Nov. 5, 1963
COMMUNICATIONS TO THE EDITOR
351 1
CHzClz indicates t h a t both high-spin and low-spin butylene proceeds by an ionic p a t h ~ a y . However, ~ species are present in this solvent. Moreover, in view several workersb have observed that chlorination of an of the above i t is tempting to postulate a planar strucolefin can induce chlorination of paraffinic solvents; ture for the low-spin compound. these reactions appear to involve free-radical cha in reactions. We wish to report some of our studies o n The infrared spectrum of solid [Co(PEt3),(NCS),] in Nujol mull shows a strong band a t 2090 cm.-' chlorination of cyclohexene which shed light on this assigned to the CN stretching vibration of the SCN apparent anomaly group and does not have absorption bands in the 2130Passage of chlorine diluted with nitrogen into cyclo2160-cm. -' region. The possibility that the metal may hexene in the absence of lzghf a t 25.0' produces in a effectively be six-coordinated through SCK groups rapid reaction6 three major products : 1,2-dichlorocyclohexane (I),' 3-chlorocyclohexene (II), and 4seeins in this case to be ruled out. I t is worth noting that the CN stretching frequency of trans- [Ni(PEta)zchlorocyclohexene (III)s; admixture with cyclohexane (NCS)2]has the same value of 2090 cm.-l. The two leads to considerable chlorocyclohexane (117). Use of strong bands observed a t 830 and iS.5 cm.-l in the oxygen as a carrier gas eliminates 111 (and I V with cyclohexane present), whereas I and I1 are produced in a spectrum of [Co(PEt3)2(XCS),] but absent in the spectra of [ C O ( P E ~ ~ ) ~and C ~ [Co(PEt3)2Br2] ?] are assigned c1 C1 to the CS (pseudo-symmetrical N-C-S) stretching vibration of the C N S group in isothiocyanates.6 The doublet structure might arise either from a cis- or from a distorted trans-planar configuration. Similar conI I1 I11 Iv clusions apply to the infrared spectra in CHzCl2 solution. In particular the spectra of 0.25 M solutions in still rapid reaction6; use of polar solvents such as acetothis solvent exhibit bands having the integral intensity nitrile also eliminates I11 and IV. This behavior sugof coordinate isothiocyanate groups. gests the existence of both an ionic pathway producing Conductivity measurements were made in CHzClz I and I1 and a radical pathway producing I, 11, and a t 25' a t a concentration of l o p 3 M. The small con111. ductivity (0.40 the reaction is the combined analytical, conductance, spectral, and almost exclusively radical, whereas a t lower concentramagnetic data is t h a t [Co(PEt3)2(NCS)*]is an example tions of olefin the ionic reaction becomes predominant. of conformational isomerism between a low-spin Consistent with this view is the fact that illumination (probably planar) and a high-spin (tetrahedral) form. (three 275-watt sunlamps a t 6-9 in. through Pyrex) did This problem is being further investigated and denot change the I11 ''11and 1/11 ratios a t high concentratailed reports will follow. tion (the expected result only if reaction were already entirely radical), but did practically eliminate the Acknowledgment.-This work was supported by the changes in I11 I1 and I ;I1 a t low concentration (curves Italian Council for Research (CNR, Rome). 3 and 4). If we assume that the radical process (6) A . Turco and C. Pecile, Future, 191, 66 (1961). produces I, 11, and 111 in the ratio 1.95:l.OO:O.GO ALDOTCRCO ISTITUTO CHIMICA GENERALE (the high concentration limit) and t h a t the ionic process CESAREPECILE ISTITUTO CHIMICA FISICA produces I and I1 in the same ratio under nitrogen as MARINOKICOLIN UNIVERSITA DI PADOVA MARIOMARTELLI under oxygen a t a given concentration of olefin (curve ITALY RECEIVED JULY 22, 1963 l), the observed I11 'I1 ratios can be used to calculate the percentage of radical reaction ( r ) a t each concentration and the corresponding value of 1)'II based on Competition Between Ionic and Free-Radical this r and the previous assumptions (Table I). The calculated results for 1/11 agree with the observed Reactions during Chlorination of Cyclohexene. values within the apparent experimental error ; to this Spontaneous Generation of Radicals extent the assumption that the radical yield ratio (1.95 : Sir: 1.00:0.60) is independent of dilution is confirmed. Although considerable evidence suggests that chlorinUse of cyclohexane as diluent gave comparable ation of olefins in polar solvents involves electrophilic results for cyclohexene-derived products and also alattack by chlorine, the mechanism of chlorination in lowed determination of the relative reactivities of a nonFolar solvents has not been as extensively studied. 4-cyclohexenyl hydrogen atom and a cyclohexyl hydroSLch chlorinations produce both addition and substi(4) W. Reeve, 1). H. Chambers, and C. S. Prickett, i b i d . , 74, ,5369 (19.52). tttion prodncts.* Taft3 has attempted to correlate (.5) (a) T. D. Stewart and I). M. Smith, i b i d , Sl, 3082 (1929), (b) T. 11. Stewart and M. H. Hanson, i b i d . , 63, 1121 (1931): (c) T. I). Stewart and the addition-substitution ratio with olefin structure on B . Weidenbaum, ibid., 67, 2036 (193.51, (d) T .I ) . Stewart and B. Weidenthe basis of ionic intermediates; indeed, labeling exbaum, i b i d . , 6 8 , 98 (1936). (e) J C. Kuriacose, I n d i a n J . A p p i . C h e m . , 28, periments strongly suggest that chlorination of iso181 (1959) (1) P. B. 1). de la Mare, Quaut. Rev. (London), 3, 126 (1949). (2) See, for example, H. P. A . Groll, G. Hearne, F. F. Rust, and W. E . Vaughan, I n d . Eflg. C h e m . , 31, 1239 (1939); I). V Tishchenko, J . Gen. C h e m . C S S R . 8 , 1232 (1938). (3) R . W. T a f t , J r . , J . A m . Chem. SOL.,7 0 , 3364 (1948).
(6) N o visual build-up of chlorine in solution
( 7 ) This product was >9YcZ lrairs in reactions described herein as ionic, but up t o 4 % cis in those described as radical. (8) Suggested as a product without isolation or positive identification by G. F. Bloomfield, J . Chem. SOL., 114 (1944).
COMMUNICATIONS TO THE EDITOR
3512
use of cyclohexene a t progressive stages of purification, ( 2 ) i t is not significantly changed by planned addition of suspected initiators such as hydroperoxides or substantially oxidized cyclohexene, and (3) the radical Such reaction still proceeds very rapidly at -78'. intermolecular "molecule-induced homolyses"l0 have been suggested in reactions of fluorine with aromatic compounds, ' l in the thermal polymerization of styrene, l 2 and in the reaction of styrene with iodine13; this chlorine-cydohexene reaction appears to be a particularly facile example of this class of reactions. Preliminary results show t h a t radical reactions also predominate during liquid phase chlorination of the isomeric butenes except for isobutylene, whose chlorination remains ionic. We hope later to report more fully on the structural factors determining the position of the ionic-radical balance for various olefins. At this point i t seems obvious, however, that prediction of identities and ratios of products from liquid phase chlorination of olefins cannot be made on the basis of a single mechanism common to all olefins and media. Acknowledgment.-The author wishes to ackriowledge the helpful discussions with and suggestions of Dr. R. L. Hinman, Dr. V. Schomaker, and Prof. C. Walling during the course of this study.
0
/=--
18
0
0
c
06
05
p,' J
111
:.:/ -
0.1
I
O
0.1
0.2
0.5
Mole fraction
Cyclohexene
0.8
0.9
Fig. 1.--Variation of product ratios with concentration during chlorination of cyclohexene a t 25.0': 0, nitrogen in the dark; 0 nitrogen with illumination; A, oxygen in the dark.
gen atom toward attack by chlorine atom; this ratio is 0.72 + 0.02 over a' 150-fold change in ratio of starting hydrocarbons. TABLE I OF PERCENTAGE OF RADICAL REACTION ( 7 ) ON DEPENDENCE CONCENTRATION IN CYCLOHEXENE CHLORINATION AT 25.0" IN THE ABSENCEOF LIGHT AND OXYGEN
Mole fraction cyclohexenea
1 00 0 50 30 20 10 050 035 020 010
r
(100) 97 89 79 60 41 31 20 12
(I/II)oaled.
2 2 2 2 2 2 2 3
01 11
26 50 72 83 94 02
Diluted with 1,1,2-trichlorotrifluoroethane of F1g 1 a
VOl. 85
1.95 1 98 2 11 2 42 2 71 2 82 2 95 3 07 From curve 2
Attempts to vary the average chlorine concentration by changing its input rate gave, a t most, minor changes in product ratios a t mole fractions of olefin (0.07-0.20) a t which radical and ionic processes are in competition. Hence, it appears that the most important factor causing the behavior shown in Table I is that the free-radical process is of higher kinetic order in cyclohexene than is the ionic process. We suggest that this high order in olefin results from participation of one or more molecules of cyclohexene in the initiation step, and that this spontaneous free-radical chlorination we have described is initiated by interaction of chlorine and cyclohexene to produce free radical^.^ T h a t initiation is not due to chance impurities is supported by several observations : (1) the percentage of radical reaction is not changed by (9) Since chlorine is stable toward cyclohexane a t 25' in the absence of light. initiation cannot be due to thermal dissociation of chlorine.
(10) J. C. Martin and E. H. Drew, J . A m . Chem. Soc., 83, 1232 (1961). (11) W. T . Miller, Jr., S . D. Koch, Jr., and F. W. McLafferty, i b i d . , 78, 4992 (1956). (12) F. Mayo, quoted in It. R . H i a t t and P. D. Bartlett, ibid., 81, 1149 (1959). (13) G. Fraenkel and P. D. Bartlett, i b i d . , E l , 5582 (1959).
~ J S I O N CARBIDERESEARCH INSTITUTE UNION CARBIDE CORPORATION
MARVIX L. POUTSMA
T A R R Y T O W N , NEW U O R K
RECEIVEDSEPTEMBER 13, 1963
Spin Densities in Tetrahedral Cobalt Complexes by Nuclear Magnetic Resonance Contact Shifts ; Evidence for a Pseudo Contact Interaction
Sir: We wish to report the observation of paramagnetic contact shifts in the n.m.r. spectra of aryl phosphine complexes of cobalt(I1) dibromide. Resonances were observed for the meta- and para-hydrogens of bis(tripheny1phosphine)-dibromocobalt(I1) and . for the meta- and methyl-hydrogens of bis-(tri-p-tolylphosphine)-dibromocobalt(I1). The peak for the parahydrogen was shifted upfield while the peaks for the meta- and methyl-hydrogens were shifted to lower fields with respect to their positions in the diamagnetic ligands. . KO observable ortho-hydrogen resonance has so far been located. Our chemical shifts are similar in character to those recently in some I m a magnetic Ni(1I) chelates. The observed half-width of our resonance peaks is about 20 C.P.S. a t room temperature and the lines become even sharper as the teniperature is lowered. Our spectra were recorded on a Varian n.rn.r. spectrometer operating a t 80 Mc. 'sec. on CDC13 solutions of the complexes with the CHCh resonance taken as an internal standard. The temperature was varied from -64' to +55O. Csing eq. 12,5 to obtain the hyperfine contact interaction constants, ai, from the temperature dependence of the spectra, and relating these to the spin densities, pi, ( I ) n R. Eaton, A D. Josey, W I). Phillips, and R . E . Renwn, J C h r m . P h y s . , 37, 317 1lHCIP). (2) D K E a t o n , A . D Josey, W. D Phillips, and R . E . Benson. Discussions F a l a d a y Sor , 54, 77 (1962). (3) D R . Eaton, A . D Josey, R . E. Benson, U' 11. Phillips, and 7'. I . Cairns, J A m . C h e m . Soc , 84. 4100 11962) ( 4 ) 11. R Eaton, M'. D . Phillips. and D . J . Caldwell. i b i d . , 85, 397 (1963). (5) H . M .McConnell and C H Holm, J . C h e m . P h y s . , 27, 311 11967)
