THERMAL DECOMPOSITION OF CHLOROCYCLOALIUNES
995
The Gas-Phase Thermal Decomposition of Chlorocycloalkanesl by J. M. Sullivan2 Rockeldyne, A Division of North American Rockwell Corporation, Carwga Park, CaliJornia 91304
and W. C. Herndon Chemistry Department, Texas Technological College, Lubbock, Texas 79406 (Received July 93, 1969)
The thermal decompositions of cyclopentyl chloride, cyclohexyl chloride, cycloheptyl chloride, and cyclooetyl chloride were investigatedover the temperature range 350470" using a gas-phase stirred flow reactor. The rates for cyclopentyland cycloheptylchloride were measured relative to cyclohexyl chloride. The Arrhenius expressions for the first-order decomposition of the CS, C6, C7, and CSchlorides were kg = 101a.44exp(-48,570/RT) exp(-47,340/RT) sec-1, ks = 101a.16exp(-44,950/RT) sec-1, kg = 1013.90 exp(-50,250/RT) sec-1, k7 = 1013.88 sec-l. The Arrhenius A factors are discussed in terms of the loose cyclic transition state postulated by O'Neal and Benson.8
Introduction Recently, O'Neal and Benson3 have demonstrated a method for estimating the Arrhenius A factors for fourand six-center unimolecular reactions via application of the transition state expression ek I' A = -- exp(AS*/R) h
The entropy of activation, AS*, is obtained from estimates based upon their model which portrays the transition state as possessing a loose cyclic structure with semiion-pair charge separation a t the reaction centers Rules are given for assigning bending, stretching, and torsion frequencies for such structures and in general the major contributions to AS* are found to come from symmetry changes and losses in hindered internal rotations. An interesting situation arises in the case of the fourcenter elimination of HX from cyclic haloalkanes. In this case there is no loss in internal rotation and the major contributions to AS* come from symmetry changes and unique features of the transition complexes such as partial loss of pseudorotation for the cyclopentyl ring and the possibility of the trans isomer in the case of the cyclooctyl compound. We have investigated the gas-phase pyrolysis of cyclopentyl, cyclohexyl, cycloheptyl, and cyclooctyl chloride which decompose to give HC1 and the corresponding cyclic olefin. The Arrhenius A factors for these unimolecular reactions are discussed in terms of the above model. e
Experimental Section The rate parameters were obtained using a gas-phase stirred flow reactor4-' which was described previously.8 Prepurified nitrogen a t 1 atm pressure was passed from a cylinder through a fine needle valve, used to adjust the flow rate, then through a magnesium perchlorate drying
tube, and into a vaporizer containing the compound (or compounds) to be investigated. The vaporizer was thermostated at 0" to ensure a low, constant partial pressure of reactant (or reactants) in the vapor phase. The gases leaving the reactor were analyzed by means of an Areograph Model A-600 "Hi-Fi" gas chromatograph equipped with a flame ionization detector and a 3-ft by '/8 -in. column packed with 20% Dow Corning #550 Silicone fluid on Embacel (60-100 mesh). The 52ml Pyrex reactor was submerged in a molten-lead bath, the temperature of which was found not to vary by more than *0.3" over a 3-hr period at 400". The flow rates of the gases leaving the reactor were measured with a soap bubble flowmeter and were corrected to reactor temperature by means of the ideal gas law.9 Cyclopentyl chloride and cyclohexyl chloride of the highest purity available were purchased. Gas chromatographic analysis indicated traces of the lower boiling cyclic olefins which were removed by bubbling prepurified nitrogen through the chlorides for several hours. Cycloheptyl chloride and cyclooctyl chloride were prepared from reaction of the corresponding cyclic alcohols with thionyl chloride using chloroform as the (1) Taken from thesis by J. M. S. in partial fulfillment of requirements for M.S., Department of Chemistry, University of Mississippi, 1963. (2) Division of Chemical Development, Tennessee Valley Authority, Muscle Shoals, Ala. 35660. (3) H. E.O'Neal and S. W. Benson, J.Phys. Chem., 71,2903 (1967). (4) W. C.Herndon, J. Chem. Edue., 41,425 (1964). (5) A. A. Frost and R. G. Pearson, "Kinetics and Mechanism," John Wley & Sons, lnc., New York, N. Y.,1956,p 185. (6) J. M. Sullivan and T. J. Houser, Chem. Ind., 1057 (1965). (7) M.F. R. Mulcahy and D. J. Williams, Aust. J . Chem., 14, 534 (1961). (8) W. C. Herndon, M. B. Henley, and J. M. Sullivan, J. Phys. Chem., 67, 2842 (1963). (9) In this investigation the partial pressure of reactant was low in
comparison with the carrier gas and no correction of flow rate for reaction was required. It is of interest to note that for a stirred flow rea* tor no such correction is necessary even at high partial pressures of reactant if the flow rate of the ezit gas is measured.
Volume '74, Nurnber 6 March 6, 1970
996
J. M. SULLIVAN AND W. C. HERNDON
solvent. The chloride products were purified by vacuum distillations and the following boiling points and refractive indices were determined: cycloheptyl chloride, b750mm = 1720110n2sD = 1.4728; cyclooctyl chloride, b i g m m = 87-88°,11n25~ = 1.4825. The cyclic olefins were purchased in the highest purity available and were used without further purification. These olefins were chromatographically identical with those obtained from the pyrolysis of the corresponding cyclic chlorides. The gas chromatographic column used in this investigation did not distinguish between the cis and trans forms of cyclooctene. Initial experiments in clean reaction vessels resulted in fast nonconsistent rates characteristic of this system of compounds.12 However, in seasoned reaction vessels much slower and more consistent results were obtained. The rates of decomposit’ion of cyclohexyl and cyclooctyl chloride were measured directly, while the rates for cyclopentyl and cycloheptyl chloride were measured relative to cyclohexyl chloride by passing the reactant pair through the reactor simultaneously. Consider the simultaneous unimolecular decomposition of two reactants in a stirred flow reactor
A-+B+C
D-+E+F
(1)
(2)
The first-order rate constant for reaction 1 is given by the stirred flow expression
where r is the residence time in the reactor and (A) and (B) are the steady-state concentrations of the reactant and one of the products, respectively, leaving the reactor. The rate of reaction 1 relative to reaction 2 is given by
I n the above expression t,he residence time has canceled and the relative rates are determined simply from the ratio of products and reactants. The first-order character of the thermal decompositions investigated is demonstrated in Figures 1,2, and 3 . Figure 1 shows a plot of (olefin)/(chloride) us. r for cyclohexyl and cyclooctyl chloride at 452 and 370°, respectively. These plots give straight lines in agreement with eq 3. I n Figures 2 and 3 plots of the concentration ratios (cyclohexene)/(cyclohexyl chloride) us. (cyclopeitene)/(cyclopentyl chloride) and (cyclohexene)/(cyclohexyl chloride) us. (cycloheptene)/ (cyc1ohept)ylchloride) at 408 and 422’, respectively, are given. The slopes of these lines are equal to lea/?& and k6/k7, I e;pectively, where k6 is the first-order rate constant for the decomposition of cyclohexyl chloride The Journal of Physical Chemhtry
1.2
--2
CHLORIDE, 452OC
W
1.0
0 w
rz
f: r
0.8
v \
5
-
0.6
lL
W
0 2
0.4 CYCLOOCTYL CHLORIDE, 37OoC
0.2
y
0
0
IO
I 20
I
I
30
40
I 50
I
60
I 70
I 80
T,SEC
Figure 1. Concentration (olefin)/(chloride),,, us. 7 for cyclohexyl and cyclooctyl chloride a t 452 and 370”.
--
0.7
w P
e
f: 0.6 u I 2
%
0.5
I W
Iu: 5 0.4
0.1
0 ( C Y CLOPENTENE) 1 (CY CLOPENTY L CHLOR I DE)
Figure 2. Concentration (cyclohexene)/(cyclohexyl chloride) us. (cyclopentene)/(cyclopentyl chloride) at 408’.
and ks and k7 are the respective rate constants for cyclopentyl and cycloheptyl chloride. Table 1gives the first-order rate constants, k6 and ks, for the pyrolysis of cyclohexyl and cyclooctyl chloride, while the relative rates, h / k s and k6/k7, for cyclohexyl chloride/cyclopentyl chloride and cyclohexyl chloride/ cycloheptyl chloride are given in Table 11. Figures 4 and 5 show plots of log and log k6/k7 us. 1/T for cyclopentyl and cycloheptyl chloride, respectively. The plots of the relative rates, ka/ks and k&7, result in relative Arrhenius parameters, Le., AE and A log A . (10) “Beilstein Handbuch der Organischen Chemie,” Band V, 450498,p 29. (11) 5. A. Miller and W. 0. Jones (to British Oxygen Go., Ltd.), Brit. 738,992,Oct 26,1955; Chem. Abstr., 50, 10768f (1956). (12) A. Maccoll, Chem. Rev., 69,33 (1969).
THERMAL DECOMPOSITION OF CHLOROCYCLOALKANES
t
Table I : Rate Coefficients of Cyclohexyl Chloride and Cyclooctyl Chloride a t Different Temperatures Temp, OC
No. of runs
Average Ice or ks X 102,sec-1
Std dev X 102, sec-1
350.0 385.0 413.8 438.8 452.2 475.8
Cyclohexyl Chloride 17 0.018 15 0.18 11 0.81 13 3.2 18 5.9 18 15.5
0.006 0.01 0.03 0.4 0.2 0.9
370.3 371.5 380.0 405.5 423.6
Cyclooctyl Chloride 22 0.75 28 0.86 5 1.93 5 4.6 5 11.7
0.13 0.10 0.08 0.2 0.5
Table I1 : Relative Rate Coefficients for Cyclohexyl Chloride/Cyclopentyl Chloride and Cyclohexyl Chloride/ Cycloheptyl Chloride a t Different Temperatures
Temp, O
C
No. of runs
Cyclohexyl 387.0 408.0 425.0 445.8 466.0
Average ks/Ics or ke/k7
1 .o
0.02 f 0.10
5.0
Figure 3. Concentration (cyclohexene)/(cyclohexyl chloride)
-0.02
0
-0.10
AE, cal/mol
1,680 rf 200 2,910 rf 300
These results are given in Table 111. Table IV gives the Arrhenius parameters and entropy of activation, as determined by the least-squares procedure, for all of the chlorocycloalkanes studied during this investigation. The parameters for cyclopentyl chloride and cyclo-
'
1.36
I
1
I
I
I
I
1.38
1.40
1.42
1.44
1.46
1.48
I/T
x 103,
1.50
OK-'
Figure 4. Log k6/ksav us. 1/T for cyclohexyl chloride/ cyclopentyl chloride.
y;::;:f#l 2
0.47 f 0.06
4.0
Std dev
Table I11 : Relative Arrhenius Parameters for Cyclohexyl Chloride/Cylopentyl Chloride and Cyclohexyl Chloride/ Cycloheptyl Chloride
Cyclohexyl chloride/ cyclopentyl cldoride Cyclohexyl chloride/ cycloheptyl chloride
3.0
2.0
( C Y C L O H E P T E N E ) / (CYCLOHEPTYL C H L O R I D E )
Chloride/Cyclopentyl Chloride 5 0.808 0.034 8 0.856 0.039 6 0.863 0.030 4 0.890 0.013 7 0.939 0.014
A log A
/8
us. (cycloheptene)/(cycloheptyl chloride) a t 422'.
Cyclohexyl Chloride/Cycloheptyl Chloride 370.4 7 0.109 0.007 384.0 5 0.115 0.004 395.4 4 0.119 0.001 398.3 8 0.116 0.004 407.6 8 0.121 0.020 416.6 6 0.130 0.008 416.7 3 0.121 0.005 421.6 14 0.130 0.017 431.9 6 0.132 0.019 452.0 6 0.143 0.015
Compound
997
\ yq
-0.90
U
-0.92
s
-0.94
-0.96 -0.96
1.38
I
I
I
I
I
I
1.40
1.42
1.44
1.46
1.48
1.50
I/T x 103,
1.52
1.54
OK-'
Figure 5. Log k6/k7ay 2)s. 1/T for cyclohexyl chloride/ cycloheptyl chloride.
heptyl chloride were obtained by subtracting the relative Arrhenius constants A E and A log A from E and log A for cyclohexyl chloride. Swinbourne has studied the pyrolysis of both cyclopentyl and cyclohexyl chloride in a static ~ y s t e m . ' ~ ~ ' ~ Volume 74, Number 6 March 6,1970
J. M. SULLIVAN AND W. C. HERNDON
998
Table IV : Arrhenius Parameters and Entropies of Activation for the Pyrolysis of Compounds
Cyclopentyl chloride Cyclohexyl chloride Cycloheptyl chloride Cyclooctyl chloride
Chlorocycloalkanw Log A
E,oal
A S , gibbs/mol
13.44 f 0.34 48,570 f 1100 -0.73 i 1.5 13.90 f 0.28 50,250
900 $1.42 i 1 . 3
13.88 f 0.38 47,340 i 1200 +l.32 i 1.7 13.16 f 0.28 44,950 i 900 -1.92 f 1.3
He found kg = 1013.4‘ exp(-48,290/RT) sec-’ for cyclopentyl chloride and ke = 101a.77 exp( -49,98O/RT) sec-l for cyclohexyl chloride. Within experimental error, his results are the same as those found in this work (Table IV).
Discussion If the procedure of O’Neal and Bensona is applied to the cycloalkyl chlorides studied during this investigation it is seen that only about +0.2 gibbs/mol of activation entropy is contributed due to the changes in molecular vibrations in passing to the transition state. All other entropy contributions must come from changes in symmetry or from specific features of the individual activated complexes. For cyclopentyl chloride an entropy contribution of R In 2 = 1.37 gibbs/mol would be predicted because of two equivalent nonsuperimposable structures for the transition complex. The experimentally obtained AS = -0.73 gibbs/mol reflects partial freezing of the pseudorotation’s of the five-membered ring due to incipient double bond formation. l6 I n the case of cyclohexyl chloride an entropy contribution of R In 2 = 1.37 gibbs/mol would result from transition states with chlorine in the axial position, while an entropy contribution of R In 4 = 2.75 gibbs/mol is predicted for reaction of an equatorial chlorine atom. Presumably the two positions are nearly equivalent;
*
The J O U To j~Physical Chemistru
hence the observed ASf = 1.42 gibbs/mol appears slightly low, but certainly within experimental error of the predicted value, R In 3 = 2.18 gibbs/mol. The larger C7 and Cs rings show considerable ring strain and the chlorine atom occupies the roomier equatorial positions. Hence for cycloheptyl chloride AS = R In 2 in excellent agreement with the observed value, 1.32 gibbs/mol. Cyclooctyl chloride is the smallest of the ring compounds whose pyrolysis off ers the possibility of an olefin with trans configuration about the double bond. The loose cyclic transition state postulated above would favor the formation of the more unstable truns-cyclooctene17with an attendant loss in entropy. Hence, the entropy contribution of R In 2 is more than overshadowed by partial formation of the trans isomer in the activated complex. In summary, the loose cyclic transition state accounts very nicely for the results observed during this investigation and is favored over the polar transition state postulated by M a c ~ o l l . ’ ~ *This ~ ~ *is~ particularly ~ true in light of the recent study by Fieldz0which showed that the activation energies for the formation of benzyl ion and t-amyl ion from protonated benzyl acetate and protonated t-amyl acetate are identical, despite the fact that the solvolysis of t-C4HsCI proceeds much more readily than does the solvolysis of CeHsCH2Cl. Acknowledgments. Gratitude is extended to the University of Mississippi and the National Science Foundation (G.P. No. 247) for their financial support.
*
(13) E.S.Swinbourne, J . Chem. SOC.,4668 (1960). (14) E.8. Swinbourne, Aust. J. Chem., 11,314 (1958). (15) J. E. Kilpatrick, K. S. Pitzer, and R. Spitzer, J. Amer. Chem. SOC.,69,2483 (1947). (16) O’Neal and Benson attribute the low A factor for cyclopentyl bromide t o the complete freezing of the pseudorotation mode.8 We do not find this to be the case with cyclopentyl chloride. (17) A. C. Cope, R. A. Pike, and C.F. Spencer, J. Amer. Chem. Soc., 75,3212 (1953). (18) A. Maccoll and P.J. Thomas, Nature, 176, 392 (1956). (19) A. Maccoll in “Theoretical Organic Chemistry,” Butterworth and Co., Ltd., Loiidon, 1958. (20) F. H.Field, J . Amer. Chem. Soc., 91,2827 (1969).
