March 5 , 1859
ISOEilERIZATION OF VIBRATIONALLY
the sonochemical production of hydrogen peroxide. It appears that there are differences between peroxide formation due to ionizing radiation and that due to ultrasonic cavitation, but further work is required to clarify the differences.
[CONTRIBUTIONFROM
THE
EXCITED cYCLOPROPANE-d*
1081
Acknowledgments.-The author wises to thank Drs. Robert L. Bowman and Charles R. Maxwell for their helpful discussions and Mr. B. Burr for mass spectrometer analyses. BETHESDA 14, MARYLAND
DEPARTMENT OF CHEMISTRY, UNIVERSITY OF WASHINGTON]
The Isomerization of Vibrationally Excited Cyclopropane-d, Produced from Methylene plus Ethylene-&' BY R. S. RABINOVITCH, E. TSCHUIKOW-ROUX AND E. W. SCHLAG RECEIVED SEPTEMBER 3, 1958 Vibrationally excited cyclopropane-dzmolecules have been produced a t 25' by addition of methylene radicals, arising from photolysis of ketene a t 3600 and 3100 A., t o Irans-ethylene-&. The addition is stereospecifically cis. The geometric and structural isomerization of the hot cyclopropane molecules has been studied at pressure up to 36 atrn. Comparison of these results with thermal studies, and between themselves, provides support for similarity of activation steps for the two isomerization processes.
trans-Cyclopropane-dz has been observed to the concomitant structural isomerization to proisomerize thermally to cis-cyclopropane-dz as well pylene. as to propylene.2 The former reaction has been Experimental shown to occur as a unimolecular process while the trans-Ethylene42 was of 99.3% isotopic purity with 0.7% latter is known to be such.3 The mechanisms of the ethylenedl. Ketene was purified by repeated bulb-to-bulb geometric cis-trans isomerization and structural distillation. Mass spectrometric analysis showed the keisomerization to propylene may involve either of tene to be essentially pure with negligible amounts of the the two molecular reaction coordinates suggested dimer. Photolysis of ketene-ethylenedz mixtures was carried originally by Chambers and K i ~ t i a k o w s k y ~ ~out ; in sealed reactors a t 25". A General Electric AH-6 namely a C-C extension, or an approach of a hydro- high pressure mercury arc lamp served as the light source. gen on one carbon atom to an adjacent carbon.4 The bands a t 3100 and 3600 A. were employed. A Corning Whatever mechanism may apply, the results of glass filter No. 5330 was employed to isolate the 3600 A. Rabinovitch, Schlag and Wiberg suggest that the region and a solution filter, made by dissolving NiS01.6H20 and CoS04.7Hz0 in water,' was used a t 3100 A. The soluactivated complexes for both geometric and tion was circulated through a Corex D glass envelope in structural isomerization are similar, particularly front of the lamp. The reactor was of Corex glass as well energy-wise. The present study was undertaken for the study a t 3100 A., while Pyrex vessels were employed to test this point in another way, by measuring a t 3600 d. A range of pressures of 0.14 to 36.6 atm. was the relative rates of the two isomerization processes covered. In each case the ratio of ketenelethylenedz was to 1/7. The amount of ketene varied from 2-20 cc. a t a range of energies appreciably different (higher) close per run, while the ketene decomposition varied from 1-20'7' from those accessible in the thermal study. in most cases, and up t o 40% in a few cases. Some polyKistiakowsky and Sauer6 pointed out that vi- merization was observed which, however, only became apbrationally excited cyclopropane is formed on preciable for the runs a t the highest pressures which necesthe longest exposure times (up to 35 min.). At addition of methylene to ethylene. Freyea re- sitated the end of each run the excess ethylenedz was pumped off ported the isolation of cyclopropane and propylene a t - 160". The remaining gases were then distilled through as principal products of the photolysis of ketene- a column of brick-dust covered with a saturated solution of ethylene mixtures, and Frey and KistiakowskyGb AgN03 in ethylene glycol and Kel-F alkane oil. This effectively removed the remaining ketene. The cyclopropresented quantitative studies of the variation of panepropylene mixture was analyzed and separated by gas product proportions with total pressure. In the chromatography, and the cyclopropane was analyzed for present work, by use of trans-ethylene-dz the stereo- the percent cis and trans-cyclopropanedz on a Beckman chemistry of the addition process has been deter- IR-2 spectrophotometer by methods used previously.z mined and the rate of geometric cis-trans isomeriResults and Discussion zation of. the vibrationally excited cyclopropaneRelevant processes are represented by the scheme dz has been measured and compared with that of where T, C and P refer to trans-cyclopropane-dz, cis-cyclopropane-d%and propylene, respectively. (1) This research was supported by the National Science Foundation. Abstracted from the M.S. Thesis of E. Tschuikow-Roux, University of
Washington, Feb., 1958. (2) Rabinovitch. Schlag and Wiberg, J . Chem. Phyr., 28, 504 (1958). (3) (a)T.S.Chambers and G. B. Kistiakowsky, THISJOURNAL, 66, 399 (1934); (b) H.0. Pritchard, R. G. Sowden and A. F. TrotmanDickenson, Proc. ROY. SOC.(London), A217, 563 (1953); (c) N. B. Slater, ibid., A218, 224 (1953). (4) These are extremes, and an intermediate type of coardinate could also be considered. See also F. T. Smith, J . Chem. Phys.. 29, 235 (1958). (6) G. B. Kistiakowsky and K. Sauer, TEIS JOURNAL, 78, 5699
(1956). (6) (a) H.M. Frey, ;bid., 79, 1259 (1957); (b) H. M. Frey and G. B. Kistiakowsky, 79, 6373 (1957).
Here k, and k, refer to the observed rate constants and the above scheme carries no further implication (7) W. A. Noyes and P. A. Leighton, "Photochemistry of Gases," Reinhold Publ. Corp., New York, N. Y., 1941, p. 69; E. J. Bowen, J . Chem. SOC.,76 (1935).
concerning mechanism. Then
where w is the specific rate of collisional deactivation. That the addition reaction of methylene to trans-ethylene-dz is stereospecifically cis as written above is shown below. The pressure dependence of cyclopropane formation is shown in Fig. 1 and that of cis-cyclopropane-dz in Fig. 2. It is evident that methylene carries excess energy arising from the photolysis process since the lifetime of the hot cyclopropane with respect to propylene formation declines] and the rate of geometric isomerization increases, with increase of light energy. This result further signifies that methylene must add to the olefin before it is “thermalized” by collisions. This I accords with Frey and Kistiakowsky’s conclu0 1000 2000 Pressure, c m . sions.6b In Fig. 1, the high pressure limit of the cyclo- Fig. 2.--cis-Cyclopropane-dz formation as function of propane/propylene ratio is e0.85, somewhat higher pressure. it was not done in every instance] since the lower pressures only encompassed a small portion of the accessible pressure range for geometric isomerization.
[ry [
SUMMARY O F
0 3600 8 3100
w
%. %.
Total press., cm.
P 0.2 04 0
L
200
600
400
Fig. 1.-Cyclopropane
800 1000 Pressure, crn
I
1 a“>
TABLE I RATECONSTANTS DETERMINED AT DIFFERENT
1600 2 4 0 0
formation as function of pressure,
10.9 20.1 41.9 64.1 94.8 154 166 300 594 738 819 855 1047 1440 1938 2730 2780
ka
wa
1010 sec. - 1
IO14
3600 .A 0.17 0.31 0.64 0.97 1.44 2.34 2.53 4.36 9.27 11.5 12.9 13.4 16.6 23.1 31.7 16.0 46.8 Av.
see.-’
10.8
k.
1010 sec. - I
0.80 .60 .61 .67 .67 .58 .55 .54 Av. .63
than that of Frey and Kistiakowsky. The extra propylene arises from addition of methylene across 9.5 the C-H bond, rather than to the C=C bond.6 The invariance of the limiting ratio with photolysis 13.0 light energy indicates independence of the relative 11.1 addition rates on methylene energy, a t least a t 12.2 these energies. 12.1 The monotonic decline of cis percentage toward 11.8 zero with increasing pressure (Fig. 2) shows that 13.1 methylene addition is stereospecifically cis and is 12.4 stereochemically similar to the addition of methyl11.9 ene to disubstituted ethylenes such as butene-2.8 11.8 A summary of calculated rate constants is 3100 A . given in Table I. At pressures larger than 3 atm. 1.39 1 9 . 7 0.30 a correction was applied to the calculated value of 1.21 18.8 66.4 1.01 the gas kinetic collision constant w to take into 0.87 149 2.27 (27.5) account the molecular volume and the partially 0.01 377 5.74 14.4 compensating shielding factor. The total cor1.07 472 7.17 16.2 rection factor is 1 0.626nb, where n = number of 1036 16.4 14.8 molecules/cc. and b = van der Waals constant. 1700 27.5 14.6 Evaluation of ks could be made accurately only in Av. 1.11 d v . 15.8 0 above total the lower pressure region since P These molecular diameters in 10-8 cm. were used in the pressures of 300 cm.; evaluation of k, was more inaccurate a t lower pressures because of the smaller calculation of w : cyclopropane, 4.92; ethylene, 4.95; ketene, 6.40. Collisional deactivation efficiency arbitrarily amounts of cyclopropane isolated for analysis; set equal to unity.
+
( 8 ) P. S. Skell and A. Y.Garner, THISJOURNAL, 78, 3409 (1956). (9) S. Chapman and T . G . Cowling, “The Mathematical Theory of Non-Uniform Gases,” Cambridge University Press. 1952, p. 274.
The average rate constant ratios are k,/ks = 18.7 a t 3600 A. and 14.2 a t 3100 A. This ratio
Xfarch 5 , 1%59
ISOMERIZATION OF VIBRATIONA1,LY
tends to be less affected by any complexities of the reaction system, such as dimerization of ketene, formation of small amounts of higher hydrocarbons, etc., than are the individual constants. A rough estimate is now desired of the internal energy of the cyclopropane molecules formed in this process and we speculate g~s follows. The energies of the 3100 and 3600 A. radiation correspond to 92 and 78 kcal./einstein, respectively. The photolysis reaction has a heat of 55 kcal./ mole a t room temperature
EXCITED CYCLOPROPAXE-d? kSE
=
10SH
const. (E+)'-'
where n is the number of vibrational modes which interact with the bond-breaking coordinate, counting degenerate modes fully. Marcus and Rice'j have given a useful representation of the classical approximation involved and the equation is better written for computational purposes E - Eo $. E , k S E = const.
(
E
+ E,*
where E z * E E r fis vibrational zero point energy"35 k ~ a l . / m o l e , Eo ~ ~ = 65 k ~ a l . / m o l e , ~E, ~= ~ 95 kcal./mole and the const. is set equal to the where AHf(CH2) = 67 kcal. is computed by adopt- o b s e r ~ e d high ~ , ~ ~pressure frequency factor, 1015.2 ing D(CHrH) = 87 kcal. as a possible valuelo sec.-l Then = 3 X lo9 set.-'; k, obsd = (and assuming the same electronic state, presum- 1.1 X olO1o sec.-l (Table I). On the same basis for ably ground singlet, of methylene in both processes) 3600 A., KSE = 9 X IOp SeC.-l, k s obsd = 6.3 X and with other heats of formation as tabu1ated.l' l o 9 sec.-l. The agreement is better than should At 3100 A., the excess light energy is 37 kcal./ be expected (the k, obsd are upper limits being based mole of which 12/19, ie., 23 kcal., is arbitrarily on unit collisional deactivation efficiency). A still assumed to reside with the methylene (if 3/4, the further substantiation of the high level of energy ratio of vibrational degrees of freedom of the of the energized cyclopropane molecules is their products, were used instead, l 2 this number would essentially mono-energetic character, as supported become 27 kcal.). The exact value is of no con- by the constancy of values of rate constants over sequence. the pressure range studied. For the reaction of interest, the heat is11Thus the energy of the vibrationally excited 67 kcal. To this may be added the 23 kcal which cyclopropane is of the order indicated, which appreciably exceeds the average energy of the re:CH2 4- CzH2Dz +Cyclo-CsHdDz acting molecules in thermal studies of cyclopropane for ad hoc purposes12 (and being associated with isomerization.16 Frey and KistiakowskyGb came the lighter and smaller association partner) is to the same conclusion earlier. assumed to appear in cyclopropane entirely as From thermal studies a t 445' of the geometrical vibrational energy, plus a few kcal of thermal and structural unimolecular isomerization of cycloenergy (at 25') of the methylene and ethylene. propane-d2,2J6the average value of the ratio corOne conceives then of an energy of the hot cyclo- responding to k,/& is 12. Relative to the valpes propane of -95 kcal./mole; since part of any error quoted earlier, 18.7 (3600 h.) and 14.2 (3100 A.), in the estimate of AHf(CH2) cancels out (12/19 the variation is negligible and supports the simihere), this number may be valid to f 10 kcal. larity of transition states for both processes. There The order of magnitude may also be checked in the is an apparent inversion of the magnitudes a t the following crude fashion. Consider one of the two wave lengths which presumably reflects exmeasured rate constants, say k,. Since the formed perimental error. cyclopropane molecules are roughly monoenergetic, SEATTLE, WASHINGTON in view both of their mode of formation and high (14) R. A. Marcus and 0. K. Rice, J. Phys. Colloid Chem., 65, 894 energy, then assuming random energy distribution (1951). we may write with Slater3bc
