Identity scrambling and isomerization networks in systems of excited

Identity scrambling and isomerization networks in systems of excited alkyl radicals. C. W. Larson, Peter T. Chua, and B. S. Rabinovitch. J. Phys. Chem...
0 downloads 4 Views 1MB Size
IDENTITY SCRAMBLING AND ISOMERIZATION NETWORKS

2507

Identity Scrambling and Isomerization Networks in Systems of Excited Alkyl Radicals1 by C. W. Laram,* Peter T. Chua, and B. S. Rabinovitch Departments of Chemical Engineering and Chemistry, University of Washington, Seattle, Washington 98196 (Received March 80, 1978) Publication costs assisted by the Petroleum Research Fund

The gas phase isomerization by H-atom transfer of vibrationally excited alkyl radicals of general structure GCC(C)bC has been studied; G was an H, methyl, ethyl, isopropyl, or tert-butyl group. The radicals were produced by chemical activation by the addition of an H atom to an appropriate a-olefin, GCC(C)C=C. This study expands on earlier findings and shows that alkyl radicals, vibrationally excited above 40 kcal mol-’, are involved in a characteristicprocess of rapid competitive, consecutive isomerization. If the radical skeleton is large enough, all possible radical sites on the skeletal framework are occupied in the process of identity scrambling. The steady-state proportions of isomeric radical species depend both on thermodynamic properties of the radicals and kinetic parameters associated with their decompositions and stabilizations. Details of Rice-Ramsperger-Kassel-Marcus (RRKM) rate calculations for the more important isomerization processes, including a general formulation for alkyl radical and activated complex frequency assignments, are described.

I. Introduction Isomerization of vibrationally excited alkyl radicals by intramolecular hydrogen atom abstraction forms an important class of alkyl radical reactions.2 However, until recently these reactions had received little s t ~ d y . ~ - BPreviously,B we reported on the isomerization of some ten large (C6 to Cg) chemically activated alkyl radicals. Exothermic and thermoneutral 1,4-, 1,5-, and 1,6-hydrogen atom shifts were the most important processes in all systems where such reaction paths existed. Related conclusions derived from the mechanistic study of the pyrolysis of c6 to C9 paraffins have been reported recently by Doue and G ~ i o c h o n . ~ Measurements of Arrhenius parameters for alkyl radical isomerization have been limited. To date, only two exothermic radical isomerizations, pentyl-1 + ~ e n t y l - 2 ~and 0 hexyl-1. --t have been studied; abnormally low A factors (-lo7 sec-l) and activation energies (8-10 kcal mol-’) were reported in both cases. These values have been questioned and higher values ~ u p p o r t e d . ~ - ~ Chemically activated alkyl radicals of any desired structure may be directly produced by the exothermic chemical reaction between a hydrogen atom and an olefin. Radicals produced in this way are characterized by a well-defined and relatively narrow energy distribution above 40 kcal mol-l. This energy range is of considerable interest in chemical kinetics : alkyl radicals containing -45 kcal mol-l internal excitation may undergo several competitive primary processes including p-carbon-carbon bond rupture, internal hydrogen atom abstraction, and collisional stabilization. olefin chemical activation technique provides The H a

+

an extremely powerful tool for the quantitative investigation of alkyl radical reactions in general. 1,4-, 1,5-, and 1,6-hydrogen (and higher) shifts are believed to occurz-7 primarily by cyclic mechanisms involving five-, six-, or seven-membered ring transition states, respectively. In this paper we describe systems in which net 1,a-H-atom shifts occur. Both endothermic and exothermic processes are reported.

11. Present Investigation : Isomerization Networks A series of five vibrationally excited 3-methylalliyl-2. radicals, GCC(C)cC*, were produced by addition of H atoms to terminal 3-methylalkenes. The series included radicals where G was hydrogen (H), methyl (Me), ethyl (Et), isopropyl (iPr), and tert-butyl (tBu). Isomerization networks, schematic diagrams which represent all paths of communication between the several radicals of a given carbon skeleton, may be constructed for each of these systems. Communication (1) This work was supported by the Office of Naval Research. (2) (a) A. Kossiakoff and F. 0. Rice, J . Amer. Chem. Soc., 6 5 , 590 (1943); (b) H. M. Frey and R. Walsh, Chem. Rev., 69, 103 (1969) ;

(c) P. G. Ashmore, F. 9. Dainton, and T. M. Sugden, “Photochemistry and Reaction Kinetics,” Cambridge University Press, New York, N. Y . , 1967, Chapter 12. (3) (a) V. B. Stefton and D. J. LeRoy, Can. J . Chem., 34, 41 (1956); (b) A. S. Gordon and J. R. McNesby, J . Chem. Phys., 31, 853 (1959) ; (e) L. Endrenyi and D. J. LeRoy, J . Phys. Chem., 70, 4081 (1966). (4) C. W. Larson, D. C. Tardy, and B. S. Rabinovitch, J . Chem. Phys., 49, 299 (1968); (b) C. W. Larson and B. 8 . Rabinovitch, ibid., 51, 2293 (1969). (5) K. W. Watkins and L. A. Ostreko, J . Phys. Chem., 73, 2080 (1969).

(6) (a) E. A. Hardwidge, C. W. Larson, and B. 5. Rabinovitch, J . Amer. Chem. Soc., 92, 3278 (1970); (b) K. W. Watkins and D. R . Lawson, J . Phys. Chem., 75, 1632 (1971). (7) F. Doue and G . Guiochon, ibid., 73, 2804 (1969). The Journal of Physical Chemistry, Vol. ’76,N o . 18, 197.9

C. W. LARSON,P. T. CHUA,AND B. S. RABINOVITCH

2508

between two radical species is greatly favored when they are linked by a five-, seven-, or especially, sixmembered ring H-atom transfer transition state; pathways with smaller transition states have not been observed to be in competition with stabilization, decomposition, or other isomerization processes;3b88 in any case, they would be negligible for the present systems and pressures. A . 3-Methylbutene-1 H System; G = H . HAtom addition to 3-methylbutene-1 produces a mixture of approximately 94% secondary (R2, parent) radical and 6% primary (RI, anti-Markovnikov) radical. The carbon skeleton is numbered as shown, where

+

cat C4-A3-C2-C1

R4 and R3) (Le., radicals where the odd electron is located on carbon number 4 or 3, respectively) are distinguishable only under isotopic labeling. The isomerization network is represented by

The circled symbols represent vibrationally excited radicals and denote the free electron position (1, 2, etc., as indicated in the labeled carbon skeleton) and the radical type (p, s, or t for primary, secondary, or tertiary, respectively). A thermoneutral isomerization pathway via a five-membered ring transition state exists between the primary R1 and RV radicals; no other H-atom transfer reactions occur appreciably in this system and R3 is not formed. B. H f 3-Methylpentene-1 System; G Me. As before, two reaction paths for H atom addition to the a-olefin provide entrance into the isomerization network from two directions. The carbon skeleton and isomerization network are

-

6% input

occurs in the G = Et system; the carbon skeleton and isomerization network are Ca'

I

Ce-CsCrC3-Cz-C1

-

6% input

t

+

D. H 3,5-Dimethylhexene-i System; G = iPr. Additional methyl substitution at the no. 5 carbon introduces no new complexities into the isomerization network. However relationships between the several isomerization paths are altered inasmuch as the Rg radical is a tertiary species; the reaction path degeneracies to form the Re radical are doubled. The carbon skeleton and isomerization network are

-

4.5-5 14.3

k~a(Ez),"mc-l-----G = iPr

102.0-6 30,8-5 67.7-3 89.0-6 24.1-5 45.2-3 >5,0-5 15.8

G

-

Et

738-6 298-5 894-3 656-6 237-5 619-3 >3.3-5 20.8

B. R6 -c R3 Process; Five-Membered Ring E'aa,

koal

mol -1

El (kcal mole")

Figure 2. Calculated energy dependence of specific rate of five-membered, 8-kea1 mol-' exothermic isomerization (k63) and six-membered, 4-kcal mol -1 endothermic isomerization (he)for the three systems: 1, G = Et; 2, G = iPr; 3, G E tBu. The Journal of Physical Chemistry, Vol. '76,No. 18,197%

E%, koal mol -1.

8.6 44.0 14.3 44.0 20.0 44.0 4.0 41.7 8.6 41.7 14.3 41.7 Measured k63 value Estimated E063, kcal mol-'

---

G = tBu

kes(Ee), sec-L----G = iPr

G = Et

82.3-6 15.2-5 18.9-3 1.50-9 63.0-6 11.0-5 5-6 12.0

302.0-6 70.5-5 114.0-3 4.53-9 230.0-6 49.0-5 18-6 12.0

1630-6 584-5 1490-3 18.4-9 1300-6 460-5 100-6 12.0

Actual reaction a Tabulated rates are for one reaction path. path degeneracies in the Rz -.+ R6 process (ff.26) are 9, 6, and 3 in G = tBu, iPr, and Et, respectively. 0 6 , 3 values are unity for each system

ON THE LIFETIMEOF CF3 IN A Q U E O U S SOLUTION the isomerizing Rz* radicals, (E#, is only slightly less than the average energy of the formed Rz*species, (EZ)'; the latter quantities are'? 46.1, 46.3, and 47.4 kcal mol-' for the G = tBu, iPr, and Et system, respectively. Neglecting energy state fractionation by reaction and collisional pertubations, the average energies of the reacting (isomerizing) Re* species, (E#, are approximately 4 kcal mol-' less than (E$. From Table IX it may be seen that the calculated rates are However, an relatively insensitive to (&)I and (.%)I. uncertainty of a factor of 3 in the experimental isomerization rate introduces a 1.bkcal mol-' uncertainty in the deduced E o a jvalue. On this basis, plausible

2517 lower estimates of thresholds in the 6ps and 5tp processes are 14 and 12 kcal mol-l, respectively. Figure 2 shows the energy dependence of the six specific isomerization rate constants obtained with E O 2 6 = 14 and Eosa = 9 kcal mol-'. The energy distribution of the formed Rz* species are also included. Figure 3 reveals the energy dependence of the equilibrium constants for C7, CS, and Cs radical isomers, R+S R,, having 4 (Le., for a ps or st process) or 8 (Le., for a pt process) kcal mol-' differences in zeropoint energy, AE. Ko(Ea)is equal t o hTi*(E, AE)/ Ni*(Ea),for uij = a,,, and for a model of identical frequency assignments for R, and R,.

+

On the Lifetime of Trifluoromethyl Radical in Aqueous Solution1 by Jochen Lilie, D. Behar, Richard J. Sujdak, and Robert H. Schuler* Radiation Research Laboratories, Center for Special Studies and Depurtment of Chemistry, Mellon Institute of Science, Carnegie-Mellon University, Pittsburgh, Pennsylvania 16915 (Received April 10, 1979) Publication costs assisted by Carnegie-Mellon University and the U.S. Atomic Energy Commission

Radiolysis of aqueous solutions of CF3C1 produces CF3 radicals which can readily be trapped as the addition product to nitromethane anion. Intercomparison of the effects of scavengers on the esr signals of this adduct radical and on the production of fluoride ion indicates that CF3 reacts with water with a period of -30 msec. Conductometric pulse radiolysis studies in the presence of 1 mM methanol, where OH radicals are rapidly removed from the system, show that no significant hydrolysis of CFs occurs on the millisecond time scale. I n the absence of methanol, however, secondary reactions which produce a large yield of H F occur a t shorter times. This H F can be accounted for by reactions involving combination of CF3 radicals with OH. It is expected that the CFgOH produced from such a combination will eliminate H F rapidly. However, it also appears from the magnitude of the increase in conductivity that CFzO must hydrolyze completely within the period of these conductivity experiments. From the limiting growth period a t high doses the period of the initial hydrolysis of CFzO is estimated to be -10 psec.

Balkas, Fendler, and Schuler have examined the radiolysis of aqueous solutions of CFaCl by steadystate and conductometric pulse radiolysis methods and have concluded that decomposition of the CFaCl occurs entirely as a result of initial attack of eaq- on the solute to produce chloride ion.2 At low dose rates, eaq-

+ CF&l+

CF3.

+ C1-

(1)

however, a yield of fluoride ion which corresponds to the complete hydrolysis of the CFa is also produced,2 and a major question exists as to the reactions responsible for the appearance of this fluoride and the time scale over which they occur. In their conductometric studies, Balkas, et al., found, as expected, that for each hydrated electron produced the conductivity increment corresponded to the production of 1 equiv of HC1 within the several microsecond resolving time of the

apparatus. They also found an additional growth of conductivity on the 10-psec time scale which corresponded to the production of -1 additional equiv of HX and suggested that this growth was the result of the initial stage of hydrolysis of the CF3. If so, this pseudo-first-order reaction would provide a convenient reference against which one could make absolute determinations of the rate of reaction of CFa with added solutes. I t was estimated, for example, that the reaction of CF3 with ethylene had a rate constant of -7 X 108 M-l sec-'.2 Bullock and Cooper3 have made an absolute pulse radiolytic determination of the rate of attack of CF3 radicals on the formate ion and have (1) Supported in part by the U. 5. Atomic Energy Commission. (2) T. I. Balkas, J. H. Fendler, and R. H. Schuler, J . Phys. Chem., 75, 465 (1971).

(3) G. Bullock and

R. Cooper, Trans. Faraday Soc., 66, 2055 (1970).

The Journal of Physical Chemistry, Vol. 76, N o . 18, 107%