides a means of elucidating free radical reactions in the complex explosive phases of combustion which have hitherto not been amenable to direct kinetic measurement. H + 0 2 = H02. (4') 1 Norrish and Portcr, Nature, 1949, 164, 658. 2 Porter, Proc. Roy. SOC. A , 1950, 200, 284. 3 Norrish, Porter and Thrush, Nature, 1952,169, 582, and Proc. Roy. SOC. A (in press). 4 Christie and Porter, Proc. Roy. SOC. A , 1952, 212, 395. 5 Porter and Wright, this Discussion. 6 Christie, Norrish and Porter, in course of publication. 7 Knox, Norrish and Porter, J. C/zem. Soc., 1952, 1477. 8 Norrish and Porter, Proc. Roy. SOC. A, 1952, 210,439. 9 Lewis and Von Elbe, J. Chem. Physics, 1942, 10, 366. 10 Porter, 4th Symp. Combustion (M.I.T., 1952). 11 Bone and Townend, Flame and Combustion in Gases (Longmans, Green, 1927). 12 Norrish, Proc. Roy. SOC. A, 1932,135, 334. 13 Porter, Faraday SOC. Discussions, 1950, 9, 60.

ISSN:0366-9033    

DOI:10.1039/DF9531400016

出版商:RSC    

年代:1953

数据来源: RSC

摘要:

STUDIES OF FREE RADICAL REACTIVITY BY THE METHODS OF FLASH PHQTOLYSIS THE PHOTOCHEMICAL REACTION BETWEEN CHLORINE AND OXYGEN BY GEORGE PORTER AND FRANKLIN J. WRIGHT Department of Physical Chemistry, University of Cambridge Received 15th April, 1952 The reaction of chlorine atoms with oxygen has been studied by the flash photolysis and flash spectroscopy method of Porter.1 The chloric oxide radical C10 is readily formed at 293" K and its concentration has been followed throughout the reaction by quantitative measurements of its absorption. The initial reactions are @1 -t 0 2 = ClOO and C1 -t ClOO r= 2CIO. The rate constant of removal of chlorine atoms by oxygen, to form both C12 and CIO, is 46 times the rate of removal in nitrogen to form Cl2. The decomposition of C10 to Clz and 0 2 occurs relatively slowly, is bimolecular with respect to C10, and the rate is independent of Cl2, 0 2 and total gas pressure.The rate constant of this reaction is where eS is the molar extinction coefficient of C10 at 2577A. than 310 and probably less than 3000. terms of the intermediate C1202. 7.2 x lo4 hJ. exp (0 f- 650/RT) I. mole-1 sec-1, The value of eS is greater The mechanism of this reaction is discussed in The extensive literature on the reactivity of chlorine atoms, produced by photo- chemical dissociation of the molecule, gives little evidence of a direct reaction with oxygen. The well-known inhibiting effect of oxygen on photochemical chlorinations is usually attributed to the removal by oxygen of hydrogen atoms, COCl radicals, hydrocarbon radicals, etc., rather than of chlorine atoms, but the rate expressions do not allow an unequivocal choice of mechanism.In a mixture of chlorine and oxygen alone there is no apparent photochemical change and no transient reactions have been detected. It was therefore somewhat unexpected when a new absorption spectrum, which was clearly that of a diatomic molecule, was discovered at high intensity during the investigation of chlorine + oxygen mixtures by the flash technique.1 The spectrum was attributed to the C10 radical, and a vibrational analysis, as well as a determination of the dissociation energy of the molecule have already been given.2 The methods of flash photolysis and flash spectroscopy were originally developed to make possible the production of labile molecules in a concentration high enough for absorption spectroscopy to be applied to the study of their kinetics, and we here describe such an investigation of the formation and reactions of the C10 radical.Direct information is also obtained in this way about the role of the oxygen in the photochemical reactions of chlorine. EXPERIMENTAL The absorption spectra of 20 free radicals, about half of them new, or previously unknown in absorption, have now been obtained in this laboratory. A detailed kinetic investigation by the flash technique involves more difficulties, however, than simply recording the spectrum of a free radical. Many hundred spectra must be taken under different conditions, flash intensities and time intervals must be accurately reproduced and the temperature rise, which is a result of the adiabatic nature of the reaction, must be eliminated. The number of intensity measurements necessary for a complete kinetic investigation by flash photolysis is best reduced by the use of photocell recording. The relative merits 2324 FLASH PHOTOLYSIS of this method, which is being used for other problems, have already been discussed,2 but in the present case, owing to the fact that the mechanism of the reaction was quite unknown, and also that a rather complex band system was being investigated, it was thought advisable to examine the whole spectral region throughout the investigation so as to be able to detect the presence of other chlorine oxides, changes in the intensity dis- tribution of the C10 spectrum and intensity changes in the spectrum of the chlorine molecule.The use of photographic recording introduces more scatter into intensity measurements but the relationships eventually obtained are particularly significant as they are the result of many experiments carried out in a quite arbitrary order. A full description of the method, and of the apparatus used for this investigation has been given elsewhere.1 The only modification necessary was to enclose the reaction vessel and the photolysis flash tube in a furnace, so that the temperature dependence of the reaction rates could be determined. The furnace, which was of the same dimen- sions as the original reflector, completely enclosed the reaction vessel, and was coated internally with magnesium oxide. To facilitate removal it consisted of two semi-cylindrical portions wound separately, and the temperature was measured by three thermocouples at the centre and at either end.An investigation of the properties of the photolysis flash lamp at temperatures up to 350" C showed that the firing characteristics are a function of the concentration of the inert gas filling rather than of the pressure and that the output is not noticeably affected by the rise in temperature. This, fortunately, makes it possible to use the same gas filling throughout and therefore to avoid intensity variations which occur from one filling to another.3 PREPARATION OF GAsEs.-Chlorine was taken from a cylinder and redistilled several times in vacuo. Oxygen was prepared electrolytically, dried over CaCI2, freed from traces of hydrogen by passing over 30 cm of platinized asbestos at 350" C and finally dried over P205. Nitrogen in which not more than 0.05 % oxygen could be tolerated, was pre- pared by heating sodium azide which gives a pure product.4 (Hilger E.1) instrument ; 25 spectra were recorded on each plate, and the intensities were measured on a non-recording microphotometer. The output of the spectroflash was constant within the accuracy of the microphotometric measurements. Each plate was calibrated separately and methyl ethyl ketone, which has a continuous spectrum in the same region as the C10 radical, was used for this purpose. The ketone was made up to about 1 atm. pressure with carbon dioxide and a range of ketone partial pressures chosen to give the same densities as the particular C10 concentrations being studied.The intensity of the C10 spectrum could then be expressed in terms of the pressure of ketone having the same extinction at a given wavelength, the plate sensitivity, path length and incident intensities being constant. The extinction curve of methyl ethyl ketone vapour was measured on a Unicam spectrophotometer. Chlorine itself has a significant absorption in the region of the C10 spectrum, and this must be allowed for in the intensity determination. It will be shown later that the amount of C10 formed is so small that the chlorine concentration does not change sig- nificantly during the experiment, and therefore this correction is a constant one. If Beer's law is obeyed we have, for a given wavelength, ln(lo/11) = FICII + €2~21, where Zo is the incident intensity, 11 is the intensity transmitted by the mixture of CI2 and CIO, I is the path length, €1 and c1 are the extinction coefficient and concentration of C10, and €2 and c2 the same quantities for chlorine.Now if ketone, at concentration c3 also transmits the intensity 11, the incident intensity being unchanged, and the extinction coefficient being €3, E l C l + €2C2 = €3C3. If 12 is the intensity through chlorine alone, and also through ketone at concentration c4 In (Zo/Z2) = €2~21 = E3Cql The values of c3 -- c4, which are determined experimentally from the calibration spectra, are therefore proportional to the ClO concentration if Beer's law applies. €3 being known, qcl can be determined but the values of €1 and c1 cannot at present be obtained separately. It is therefore necessary to express the concentrations of C10 as some function of its extinction coefficient and all concentration measurements are expressed as E~G, where c is the true pressure of C10 in mm/Hg and eS is the molar decadic INTENSITY AND CONCENTRATION MEASUREMENTS.-The spectrograph Was a Littrow and €3(C3 - c4) = qc1.(1)GEORGE PORTER AND FRANKLIN J . WRIGHT 25 extinction coefficient at 2577 A. This wavelength is in the continuous region and should therefore be unchanged if a different resolving power is used. VALIDITY OF BEER'S LAW.-B~IS'S law has been assumed to apply both to the ketone and the ClO in the derivation of eqn. (1). No deviations were expected for methyl ethyl ketone, which was chosen for the lack of fine structure in its spectrum,S and the law was verified by measuring the extinction spectrophotometrically over the range of total pressures used in the calibrations.The following precautions were taken to assure the applicability of the law to the C10 spectrum. Firstly, possible deviations due to pressure effects were eliminated by keeping the total pressure many hundred times greater than the pressure of ClO. Except in one case, which is discussed separately, no comparisons are made at different temperatures. Finally, errors which might be caused by incomplete resolution of fine structure were first reduced by taking all measurements in the continuum or near to the heads of the predissociated bands where most of the absorption is due to lines of a width greater than the resolving power of the spectrograph.Each position was then checked by plotting c3 - c4 for a number of C10 spectra taken at different concentrations against the same quantity for a different wavelength. In one case only, the 13, 0 band at 2751 .$, there was a small deviation at high concentrations. In all other cases a linear plot was obtained and as this included measurements in the con- tinuum, where the spectrum is completely continuous, the validity of Beer's law is con- firmed for the measurements to follow, in which the 13, 0 band is not used. FIG. 1.-Plot of c3 - c4 at 2773 8, against c3 - c4 at 2797 A. From the gradient of these plots, one of which is shown in fig. 1, the relative ex- tinction coefficients at the different wavelengths were obtained so that all measurements could be expressed in terms of eS.For the three bands which were used in addition to the continuum they are as folIows : band 12,o 1 l,o 10,o continuum wavelength X (A) 2773 2797 2824 2577 EL/% 1 -50 1.44 1 *44 1-00 REDUCTION OF THE ADIABATIC TEMPERATURE EmxT.-Unless precautions are taken, the heat liberated by the reactions of the atoms and radicals formed may produce a temperature rise of over 1000", the rate of thermal diffusion to the walls being less than that of the chemical reactions. Preliminary experiments showed that the C10 radical was still present in pressures of oxygen or nitrogen as high as 1 atm, and in most of the experiments to follow the total pressure used was 600 mm the pressure of chlorine being about 5 mm.Under these conditions the temperature rise, estimated by calculation and by use of the concentration effect, is 1 or 2" C and it will be shown later that a tem- perature rise of 100" C has no effect on the measured constants. The photolysis flash was operated from 168 pF at 4000 V unless otherwise stated and the absorption tube was 1 m in length and 2 cm diam.26 FLASH PHOTOLYSIS RESULTS A typical series of spectra, taken at increasing times after the flash, is shown in plate 1. It was first nesessary to ascertain that the whole of the spectrum being measured was that of C10 and that there was no overlying spectrum of another molecule. It was found that the rates determined in each part of the spectrum were the same, as would be expected in view of the relationships between the extinction coefficients mentioned previously.A search was then made for spectra in other regions, in particular those of the C102 and C120 molecules which have high extinction coefficients in the region of 2 8 0 0 k None was found. Finally there was no difference between the results of experiments on newly mixed gases and those which had been flashed up to 50 times. Under all conditions, the rate of reaction of C10 was very much less than the rate of its formation during the flash. Further, at the shortest times, immediately after the flash, the observed rate of reaction of C10 at a given concentration was the same as at longer times at the same concentration, showing that the reactions by which C10 is formed occur in a time which is short compared with the duration of the flash.It is therefore possible to divide the investigation into two parts : the dark reactions of CIO occurring after the flash and the photochemical reactions by which C10 is formed during illumina- tion. The times involved in a typical experiment are illustrated by fig. 2 which gives the FIG. 2.-Light intensity and C10 concentration against time. intensity against time and the C10 concentration against time curves. The latter, at times greater than 1.4 msec gives the rate law for the dark reaction and will be studied first, the reactions during the first msec being discussed later. THE KINETICS OF C10 DISAPPEARANCE.-The concentration of CIO a sa function of time was determined for a given mixture by recording a number of spectra, in an arbitrary order, at different times after the flash.Measurements were made at several wavelengths and concentrations determined, via the ketone calibrations, in the manner described. DEPENDENCE ON c10 CONCENTRATION.-Fig. 3 shows the relationship between the reciprocal of the C10 concentration and the time, using a mixture of 10 mm Clz and 600 mm 0 2 . A linear plot is obtained and, in the course of this work, over 60 such graphs were drawn from measurements on a wide variety of mixtures, all of which showed a direct proportionality within the experimental error. We conclude that, under all conditions described below, - d(ClO)/dt =; k(Cl0)z. If k’ is defined by the equation - d((Cl0k)dt = k’((C10)Es)2, the value of k’ obtained from the gradient of fig.3 is 4.9 mm-1 sec-1 cS units-1, and the absolute rate constant is given by k = k‘es mm-1 sec-1 = 1.70 x 104 k’ES 1. mole-1 sec-1.0 2700 A i 2800 A" 0.0 5.2 Bimde c ulur Dlsup-eurunce of Cblorlc Oxide (Clq PLATE 1. [To face page 26.GEORGE PORTER AND FRANKLIN J . WRIGHT 27 DEPENDENCE ON OXYGEN PRwm.-The total pressure of oxygen plus nitrogen was kept constant and the relative pressure of oxygen varied. Owing to the similar specific heats and collision diameters of the two gases, oxygen pressure is the only significant variable. The rate constants for different mixtures (pressures in mm Hg), determined as before, are given in table 1. About 10 spectra were used in each determination of these and all subsequent rate constants.The values of k' are constant to within 10 % whilst the oxygen pressure is changed by a factor of 60. TABLE RATE CONSTANT k' AT DIFFERENT OXYGEN PRESSURES (P IN MM Hg) P ( W P(O2) PW2) PW2f N?) k' (mm-1 Sec-1 eg units-1) mean k' 10 600 0 600 4.0 3.7 3.8 4.9 4 1 10 100 500 600 4.3 4.1 4.3 - 4-2 10 10 590 600 3.8 4-7 3.5 4.1 4.0 FIG. 3.-Dlot of l/(CIO)es against time. DEPENDENCE ON CHLORINE PRESSURE.-The results given in table 2 show that the rate constants are invariant over a tenfold range of chlorine pressure. TABLE 2.-RATE CONSTANT k' AT DIFFERENT CHLORlNE PRESSURES (P IN MM Hg) P(CW P(0Z) k' (mm-1 sec-1 as units-') mean k' -_ 20 600 4.3 3.7 3.8 - 3.9 10 GOO 4.0 3.7 3.8 4.9 4.1 5 600 4.5 4.4 4.0 4.0 4.2 2 600 3-9 3.7 4.3 - 4.0 DEPENDENCE ON TOTAL PRE.ssuRE.-At low total pressures the reaction is no longer even approximately isothermal but reference to the next section shows that the effect of temperature rise on the rate can be ignored over a wide range and it was therefore possible to vary the total pressure by a factor of 10 without introducing a temperature change large enough to affect the rate.The results given in table 3 show that over this range the value of k' is independent of total pressure. TABLE 3 . - h T E CONSTANT k' AT DIFFERENT TOTAL PRESSURES (P 1N MM Hg) P(C12) P(O2) P(N9 P(tota1) k' (mm-1 sec-1 es units-1) mean k' 10 100 500 610 3.8 4.3 5.2 4 4 10 100 300 410 4 9 4.9 6- 1 5.3 5 50 100 155 4 0 4-8 4.4 4 4 10 100 0 100 3.1 3.2 3.4 3.2 5 50 0 55 4.8 4-5 5-2 4.828 FLASH PHOTOLYSIS DEPENDENCE ON TEMPERATURE.-The rate constants measured at higher temperatures are given in table 4, the pressures being referred to 293" K. TABLE 4.-RATE CONSTANT k' AT DIFFERENT TEMPERATURES (P IN MM Hg) To K P(C12) P(0d k' (mm-1 sec-1 as units-1) mean k' 293 4.2 433 5 300 4.0 4.2 4-0 4.1 473 10 400 6.2 6.3 7.3 6.6 mean of values in tables 1, 2 and 3 The values of k' are constant in the temperature range 293" K to 433" K but increase slightly at 473" K.Measurements at higher temperatures showed a further increase but they are of doubtful significance owing to the changed intensity distribution (see results on temperature dependence of C10 formation). If there were a true rate increase at higher temperatures it might imply a change in mechanism at about 450" K but it is also possible that the ratio of cs to eAdh is no longer constant. In view of this uncertainty we cannot say anything about the higher temperature mechanism at present.In the range between 293" K and 433" K there is no significant change in the intensity distribution and the measured rates are constant. Assuming a possible variation in k' of f 40 %, in order to allow for any slight changes in the intensity distribution, we con- clude that if the temperature coefficient of CIO removal is expressed as exp (- E/RT) then the activation energy in the range of temperature between 293" K and 433" K is 0 f 650 cal. The introduction of a T i dependence of the non-exponential term would reduce the value of E by 350 cal. the rate law of ClO removal in the dark reaction is 1: The results of this section are summarized as follows : - d(ClO)/dt = k(ClO)2(O~)O(Cl~)O(N~)O ; the mean value of k' = k/eS at 293" K is 4.2 mm-1 sec-1 cs units-1 ; the rate constant k, in the temperature range 293" K to 433" K, is 650/RT)I.mole-1 sec-1. 7.2 x 104 eS exp (0 ABSOLUTE VALUE OF cS AND k . 4 n view of the very intense absorption by the C10 radical it seemed probable that a proportional decrease in the intensity of the chlorine molecule spectrum would be observed and that this would give the absolute concentra- tions of CIO and values for cS and k. Experiments designed for this purpose have shown no such decrease and therefore, at present, it is only possible to give a lower limiting value for es. In order to increase the sensitivity of the method 10 similar spectra were taken of chlorine in the presence of a high concentration of C10, each accompanied by a blank of the chlorine alone, the voltage being 6OOO V in this case.In mixtures of 5 mm C12 with 700 mm 0 2 a slight decrease in Cl2 absorption was observed which control experi- ments, using nitrogen in place of oxygen, showed to be due to the temperature-concentra- tion effect alone. A mixture of 1 mm Cl2 with 700 mm 0 2 showed no decrease whilst calibration spectra with known chlorine pressures showed that 0.05 mm pressure decrease would have been detected. The average value of (CIO)eS was 31 mm which gives the minimum value for of 310. It is helpful to have some idea of the maximum possible value of eS for the purpose of discussion and we have two reasons for supposing that it is not greatly different from the above minimum.Firstly an examination of the known extinction coefficients of similar molecules and radicals having partly continuous spectra make it improbable that the value of eS, which is by no means the maximum extinction coefficient, would exceed 3000. Secondly, if chlorine atoms recombine at a rate equal to or less than the three- body collision rate in the presence of nitrogen the later investigations on the relative rate constants of this reaction and C10 formation, coupled with the fact that the formation of C10 is not observable after 1.5 msec lead to a lower limit for the C10 concentration which is again within a factor of 10 of the above minimum. We hope to be able to deter- mine cs experimentally by other methods ; it seems very probable, however, that the true values of the constants are not more than a factor of 10 greater than the followingexperi- inentaI minima : cS > 310, k293 y 2.2 x lo7 I.mole-1 sec-1.GEORGE PORTER AND FRANKLIN J . WRIGHT 29 THE INITIAL PHOTOCHEMICAL REACTION The times involved are too short for the investigation of the C10 concentration changes during the flash, though a simple modification of the apparatus would make this possible. A different approach was used here, however, the reactions during the flash being studied by measurements of the concentration ((Clo)~,) formed during a given time, the time chosen being 0.7 msec. During this time a fraction of the C10 formed will have reacted by the mechanism already studied and, using the known rate constant, it is possible to correct for this and to calculate, for any measured concentration at time 0.7 msec, the total C10 formed during this period.The calculation was performed as follows : (i) The intensity against time curve was taken from an oscillograph of the flash and a graph of total light output against time was constructed, from the area beneath, in arbitrary units. (ii) As the concentration changes are small the light absorbed, and therefore, owing to the high rate of C10 formation, the C10 present at a given time will follow the same curve if the reactions by which the ClO is removed are ignored. Taking these reactions into account it is possible to calculate the true ClO concentration at any time by using the known rate constant k'. A step method was found most convenient and the cal- culation was simplified by the almost linear nature of the intensity curve up to 0.7 msec.A corrected curve is shown in fig. 2. (iii) For the series of observed concentrations at time 0.7 msec the total C10 formed is determined in this way. It was found that even at the highest concentrations the correction was only about 15 % and no significant errors are likely to be involved in these corrections. C10 formed during the first 0.7 msec was determined for a range of oxygen pressures, the total pressure being kept constant at 600 mm by the addition of nitrogen. The chlorine pressure was 5 mm and the flash intensity and other conditions were kept as constant as possible. The concentrations, after correction, are plotted as a function of oxygen pressure in fig.4. It will be seen that only when the oxygen concentration is reduced to about 1/50 of the nitrogen concentration is the C10 formed reduced by Q. This suggests that the rate constants of the reactions in oxygen and nitrogen respectively are also in this ratio and a more detailed derivation of this relation will be given later. EFFECT OF TEMPERATURE ON C10 FORMATION.-The intensity of ClO at 0.7 msec was investigated as a function of temperature between 293°K and 593°K using a mixture of 5 mm of chlorine and 300 mm of oxygen measured at 293" K. At the highest tem- peratures the rotational structure became very extended and the bands were less distinct, but intensity measurements in the continuum and even in the diffuse bands at our standard wavelengths were independent of temperature to within i- 20 %.The only true measure of relative concentrations under these conditions is EAdh which could not be evaluated exactly, but intensity comparisons at a number of wavelengths showed that the value of the integral was almost constant and could hardly have varied by a factor of more than 2 over the whole temperature range. If the temperature coefficient of C10 formation is expressed as exp (- E/RT) the value of E is therefore 0 rlr 0.8 kcal. RELATION BETWEEN c10 FORMED AND OXYGEN PRESSURE.-The COnCentratiOn Of lorn30 FLASH PHOTOLYSIS DISCUSSION The reactions studied occur exclusively in the homogeneous gas phase. This is shown by the nature of the observations, which are made on the gas near to the centre of the vessel, the independence of the rates on the total pressure and by simple calculation, which gives times of diffusion to the wall greatly in excess of the time intervals observed.The problem is further simplified by our accurate knowledge of the bond energies of C10 (63 kcal)2 CI2 (57 kcal) and 0 2 (117 kcal)6 and the heats of formation of the stable chlorine oxides.7 Consider, for example, the reactions (1) (2) 2CIO = CI02 + C1 - 7 kcal (3) 2C10 = C120 + 0 - 32 kcal (4) Cl + 0 2 = c10 + 0 c1 + CIO = c12 + 0 - 54 kcal - 6 kcal Reaction (1) is very endothermic and cannot be a significant mechanism of C10 formation which has nearly zero temperature dependence. Reaction (3) and (4) are similarly eliminated as mechanisms of C10 removal for which the activation energy is zero within a few 100 cal.Reaction (2) must be considered a little more carefully for, if the oxygen atom formed reacts with C10, a chain mechanism is possible and with suitable activation energies for the termination reaction, this might lead to a temperature independent rate constant for C10 removal. Apart from the fact that no such termination step can be found which gives the observed rate law, and also that the rate constant is too high to involve reaction (2) as a propagation step, the use of chlorine atoms in this mechanism is entirely incompatible with the fact that C10 is formed very rapidly by their reaction with oxygen but the rate of removal of C10 is completely independent of oxygen pressure.The probable fate of any oxygen atoms formed by this or any other reaction would be a reaction with chlorine by the reverse of (2). We are now in a position to consider the few remaining possibilities which, in conjunction with the observed rate expressions, give the mechanism in some detail. THE MECHANISM OF C10 FORMATION.-Reaction (1) having been excluded, the only possible reaction by which C10 can be formed is 2c1+ 0 2 = 2c10, ( 5 ) although this may proceed in two stages as follows : c1 + 0 2 = ClOO CI + ClOO = 2c10. The radical ClOO is not to be confused with the stable radical 0-Cl-0. Using this mechanism it should now be possible to interpret the results of the experiments on 0 2 + N2 mixtures given in fig. 4. As nitrogen can play no chemical role its effect must be ascribed to the reaction 2C1+ N2 = Clz + N2* (8) and we must also consider the reaction 2CI + 0 2 = Cl2 + 02* (9) which again may proceed via the intermediate C100.The reactions by which C10 is removed are automatically eliminated by the method of calculating the total C10 formed. Ignoring the intermediate ClOO for the moment the only reactions by which CI atoms are removed will now be (3, (8) and (9). Then d(2ClO)/dt = kg(C1)2(02) and d(C12)ldt = kg(C1)2(02) + kg(Cl)2(N2).GEORGE PORTER AND FRANKLIN J . WRIGHT When the removal of CI atoms is complete we have (2~10) = Jrm 3c kg(C1)2(02)dt r-=o and t==m t=O t=o (c12)f = 1 kg(C1)2(02)dt + jt=m kg(C1)2(Nz)df, 31 where (Cl& is the total (C12) formed from C1 atoms. and (N2) are constants and 1 Therefore For any given mixture (02) (C1)2 dt has the same value in both expressions.t=m t=o If (2C10) and (C12)~ refer to the concentrations at a given time, say 0.7 msec, then the same expression is obtained by eliminating the integral t=0.7 t=o 1 (Clpdt. For all (02)/(N2) ratios, twice the number of C1 atoms formed during the flash is a constant and is equal to (2CIO) + (C12)f = K (say). (ii) This again applies to any time interval if the reactions removing C1 atoms are fast When (N2) = 0, and putting (2ClO) = (2CIO),a,, then (iii) Eliminating K and (Cl& from (i), (ii) and (iii) we obtain A plot of (ClO)max/(ClO) against (N2)/(02), where (C1O)max is the C10 formed in the absence of nitrogen, should therefore give a straight line of slope k8/(k5 + kg).The results already shown in fig. 4 are plotted in this way in fig. 5 and a linear plot is obtained confirming the mechanism suggested. Further, the gradient of this line gives the above ratio of rate constants and is found to be 1/46. The sum of k5 and k9 is the rate constant of removal of CI atoms by oxygen and k8 by nitrogen, i.e. - d(Cl)/dt = 2(k5 + kg)(Cl)2(02) in oxygen, and - d(Cl)/dt = 2(kg)(Cl)*(N2) in nitrogen. Therefore the reaction of chlorine atoms with oxygen to form C12 and C10 occurs at a rate 46 times that of the reaction in nitrogen to form Cl2. This result has several interesting consequences. Firstly, there is every reason to suppose that the recombination of chlorine atoms in a gas such as nitrogen occurs at every termolecular collision.The collision diameters of oxygen and nitrogen being very similar, for example in the recombination of other halogen atoms,8 it follows that reactions ( 5 ) and (9) as written are insufficient to account for the rate and these reactions must involve the formation of a relatively stable complex, which can only be C100. The inclusion of this intermediate leads to the same rate expressions if reaction (6) is reversible and the equilibrium main- tained. It is not possible to say whether this implies an activated complex C100*32 FLASH PHOTOLYSIS with a lifetime 46 times greater than the CIN2* complex or equilibrium involving stabilization by a third body and bimolecular dissociation. In either case, if the collision efficiency of C1 atom recombination in nitrogen at 1 atm.pressure is taken as 1/900, the equilibrium constant (C1)(02)/(CIOO) is less than 20 atm even if the reaction of C1 with ClOO occurs at every collision. The very small temperature coefficient of C10 formation is fully in accordance with our mechanism and confirms that reactions (I), (2), (3) and (4) are unim- portant in the photochemical part of the reaction. If reaction (9) occurs to a significant extent it follows that E g = E5 f 0.8 kcal. I zoo 60 /zu 160 FIG. 5.-(ClO),,,/(CIO) against the ( N 2 ) / ( 0 2 ) ratio. THE MECHANISM OF CIO REMOVAL.-The rate IaW shows conclusively that the subsequent reactions by which CIO is removed are unaffected by the pressure of any gas other than CIO. Reasons have already been given for excluding chlorine atoms, and reactions (3) and (4), and the only remaining possibility is 2c10 = c12 + 02.(10) There is a strong objection to this reaction in its simple form in that all known double decompositions of this type are associated with high energies of activation, a useful empirical rule being that the activation energy is about 1/4 of the sum of the energies of the bonds broken? There is one important difference in the radical reaction, however, and that is the possibility of dimerization. We must therefore consider the reactions 2c10 = a 2 0 2 (1 1) c1202 = 2c10 (12) a 2 0 2 = c12 -l- 0 2 (1 3) which lead to the rate expression if we assume a stationary concentration of C1202. Two limiting cases may be considered. 6) k13 > k12- Reaction (11) now becomes rate determining and a zero activation energy is quite probable.We then have the difficulty of explaining the very low rate constant, for it has been shown that the rate is independent of total pressure andGEORGE POK'TER AND PKANKLlN J . WRIGHT 33 a third-body collision cannot therefore be necessary. We should have to explain the low rate entirely by means of a steric factor which could hardly be greater than 10-3, whereas such rate constants as are known for simple radical recom- binations have steric factors very near to unity.10 This explanation therefore seems improbable. In this connection it should be mentioned that the possibility of formation of a " stable " C1202 molecule is not entirely eliminated by the experimental evidence, in which case the observed rate would be that of formation of Cl202 which might then decompose more slowly to C12 and 0 2 .This would only be true if C1202 had a very low extinction over the whole region investigated and had a lifetime of the order of seconds. In addition exactly the same objections apply as were given in the last paragraph. (ii) k12 > k13- The effective rate constant is now kllk13/k12 and the observed activation energy will be El1 - El2 + E13. The difference El2 - El1 is equal to the heat of forma- tion of a 2 0 2 from two C10 radicals and must be very nearly equal to E13, the activation energy of dissociation to C12 and 0 2 in order to explain the observed temperature independence. This is not unlikely if both energies are small and the low rate is then explicable. We therefore think that the reaction is best explained in terms of an equilibrium between C10 and C1202, the latter decomposing to C12 and 0 2 .It is then possible to see why C10 might behave differently from both NO and OH which are in many ways similar radicals. The former has a very unstable dimer which dis- sociates to 2N0 much more readily than to N2 and 0 2 whilst the latter forms a stable dimer which dissociates only very slowly. C10 is an intermediate, semi- stable radical because its dimer is also of intermediate stability. We are in- vestigating this reaction further by the transition state method but it seems doubtful whether, in the absence of data about the C1202 molecule, anything much more quantitative can be said. results show unequivocably that C1 atoms react rapidly with oxygen, and these reactions must occur to some extent in the presence of other gases.This is in accordance with the complete rate expressions of Thon 11 and of Bodenstein and Schenk 12 for the H2 + 0 2 + C12 system. We have carried out some pre- liminary experiments on C10 formation in the presence of hydrogen and have found that in excess oxygen (10 mm C12, 10 mm H2 and 300 mm 02) the C10 formed and the rate of its reactions are not greatly changed and that little HCl is formed. In excess hydrogen (10 mm C12, 25 mm 0 2 and 350 mm H2) no C10 was observed and most of the C12 reacted permanently. Excess carbon monoxide, on the other hand, resulted in a very high C10 con- centration and C10 is almost certainly the long-lived intermediate which resulted in the slow approach to the steady state observed by Bodenstein, Brenschede and Schumacher.13 In this, and similar oxidations photosensitized by chlorine, the ClOO radical may play an important part, in addition to C10. For example, the following mechanism for the sensitized oxidation of C02 would now appear probable : THE CHLORINE + OXYGEN REACTION IN PHOTOSENSITIZED OXIDATIONS.-oUr Cl + 0 2 + ClOO ClOO 1- co -+ c02 + c10 ClO + co 3 c02 + c1. A more detailed discussion of these reactions will be given elsewhere along with an investigation of the reactions of C10 with other gases. One of us (F. J. W.) is indebted to the Anglo-Iranian Co. for financial support B during the tenure of which this work was carried out.34 M ASS S 1' E C'I' RO M E'T K Y 1 Porter, fruc. Roy. Soc. A, 1950, 200, 284. 2 Porter, Faraday SOC. Discussions, 1950, 9, 60, 3 Christie and Porter, Proc. Roy. SOC. A , 1952, 212, 398. 4 Justi, Ann. Physik, 1931, 10, 983. 5 Duncan, Ells and Noyes, J . Anzer. Chem. Soc., 1936, 58, 1454. 6 Gaydon, Dissociation Energies (Chapman and Hall, 1947). 7 Goodeve and Marsh, J. Chem. SOC., 1939, 1332. 8 Rabinowitch and Wood, Trans. Farads-v SOC., 1936, 32, 907. 9 Glasstone, Laidler and Eyring, The Theory of Rate Processes (1941). 10 see, for example, Dodd, Trans. Faraday SOC., 1951, 47, 56. 11 Thon, 2. physik. Chem., 1926,124, 327. 12 Bodenstein and Schenk, 2. physik. Chern. B, 1933,20,420. 13 Bodenstein, Brenschede and Schumacher, 2. physik. Chem. B, 1937,35, 382.

ISSN:0366-9033    

DOI:10.1039/DF9531400023

出版商:RSC    

年代:1953

数据来源: RSC

摘要:

34 M ASS S I' E C'I' RO M E'T K Y FREE RADICALS BY MASS SPECTROMETRY PART I1 --THE THERMAL DECOMPOSITION OF ETHYLENE OXIDE, PROPYLENE OXIDE, DIMETHYL ETHER, AND DIOXANE" BY F. P. LOSSING, K. U. IN GOLD^ AND A. W. TICKNER~ National Research Council, Ottawa, Canada Received 3rd March, 1952 The thermal decompositions of ethylene oxide, propylene oxide, dimethyl ether and dioxane in a stream of helium have been studied by means of a mass spectrometer de- signed to permit the measurement of free radical concentrations. With ethylene oxide it was found that each molecule decomposing gave rise to about 0.6 methyl radicals, and in propylene oxide the corresponding number had a lower limit of 0.36. Methyl radicals were also abundant in the decomposition of dimethyl ether. In the decom- position of dioxane only a small number of methyl radicals were found, and they probably come from a secondary reaction.No other radicals were detected, although formalde- hyde was found to be a comparatively long-lived intermediate in the decomposition of dimethyl ether. The effect of the addition of nitric oxide on the decomposition of ethylene oxide was examined. Average values for the collision efficiency of the nitric oxide-tmethyl radical reaction at 950", and the methyl radical recombination reaction from 925" to 975" were found to be 2 x 10-4 and 2.5 x 10-2 respectively. A number of studies of reaction intermediates in gaseous reactions have been made using mass spectrometers. A study of relatively long-lived intermediates in the thermal decomposition of dimethyl ether and acetaldehyde was carried out by Leifer and Urey.1 The detection of free radica!s such as methyl from the pyrolysis of methane and lead tetramethyl, methylene from diazomethane, ally1 from propylene, and ethyl from ethane, as well as a study of the role of inter- mediates in low pressure flames, was made by Eltenton.293 The measurement of the ionization potential of methyl radicals produced by the thermal decomposi- tion of lead tetramethyl was carried out by Hipple and Stevenson.4 Robertson 5 detected methyl and ethyl radicals in the pyrolysis of methane and ethane on a platinum filament.For the detection of radicals, the authors mentioned above made use of the fact that the ionization potential of a free radical is less than its appearance potential from a.@ven compound by an amount of energy equal to the sum of the strength of the bond broken and the kinetic energy of the fragments formed by electron impact.In most cases the ion current due to R+ formed by electron * N.R.C. contribution no. 2770. -t National Research Council Post-graduate FelIow.F . P . L O S I N G , K . U. INGOLD AND A . W. TICKNER 35 dissociation of RR' will disappear at an electron energy a few volts higher than the ion current due to Rf produced by ionization of the radical R already present from the thermal dissociation of RR'. By careful selection of an electron energy greater than the ionization potential of R but less than the appearance potential, the presence of the radical was demonstrated. However, because of the more or less exponential shape of the feet of the ionization efficiency curves a clear separation of the two processes results, in most cases, in a very low ionization efficiency (sensitivity) for the process R + e -+ R-' + 2.Since the ionization efficiency is not only low at these electron energies, but varies with the energy in a manner that is not easily calculated, these conditions are not very suitable for the quantitative determination of the concentration of the radicals. The reproducibility of the ionization efficiency at low electron energy is complicated further by the difficulty of maintaining constant the contact potential and the energy distribution of the electrons. In the present work the method outlined above has been used for the quali- tative detection of free radicals, but the quantitative measurements have been made by measuring the mass spectra using an electron accelerating potential of 50V.Under these conditions the ionization efficiencies and the mass spectral patterns are almost independent of small changes in the electron energy. If all the reaction products which contribute to the radical peak are identified, these contributions can then be subtracted by reference to calibration spectra in the manner well known in the analysis of hydrocarbons.6 In this way a net peak height is obtained which is a measure of the partial pressure of the radical. The use of this method to obtain the sensitivity for methyl radicals formed by the thermal dissociation of mercury dimethyl and di-tert.-butyl peroxide has been described.7 A brief summary of this method is given below under the heading of Calibration for methyl radical sensitivity.In the present work the thermal decompositions of ethylene oxide, propylene oxide, dimethyl ether, and dioxane have been examined by this method to deter- mine the extent to which free radicals play a part. An analysis of the stable products was also made. EXPERIMENTAL THE REACTOR AND MASS SPECTROMETER.-A diagram of the reactor and the ion source of the mass spectrometer used in this work is shown in fig. 1. The reactor consisted of a quartz tube of 7 mm internal diameter which was heated by an element consisting of a sheet of tantalum 0.005 in. in thickness rolled into a cylinder 12.7 mm in diameter and held in this shape by lugs spotwelded at the edge.The upper end of the heating element fitted over, and was spotwelded to a steel tube which was bored out to make a sliding fit over the upper end of the quartz tube. The lower end of the element was spotwelded to the inside of a ring which made electrical contact to the outer steel jacket through four heavy iron legs. The legs were soft-soldered to the inside of the steel jacket after assembly. The lower end of the quartz tube extended to the lower edge of the ring. The expansion of the heating element was taken up by a bellows which was silver-soldered over a sliding joint in the upper steel tube. The furnace was surrounded by a radiation shield of thin quartz coated with aquadag. A thermocouple well of thin quartz tubing was sealed into the side of the quartz tube 2 cm from the lower end, and passed out through holes in the heating element and the radiation shield.A chromel-p alumel thermocouple was used to measure the temperature. The outer steel jacket was provided with a cooling coil of copper tubing through which water was passed. Electrical contact to the lower end of the heating element was made through this coil and to the upper end through the inner steel tube. A ring of glass sealed to two rings of Kovar prevented the short-circuiting of the element by the outer steel jacket. The current required to heat the reactor to 1200" C was about 250 A at 4 V. This was supplied from a transformer consisting of eight turns of in. copper tubing wound on the core of a 50 A Variac. Electrical connection from this winding to the heating36 MASS S 1'E C T K 0 ME I' R Y element was made using & in.copper tubing connected by pressure fittings. This enabled the transformer secondary to be cooled by circulating water, to prevent large temperature changes which would cause the temperature of the reactor to drift. The primary current to the 50 A Variac was controlled by a 10 A Variac operated from a 110 V line stabilized Q W . C L FIG. 1,-Reactor and ion source. by a Sorenson electronic regulator. This form of reactor was found to be quite satisfactory and did not have a large temperature lag. The temperature could be held within 4: 1" C for long periods. The effective length of the furnace, although rather uncertain, was taken to be 7 cm. The flow of gas from the reactor into the mass spectrometer took place through a hole in a thin quartz thimble which pro- jected into the heated zone of the reactor.The diameter of the hole was 30 microns and the thickness of the wall surrounding the hole was about 16 microns. The hole was formed by means of an electric spark as previously described,7 and was almost perfectly circular with smooth edges. The bottom of the thimble was thickened and ground flat, and was cemented to a silver disc with silver chloride. The disc was soft-soldered to a threaded steel cap which closed the end of the reactor. The major part of the gas entering the reactor passed out through the space around the outside of the heating element, but a small portion issued from the orifice in the quartz thimble into a region of very low pressure as a mole- cular beam which was directed into a hole 2 mm in diameter in the plate covering the ionization chamber.The distance from the orifice to theelectron beam was about 1.5 cm. The mass spectrometer itself 7 was a 90" sector type instrument using an ion source with drawing-out potentials as described by Nier,8 but with a modification to provide for differential pumping of the filament chamber. The analyzer tube and the chamber into which the reactor and the ion source projected were pumped separately by high speed mercury diffusion pumps. The electronic controls for the mass spectrometer were similar to those described by Thode.9 Magnetic scanning and pen recording 10 were employed. The sensitivity of recording was such that an ion current of 3 x 10-14 A gave a deflection of 1 cm on the Speedomax recorder.THE now SYSTEM.-AIl the decomposition experiments were carried out in a stream of helium at pressures in the region of 10 mm. The flow of helium was regulated by a small metal needle valve opening from a 3 1. flask in which a pressure of about 10 cm of helium was maintained by a rough manostat. The flow of reactant gas was con- trolled by a metal needle valve opening from a trap in which a fixed pressure in the region of 5-15 cm of reactant was maintained by means of a freezing bath of an appropriate substance. By this means a flow of helium at about 10 mm pressure containing a partial pressure of a few microns of reactant could be held constant for an hour or more with a drift of only a few per cent.This could be allowed for by timing all measurements and using a linear time correction. METHOD OF OPERA=TION.-A flow of helium containing a small partial pressure of reactant was passed through the reactor as described above, and mass spectra were obtained at several reactor temperatures below and through the range of temperature in which the reactant decomposed. The method of detection of methyl radicals from mercury dimethyl using an electron energy of about 11 V is shown in fig. 2. The mass 15 peak in the upper spectrum, which was obtained at 200°C where no decomposition occurred, was due to CH3+ formed from mercury dimethyl by electron dissociation.F . P . LOSSING, K. U. INGOLD AND A . W. TICKNER 37 This was the remainder of a peak which was over 100 times as large when measured using 50 V electrons.The peaks at 18 and 28 are those of residual water and carbon monoxide which are present in all spectra. The lower spectrum was obtained with the same electron energy but with the reactor at 800" C and with a contact time of 2.1 x 10-3 sec. The increase in the mass 15 peak corresponds to CH3+ formed by the ionization of methyl radicals produced in the thermal decomposition of the mercury dimethyl. A trace of the ethane spectrum can be seen at the mass 28-30 region. This method was used for detecting radicals, but the quantitative measurements were made by measuring the spectra obtained with 50 V electrons, and subtracting the superposed spectra of the reaction products as described above.- 18 R FIG. 2.-Detection of methyl radicals from mercury dimethyl. The upper spectrum was obtained at 200" C, the lower at 800" C. The electron energy was 11 V. CALIBRATION FOR STABLE PRoDuCTS.-The mass spectral patterns of the reactants and the stable products were determined in separate experiments by passing through the reactor a stream of helium containing a convenient amount of the gas to be examined. By comparing spectra obtained at different reactor temperatures, it was found that in many cases the spectra changed slightly with temperature as previously reported for hydrocarbons.ll.12 For the reactants, this effect could not be measured over the whole temperature range since decomposition occurred. The mass spectral patterns of the reactants at the higher temperatures were obtained by extrapolation from lower temperatures.In addition to the change in the patterns, it was found that the sensitivity for a given substance decreased with increasing temperature. This effect could be ascribed partly to the effect of temperature on the flow of gas through the orifice, and partly to the effect of increased molecular velocity on the ionization probability. Since the flow through the orifice was partly molecular and partly viscous at these pressures, the relation between sensitivity and temperature was not simple, and was obtained by calibration. The " tem- perature coefficient of sensitivity " measured in this way was applied as a correction to the partial pressures of the products measured at any temperature.This coefficient was in some cases as large as 0.50 for a range of 800" C. The sensitivities for the reactants and products relative to that for methane were measured in separate experiments. Known proportions of methane and the gas in question were passed in a stream of helium through the reactor at a fixed temperature. These relative sensitivities were then related to the actual pressure in the reactor by measuring the sensitivity for a known partial pressure of methane, and could be expressed as a peak height per micron pressure. methyl radicals to that for methane with 50 V electrons was previously determined 7 by decomposing a known quantity of mercury dimethyl and by making a continuous analysis of the products, of which methyl was the major component. The stable pro- ducts were ethane and methane. The net peak height at mass 15 due to methyl was obtained by subtracting from the total mass 15 peak the known contributions from methane, ethane, and mercury dimethyl.The sensitivity to methyl was then obtained by assuming a 100 % carbon balance. In order to be more independent of instrumental factors, this sensitivity was reported relative to that for methane. The value of sensitivity for CH3/sensitivity for C& was 0.47 f 0.07. MAmmLs.-The helium was of 99.8 % purity or better and was further purified by passing it through a charcoal trap immersed in liquid nitrogen. The oxygen content was reduced to less than 0.005 % by this means. The ethylene oxide was obtained from CALIBRATION FOR METHYL RADICAL sENsImvlTY.-The ratio of the Sensitivity for38 MASS SPECTROMETRY the Matheson Co., the propylene oxide from the Eastman Kodak Co., the 1 : 4-dioxane from Anachemia Chemicals Ltd., and the dimethyl ether from the Ohio Chemical and Manufacturing Co.All these materials were purified by several distillations at reduced pressure. RESULTS AND DISCUSSION ETHYLENE oxIDE.-The decomposition of ethylene oxide was investigated in the temperature range 800" to 1000". The rate constants given (at 850') in table 1 show the reaction to be first order with respect to the partial pressure of the ethylene oxide, but the order appears to depend slightly on the helium pressure. The activation energy is independent of both ethylene oxide and helium pres- sures ; a mean value of 421 kcal was obtained which is in closer agreement with that of Rice and Johnston 13 value for the split into free radicals (44 kcal) than with activation energies measured in static systems for the overall reaction.TABLE EF EFFECT OF HELIUM AND ETHYLENE OXIDE PRESSURES ON k AND E helium ethylene oxide k at 850' C activation pressure (mm) pressure (/A) (SeC-1) energy (kcal) 5.0 5.68 42.4 41.6 10.0 1-64 54.6 41.0 15.0 1.37 55.7 41.2 15.0 1.58 54.1 42.5 15.0 10.98 55.2 43.8 The products of reaction, at three different temperatures and slightly differing times of contact, are given in table 2. The partial pressures of the products are given in microns. The reliability of the data is evaluated in table 3 which shows the amounts of carbon, hydrogen, and oxygen found in the products expressed as a percentage of the amount which should be present as calculated from the amount of ethylene oxide decomposed, e.g., The carbon and oxygen balances are fairly good, but the hydrogen balance is particularly susceptible to slight changes in the relative amounts of the different products.The last column in table 3 shows the number of methyl radicals formed per ethylene oxide molecule decomposed. The values are based on the assumption that all the ethane formed comes from the recombination of two methyl radicals. All the products listed have been reported previously.14-19 In particular methyl radicals were specifically reported by Rice and Johnston.13 Methylene radicals, first postulated by Fletcher and Rollefson,l~ have never been detected TABLE 2.-THE THERMAL DECOMPOSITION OF 9 .2 4 ~ OF ETHYLENE OXIDE IN 15 MM HELIUM (Partial pressures of products are given in microns) temp. ("C) contact time % (secx 10-3) decomp co CH4 H2 CH3 CH2=CO 875 0.90 16-7 1-49 0.40 0.055 0.73 0.70 0-087 925 0.86 32.0 2.76 0.74 0.11 1.77 1.63 0.17 975 0.83 50.1 4.28 1.19 0.20 2.84 2.65 0.30 TABLE 3.-THERMAL DECOMPOSITION OF 9 . 2 4 ~ OF ETHYLENE OXIDE % carbpn % hydrogen % oxygen cH3 formed per found in found in found in ("C) products products products G-O decomp. rnol* temp. 875 92 91 I01 0.520 925 96 105 99 0 626 975 99 110 100 0.663F . P . LOSSING, K . U. INGOLD AND A . W. TICKNER 39 experimentally and we too were unable to find them. Nor were any hydrogen atoms detected, though the large amount of hydrogen gas in the products indicates that these may play some part in the reaction.They will, however, recombine very readily and are therefore not very susceptible to detection by the present method. We were not able to discover any propane,l8 formaldehyde (a postulated intermediate 15) nor any acetaldehyde.18 In a separate experiment it was shown that acetaldehyde did not decompose under our experimental conditions. At 850" there were no products of greater molecular weight than ethylene oxide in the mass range 50 to 100. ADDITION OF NITRIC OXIDE.-Although some previous measurements have been made on the inhibiting effect of both nitric oxide 15s 16Cs 20 and propylene 19s 21 on the decomposition of ethylene oxide, no direct measurements have previously been made on the change in methyl radical concentrations caused by the addition of an inhibitor.So in order to obtain additional information as to the extent to which methyl radicals play a part in the decomposition, some experiments were carried out in which nitric oxide was added to the reacting gas stream. The effect of added nitric oxide (table 4) is to decrease slightly the amount of ethylene oxide decomposed, i.e. there is partial inhibition of the reaction. Methyl radicals are still present but are not so abundant and the concentration of methane in the products is also decreased. The decrease in methane and methyl concentra- tions are given as a percentage of the amounts present in the absence of nitric oxide. TABLE 4.-THERMAL DECOMPOSITION AT 950" c OF 28 p OF ETHYLENE OXIDE IN (Carrier gas 15 mm helium ; contact time 0.83 x 10-3 sec) PRESENCE OF NITRlC OXIDE decrease of decrease of of NO of NO pressure of NO $52: [CH4] caused [CHJ] caused -As position by addition by addition ACHj (4 (1) 186 3.8 % 12.8 % 22-9 % 0.56 (2) 402 3.5 % 18.7 % 28.6 % 0-65 The final column in table 4 gives the ratio of the decrease in methane to the decrease in methyl concentration.This ratio, virtually constant for a 2.3 fold change in nitric oxide concentration, has a value of 0.6 or 3/5. There are two fundamentally different reactions by which methane can be formed during the thermal decomposition of ethylene oxide, viz., (1) (2) If the concentration of RH is not affected by the addition of nitric oxide, i.e. if it is not a radical itself, then 2/5 of the methane will be formed by reaction (1) and 3/5 by reaction (2).From the analytical data at 975" it is seen that 26 % of the ethylene oxide decomposed gives rise to methane. Assuming RH to be ethylene oxide, this can be divided between reactions (1) and (2), i.e. at 975" reaction (1) occurs to 26 x 215 = 10.4 % and reaction (2) to 26 x 3/5 = 15-6 % of the ethylene oxide decomposed. On the other hand, if RH is wholly or in part a radical whose concentration is decreased by the addition of nitric oxide (a H atom, for example) then the amount of methane formed in reaction (1) will be greater than 10.4 %, by an amount depending on the decrease of the concentration of RH. The addition of nitric oxide decreased the methyl concentration by 22.9 % and 28.6 % in two separate experiments.On the simple assumption that RH is wholly ethylene oxide we can calculate the expected degree of inhibition of the 15'6 22*9 . 3.6 % and -- ;-- 4.4 % respectively. reaction to be --- (CH2)20 3 CH4 (+ CO) CH3 + RH -+ CH4 (+ R) 15.6 x 28.6 100 10040 MASS SPECTROMETRY These results are in good agreement with the experimental measurements given in table 4, i.e. 3.8 % and 3.5 % respectively. REACTION MECHANISM.-An elaboration of Sickman's 22 free radical mechan- ism for the decomposition of ethylene oxide has been put forward by Fletcher and Rollefson.1~~ It involves induced decomposition by a methylene radical formed in the primary reaction to two methyl radicals and carbon monoxide, CH2 + (CH2)20 --t 2CH3 + CO. Although the methylene radical should be susceptible to detection by the present method 3 none was in fact found among the decomposition products of ethylene oxide.Moreover, Fletcher and Rollefson's mechanism does not postulate any reaction between methyl radicals and ethylene oxide. It has recently been shown, however, by both thermal 19 and photochemical 23 studies, that the reaction CH3 + (CH2)20 + products occurs readily. This, in conjunction with our own experiments in which methyl radicals were the only free radicals to be detected, leads us to the view that not only do methyl radicals induce the decomposition of ethylene oxide, but also the primary split of the parent molecule gives a methyl radical (CH2)20 +- CH3 + CHO. (3) The number of methyl radicals formed per ethylene oxide molecule decom- posed appears to be considerably greater than under Fletcher and Rollefson's experimental conditions (see table 3).Fletcher and Rollefson reported a value of 0-28. In the calculation of our values it is assumed that all the ethane arises from the recombination of two methyl radicals. This assumption is almost certainly true, but since the ethane concentration is only 13-15 % of the methyl concentra- tion, if a part of the ethane arises in some other way the values given in table 3 will not be greatly affected. It should be noted that the sensitized decomposition of ethylene oxide will not appreciably affect the radical concentration, since apart from the small amount of side reaction to ketene, each methyl radical removed by reaction is regenerated. This will become clear from an examination of the proposed reaction mechanism.In actual fact the values given in table 3 are mini- mum values since some of the methyl radicals will disappear by reaction with other free radicals, e.g. CH3 + H(+ M) -+ CH4(+ MI. (4) (CH2)20 3 H . CHO + CH2 The mechanism proposed below is similar in many respects to that proposed by Gomer and Noyes 23 for the photochemical decomposition of ethylene oxide at 175", the major differences being that no formaldehyde18 or acetaldehyde was found. Moreover the C2H3O radical, if it exists at all as a separate entity at our elevated temperatures, has too short a life to combine with other atoms or radicals. Thus we find no products of greater molecular weight than ethylene oxide and in particular we find no propionaldehyde.The most important primary reactions and most of the possible radical re- moving steps are listed below : (CH2)20 --+ CH4 + CO (CH2)20 + CH3 + CHO CHO -+ CO + H (1) (3) (5) (6) (7) (8) CH3 + (CH2120 + CH4 + C2H30 H 3- (CH2120 + H2 + C2H30 C2H3O --f CH3 + COF. P. LOSSING, K. U. INGOLD AND A . W. TICKNER 41 C2H3O -+ CH2CO + H CHO + CHO -+ 2CO + H2 H + CHO -+ CO + H2 CH3 + CHO + CO + CH4 CH3 + CH3 + C2H6 (M+)H -I- CH3 -+ CH4(+M) (M+)H + H 3 H2(-t-M) and possibly (CH2)20 -+ CH2CO f- H2 The radicals H, CHO and C2H3O were not detected experimentally and therefore, if they exist in the free state, have a shorter life than the methyl radical. The CHO radical has become quite popular in reaction kinetics; its heat of dissoci- ation to CO and H is not large.Its detailed reactions are not known, but all those postulated are quite probable. If all the methane comes from reactions (1) and (6) and all the hydrogen atoms are removed by reaction (7),24 a minimum value to the extent of reaction (3) can be given. At 975" this will be $(lo0 - 26) = 37 %. The half arises because effectively all the C2H3O produced in (7) gives a methyl radical by reaction (8). The amount of ethylene oxide decomposing by reaction (3) will be increased above 37 % in the event that either of the following reactions occur; that some CHO disappears by reactions not forming hydrogen atoms or that hydrogen atoms are removed by reactions not regenerating CH3. COLLISION EFFICIENCY OF REACTION BETWEEN METHYL RADICALS AND NITRIC OXIDE.-The collision efficiency for the reaction CH3 + NO --> products is given by the number of collisions between methyl radicals and nitric oxide which lead to reaction, divided by the total number of collisions.The pressure of nitric oxide was converted to molecules cm-3 and then substituted in the equation which gives the number of collisions each methyl has with nitric oxide per cm3 per sec. The mean molecular diameter was taken as 3-75 x 10-8 cm.25 The collision efficiency was obtained by multiplying the collision number 2 by the contact time and dividing into the proportion of methyls which react (see table 4). The results obtained in the two experiments and the values obtained by earlier workers are given in table 5. The close agreement obtained over such a large temperature range may indicate that the activation energy of the reaction is small, or zero.TABLE 5.-cOLLISION EFFICIENCY OF REACTION C H 3 -1- NO -+ PRODUCTS workers Forsyth 25 Staveley and Hinshelwood 26 Durham and Steacie 27 present experiment (1) [CH,] = 6.0 p present experiment (2) [CH3] = 5.2 p temp. collision ("C) efficiency 540 6.6 X 10-4 800 1.4 x 10-5 25 1.5 x 10-4 950 2.6 x 10-4 950 1.4 x 10-4 COLLISION EFFICIENCY OF METHYL RECOMBINATION.-If it iS assumed that aII the ethane formed in the ethylene oxide decomposition arises from methyl re- combination, the data for the partial pressure of methyl and ethane given in table 2 can be used to calculate a collision efficiency for the recombination. The calcu- lated values are 2.1 x 10-2 at 975" and 2.9 x 10-2 at 925".These are in good agreement with the average value of 2.9 x 10-2 previously calculated in the same way from the data for mercury dimethyl decomposition.742 MASS SPECTROMETRY PROPYLENE oxIDE.-The major products of the thermal decomposition of 3 . 2 ~ of propylene oxide at a temperature where decomposition is essentially complete (950") are shown in table 6.16~ The partial pressures of the products are given in microns. In addition to the products listed, traces of water, ketene, propylene and allene were found. Since the only important oxygenated product was carbon monoxide it was assumed that under these conditions of complete decomposition the amount of propylene oxide originally present was equal to the amount of carbon monoxide formed.The carbon and hydrogen balances were worked out from this basis. The activation energy of the reaction was 31.8 kcal, and the rate constant at 650" was 165.1 sec-1. TABLE 6 . T H E THERMAL DECOMPOSITION OF 3.2 p PROPYLENE OXIDE AT 950°C IN 1 0 m m HELIUM (Contact t i m e = 1 - 8 5 x 10-3 sec) 0.36 1.93 0.27 0.44 1 . 0 4 3.25 3 - 2 1 C2H6 c2H4 C2H2 CH4 CH3 H2 co % C found in products =: 103 % H found in products = 1 1 3 It is apparent from the large quantity of methyl radicals present in the pro- ducts that a free radical mechanism is of importance in the decomposition of propylene oxide. The compound was, however, not examined in sufficient detail for this mechanism to be thoroughly elucidated; it will, however, probably be quite similar to that for ethylene oxide, and the following is suggested : CH3CHCH2O -+ CH3 + C2H3O.I I Since no formaldehyde was detected reaction (2) must be very fast at this tem- perature. The two radicals formed in reaction (3) will follow the same paths as in ethylene oxide. The acetylene probably arises from dehydrogenation of ethylene. Calculations based on the collision efficiency of the CH3 + CH3 reaction show that only 0*03p of ethane arises from this reaction, the remaining ethane must arise from the hydrogenation of ethylene. The minimum value for the number of methyl radicals formed per propylene oxide molecule decomposed is therefore (1.04 + 2 x 0.05)/3.2 = 0.36. If it is assumed, as is quite probable, that all the methane arises from the reaction of methyl radicals with propylene oxide a value of 0.48 is obtained.CH3 -t CHJCHCH~O -+ CHq + C3H50 I I (4) I--I C3&0 -+ C2H4 + CHO. ( 5 ) On the basis of these simplifying assumptions, the extent of reaction (1) is then 64 %, reaction (3) 24 % (i.e. & x 0.48 x 100, since two methyl radicals are formed for every propylene oxide molecule decomposing by reaction (3)), and reaction DIMETHYL ETHER.-The products of the thermal decomposition of 2.23 p of dimethyl ether are shown in fig. 3 as a function of temperature, together with the decrease in the amount of dimethyl ether remaining. In table 7 these results are summed up in the carbon and oxygen balances. The oxygen balance ap- proaches 100 at the higher temperatures, but it is apparent from the carbon balances that about 20 % of the carbon is disappearing to compounds we have not detected.This result we have so far been unable to account for satisfactorily. At the lower temperatures it might be due to polymerization of the formaldehyde. (4) 12 %.F . P. LOSSING, K . U. INGOLD A N D A. W. TICKNER 43 The activation energy of the reaction was found to be 32.4 kcal with a rate constant at 650" of 8.7 sec-1. An activation energy of 17.8 kcal was obtained for methane production, using the results at the lower temperatures where secondary reactions are comparatively unimportant. This compares favourably with a value of 16 kcal obtained by Leermakers 28 (who, by photolyzing acetone as a source of methyl radicals, sensitized the decomposition of dimethyl ether at temperatures from 270" to 400"). There is abundant evidence 29-31 for the existence of free radicals and chain reactions in the decomposition.The build up and subsequent decay of the inter- mediate formaldehyde during the decomposition at 500" has been studied pre- viously by Leifer and Urey 1 using a mass spectrometer, which was not, however, FIG. 3.-Formation of products from the decomposition of 2-23 p of dimethyl ether. designed to study free radicals. The same build up and decay of formaldehyde is apparent in fig. 3. The formaldehyde decomposes to hydrogen and carbon monoxide, as is clearly shown by the rapid increase of these two products. The concentration of ethane and methyl radicals also pass through a maximum, the former because of some decomposition to ethylene and hydrogen (at the highest temperature, 1050", the ethylene has increased to an amount equal to the ethane).The decrease of the methyl concentration may be due in part to reaction with the rapidly increasing concentration of hydrogen. Except for the absence of the CH30 radical, which may well be unstable at these temperatures, our results appear to be in accord with the mechanism of Rice and Herzfeld.32 TABLE 7.-cARBON AND OXYGEN BALANCES IN DECOMPOSITION OF DIMETHYL ETHER temp. ("C) 775 850 00 950 1000 % decomposition 32-0 62-0 79.0 92.0 99.0 contact time (sec x 10-3) 2.16 2-0 1 1 -93 1-85 1.78 % C found in products 75.0 80.0 80.0 83.0 82.0 % 0 found in products 8 7 . 0 94.0 95.0 98.0 104.0 DIoxAm-The main products of decomposition of dioxane at 915" were carbon monoxide, hydrogen and ethane, with smaller amounts of water, formal- dehyde, ketene, ethylene and acetylene.Methyl radicals were present only in44 MASS SPECTROMETRY amounts corresponding to about 1 % of the dioxane composed. This quantity may well arise merely from the decomposition of ethane. No methylene or other radicals were detected. This is interesting in view of the strongly inhibiting action of nitric oxide reported by Kiichler and Lambert.33 However, these workers report the reaction to be not catalyzed by methyl radicals from the decomposttion of azomethane, and moreover, the para-ortho hydrogen method gave negative results. The radicals reported by Forsyth 25 are undoubtedly methyl. The activation energy of the decomposition was 433 kcal, and a value of 9.8 sec-1 was obtained for the rate constant at 650".Since practically no radicals were found, this compound was not studied in detail. The authors wish to thank Dr. E. W. R. Steacie for encouragement and advice during the course of the work. They also wish to express their appreciation to Mr. G. Ensell for his aid in many glass-blowing problems which arose in the construction of the reactor. 1 Leifer and Urey, J. Amer. Chem. SOC., 1942, 64, 994. 2 Eltenton, J. Chem. Physics, 1942, 10,403. 3 Eltenton, J. Chem. Physics, 1947, 15, 455. 4 Hipple and Stevenson, Physic. Rev., 1943, 63, 121. 5 Robertson, Proc. Roy. SOC. A, 1949, 199, 394. 6 see for instance Brewer and Dibeler, J. Res. Nat. Bur. Stand., 1945, 35, 125. 7 Lossing and Tickner, J. Chem. Physics, 1952, 20, 907. 8 Nier, Rev. Sci. Znstr., 1947, 18, 398. 9 Graham, Harkness and Thode, J. Sci. Znstr., 1947, 24, 119. 10 Lossing, Shields and Thode, Can. J. Res. B, 1947, 25, 397. 11 Fox and Hipple, J. Chem. Physics, 1947, 15, 208. 12 Berry, J . Chem. Physics, 1949,17, 1164. 13 Rice and Johnston, J. Amer. Chem. Soc., 1934, 56,214. 14 Heckert and Mack, Jr., J. Amer. Chem. SOC., 1929,51,2706. 15(a) Fletcher, J. Amer. Chem. SOC., 1936, 58, 534; (6) Fletcher and Rollefson, J. Amer. Chem. SOC., 1936, 58, 2135. 16(a) Thompson and Meissner, Nature, 1936, 137, 870; (b) Trans. Fara&y SOC., 1936, 32, 1451 ; (c) Trans. Faraday SOC., 1938, 34, 1222. 17 Seddon and Travers, Proc. Roy. SOC. A, 1936, 156, 234. 18 Simard, Steger, Mariner, Salley and Williams, J. Chem. Physics, 1948, 16, 836. 19 Mueller and Walters, J. Amer. Chem. Soc., 1951, 73, 1458. 20 Steacie and Folkins, Can. J. Res. B, 1939, 17, 105. 21 Rice and Polly, J. Chem. Physics, 1938, 6, 273. 22 Sickman, J. Chem. Physics, 1936, 4, 297. 23 Gomer and Noyes, Jr., J. Arner. Chem. SOC., 1950, 72, 101. 24 Trost, Darwent and Steacie, J. Chem. Physics, 1948, 16, 353. 25 Forsyth, Trans. Faraday SOC., 1941, 37, 312. 26 Staveley and Hinshelwood, Proc. Roy. SOC. A, 1937,159, 192. 27 Durham and Steacie, J. Chem. Physics (in press). 28 Leermakers, J. Amer. Chem. SOC., 1934, 56, 1899. 29 Steacie, Atomic and Free Radical Reactions (Reinhold Publishing Corporation, 30 Klute and Walters, J. Amer. Chem. SOC., 1945, 67, 550. 31 Goldsanskif, Upsekhi Khim., 1946, 15, 63. 32 Rice and Herzfeld, J. Amer. Chem. Soc., 1934, 56, 284. 33 Kuchler and Lambert, Z. physik. Chem. B, 1937, 37, 285. New York, 1946), p. 110.

ISSN:0366-9033    

DOI:10.1039/DF9531400034

出版商:RSC    

年代:1953

数据来源: RSC

摘要:

THE REACTIONS OF METHYL RADICALS WITH THE HYDROGEN ISOTOPES BY T. G. MAJURY * AND E. W. R. STEACIE Division of Chemistry, National Research Council, Ottawa, Canada Received 3rd March, 1952 CH3 and CD3 radicals were prepared by the photolysis of the appropriate acetone and their reactions with hydrogen and deuterium studied in the range 130-300". The energies of activation and steric factors for the four reactions were evaluated ; the former lie in the range 9-12 kcal and the latter are of the order of 10-3. The main influence on reaction rate is due to the isotope of hydrogen concerned, the nature of the methyl radical exerting relatively little effect. In a recent series of papers,l we have attempted to obtain information on the reactions of methyl radicals with a variety of hydrogen-containing substances, of the type The results obtained are of some interest in connection with the effect of structure on the reaction rate. They are also of interest in that they show that the steric factors for these reactions are of the order of 10-3 to 10-4, which is considerably lower than the value of 0.1 often arbitrarily assumed in the past.From a similar point of view it is of interest to investigate the reaction of methyl radicals with hydrogen, especially in view of the importance of this re- action, and of the wide discrepancies in previous results? It is also of interest in connection with isotope effects to extend such an investigation to include the reactions of CH3 and of CD3 with H2 and with D2. Finally, as discussed later, the results are of considerable interest in connection with the reverse reaction CH3 + RH + CH4 + R.H + CHq --+ CH3 + H2. EXPERIMENTAL METHoD.-It has been shown3 that in the photolysis of acetone between 130" and 300" the production of methyl radicals and the formation of methane and ethane are accounted for by the following reactions : (1) (2) (3) CH3. CO . CH3 -+ 2CH3 + CO CH3 + CH3 -> C2H6 CH3 + CH3. CO . CH3 -> CH4 -1- CH2. CO . CH3 Assuming a steady concentration of methyl radicals, it follows that In the presence of hydrogen the methyl radicals also react as follows whence CH3 + H2 -+ CH4 + H (4) where (RCHJ4 represents the rate of production of methane by reaction (4). The value of (RCH$4 can be obtained by subtracting (RcHo)3 from the total methane, the value of * National Research Council of Canada Postdoctorate Fellow, 1950-51.4546 REACTIONS OF METHYL RADICALS (&H4)3 being obtained from experiments with acetone alone. Alternatively, for CH3 and D2, or for CD3 and Hz, (RCH4)4 can be obtained from an isotopic analysis of the methanes produced, using a mass spectrometer. The fate of the H atom produced in reaction (4) is presumed to be H + CH3 CO . CH3 --> Hz + CHz . CO . CH3. ( 5 ) The alternative (6) appears to be ruled out as a significant process by evidence obtained in the course of the present work. The apparatus, experimental procedure, etc., were similar to those described in previous papers.1 No details will be given here, since the work will be published in full shortly. H + CH3 -+ CH4 RESULTS Six sets of experiments were performed in all, viz.the photolyses of normal and of deutero-acetone alone, and of each of these in the presence of hydrogen and of deuterium. The results for acetone and deutero-acetone done are in excellent agreement with previous work.1 e.0 1'5 1'0 0 5 0 I I I I \ I 1-9 2'1 2.3 2'5 FIG. 1. CH3 -F CH3COCH3 open circles CH3 + H2 open squares CH3 i- D2 open triangles CD3 --I- CD3COCD3 filled circles CD3 + H2 filled squares CD3 + D2 filled triangles In investigating the reaction of CH3 with D2 (the photolysis of ordinary acetone in the presence of deuterium), and of CD3 with Hz (deutero-acetone in the presence of hydrogen) it was possible to use two methods, as mentioned above, one depending only on gas analysis for methane, and the other involving an isotopic analysis of thc methanes.In both cases the results agreed to within the experimental error, but the latter method was considerably more precise. It was necessary t o make corrections for isotropic impurities in the deuterium and deutero-acetone used. In the deutero-acetone, 93.2 % of the molecules present had 6 D atoms. Roughly 96 % of the methyl radicals formed were therefore CD3, The deuterium used contained ap- proximately 95 % D2, 5 % HD. In the reaction of CD3 with D2, by basing all results on fully deuterized products no correction was necessary. In the reactions CD3 + H2 The remaining 6.8 % had 1 or 2 H atoms per molecule.T. G . MAJUKY AND E. W. K. STEACIE 47 and CH3 -1. D2 corrections were applied for isotopic impurities.These were, however, small enough so that any uncertainty in the corrections introduces no appreciable error into the results. The fate of the H atom resulting from reaction (4) introduces some,uncertainty into the results. It seeins most likely that this is removed by (5). If so, no complications will be introduced. If the H atom were always removed by (6) it would make little differ- ence, since methane formation would merely be twice too high throughout. If, however, the relative importance of (5) and (6) varied greatly with temperature an error would be introduced into the observed temperature coefficient. It seems highly probable that reaction (6) would be affected by third-body restrictions. It was found, however, that variations in hydrogen concentration and the addition of carbon dioxide had no effect on the kinetics.It therefore seems safe to assume that no appreciable error is introduced by the occurrence of reaction (6). For the present they are merely plotted in fig. 1 and a summary of activation energies and steric factors is given in table 1. The results will be reported in detail later. T A A u , s . 1 ,7-*fiF,DAF~m7WSu w* CHq- 4 N L ~ . l ~ - ~ . B ~ r ~ I s ; u I I T H , Y Y ~ ~ ~ ~ . , ~ ~ ~ ~ l ~ I M AND THE PARENT ACETONE kA/kBa X loi3 EA - WB PA,/PB* x 103 kcal reaction 130" 210" 290" CH3 + CH3. CO . CH3 5.2 37.0 151 1.9 f 0.3 9.5 f 0.1 CH3 + H2 3.3 22.0 87 0.7 f 0-2 9.2 f 0.3 CH3 3. D2 0.6 6.6 37 3.5 f 0.5 11.7 f 0.1 CD3 + CD3. CO . CD3 1.2 10.7 51 1.8 j: 0.5 10.6 -j= 0.3 CD3 -t D2 0.9 8.9 46 2.0 3; 0.6 10.9 f 0-3 CD3 + H2 3.7 31.0 140 2.5 -i- 0.5 10.2 0.2 k A is the rate constant for H or D abstraction ; kB is the rate constant for methyl recombination ; rate constants are expressed in terms of molecules, cm3 and sec.DISCUSSION The values of the ratios of the steric factors to that for methyl combination are given in the second to last column of table 1. The last column of the table gives values of EA - QEB, where A refers to H abstraction and B to methyl com- bination. Recent work on the combination of CH3 radicals4 indicates that median values for PB and En can be accepted as ca. 1 and zero respectively, and these values are no doubt valid also for the combination of CD3 radicals. The figures in the last column of table 1 can therefore each be taken as the actual energy of activation of the reaction in question, and the absolute steric factors will differ little from the values in the preceding column.The steric factors are thus of the order of 10-3 and, like those for other reactions of methyl radicals, are thus considerably lower than the value of 0.1 which has frequently been arbitrarily assumed in the past.5 The values of PA and EA for the reactions of CH3 radicals with hydrogen and deuterium differ appreciably from those recently published by Anderson, Davison and Burton.6 They investigated the reaction of methyl radicals with deuterium by a method similar to that of the present work, and concluded that the activation energy of CH3 + D2 -+ CH3D 4- D (7) was 4.6 kcal higher than that of CH3 + CH3COCH3 -+ CH4 + CH3COCH2 (3) i.e.However, the Arrhenius plot from which the difference was deduced consisted of two portions, corresponding to activation energy differences of 3-0 and 4.6 kcal. E7 = 9.7 + 4.6 = 14.3 kml.48 KEACTIONS OF METHYL RADICALS No real reason was advanced for the choice of 4.6 kcal. In view of this and of other uncertainties we think that their results can be fairly expressed by the average difference, with an uncertainty of about 1-5 kcal, i.e. Whence, allowing for the zero-point energy difference, their results for the re- action CH3 + H2 --t CH4 -j- H (4) become approximately E4 = 11.7 & 1.5 kcal. As discussed later, we believe that our results have an uncertainty of about 0.5 kcal, i.e. E4 = 9.7 f 0.6, which virtually agrees with their value.We therefore feel that their results are quite compatible with the lower value of E4 which we have obtained. ISOTOPE EFFECTS.-The standard errors quoted for the activation energies indicate their experimental precision, but having regard to small uncertainties of mechanism their true uncertainty is probably of the order of 0.3 to 0-5 kcal. It is clear that this variance, being fairly large relative to the experimental differ- ences, precludes any very precise interpretation of the results. However, it seems possible to draw two general conclusions regarding isotope effects. First, com- paring the reactions of the CH3 radical with those of the CD3 radical, there is no consistent evidence of any significant difference in their behaviour. Secondly, comparing the reactions of hydrogen with those of deuterium, it will be seen that the latter require the higher activation energies, the average value of the differ- ence being 1.6 -l 0.6 kcal.This is in agreement with theory7 which predicts that at low temperatures the difference in activation energies for reactions of isotopic molecules should approach the difference in their zero-point energies, which in this case is 1.8 kcal. The values of both P and E for the reaction of CH3 with D2 are somewhat out of line with the others. It is probable that in this case errors in E and P compensate one another, and that the real values of both may be somewhat lower. All that can be said for the moment is that for both CH3 and CD3 the results can be expressed by the average values E7 = 13.5 f 1.5 kcal.CH3 (or CD3) + H2 CH3 (or CD3) + D2 E = 9.7 f 0.6 kcal E = 11.3 i- 0.6 kcal. It is also instructive to consider the relative values of the rate constants at a given temperature, since here the results are not affected by compensating errors in E and P, and such relative values should therefore be somewhat more precise. Thus, at 210" C, for the ratios of values of kA/k3, we have CD3 + D2 3.3 1 3.5 j average 3.4 I average = cH3 + D2 = 0.74 I CD3 + D2 0.7. This gives further confirmation of the considerable effect of the substitution of D2 for H2 and the relatively small effect of the substitution of CD3 for CH3. In one case information is available on the corresponding reaction of ethyl radicals, since Wijnen and Steacie 8 have investigated the reaction C~HS + D2 -+ C2H5D + D CH3 + D2 -+ CH3D + D Comparing this withT. G .MAJURY AND E. W. R . STEACIE 49 we have E kcal PIP,* C2H5 f D2 13.3 31 0.5 10-3 CH3 + D2 11.7 f 0.5 3.5 x 10-3 There thus appears to be little difference in the steric factors of the two reactions. The fact that Es is somewhat higher than E7 is in line with the lower value of the bond dissociation energy of C2H6, which makes reaction (8) 3 or 4 kcal more endothermic than (7). of the bond dissociation energy D(CH3-H) is well established at 101 f 1 kcal9 and that D(H-H) is accurately known, we may write THE REACTION OF H ATOMS WITH METHAF4E.h View of the fact that the Value E 9.2 f 0.5 k u l Es CH3 + H2 CHq + H - 2.2 & 1 kcal. Hence, for the activation energy of the reverse reaction Eg = 9.2 - 2.2 = 7.0 f 1.5 k a l .Considerable work has been done on this reaction, almost all by producing H atoms by means of a discharge. Under these circumstances, it is impossible to obtain sufficient accuracy to determine a temperature coefficient with any precision. The results thus obtained are consistent with values of Eg -= 13 j, 2 kcal, assuming a steric factor of 0.1. In all probability the results obtained at higher temperatures should be disregarded, and the true activation energy may be much lower than this if P is also lower. In view of the many qualitative results which show that the reaction of H atoms with methane is much slower than that with other hydrocarbons, it appears, however, that if Eg is significantly lower than the above figure P must be lower also.If 8-5 kcal is the extreme upper limit of Eg from the present work, there remains a considerable discrepancy in E which must be taken care of by a lower steric factor. It may, therefore, be concluded that the present results can only be reconciled with the results on the rate of reaction (9) if P g is between 10-3 and 10-4. It is of interest to note that the results on the reaction which Professor LeRoy is presenting to this Discussion indicate a lower activation energy, together with a lower steric factor for this reaction also than had pre- viously been inferred from experiments by the discharge-tube method. 1 Trotman-Dickenson and Steacie, J. Chem. Physics, 1950, 18, 1097 ; 1951, 19, 169, 329. Trotman-Dickenson, Birchard and Steacie, J . Chem. Physics, 1951, 19, 163. Raal and Steacie, J . Chem. Physics, 1952, 20, 578. 2 for references to previous work, see Steacie, Atomic and Free Radical Reactions (Reinhold Publishing Corporation, New York, 1946). 3Davis, Chem. Rev., 1947, 40, 201. Noyes and Dorfman, J. Chem. Physics, 1948, 16, 557, 788. 4Dodd, Trans. Farahy SOC., 1951, 47, 56. Durham and Steacie, J. Chem. Physics, 1952, 20, 582. Gomer and Kistiakowsky, J. Chem. Physics, 1951, 19, 85. Lucas and Rice, J. Chem. Physics, 1950, 18, 993. Miller and Steacie, J . Chem. Physics, 1951, 19,73. 5 for a discussion of the status of such frequency factors, see Steacie and Szwarc, J . Chem. Physics, 1951, 19, 1309. 6 Anderson, Davison and Burton, Faraduy SOC. Discussion, 1951, 10, 136. Davison and Burton, J. Arner. Chem. SOC. (in press). We are very much indebted to Dr. Burton for communicating his full results to us prior to publication. 7 Bigeleisen, J . Chem. Physics, 1949, 17, 675. 8 Wijnen and Steacie, J. Chem. Physics, 1952, 20, 205. 9 Kistiakowsky and Van Artsdalen, J . Chem. Physics, 1944, 12,469.

ISSN:0366-9033    

DOI:10.1039/DF9531400045

出版商:RSC    

年代:1953

数据来源: RSC

摘要:

THE REACTION OF ATOMIC HYDROGEN WITH ETHANE BY M. R. BERLIE AND D. J. LE ROY Department of Chemistry, University of Toronto, Toronto, Canada Received 18th February, 1952 The reaction of atomic hydrogen with ethane has been studied in a flow system over the temperature range 80" to 163" C. Under the conditions used, methane was the only product. By measuring the rate of consumption of H atoms, as well as of ethane, evidence was obtained in favour of the mechanism : (1) (2) (2d (3) H + C2H6 =-' H2 + C2H5 H + C2H5 = 2CH3 H + C2H5 = C2H6 H + CH3 = CH4. The rate constant of the primary process is given by the relation kl = 3.4 x 1012 exp (- 6800/RT) cm3 mole-1 sec-1 ; the steric factor, calculated from collision theory, is 4.8 x 10-3. The activation energy of 6.8 kcal/mole is shown to be in reasonably good agreement with predictions based on the activation energy of the reverse reaction and the C-H bond dissociation energy of ethane.Although it is generally agreed that the primary reaction of atomic hydrogen with a paraffin can be represented by equations of the type H + CzH6 = H2 + Czb, (1) there has been considerable uncertainty as to the activation energies of such reactions.1 This uncertainty has arisen from the fact that most of the values have been calculated from data obtained for a single temperature. Steacie, Darwent and Trost 1 . 2 have pointed out that the customary value (0.1) assumed for the steric factor, on which the activation energies are based, is undoubtedly too large. In the present investigation we have attempted to overcome this difficulty by studying the reaction between atomic hydrogen and ethane over a range of temperature .In principle, the temperature coefficient of the rate constant should yield a reliable value of E. However, the rate constants obtained by the discharge tube method at any particular temperature are subject to some uncertainty. Aside from random variations, which may approach 50 % or more, there are a number of sources of error which may yield low values of the rate constant. Among these are (i) the assumption that the effective H atom concentration in the presence of reactant is the same as that determined in the absence of reactant, (ii) failure to take into account the removal of H atoms by recombination on the wall or in the gas phase, (iii) failure to take 'into account the manner in which the con- centration of the reactant decreases during the course of the reaction.A fourth source of error is the practice, followed in many cases, of estimating the rate constant from the rate of consumption of paraffin; if any paraffin is re-formed during the course of the reaction the calculated value of the rate constant will be too smalI. This error can, of course, be eliminated by the use of isotopic tracers provided the necessary corrections for mass effects can be made. 50M . R . BEKLlE AND D. J . LE ROY 51 EXPERIMENTAL METHOD.-The H atom technique of Dingle and Le Roy 3 ~ 4 is better adapted than the discharge tube method for the elimination of the errors listed above, and it has been used in the present investigation. Of particular interest is the possibility of following the H atom, as well as ethane, consumption during the reaction.In a preliminary report 5 we have compared the values of the rate constants obtained by the method of Dingle and Le Roy with those of other workers, who used the Wood-Bonhoeffer method ; in every case our values were larger, as predicted. In our earlier work we measured a quantity proportional to the H atom concentra- tion by the use of a platinum or tungsten detector on which H atoms recombined with the liberation of a measurable quantity W of heat per unit time. For the present work we enlarged the area of the detector in the hope of removing all of the H atoms from the gas stream at that point. The corresponding value of W, together with the flow rate and pressure, would then permit the calculation of the actual H atom concentration. The value of W for a fixed set of conditions was found to reach a maximum for a platinum detector 0.0030 in.diam. and 7.7 cm long. The compIete removal of H atoms from the gas stream at any desired point made it possible to follow the ethane concentration as a function of time (or distance down the reactor for a known linear flow rate). Methane was found to be the only product. In the course of an experiment, after measuring the H atom concentration at a number of positions in the reactor in the absence and in the presence of ethane, the detector would be placed at one of four positions, after which the methane content of the gas leaving the reactor was determined.The ethane concentration at any time could then be calculated. RESULTS A N D DISCUSSION In order to calculate the rate constant for reaction (1) it is necessary to assume a mechanism for the overall process and to make proper allowance for the con- sumption of H atoms by recombination. It has been shown previously 4 that under the conditions used the recombination may be either first or second order, but that if the rate of recombination is not excessive the results are not sensitive to the particular order assumed. For the present work we have assumed the first order recombination process, €4 = W2Y (b) and have calculated the corresponding rate constant kb from measurements of H atom decay in the absence of ethane. If reaction (1) is assumed to be rate-controlling, and is followed by H + C2H5 = 2CH3 H f C H 3 =CH4, the rate of consumption of H atoms will be given by the relation (9 in which k’ = 4kl.The H atom concentration at any time is undoubtedly pro- portional to Wy the measured rate of liberation of heat on the detector, and hence 2.303 log ( WO/ Wt) - kb? = k’ [C2H6]0t - 4J‘ [CH4ldt] . (ii) The integral of the methane concentration is determined graphically from the four analyses. To avoid uncertainty in the H atom and ethane concentration in the vicinity of the ethane inlet the time t = 0 was chosen to correspond with the position of the detector at which the first sample was taken for methane analysis. On the basis of the same mechanism the rate of consumption of ethane is given by the relation - d[CzHs]/dt = kf’[H][C2H6], (iii) in which k” = kl.Iff is the constant of proportionality between [HI and W, then - d[Hl/dt = kdH1 + k’[H1[C2H6’1, { 052 REACTION OF ATOMIC HYDROGEN If the mechanism is correct it follows that k' will be equal to 4k". However, on solving eqn. (ii) and (iv) for these quantities it was found that k' was, in general, greater than 4k". It was therefore necessary to assume the occurrence of a reaction leading to H atom consumption without methane formation. The most probable reaction would appear to be (2a), (24 In their investigation of the Hg(3P1) photosensitized decomposition of ethane H + C2H5 = C2H6. at low pressures, Darwent and Steacie 6 found evidence for the reactions H + C2H5 = C2H6* c2H6* = 2CH3 These reactions would be equivalent to (2) and (h), provided the latter is assumed to require a third body.The inclusion of (2a) would also account for ethane exchange in the presence of D atoms.L7 The exchange reaction C2H5 + H2 = CzH6 + H (4) would generate H atoms; if it occurred to any appreciable extent the value of kl calculated from (ii) or (iv) would be too small. However, Kahn and Le Roy 8 found E4 to be greater than 10.5 kcal/mole. The results of Wijnen and Steacie on the reaction of ethyl radicals with deuterium, to which Dr. Steacie has referred in this Discussion, also suggest that reaction (4) is not important under the conditions used in our experiments.15 In addition to reaction (3), which probably requires a third body, methane formation might be attributed to the reactions CH3 + H2 = C& + H ( 5 ) CH3 + C2H6 = C& + C2H5. (6) Trotman-Dickenson, Birchard and Steacie 9 have obtained values for the ratio k6/k74 at various temperatures ; k7 is the rate constant for the reaction 2CH3 = c2H6.(7) In view of the results reported by Majury and Steacie at this Discussion for the ratio ks/k7*, and the fact that in our experiments [Hd was of the order of 50 times [C2H6], it is logical to conclude that reaction (6) may be eliminated as a source of methane. It also seems justifiable to eliminate reaction (5). If it occurred to an ap- preciable extent, k' would be much less than 4k", unless the rate of exchange of ethane by reactions involving H or D atom consumption was, at the same time, considerably greater than the rate of methane formation.This does not appear to be the case; the relatively small amount of deuterium found in the " un- reacted" ethane1.7 can, in view of our results, be accounted for by reaction (2a) or the equivalent process suggested by Danvent and Steacie.6 If reaction (2a) is assumed to occur, as well as (l), (2), (3) and (b), eqn. (ii) and (iv) are still valid, but (v) (vi) and hence ki = k'i2 - k", (Vii) (viii) In table 1 are shown the experimental conditions used and the results obtained over a temperature range of 80" to 163" C. Accurate results can only be obtained if the rate of consumption of H atoms is appreciably greater in the presence than in the absence of ethane. At lower temperatures the decrease in kl can be com- pensated to some extent by increasing the ethane pressure, but in the experiments C2H6* + M = C2H6 + M.k' = 2ki(l + k2/(k2 + k d ) , k" = klk2/(k2 + kza), k2/(k2 + k2a) = 2k''/(k' - 2k").M. R . BERLIE AND D . J . LE ROY 53 reported here the modification in the ethane flow system required for accurate measurements at temperatures lower than 80" C had not been made. The upper limit of the temperature range was fixed by the softening point of the phosphoric acid poison.4 Dingle 10 has shown the reaction of atomic hydrogen with ethane to be first order in ethane over the concentration range 0.4 x lo15 to 4.4 x 101s molecules cm-3. In fig. 1 the average values of log kl for each temperature are plotted against 1/57 The " best " straight line through the points was calculated by the method of least squares. (iX) are 5.7 x 10-12 cm3 molecules-1 sec-1 (3.4 x 1012 cm3 mole-1 sec-1) and 6.8 kcal/ mole respectively.The corresponding values of A and E in the equation kl = A1 exp (- E1/RT) FIG. 1. If the results are interpreted in terms of the collision theory expression in which 2 is proportional to T*, we obtain the values P1 = 4.8 x 10-3, El = 6-4 kcal/mole. This value of P1 is based on the collision diameters used by Trost and Steacie.1 If the results are interpreted in terms of activated complex theory it is readily seen that the pre-exponential factor in the expression for kl should be proportional to T-*; in this case El becomes 7.2 kcal/mole. TABLE 1 The units of concentration and time are molecules cm-3 and seconds respectively. k2Wz +h) 1.00 0.98 087 079 080 0.67 0.66 088 0.97 0-86 0.60 kl x 1015 0350 0.355 0.528 0.580 0-576 1-53 1.38 1.44 2.15 2.14 2.2254 REACTION OF ATOMIC HYDROGEN In view of the precision with which rate constants may be determined in the future, it seems appropriate at this time to consider the form in which the results should be presented. So long as the values of the " constants " are used merely as short-hand representations of k values, it makes little difference which of the three procedures is followed, provided it is stated explicitly.However, if equations of the type (xi) are to be accepted and used in the calculation of activation energies or bond dis- sociation energies, the method of evaluating activation energies becomes of some importance. From this point of view, the only procedure which has a logical, even if not infallible, foundation is that based on activated complex theory.Steacie 11 has estimated that E4 = 9 + 2 kcal/mole. More recently Kahn and Le Roy 8 obtained a minimum value of 10.5. On the basis of their mechanism for the Hg(3P1) photosensitized hydrogenation of ethylene, the actual value of E4 would be equal to 10.5 plus one-half the effective " activation energy " cor- responding to the quantity kg + kg, 2C2H5 = C2H6 + C2H4 (8) 2C2H5 = C4H10. (9) E4 - El = AHQ They found E8 - Eg to be approximately equal to 1-2 kcal/mole. Dorfman and Sheldon 12 obtained a value of 4.8 kcal/mole, but more recent work 13.14 indicates that (8) and (9) have essentially the same activation energy and pre- exponential factor. Furthermore, Ivin and Steacie 13 have estimated that Eg < 0.65 kcal/mole.It would therefore appear that the results of Kahn and Le Roy could be interpreted as giving a value of E4 in the range 10.5 to 11.2 kcal/mole. Modern values of the C-H bond dissociation energy of ethane suggest that in the temperature range used in our present experiments, AH0 would be approx- imately 5 to 6 kcal/mole. If, in view of the uncertainty arising from the method of calculating El, we express our results in the form El = 6.8 & 0.4 kcal/mole, the value of E4 calculated from eqn. (xi) becomes 12.3 & 1 kcal/mole. The agreement with the results of Kahn and Le Roy, and with those of Wijnen and Steacie,ls to which Dr. Steacie has referred in this Discussion, is as satisfactory as could be expected. The authors are grateful to Dr. E. W. R. Steacie for his kindness in making available to them the results of recent investigations in his laboratory. They would also like to acknowledge the financial assistance received from the Associate Committee on Scientific Research of the University of Toronto. One of them (M. R. B.) expresses his appreciation of a National Research Council of Canada Studentship. 1 Trost and Steacie, J. Chem. Physics, 1948, 16, 361. 2 Steacie, Darwent and Trost, Faraday Soc. Discussions, 1947, 2, 80. 3 Tollefson and Le Roy, J. Chem. Physics, 1948, 16, 1057. 4 Dingle and Le Roy, J. Chem. Physics, 1950, 18, 1632. 5 Berlie and Le Roy, J. Chem. Physics, 1952, 20, 200. 6 Darwent and Steacie, J. Chem. Physics, 1948, 16, 381. 7 Trenner, Morikawa and Taylor, J. Chem. Physics, 1937, 5, 203. 8 Kahn and Le Roy, J. Chem. Physics, 1947,15, 816. 9 Trotman-Dickenson, Birchard and Steacie, J. Chem. Physics, 1951, 19, 163. 10 Dingle, Thesis (University of Toronto, 1949). 11 Steacie, Atomic and Free RadicaZ Reactions (Reinhold Publishing Corporation, 12 Dorfman and Sheldon, J. Chem. Physics, 1949, 17, 5 1 1. 13 Ivin and Steacie, Proc. Roy. Soc. A, 1951, 208, 25. 14 Kutschke, Wijnen and Steacie, J. Amer. Chem. SOC., 1952, 74, 714. 15 Wijnen and Steacie, J. Chem. Physics, 1952, 20, 205. New York, 1946).

ISSN:0366-9033    

DOI:10.1039/DF9531400050

出版商:RSC    

年代:1953

数据来源: RSC

摘要:

THE REACTIONS OF HYDROGEN ATOMS WITH HYDROCARBONS * BY B. DE B. DARWENT AND R. ROBERTS f National Research Laboratories, Ottawa, Canada Received 3rd March, 1952 Hydrogen (or deuterium) atoms were produced by the photochemical decomposition of H2S (or D2S). The rates of abstraction reactions of these atoms with ethane, propane, n- and isobutane, ethylene, propylene and 2-butene and of the abstraction and addition reactions with ethylene and propylene were investigated at temperatures between 25" C and 300" C. The activation energies of the abstraction reactions follow a trend similar to that previously found for the methyl radical reactions. The results with propane and rr-butane are in good agreement with the discharge tube data but the rate with ethane was lower and with isobutane higher than found in the discharge tube experiments.The rates of addition of hydrogen atoms to olefines are in essential agreement with the results of Melville and Robb. The steric factors are between 0.1 and 1 .O and appear to be constant for each type of hydrocarbon. The kinetics of the reactions of methyl radicals with hydrogen and with simple hydrocarbons have been investigated 1-5 and discussed.6B 7 The experimental results have tended to confirm the suggestion 6 that the steric factors are about 10-4 instead of about 10-1 required by transition state calculations.7 Although the similar reactions involving hydrogen atoms have been the subject of many investigations,89 99 10 the interpretation of the results of those experiments is not as clear as with the methyl radical reactions.The hydrogen atoms were produced by an electrical discharge and mixed with the hydrocarbon in a high speed flow system. It was found that the activation energy, obtained from the effect of temperature on the rate, increased with increasing temperature and this may have been caused by a progressive, but unmeasurable, change in the concentration of hydrogen atoms with increasing temperature. To circumvent this difficulty the activation energies were calculated by assuming a steric factor of 10-1 and the results at room temperature weighted heavily. The activation energies derived in this manner did not show the same trend that was observed in the abstraction reactions involving methyl radicals. Steacie, Darwent and Trost 6 showed that, if small and variable steric factors were assigned to the reactions of hydrogen atoms with paraffins, the results of the discharge tube experiments and the methyl radical experiments could be reconciled.The experiments described in the present paper were designed to test this suggestion and to obtain more precise data on the reactions of hydrogen atoms with simple paraffins and olefines. The hydrogen atoms were produced by the photolysis of hydrogen sulphide, which reaction has recently been re-investigated,ll and their reactions with olefines and paraffins studied by the following methods. THEORY OF THE METHODS.-The following reactions probably constitute the important steps in the photochemical decomposition of hydrogen sulphide :-*I H2S + h -+ H + HS (0) H + H2S -+ H2 4- HS.(1) * Contribution No. 2758 from the National Research Laboratories, Ottawa, Canada. t N.R.C. Post-Doctoral Fellow (1950-51) ; Present address : Monsanto CheniicaIs 55 Ltd. (Research Department), Ruabon Nr. Wrexham, N. Wales.56 REACTIONS OF HYDROGEN WITH HYDROCARBONS OLEFINES.-Tn the presence of olefines (e.g. propylene) the following processes may also be important : H 4- C~HG -+ C3H7 (2) -> C3Hs + H2 (3) and it may be shown that where Ri and Ro are the rates of production of hydrogen from H2S alone and in the presence of the olefine respectively and the k’s the rate constants. Hence by plotting Ro/(Ri - Ro) against [H~S]/[C~HG] a straight line of slope equal to kl/k2 and intercept k3/k2 should be obtained. If one of those constants is known the others may be calculated.PARAFFINS.-Method I cannot be applied to paraffins since there is now no reaction which removes hydrogen atoms, i.e. kz = 0, and Ri =: Ro and is inde- pendent of pressure. In these cases the reaction was investigated by using deuterium sulphide, the important steps now being (0) D2S + h~ -+D + SD At any time, D + D2S + D2 + SD D + RH -+ HD + R (1’) (3’) and by plotting [HD]/[D2] against [RH]/[D;?S], a straight line of slope k3’/k,’ should be obtained. Thus, in theory, the rate constants for the reactions of deuterium atoms with paraffins and olefines and for the reactions of hydrogen atoms with olefines may be measured relative to those for the reactions involving hydrogen (or deuterium) sulphide. EXPERIMENTAL Hydrogen sulphide and the hydrocarbons were obtained commercially as samples of high purity and used after only simple trap-to-trap distillations t o remove air and water.Hydrogen was purified by passage through a heated palladium thimble. Deuter- ium sulphide was prepared from purified aluminium trisulphide by the reaction with either high purity D20 or deuterium at 400” C. The photolysis was carried out in a quartz cell enclosed in a furnace and connected to a standard type of high vacuum system. Light of suitable wavelength was obtained from a low pressure cadmium lamp (A 2288 A) or from a medium pressure mercury lamp (a broad band at about 2550A). The progress of the reaction was followed either by measuring the hydrogen produced (method I) or by mass-spectroscopic analysis for H2, HD and DZ (method 11).In these experiments (method I) a known amount of hydrogen suIphide was irradiated for successive short intervals ; at the end of each irradiation the hydrogen sulphide was frozen in a trap cooled in liquid nitrogen and the pressure of hydrogen measured. In this way the rate Rj was established ; it was found to be independent of time. After the final irradiation the hydrogen was pumped off and a known amount of the olefine added t o the residual hydrogen sulphide. The rate was measured as before; it was also independent of time and this rate gave the quantity Ro. The extent of the reaction was restricted to such small values that, even after the final exposure in the presence of the olefine, the total amount of H2S reacted never exceeded 5 % and was usually very much less. The experi- ments in which D2S was used were carried out in a similar manner and, in addition, the percentages of H2, HD and D2 were determined.B. D E B.DARWENT AND R. ROBERTS 57 RESULTS THE REACTIONS OF HYDROGEN ATOMS WITH ETHYLENE AND PROPYLENE-METHOD I.- The effect of the ratio of hydrogen sulphide to olefine on the rate of production of hydrogen was studied with ethylene at 27" C and 190" C, and with propylene at 25" C and 205" C. The results obtained by plotting Ro/(Ri - Ro) against [H2S]/[olefine] are shown in fig. 1. The points for any series of experiments fall on one straight line as demanded by theory (eqn. A). With ethylene some of the experiments were done with the mercury lamp ( A W 2550 A) and the others with the cadmium lamp (A = 2288 A) ; in both cases the points fall on the same line, hence the frequency of the light does not influence the results.1 2 3 4 Fro. 1.-The dependence of &/(Ri - Ro) on [HzSl/blefinel for ethylene and propylene: curve A-ethylene at 27" C ; curve B-ethylene at 190" c ; curve C-propylene at 25" C (+) and 205" C (0) with 50 mm &s. (9) at 205" C with 100 mm H2S. The effect of pressure was studied with propylene at 205" C and it is seen that a two-fold variation in pressure did not affect the results. The results are summarized in table 1. in which the values of kllk2 and k3/k2, obtained as the slopes and intercepts respectively in fig. 1, are given for the temperatures used. The differences in activation energies ( , 3 - ~ ~ and E3-E2) were calculated in the usual manner from the effect of temperature TABLE I,-THE ADDITION OF HYDROGEN ATOMS TO ETHYLENE AND PROPYLENE temp."C ethylene 130 ... - ~ 27 II__- - 1.60 2.82 0.00 0.00 large 0.75 9.8 0.90 prop ylene - . - - ~ 25 205 0.96 0.96 0.42 0.42 0.00 0.00 0.72 1.34 0.42 collision diameters : d ~ , H ~ S = 4.5 A ; d ~ . Q H ~ = 4.8 A ; CIH, C,H, = 5.4 A. collision frequencies (molecules cm-3 sec-1) :- H, H2S = 3.6 X 1014; H, C2H4 = 3.8 X 1014; H, C3H6 = 5.0 x 1014.58 REACTIONS OF HYDROGEN WITH HYDROCARBONS on the slopes and intercepts. The ratios of the probability factors (PI/& and P3/P2) were obtained by using those activation energy differences and the values of the collisions diameters shown in the footnote to table 1.The activation energies were derived from results at only two temperatures, but the differences (E1-E2 and E3-Ez) are small and are probably accurate to a few hundreds of calories. The uncondensable gas consisted entirely of hydrogen, except in some experiments with propylene in which a little methane was formed. THE REACTIONS OF DEUTERIUM ATOMS-METHOD rI.-The afore-mentioned experiments (method I) show that the ratios kl/k2 and k3/k2 may be measured. If any one of these rate constants is known an absolute value may be assigned to the others. The reaction of deuterium atoms with hydrogen, D + H2 + HD + H, has been investigated 12 and shown to have an activation energy of 4.85 kcal/mole-1. There is, however, some uncertainty attached to the exact magnitude of this reaction 13 and we have adopted, empirically, the value of 5.0 kcal/mole-1. 0't 1.0 200 3-0 40 FIG.2.-The dependence of HD/D2 on [H2]/[DzS] between 25" C and 178" C. symbol + o + a - @ temp. ("C) 28 94 133 163 178 Deuterium sulphide was photolyzed in the presence of varying amounts of hydrogen at different temperatures and the ratio &[HD]/[D2] is plotted as a function of [H2]/[D2S] in fig. 2 ; the factor 4 is necessary here since the radical R produced in reaction (3') is now a hydrogen atom which must end up as HD, so that two molecules of HD are pro- duced as a result of each reaction of a deuterium atom with H2. The precision of these experiments is not high, since it is difficult to measure accurately small concentrations of HD in the presence of a large amount of H2, but it is evident that there is no systematic change in the slope of the line as the temperature was increased from 28" to 178" C.Hence it may be assumed that the activation energies for the reactions D + H2 -> HD -I- H and D -1 D2S + D2 t- DS are about equal. We have, therefore, ,assigned the value of 5.0 kcal/mole-l to El'.B . DE B . DARWENT AND R . ROBERTS 59 Farkas and Farkas found that the probability factor for the D + H2 reaction was 0-07; the slope of the line (fig. 2) is 0.099, we then have By using the values 2.8 A and 4.5 A for d3’ and d1’ (the collison diameters of D atoms with H2 and D2S respectively) we calculate 2 3 ‘ = 3.7 X 1014 and 21’ = 3-6 X 1014 molecules-1 cm-3 sec-1, and since E3’ = El‘, we find Pi’ = 0.73. The reactions of D atoms with ethane, propane, n- and isubutanes, ethylene, propylene and isobutene were investigated by method 11.With ethylene the addition reaction and the D atom induced polymerization were much faster than the abstraction reaction. - 200 - 1.75 - 1.50 - 1.25 - 1.00 - -. HD D2 0.75 - 5 10 I5 20 25 30 0 [C2H6l/[D2Sl FIG. 3.-The dependence of HD/D2 on [C2&]/[DzS] between 31” C and 306” C. symbol + o a + temp. (“C) 31 104 204 306 Hence very little hydrogen was formed when the extent of the conversion was restricted to small values ; therefore, the results with ethylene were inconclusive and have not been included. With the other hydrocarbons good straight lines are obtained in accordance with theory (eqn. B). The results obtained with ethane and propylene, shown in fig.3 and 4, illustrate the agreement between experiment and theory‘ With the four paraffins the slopes (k3’lkl’) increased with increasing temperature, indicating that E3’ > El’. With the olefines no systematic change of slope occurred as the temperature was increased and it may, therefore, be assumed that, for the olefines, E3’ w El’. The influence of temperature on k3’/kl‘ for the paraffins is shown in fig. 5 and 6; no such plots were made for the olefines since there was no definite trend in k3c/kl’ with temperature. The values of E3’ - El’ for the paraffins were obtained from the slopes of the Arrhenius lines in fig. 5 and 6 and, by putting El‘ = 5.0 kcal/mole-l and PI’ = 0.73, the values of E3‘ and P3’ for the hydrocarbons were calculated in a manner analogous to that described for H2.The values so obtained and the collision diameters used in the calculation are given in table 2.60 REACTIONS OF HYDROGEN WITH HYDROCARBONS TABLE 2 . T H E ABSTRACTION OF HYDROGEN FROM PARAFFINS AND OLEFINES BY DEUTERIUM ATOMS E J collision no. reactant (kcal /mole-1) (A) (molecules-1 cm-3 sec-1) x 1014 C2H6 9.0 5-0 4.2 0.6 C3H8 7.2 5.4 4.8 -0.6 n-C4HlO 7.1 5.8 4.6 0.6 ~so-C~H~ 0 6.3 5.9 4.6 0.6 C3H6 5.0 5.4 5.0 0.1 St?C.-C4Hg 5.0 5.6 4.6 0.3 The activation energies E are based on the value of 5.0 kcaI/mole-1 for the reaction D + H2 + HD + H. 3.4 30 2% 2 2 I -8 0 1.4 I *o 0.6 0.2 0 5 10 I5 20 25 FIG. 4.-The dependence of HD/D2 on [C3H6]/[D2S] between 28" C and 230" C. symbol a - 4 + + temp. ("C) 28 65 95 230 The straight lines obtained may be taken as substantiation of the adequacy of the theory and the applicability of the experimental results to the reactions in question.However, positive intercepts, which are not predicted by theory, were obtained. It was found that the photolysis of the deuterium sulphide alone gave HD/D2 = 0.10 0.01, and H2/D2 = 0.045. The simple theory applied to these experiments requires the inter- cepts to be 0.10 f 0.01, to be the same.for all the compounds and to be independent of temperature. The intercepts actually found (table 3) do appear to be independent of temperature but are much larger than 0.10. TABLE 3.-INTERCEPTS OBTAINED IN GRAPHS OF HD/D2 AGAINST [RH]/[D2S]B . DE B. DARWENT AND R. ROBERTS 61 The points in the Arrhenius plots appear to fall on straight lines, except for propane at the highest temperature (314" C) where the point lies considerably above the line.This point was not taken into consideration when the line was drawn since at such a hi& temperature there were probably complications caused by the decomposition of the propyl radicals. -04 -0.8 - 1.2 -1.6 -1.4 4.8 . 2 2 -2.6 I 2 3 Fm. S.-Arrhenius plots for ethane (B) and propane (A). -0.4 -0.8 -1.2 -1.6 0 I 2 3 FIG. 6.-Arrhenius plots for n-butane (B) and isobutane (A). DISCUSSION The validity of the methods used in this investigation depends on the absence of any reaction, other than those of types (01, (11, (2) and (3, in which hydrogen (atoms or molecules) are produced or removed. It has been assumed that this condition was fulfilled and this is justified by the facts that straight lines are62 KE A C'T I 0 N S 0 F H Y 1) K 0 G EN W 1T H HYDRO C A RB 0 N S obtained (fig.1 to 4) as required by eqn. (A) and (B), and that the Arrhenius plots (fig. 5 and 6) appear to be reasonably straight lines. On the other hand the intercepts in fig. 3 and 4 (table 3) are two to four times as large as predicted. These large intercepts may be explained on the basis of the differences in the rates of reaction of D and H atoms,97 10 but this is rather speculative. The results, based on the data of Farkas and Farkas,12 appear to be reasonable. The activation energies for the reactions of deuterium atom with the paraffins follow a simiIar pattern to those of the similar reactions involving methyl radicals 4 so that, although it is obvious that the assumed mechanism has not been established, the results appear to be of some significance.The steric factors P for the abstraction reactions involving the paraffins are about 0.6. The results of the discharge tube experiments, based on P = 0.1, are compared with the values obtained by method I1 in table 4. The activation energy quoted for isobutane was that found by White, Winkler and Kennalty 14 for the reactions of H atoms. TABLE 4.-cOMPARISON OF STERIC FACTORS P AND ACTIVATION ENERGIES E IN THIS INVESTIGATION WITH THOSE DERIVED BY THE DISCHARGE TUBE EXPERIMENTS ; REACTION D + RH + HD 4- R photolysis of DaS discharge tube experiments (P = 0'1) E P E 9.0 0.6 8.7 9 7.2 0.6 8.0 10 7.1 0.6 7.9 10 6.3 0-6 9 .p 14 * H atoms + isobutane. The results of the present experiments show that E, for the abstraction re- action with paraffins, decreases in the order ethane > propane > isobutane in a manner similar to that found for the methyl radical reactions, but no indication of variable or abnormally low steric factors have been found. The results of the abstraction reactions with the olefines are included in table 2. It is interesting to note that, with the exception of ethylene, the abstraction reaction requires approximately the same activation energy as does the addition. Further indication that, in the present investigation, the quantities measured are related to the reactions of hydrogen atoms with the hydrocarbons is obtained by a comparison (table 5 ) of the activation energies found with those ascribed to the similar reactions involving methyl radicals.The differences in activation energies (ECH3 - ED) is 1.3 kcal/mole-l for the paraffins and 27 kcal/rnole-l for the olefines. The average of these differences (2.0 & 0.7 kcal/moIe-l) is just about equal to the difference between the bond energies in D2 and methane. TABLE TH THE REACTIONS OF DEUTERIUM ATOMS AND METHYL RADICALS WITH HYDROCARBONS ED %H, - ED hydrocarbon C2H6 9.0 10.4 1.4 n-c4H1 o 7-1 8.3 1.2 ~SO-C~H~O 6-3 7-6 1.3 C3H6 5-0 7.7 2.7 sec.-C4Hs 5.0 7.7 2.7 The results obtained for the addition reactions may be compared with the data of Melville and Robb,l5 who found collision efficiencies between 1.4 x 10-4 and 8.3 x 10-4 for a large number of olefines.The activation energies and steric factors found in the present investigation are given in table 6, in which the col- lision yield at 18" C is calculated and compared with the results obtained by theB . DE B . DAKWENT A N D K. HOLlEKI'S 63 molybdenum oxide technique.15 Melville and Robb found the collision yield (assumed to be at 18" C) for propylene to be 1.4 x 10-4; this probably represents the sum of the collison yields for addition and abstraction, since they measured the sum of the rates of all processes that caused hydrogen atoms to be removed. TABLE 6.-THE ADDITION OF HYDROGEN ATOMS photolysis of HIS TO OLEFINES molybdenum oxidela collision (18' C) olefine E P efficiencies x lo* C2H4 4.1 0.09 0-8 1 C3H6 5.0 0.52 0.96 collision efficiencies x 104 (18" C) 8-3 1-4 We find collision yields of 0.96 x 10-4 for addition and 0.18 x 10-4 for abstrac- tion giving a total collision yield of 1.14 x 10-4 in excellent agreement with Melville and Robb. However, with ethylene, the results of the two methods differ by a factor of ten ; Melville and Robb found the collision yield with ethylene (8.3 x 10-4) to be larger, by a factor of six, than with propylene whereas we find that the reaction with ethylene to be somewhat slower than with propylene. 1 Cunningham and Taylor, J. Chem. Physics, 1938, 6, 359. 2 Smith and Taylor, J . Chem. Physics, 1939, 7, 390. 3 Phibbs and Darwent, Trans. Faraday SOC., 1949,45, 541 4 Trotman-Dickenson, Birchard and Steacie, J . Chem. Physics, 1951, 19, 163, 169. 5 Raal and Danby, J. Chem. SOC., 1949, 2219,2222,2225. 6 Steacie, Darwent and Trost, Faraday SOC. Discussions, 1947, 2, 80. 7 Evans and Szwarc, Trans. Faraday SOC., 1949, 45, 940. 8 see Steacie, Atomic and Free Radical Reactions (Reinhold, New York, 1946) for 10 Schiff and Steacie, Can. J . Chem., 1951, 29, 1. 11 Roberts and Darwent (in press). 12 Farkas and Farkas, Proc. Roy. SOC. A, 1935, 152, 124. 13 van Meerssche, Bull. Soc. Chim. Befg., 1951, 60, 99. 14 White, Winkler and Kennalty, Can. J. Res. B, 1942, 20, 255. 15 Melville and Robb, Proc. Roy. SOC. A, 1950, 202, 181. early references. 9 Trost and Steacie, .I. Chem. Physics, 1948, 16, 361.

ISSN:0366-9033    

DOI:10.1039/DF9531400055

出版商:RSC    

年代:1953

数据来源: RSC