803
J. Phys. Chem. 1982, 8 6 , 803-805
-
Thermal Isomerization of Cyclopropane to Propene in the Presence of Boron Trichloride at T 1000-1300 K David K. Lewis,’ H. William Bosch, and Jeanne M. Hossenlopp Depaffment of Chemistry, Colgate University, Hamilton, New York 13346 (Received: JuW 28, 1981; I n Final Form: October 5, 1981)
Rates of thermal isomerization of cyclopropane (CP) to propene have been measured in the presence of the Lewis acid BC13 Sample mixtures containing 1%CP in argon, and 1%CP plus 1%BC13in argon were heated to 983-1333 K at P 3 atm in a single pulse shock tube. The BC13 was found to have no catalytic effect on the subject reaction under these conditions. This finding contrasts sharply with lower temperature studies, in which CP isomerization was found to be greatly accelerated by a variety of Lewis acids. However, BC1, was found to alter the concentrations of other products produced at temperatures of 1300-2000 K, probably from subsequent pyrolyses of the product propene.
-
Introduction A number of decomposition or rearrangement reactions of gaseous hydrocarbon molecules are known to be subject to homogeneous catalysis by strong Lewis acids. For example, Stimson and co-workers have shown that the structural isomerization of cyclopropane to propene, normally a unimolecular reaction, becomes a profoundly more rapid bimolecular reaction in the presence of relatively small amounts (2011 ratio of CP to Lewis acid) of HBr, BCl,, and BBr3 at temperatures of 525-750 K.’ Since two of those Lewis acids do not contain hydrogen, the mechanism of catalysis cannot be by protonation of the reactant. Instead, a polar transition state with an internal 1 2 shift of a hydrogen atom was proposed.l* The cyclopropane propene reaction has long been considered an important test reaction for comparing experimental reaction rates with theoretical excitation models.* Readers may recall the considerable controversy that arose over the interpretation of anomalously low rates of this reaction at temperatures above 1100 K, as measured in single-pulse shock tubes.3 Recently, three groups have also reported studies of this reaction via multiphoton excitation.* Hall and K a l d ~ reported r~~ that pumping of certain (but not all) vibrational modes of cyclopropane with intense infrared radiation could induce a product mixture distinctly different from that available thermally. It seems only logical that the effects of Lewis acids on the reactions of cyclopropane subjected to intense laser beams will be examined, and we are aware of one group that is presently doing SO.^ Multiphoton excitation experiments are believed to produce transitory vibrational temperatures on the order of 1000-2500 K.6 Specific effects of laser irradiation will be most easily discovered and interpreted if the results of such experiments can be compared with results of thermal
- -
(1)(a) Johnson, R. L.; Stimson, V. R. A u t . J. Chem. 1975,28,447.(b) Stimson, V. R.; Taylor, E. C. Ibid 1976,29,2557. (2)See, for examples, reviews presented by (a) Robinson, P. J.; Holbrook, K. A. ‘Unimolecular Reactions”; Wiley-Interscience: New York, 1972. (b) Forst, W. ‘Theory of Unimolecular Reactions”;Academic Press: New York, 1973. (3)(a) Barnard, J. A.; Parrott, T. K. Int. J. Chem. Kinet. 1977,9,387. (b) Lewis, D.K.; Giesler, S. E.; Brown, M. S. Ibid. 1978,10,277. See also earlier experimental work referenced therein. (4)(a) Lesiecki. M. L.: Guillorv. W. A. J. Chem. Phvs. 1977.66.4317. (b)’K&ny, Z.; Zke, R. N. Chek.’Phys. 1977,23, 32i. (c) Hall, R. B.; Kaldor, A. J. Chem. Phys. 1979,70,4027. (5)Zare, R. N., private communication. (6)(a) Preses, R. E.; Watson, Jr., R. E.; Flynn, G. W. Chem. Phys. Lett. 1977,46,69.(b) Grant, E.R.; Schultz, A. S.; Sudlo, A. S.; Shen, Y. R.; Lee, Y. T. Phys. Reu. Lett. 1978,40,115.(c) McNair, R. E.; Fulghum, S. F.: Flynn, G. W.; Feld, M. S.; Feldman, B. J. Chem. Phys. Lett. 1977, 48, 271. 0022-3654/82/2086-0803$01.25/0
excitation experiments carried out within the same temperature range. However, we know of no reports of the effects of Lewis acids on thermally induced reactions of hydrocarbons at temperatures approaching or exceeding 1000 K. Accordingly, we have begun reaction rate/ mechanism studies of a number of normally unimolecular reactions in the presence of Lewis acids over the temperature range 1000-2000 K. In this first report, we describe the effects of BCl, on the CP propene isomerization.
-
Experimental Section The 1-in. i.d. Pyrex single-pulse shock tube used in these experiments was patterned after the design of Lifshitz, Bauer, and Resler.’ Dimensions and standard operation procedure were described in a previous report.3b This shock tube has previously been used only for comparative rate single-pulse experiments, in which reaction temperatures were determined from the extent of another noninteractive standard reaction. Since the effects of BCl, on any standard reactions that might be suitable for such a study are unknown, the comparative rate technique is inapplicable for the present study. Accordingly, we have added a three-stage laser schlieren system for measuring shock wave velocities. A helium-neon laser beam (Spectraphysics Model 120) is divided by partially reflecting mirrors into three beams, which are directed through the shock tube (normal to the shock tube axis) at separations of 10 cm, near the downstream end. Each beam is partially blocked by a knife edge, and directed to a phototube. Passage of the shock front through the beam causes deflection of the beam due to the gas density gradient, thereby producing a signal which triggers or is recorded on an oscilloscope (Tektronix 7403). Elapsed times between arrival of the shock front at the three laser beams can be measured, hence shock speeds and shocked gas temperatures can be calculated by well-established methods.8 Analysis of reactant and product gas samples was performed on a Varian 1440-20 flame ionization chromatograph equipped with a 5-ft by ll8-in. Poropack Q column (for cyclopropane, propene, and other organic side products) and a Varian 90 P thermal conductivity chromatograph equipped with a 5-ft by ‘I4-in. Poropack Q column (7)Lifshitz, A.; Bauer, S. H.; Resler, E. L. J. Chem. Phys. 1963,38, 2056. (8)See, for example, (a) Bradley, J. N. “Shock Waves in Chemistry and Physics”; Wiley: New York, 1962;Chapers 2 and 3. (b) Greene, E. F.; Toennies, J. P. “Chemical Reactions in Shock Waves”; Academic Press: New York, 1964; Chapter 3.
0 1982 American Chemical Society
804
The Journal of Physical Chemistry, Vol. 86, No. lOOC
I50 K
r--i
K
7
I
~
Lewis et al.
5, 7982
'\\
\
e
-
\
I
I
1-
-
\, '\
8
5
10
IO~T(K)
Flgure 1. First-order rate constants for cyclopropane isomerization to propene: ( 0 )mixture A (1 % CP in argon): (0)mixture B (1 % CP 1 YO BCI, in argon); (line) eq 1. Arrows give representative falloff corrections, k k, (see text).
+
-
(for BCl,). Matheson cyclopropane (99%) and BCl, (99.9%) were used as obtained from the tanks. Gas chromatographic analysis indicated that the CP was actually 99.7% pure, with 0.3% propene as the only detectable impurity. Gas mixture A (1% CP in argon) and mixture B (1% CP + 1% BCl, in argon) were prepared on a high-vacuum manifold and stored in Pyrex bulbs sealed with stopcocks lubricated with Kel-F lubricant. Matheson ultrahigh-purity argon (99.999%) was used as the diluent. The gas mixtures were homogenized thoroughly and checked chromatographically for reactant concentrations before use. Calibration mixtures containing 1% CP, 1% propene, or 1% BC13 were also prepared, as described above. For the shock tube experiments, reactant gas pressures ranged from 120 to 30 torr, and temperatures of 1000-2OOO K were achieved behind the reflected shock waves. Product gas samples were withdrawn immediately after the shocks and subjected to duplicate analyses on the 1440 instrument for CP, propene, and other organic materials. A single analysis for BCl, was performed on the 90-P instrument after some runs to be certain the BC1, could be recovered intact. A fresh GC calibration curve was prepared for each principal reactant and product on each day of shock tube experimentation.
Data and Results A series of shock tube experiments was performed on 80-120-torr samples of reaction mixtures A and B. The reflected shock waves heated the samples to 983-1333 K for 800 f 100 ps, and produced reaction zone pressures of 2.5-3.0 atm. Under these conditions, the dominant reaction in uncatalyzed mixtures is conversion of cyclopropane to propene; that product concentration exceeds all side products concentrations by >50:1. Rate constants were calculated by use of the integrated first-order rate expression for an irreversible reaction, and t = 800 ps; they are displayed in Figure 1. Included in Figure 1is the line log k (s-') = 15.2 - 65000/4.576T (K)
(1)
which we believe best represents the high-pressure limit for this reaction.3b The arrows attached to two points give k,. These correpresentative corrections for falloff, k rections were estimated from calculations performed by
-
Frey and reported by Barnard et al.;9 we used p, = 1.00 and 0.07 as the collision efficiencies of cyclopropane and argon, respectively. The rate constants from reactants A (CP only) and B (CP plus BCl,) are in good agreement with each other throughout the temperature range covered, indicating that the Lewis acid has no detectable effect on the rate of CP structural isomerization under these conditions. Both sets of rate constants are also in good agreement with the preferred high-pressure limit (line in Figure 1)up to 1100 K, but are somewhat low a t higher temperatures. This observation is familiar; we believe it is largely due to experimental errors (inclusion of small amounts of improperly heated gas, etc.)l0 that become significant as k approaches and exceeds lo3 s-'. Although BCl, does not appear to catalyze the cyclopropane propene reaction at T > lo00 K, we did observe an effect on the abundances of other products produced, probably from propene pyrolysis, a t T > 1300 K. Pairs of shocks were run on the two mixtures at initial pressures of 70 (-1350 K), 50 (-1600 K), and 30 torr (-2000 K). The dominant gaseous products observed from mixture A (CP only) were, in order of decreasing concentrations, propene, ethene, methane, and acetylene, in good agreement with the earlier study of Bradley and Frend under similar conditions.l' Ethene/methane/acetylene ratios were -1/0.25/0.05 at 1300-1600 K, and 1/0.25/-0 at 2000 K. Only propene, ethene, and methane were produced from mixture B (CP + BCl,), with ethene/methane ratios approximating 1/1 at 1300-1600 K and 1/0.25 at 2000 K. Thus it appears that the BC13 enhanced the production of methane considerably, and possibly also increased the production of ethene by a lesser amount, but the acetylene production was effectively quenched at temperatures of 1300-1600 K. Around 2000 K there appears to have been no effect. We expect to study the pyrolysis of propene directly, in the presence of BCl,, in order to determine the mechanism by which these changes in product concentrations are produced. Mass spectrometric analysis of products of CP + BCl, mixtures heated to T I1350 K has shown no evidence of inclusion of chlorine atoms in the hydrocarbon products, but similar studies on propene a t T > 1350 K have not yet been performed.
Discussion It is interesting to compare the rates of cyclopropane isomerization to propene expected from the unimolecular reaction at k , (eq 1)and the BC13 catalyzed bimolecular process (eq 2). This comparison is shown in Figure 2, for log k (cm3/(mole 5)) = 11.35 - 25480/4.576T (K)la (2) reactant gas pressures and temperatures approximating those used in the present study. Barring a second catalytic mechanism, the considerably higher A and E, of the unimolecular reaction causes it to dominate at T k 900 K. Thus our observation of negligible effect of BC1, on CP isomerization to propene at T k 1O00, even in the presence of some unimolecular falloff, comes as no surprise. While the CP propene rates in this study are as predicted from extrapolation of lower temperature work, we would caution against generalizing too widely from this result. First, mechanisms other than those observed at lower temperatures, hence having different Arrhenius
-
(9) Barnard, J. A.; Cocks, A. T.; Lee, R.Y-K. J . Chem. Soc., Faraday Trans. 1 ~ 1974. ~ 70. 1782. --~ ~ . (10) (a) Skinner, G. B. I n t . J . Chem. Kinet. 1977,9, 863. (b) Jeffers, P. M.; Northing, J. Ibid. 1979, 11, 915. i l l ) Bradley, J. N.; Frend, M. A. Trans. Faraday SOC.1971, 67, 7 2 . ~I~
805
J. Phys. Chem. 1982, 86, 805-807
-3t x
fective laser temperatures. For example, Danen12irradiated ethanol in the presence of 2-propanol and HBr, and found that dehydration of ethanol dominated both the dehydration of 2-propanol and acid-catalyzed dehydration of either alcohol. However, his energetic analysis was referenced to a study of 2-methyl-2-propanol dehydration which had reported an activation energy 5 kcal lower than the value we believe from a recent shock tube study.13 The higher E, (and therefore A factor) suggest considerably larger rates for unimolecular dehydration of ethanol or 2-propanol a t the effective vibrational temperature (1000-2500 K) than those apparently estimated by Danen. Thus, the reported observation that the ethanol dehydration dominated the lower energy 2-propanol dehydration during ethanol irradiation is remarkable, but the suggestion that the acid-catalyzed reaction would have been expected to be predominant is incorrect.
l
\
-5 -6
-104
I
~
10
8
12
I$'T(K)
-
Figure 2. Rates of unimolecular and BC1,-catalyzed conversion of cyclopropane to propene expected in 1% CP plus 1% BCI, at p 2.5 atm: (-) unimolecular rate, given by eq 1 (note: this assumes reaction is at the higbpressure limit; actual rates are somewhat lower); (---) bimolecular rate, given by eq 2.
parameters, might come into play at higher temperatures, particularly in the presence of significant concentrations of radicals. We believe this is happening in the case of propene + BCl, at T 2 1300 K. Second, minor errors in activation energy values evaluated at lower temperatures can lead to large errors in predicted reaction rates at ef-
Acknowledgment. D.K.L. thanks Professor S. H. Bauer, his Ph.D. research mentor, for suggesting the series of studies of which this report is the first result. We also thank Mr. Mark Greaney and Mr. Stephen Incavo for assisting with the mass spectrometric analysis of gaseous products, and Mr. Stephen Didziulis for assisting with the collection and reduction of some of the data in Figure 1. This work has been supported by a Cottrell College Science Grant from the Research Corporation, and by Colgate University. ~
~~~~~~
(12)Danen, W.C. J . Am. Chem. SOC.1979,101,1187. (13)Lewis, D. K.; Keil, M.; Sarr, M. J. Am. Chem. SOC.1974,96,4398.
Reaction of Hydrogen and Bromine behind Reflected Shock Waves' J. M. Bopp, A. C. Johnson, R. D. Kern;
and 1. Nlkl
&pat?ment of Chemistry, University of New Orleans, New Orleans. Loulsiena 70122 (Received: July 28, 1981; I n Final Form: September 2 1, 198 1)
The reaction of equimolar mixtures of hydrogen and bromine diluted by inert gases was studied in the reflected shock zone over a temperature and total density range of 1400-2000 K and 1.5 X 104-3.3 X lo4 mol ~ m - ~ , respectively. Infrared emission from HBr passing through a narrow interference filter centered at 3.60 pm was recorded during observation periods typically of 500-ps duration. Conversion of the emission intensity traces to concentration-time data revealed nonlinear product growth for the low-temperature runs and near-linear product profiles at the higher temperatures. The individual experimental profiles were matched with the corresponding model calculations which employed a modern set of rate constants for the various elementary reactions comprising the atomic mechanism. The average percent deviation of 62 experiments from the calculated profiles was 5.4%.
Introduction Professor Bauer has reviewed recently the status of four-center metathesis reactions2paying particular attention to exchange reactions which he3-10 and his col(1),(a)paper presented at the 178th National Meeting of the American Chemical Society, Washington, DC, Sept, 1977. (b) Support of this work by the National Science Foundation, Grant CHE-7608529,is gratefully acknowledged. (2)Bauer, S. H. Annu. Reu. Phys. Chem. 1979,30,271. Ossa, E. J. Chem. Phys. 1966,45,434. (3)Bauer, S. H.; (4)Burcat, A.; Lifshitz, A.; Lewis, D.; Bauer, S. H. J. Chem. Phys. 1968,49,1449. (5)Lifshitz, A.;Lifshitz, C.; Bauer, S. H. J. Am. Chem. SOC.1965,87, 143. (6)Watt, W.; Borrel, P.; Lewis, D.; Bauer, S. H.J.Chem. Phys. 1966, 45,444.
0022-3654/82/2086-0805$01.25/0
leagues11-16have studied for a number of years. These investigations employed the single-pulse shock tube technique. The major finding of these reports is that these homogeneous exchange reactions occur via complex mechanisms that do not involve significant atomic con(7)Carroll, H. F.;Bauer, s.H.J. Am. Chem. SOC.1969,91, 7727. (8)Lewis, D.; Bauer, S.H. J. Am. Chem. SOC.1968,90,5390. (9)Bauer, S. H.;Lederman, D. M.; Resler, E. L. Jr. Int. J. Chem. Kinet. 1973,5,93. (IO) Bauer, S. H.; Hilden, D.; Jeffers, P. J. Phys. Chem. 1976,80,922. (11)Burcat, A.;Lifshitz, A. J . Chem. Phys. 1967,47,3079. (12)Burcat, A.;Lifshitz, A. J. Chem. Phys. 1970,52, 337. (13)Bar-Nun, A.;Lifshitz, A. J. Chem. Phys. 1967,47,2878. (14)Burcat, A.;Lifshitz, A. J . Chem. Phys. 1967,51,1826. (15)Burcat, A.;Lifshitz, A. J. Chem. Phys. 1970,52, 3613. (16)Lifshitz, A.;Frenklach, M. J. Chem. Phys. 1977,67,2803.
@ 1982 American Chemical Society