Australian Journal of Chemistry (1975), 28(10), 2227-54

Aust. J. Chem., 1975, 28, 2227-54

The Acetolysis of Some a-Cyclooctatetraenylalkyl p-Nitrobenzenesulphonates

George E. Gream and Michael Mular

Organic Chemistry Department, University of Adelaide, P.O. Box 498D, Adelaide, S.A. 5001.

Abstract

The synthesis and solvolytic behaviour (in buffered acetic acid) of the p-nitrobenzenesulphonate esters of 2-cyclooctatetraenylethanol, 3-cyclooctatetraenylpropan-1-01 and 4-cyclooctatetraenylbutan-1-01 are described. For comparison with the cyclooctatetraenylalkyl derivatives, kinetic data for the corresponding w-(cyclooct-1'-eny1)alkyl and w-cyclooctyl p-nitrobenzenesulphonates are reported.

Kinetic and product studies have shown that rt-bond participation occurs in the acetolysis of 2-cyclooctatetraenylethyl p-nitrobenzenesulphonate. The nature of the cyclized products, which amount to 95 % when 1.1 equiv. of sodium acetate (the buffer) are used, depends markedly on the concentration of buffer.

Acetolysis of 3-cyclooctatetraenylpropyl p-nitrobenzenesulphonates gives at least 98% non- cyclized products. A probable cyclized product (<2%) is tentatively identified as bicyclo[6,3,0]- undeca-1,3,5,7-tetraene.

Cyclized products (44 %) are formed in the acetolysis of 4-cyclooctatetraenylbutyl p-nitrobenzenesulphonate.

Introduction

As an extension of an interest in n-routes to carbonium and in the chemistry of cycl~octatetraene,~,~ it

was decided to investigate the acetolysis of some o-cyclo- (CH&X

octatetraenylalkyl derivatives (1) (where n = 1-4, and X = good leaving group)

Interest in a study of the capacity of the double bonds in a cyclooctatetraenyl moiety to behave as nucleophiles

0 ,, in the solvolysis of o-cyclooctatetraenylalkyl derivatives is enhanced by the following unique physical and chemical properties of cyclooctatetraene and some of its derivatives (for some reviews,

' Gream, G. E., Aust. J. Chem., 1972, 25, 1051. Gream, G. E., and Serelis, A. K., Aust. J. Chem., 1974, 27, 629. Gream, G. E., Serelis, A. K., and Stoneman, T. I., Aust. J. Chem., 1974, 27, 1711. Gream, G. E., Mular, M., and Wege, D., Aust. J. Chem., 1974, 27, 567. Huisgen, R., Konz, W. E., and Gream, G. E., J. Amer. Chem. Soc., 1970, 92, 4105. Gasteiger, J., Grearn, G. E., Huisgen, R., Konz, W. E., and Schnegg, U., Chem. Ber.,

1971, 104, 2412. Schroder, G., 'Cyclooctatetraen' (Verlag Chemie: Weinheim 1965); Craig, L. E., Chem. Rev.,

1951, 49, 103; Scott, L. T., and Jones, M., Chem. Rev., 1972, 72, 181. * Figes, H. P., in 'Topics in Carbocyclic Chemistry' (Ed. D. Lloyd) Vol. 1, p. 296 (Plenum: New York 1969).

G. E. Gream and M. Mular

Cyclooctatetraene is a tub-shaped molecule with alternating single and double bonds of lengths 1 -476 A and 1.340 A, respec~ively.~ A recent analysisi0 of heats of hydrogenation and other data indicates that the strain and resonance energies in cyclooctatetraene are c. 50 and 60 kJ mol-', respectively. The compound is thus adequately described as 'a slightly strained, weakly conjugated cyclic tetra-olefin'.'

An important property of cyclooctatetraene and its derivatives is the 'mobile' nature of the molecules at room temperature. Extensive n.m.r. studies have established that both bond-shifts and ring-inversions take p1a~e . l ' ~ ' ~

A kinetic examination of the reaction of cyclooctatetraene and some derivatives with dienophiles enabled Huisgen and coworkers13 to establish yet another important feature of these compounds, namely the existence of bicyclic valence tautomers such as bicyclo[4,2,0]octa-2,4,7-triene (2) for cyclooctatetraene itself. By the use of a range of dienophiles, and in particular the very reactive N-phenyltriazoledione, Huisgen5 and Paquettei4 have established that monosubstituted cyclooctatetraenyl derivatives (3) can exist in equilibrium with one or more of the four possible bicyclic valence tautomers (4)-(7). The nature of the substituent R in (3) is however very important in determining which of the valence tautomers (4)-(7) will be favoured.14

Traetteberg, M., Acta Chem. Scand., 1966, 20, 1724. l o Turner, R. B., Mallon, B. J., Tichy, M., Doering, W, von E., Roth, W. R., and Schroder, G., J. Amev. Chem Soc., 1973, 95, 8605.

Anet, F. A. L., J. Amev. Chem. Soc., 1962,84,671; Anet, F. A. L., Bourn, A. J. R., and Lin, Y. S., J. Amer. Chem. Soc., 1964, 86, 3576; Gwynn, D. E., Whitesides, G. M., and Roberts, J. D., J. Amev. Chem. Soc., 1965, 87, 2862; Bryce-Smith, D., Gilbert, A., and Grzonka, J., Angew. Chem., Znt. Ed. Engl., 1971, 10, 746; Allinger, N. L., Sprague, J. T., and Finder, C. J., Tetrahedron, 1973, 29, 2519; Oth, J. F. M., Puve Appl. Chem., 1971,25, 582.

Oth, J. F. M., Merknyi, R., Martini, T., and Shroder, G., Tetvahedvon Lett., 1966, 3087. l3 Huisgen, R., and Mietzch, F., Angew. Chem., Int. Ed. Engl., 1964, 3, 83; Huisgen, R., Mietzch, F., Boche, G., and Seidl, H., 'Organic Reaction Mechanisms' Chem. Soc. Special Publ. No. 19, p. 3 (Chemical Society: London 1965). l4 Paquette, L. A., James, D. R., and Birnberg, G. H., J. Amev. Chem. Soc., 1974, 96, 7454.

An additional characteristic property of cyclooctatetraene is the ease with which it reacts with a proton and other electrophiles to form stabilized homotropylium cations1' [e.g. ion (8) from the action of strong acid on cyclooctatetraene itself]. The formation of homotropylium ions from substituted cyclooctatetraenyl derivatives has been reported.16

When the work described in this paper was commenced (1970), there were no reported examples of the cyclooctatetraene ring behaving as a nucleophile in the solvolytic reactions of cyclooctatetraenyl derivatives. In 1973 however, Paquette and H e n ~ e 1 ~ ' ~ l ~ described the behaviour of cyclooctatetraenylmethyl chloride (9) in aqueous ethanol and 2-cyclooctatetraenylethyl p-bromobenzenesulphonate (15)Vn acetic acid. A planned investigation of the solvolytic behaviour of cyclooctatetraenylmethyl bromide (10) (which had already been preparedlg) was therefore abandoned by us. A study of the behaviour of 2-cyclooctatetraenylethyl p-nitrobenzenesulphonate (14) in acetic acid was however well underway at the time and the results are reported here and compared with those of Paquette and Henzel.

In this paper, the results of a study of the acetolysis of 3-cyclooctatetraenylpropyl p-nitrobenzenesulphonate (19) and 4-cyclooctatetraenylbutyl p-nitrobenzenesulphonate (25) are also reported. For comparative purposes, kinetic data for the w-(cyclooct-1'-eny1)alkyl (11, n = 2-4) and w-cyclooctylalkyl (12, n = 2-4) p-nitrobenzenesulphonates corresponding to (14), (19) and (25) have been determined.

Synthesis of Required Compounds

w-Cyclooctatetraenylalkyl Derivatives

Treatment of a solution of cyclooctatetraenyllithium in ether (prepared readily from bromocyclooctatetraene6 and butyllithium) with ethylene oxide a t low temperature gave a good yield (70%) of 2-cyclooctatetraenylethanol (13). The use of cyclo-

* Abbreviations used in the formulae: Bs, p-bromobenzenesulphonyl; Ns, p-nitrobenzenesulphonyl; Ts, p-toluenesulphonyl; Ac, acetyl.

I5 Story, P. R., and Clark, B. C., in 'Carbonium Ions' (Ed. G . A. Olah and P. von R. Schleyer) VoI. 3, p. 1084 (Wiley-Interscience: New York 1972). I6 Hehre, W. J., J. Amer. Chem. Soc., 1973, 95, 5807, and references therein; Huisgen, R., and Gasteiger, J., Tetrahedron Lett., 1972, 3661, 3665; Paquette, L. A., Broadhurst, M. J., Warner, P., Olah, G. A., and Liang, G., J. Amer. Chem. Soc., 1973, 95, 3387; Ahlberg, P., Harris, D. L., Roberts, M., Warner, P., Seidl, P., Sakai, M., Cook, D., Diaz, A,, Dirlam, J. P., Hamberger, H., and Winstein, S., J. Amer. Chem. Soc., 1972, 94, 7063; Brookhart, M. S., and Atwater, M. A. M., Tetrahedron Lett., 1972, 4399.

Paquette, L. A , and Henzel, K. A., J. Amer. Chem. Soc., 1973, 95, 2724. Paquette, L. A., and Henzel, K. A., J. Amer. Chem. Soc., 1973, 95, 2726.

I9 Bowie, J. H., Gream, G. E., and Mular, M., Aust. J. Chem., 1972, 25, 1107.


octatetraenylmagnesium bromide, rather than the lithio reagent, gave a considerably reduced yield (21 %) of the required alcohol.

Conversion of the alcohol into its p-nitrobenzenesulphonate ester (14) was achieved in good yield with p-nitrobenzenesulphonyl chloride in the presence of a slight excess of pyridine. When pyridine was used in large excess, water-soluble derivatives not extractable into ether were obtained. The preparation of 2-cyclooctatetraenylethyl acetate (16) was readily carried out by treating the alcohol (13) with acetic anhydride in pyridine.

(13) X = O H (17) X = O H (14) X = OKs (19) X = OSs (15) X = OBs (20) X=0.4c (16) X = OXc (22) X = CN (18) X = C0,Et (23) X = C0,H

(24) X = C0,Me

Initially, 3-cyclooctatetraenylpropan-1-01 (17) was prepared by reduction of ethyl 3-cyclooctatetraenylpro~ate (18) with lithium aluminium hydride. The required intermediate ester (18) was however not readily available. Treatment of cyclooctatetraenyllithium with ethyl acrylate in the presence of cuprous chloride gave a very low yield (5 %) of the ester. A similar reaction with cyclooctatetraenylmagnesium bromide gave an increased yield (20%) of (18).

In later work, it was found that 3-cyclooctatetraenylpropan-1-01 could be prepared more conveniently and in better yield (51 %) in a one-step reaction by heating a mixture of cyclooctatetraenylmagnesium bromide and two equivalents of oxetan. When an equivalent of oxetan was used, 3-bromopropan-1-01 (31 %) was the major product; 3-cyclooctatetraenylpropan-1-01 was formed in only 3 % yield. Bromide ion (from magnesium bromide present in the solution of the Grignard reagent) apparently reacts with oxetan to give the above bromo alcohol.

When cyclooctatetraenyllithium was treated with oxetan at room temperature, no hydroxylic compounds were formed. At 80" (in boiling benzene), the major product was heptan-1-01 (30%); none of the desired alcohol (17) could be detected in the product. The above results can be accounted for when one considers that the formation of cyclooctatetraenyllithium from bromocyclooctatetraene and butyllithium is a reversible rea~t ion.~ ' At room temperature, neither organolithium compound is sufficiently reactive to cause ring opening of oxetan and thus no alcoholic products are formed. At SO0, however, butyllithium is apparently a more reactive nucleophile than cyclooctatetraenyllithium and reacts readily with oxetan. It is interesting to note that SearlesZ1 has prepared heptan-1-01 in 30% yield by heating a mixture of butyllithium and oxetan in ether-benzene.

3-Cyclooctatetraenylpropan-1-01 was readily converted into 3-cyclooctatetraenylpropyl p-nitrobenzenesulphonate (19) and its acetyl derivative (20) by procedures already mentioned for the preparation of the corresponding esters of 2-cyclooctatetraenylethanol.

4-Cyclooctatetraenylbutan-1-01 (21) was prepared in a conventional manner from 3-cyclooctatetraenylpropyl p-nitrobenzenesulphonate (19). Treatment of the latter

20 Cope, A. C., Burg, M., and Fenton, S. W., J. Amer. Chem. Soc., 1952, 74, 173. 21 Searles, S., J. Amer. Chem. Soc., 1951, 73, 124.

compound with sodium cyanide in dimethylformamide gave a quantitative yield of 4-cyclooctatetraenylbutyronitrile (22) which on basic hydrolysis was converted into 4-cyclooctatetraenylbutanoic acid (23) (67 %). The crude ester (24) obtained from the last-mentioneri compound by the action of ethereal diazomethane was then reduced with lithium aluminium hydride to give the required 4-cyclooctatetraenylbutan-1-01 (21) (91 %). Conversion of the alcohol into 4-cyclooctatetraenylbutyl p-nitrobenzenesulphonate (25) and the acetyl derivative (26) was achieved without difficulty by the methods already mentioned for the alcohols (13) and (17).

(21) X = OH (25) X = ONs

X (26) X = OAc

o-(Cyclooct-1'-eny1)alkyl Derivatives

1-Bromocyclooctene, an obvious starting material for the synthesis of at least some of the required w-(cyclooct-1'-eny1)alkyl derivatives, was prepared most conveniently in 60% yield by heating 1,2-dibromocyclooctane with diethylamine in a sealed tube a t 100". In the light of earlier work by MaittezZ and B r a ~ d e , ' ~ it was expected that the formation of the Grignard reagent and the lithio derivative from l-bromocyclooctene could be troublesome, especially if impurities were present in the starting bromide. Accordingly, the bromocyclooctene used in the present work was purified by chromatography on alumina followed by careful fractional distillation immediately before use. Unless these steps were taken the lithiation reaction often failed, or resulted in low yields.

Treatment of cyclooct-1-enyllithium or cyclooct-1-enylmagnesium bromide with ethylene oxide gave 2-(cyclooct-1'-eny1)ethanol (27) in yields ranging from 21-25 %."

Unexpectedly, the preparation of 3-(cyclooct-1'-eny1)propan-1-01 (29) presented difficulties. The reaction of cyclooct-1-enyllithium and cyclooct-1-enylmagnesium bromide with ethyl acrylate in the presence of cuprous chloride gave polymeric material, and none of the desired ethyl 3-(cyclooct-1'-eny1)propanoate (30), in both cases. It should be noted that ~ a i t t e " found that the reaction between cyclooct-l- enylmagnesium bromide and acrolein yielded a resin.

* Cyclooct-1-enyllithium has been shownzos24 to undergo reaction with a variety of substrates to give products in yields of 16-75 %. 22 Maitte, P., Bull. Soc. Chim. Fu., 1959, 499. 23 Braude, E. A., in 'Progress in Organic Chemistry' (Ed. J. W. Cook), Vol. 3, Ch. 4 (Butterworths: London 1955). 24 Braude, E. A., Forbes, W. F., Gofton, B. F., Houghton, H. P., and Waight, E. S., J. Chem. Soc., 1957, 4711; Cope, A. C., and Marshall, D. J., J. Amer. Chem. Soc., 1953, 75, 3208.


An attempt to prepare 3-(cyclooct-1'-eny1)propan-1-01 via I-allylcyclooctene (32) (for an analogous synthesis of 3-(cyclohex-1'-eny1)propan-1-01 via I-allylcyclohexene, see3) was abandoned when it proved not possible to obtain a reasonably pure sample of the diene. Treatment of cyclooctanone with allylmagnesium bromide gave 1-allylcyclooctanol (4340%). Attempted dehydration of the tertiary alcohol with p-toluenesulphonic acid in benzene to give the required diene led to a complex mixture of olefinic compounds which was not further investigated.

3-(Cyclooct-1'-eny1)propan-1-01 (29) was successfully prepared, however, in modest yield (38-48 %) by the reaction of cyclooct-1-enyllithium with oxetan. When cyclooct-I-enylmagnesium bromide was allowed to react with oxetan, the required alcohol was also obtained; the yield was however very low (6-10%).

Successful syntheses of 4-(cyclooct-1'-eny1)butan-1-01 (33) and its p-nitrobenzenesulphonate ester (34) were achieved by the routes shown in Scheme 1. Analogous sequences had been used previously to prepare other 4-(cycloalk-1'-eny1)butyl derivative^.^ a 3

Scheme 1. Synthetic route to 4-(cyclooct-1'-eny1)butyl p-nitrobenzenesulphonate (34): (i) CH2=CH(CH2)2MgBr; (ii) oxalic acid, dimethyl sulphoxide; (iii) [Me2CHCH(Me)l2BH; (iv) H202/-OH; (v) NsCI, pyridine.

In the dehydration of 1-(but-3'-enyl)cyclooctanol (35) with oxalic acid in dimethyl s u l p h o ~ i d e , ~ ~ the required diene (36)* was contaminated with a component (15 %) which was assumed to be the isomeric diene (37). The n.m.r. spectrum of the mixture showed a signal at 6 2.65, integrating for 15%, corresponding to the presence of two bis-allylic protons. In 4-cyclohexylidenebut-1-ene,' the two bis-allylic protons give a signal at 6 2.67. The dehydration with oxalic acid of 1-(but-3'-eny1)-

* In attempts to obtain an isomerically pure sample of (36), three methods were investigated: (a) I-bromocyclooctene was treated with the lithium dialkenylcupratez6 prepared from 4-bromobut-l- ene, lithium and cuprous iodide; (b) 1-bromocyclooctene (in tetrahydrofuran) was treated with but-3-enylmagnesium bromide in the presence of ferric ch l~r ide ;~ ' (c) cyclooct-1-enyllithium was treated with 4-bromobut-1-ene under a variety of conditions.

In all three cases, the required olefin was formed in very low yields (1-7%).

25 Beckwith, A. L. J., Gream, G. E., and Struble, D. L., Aust. J. Chem., 1972, 25, 1081. Whitesides, G. M., Fisher, W. F., San Filippo, J., Basche, R. M., and House, H. O., J. Amer.

Chem. Soc., 1969, 91, 4871. 27 Tamura, M., and Kochi, J., J. Ameu. Chem. Soc., 1971,93, 1483, 1485, 1487; Synthesis, 1971, 303.

cycl~hexanol,~ l-(pent-4'-enyl)cyclohexano125 and 1-(but-3'-enyl)cyclopentano13 also give diene mixtures containing small amounts of the exocyclic isomers corresponding

to (37). The presence of (37) in (36) led to the alcohol (33) being contaminated with an impurity (5%) presumably (38). Evidence that the 4-(cyclooct-1'-eny1)butyl p-nitrobenzenesulphonate used in the present work was contaminated with (39) (c. 5%) came from the finding that only 95% of the theoretical amount of p-nitrobenzenesulphonic acid was liberated after 10 half-lives in the kinetic determinations involving (34).

w-Cyclooctyl Derivatives 2-Cyclooctylethanol (40), 3-cyclooctylpropan-1-01 (41) and 4-cyclooctylbutan-1-01

(42) were conveniently prepared by catalytic hydrogenation of the respective w-(cyclooct-1'-eny1)alkan-1-01s (27), (29) and (33). The first-mentioned alcohol was

also prepared by catalytic hydrogenation of 2-cyclooctatetraenylethanol (13). Con- version of each of the saturated alcohols into their respective p-nitrobenzenesulphonate esters (43), (44) and (45) was achieved without difficulty by the action of p-nitrobenzenesulphonyl chloride in pyridine.

Kinetic and Product Studies 2-Cyclooctatetvaenylethyl, 2-(Cyclooct-1'-eny1)ethyl and 2-Cyclooctylethyl p-Nitro- benzenesulphonates (14), (28) and (43)

Kinetic determinations (Table 1) as well as the formation of cyclized products (Table 2) clearly show that the cyclooctatetraenyl moiety assists in the ionization process in the acetolysis of 2-cyclooctatetraenylethyl p-nitrobenzenesulphonate (14). At 100°, the ester (14) is 10.3 times* more reactive than its saturated analogue (43).

* For the corresponding p-bromobenzenesulphonate derivatives, Paquette and Henzel18 found the factor to be 4.6 at 65'. 2-Cyclooctatetraenylethyl and 2-(cyclooct-1'-eny1)ethyl p-nitrobenzenesulphonates (14) and (28), respectively were found to be 3 .2 and 3.5 times, respectively more reactive than the corresponding p-bromobenzenesulphonates (using data of Paquette and Henzells). Values of 3 to 4 Can be expected (from ~ a l ~ e ~ ~ ~ ' ~ ~ given for ~ R O N S / ~ R O T S and ~ R O B S ~ ~ R O T S ) for ~ R O N S / ~ R O B S

(for caution however about the use of values for these types of ratios, see30).

28 Bartlett, P. D., Bank, S., Crawford, R. J., and Schmid, G. H., J. Amer. Chem. Soc., 1965, 87, 1288; Krapcho, A. P., and Johanson, R. G., J. Org. Chem., 1971, 36, 146; Muneyuki, R., and Tanida, H., J. Amer. Chem. Soc., 1968, 90, 656. 2 9 Winstein, S., and Heck, R., J. Amer. Chem. Soc., 1956, 78, 4801. 30 McDonald, R. N., Wolfe, N. L., and Petty, H. E., J. Org. Chem., 1973, 38, 1106.


The double bonds in the cyclooctatetraenyl ring are however not as effective as a similarly located isolated double bond in assisting ionization. This is clearly shown by the finding that 2-(cyclooct-1'-eny1)ethyl p-nitrobenzenesulphonate (28) is 62 times more reactive than its cyclooctatetraenyl analogue (14) at 65" and 398 times more reactive than its saturated analogue (43) at 100". On the other hand, the cyclooctatetraenyl ring is more effective than a phenyl ring [despite its larger adverse inductive effect (see later)] as a nucleophile since 2-cyclooctatetraenylethyl p-bromobenzenesulphonate (15)18 undergoes acetolysis at 75" 22 times faster than does 2-phenylethyl p-bromobenzenes~lphonate.~~

Table 1. Rates of acetolysis of 2-cyclooctatetraenylethyl and related p-nitrobenzenesulphonates The solutions were initially 0 O ~ M in ester and 0 . 0 2 ~ in sodium acetate

p-Nitrobenzene- Temp. lo5k AH4 AS$ kunsat/ sulphonate ('c) (s-l) (kJ mol-I) (J K-' mol-') k,,,

2-Cyclooctatetraenylethyl (14) 65.0 2.09, 2.16 75.0 6.74, 6.90 106.7 (0.51)A -19.6 (l.4)A 85.0 18.8, 18.8

100.0 83.9B 10.3 2-(Cyclooct-1'-eny1)ethyl (28) 35.0 4.46, 4.54

45.0 12.0, 12.3 97.2 (0.42)A -13.6 (l.3)A 55.0 48.8,49.1 65.0 132B

100.0 3230B 398 2-Cyclooctylethyl (43) 100.0 8.11, 8.20

A Value in brackets is the standard deviation. Extrapolated value based on activation parameters.

Table 2. Products from soivolysis of 0 . 0 1 ~ solution of 2-cyclooctatetraenylethyl p-nitrobenzenesulphonate (14)

in acetic acid containing sodium acetate at 85"

Sodium acetate Products (%) of solvolysis (MI (46) (16) (47)

Provided that the cyclized cations (48) and (49) (Scheme 2) do not ring-open to give the non-cyclized acetate (16), it should be possible to predict the percentage of cyclized products formed from the p-nitrobenzenesulphonate (14) on the basis of kinetic data given in Table 1. Fortunately, Paquette and Henzel18 have provided experimental evidence that ring-opening of the cyclized cations formed from the acetolysis of (15) to generate (16) is not a significant process. The relation3'

where k,,, is the observed rate of solvolysis, kc is the rate of cyclization of an

31 Roman, S. A., and Closson, W. D., J. Amer. Chem. Soc., 1969, 91, 1701.

unsaturated derivative and k, is the rate of unassisted solvolysis, can be rewritten as

where k, is the rate of solvolysis of the saturated analogue and x represents the rate retardation attributable to the double bond. Unfortunately, a reliable value of x for homoallylic systems is not a~ai lable .~ ' An estimate of 0 .5 has been made33 (also see34) and this is close to the value of 0.47 found for 3-methylenecyclohexy1 p-toluenesulphonate, an homoallylic derivative which undergoes acetolysis without double bond parti~ipation.~' If it is assumed that the retardation factor for homoallylic systems is 0.5, it must be corrected to 0.38 (0.5 x 0.6/0-79) for 2-cyclooctatetraenylethyl p-nitrobenzenesulphonate (14) since the kinetic data (Table 3 below) for 3-cyclooctatetraenylpropyl p-nitrobenzenesulphonate (19) show that the cyclooctatetraenyl ring exhibits a larger adverse inductive effect than does a similarly located lone double bond. With the above expression, it can be calculated that the yield of cyclized products from (14) should be 96%, which is in reasonable agreement with 93 % [the interpolated value from Table 2 (for a solution initially 0 . 0 1 ~ in ester and 0 . 0 2 ~ in sodium acetate)].

In the acetolysis of 2-cyclooctatetraenylethyl p-nitrobenzenesulphonate (14), 1,2-dihydronaphthalene (46), 2-cyclooctatetraenylethyl acetate (16)" and the tetrahydroazulenyl acetate (47) (as well as traces of azulene) were formed in varying yields depending on the concentration of sodium acetate present (Table 2). The marked dependence of product composition on sodium acetate concentration found in the present work is in accord with the findings of Paquette and Henze1l8 for the acetolysis of 2-cyclooctatetraenylethyl p-bromobenzenesulphonate (15). In one

aspect, however, the results of the American workers differ from ours. When a molar equivalent of sodium acetate was used, they identified naphthalene (7 %) and tetralin (7%) in the product from (15). These two compounds were not detected in the present work when (14) was acetolysed in the presence of 1 .1 molar equivalents of sodium acetate. Although 1,2-dihydronaphthalene (46), disproportionates at high temperatures and also under acidic conditions into naphthalene and t e t r a l i ~ ~ , ~ ~ it was found to be stable under the conditions of acetolysis used for (14).

To rationalize the formation of 1,2-dihydronaphthalene (46) and the acetate (47) from (15), Paquette and ~ e n z e l ' ~ proposed the pathway(s) shown in Scheme 2. On

* The acetate (16) is assumed to have been formed by direct nucleophilic displacement by the solvent and added acetate ion. Paquette and Henzel18 have provided experimental evidence which is consistent with this assumption.

32 Closson, W. D., and Kwiatkowski, G. T., Tetrahedron, 1965, 21, 2779. 3 3 Servis, K. L., and Roberts, J. D., J. Amer. Chem. Soc., 1964, 86, 3773. 34 Bergstrom, C . G., and Siegel, S., J. Amer. Chem. Soc., 1952, 74, 145; Laughton, P. M., and Robertson, R. E., Can. J. Chem., 1955, 33, 1207; Roberts, J. D., and Mazur, R. H., J. Amer. Chem. Soc., 1951, 73, 2509; Vernon, C . A., J. Chem. Soc., 1954, 4462. 3 5 Gill, G. B., and Hawkins, S., Chem. Commun., 1974, 742, and references therein.


the available evidence, we believe that this mechanism accounts satisfactorily for the rearrangements which take place when (14) is acetolysed. In order to define the mechanism more precisely, one would like to know whether (i) the cation (48) is preceded by the non-planar species (51), (ii) the cation (49) is converted into the planar cyclooctatetraenyl type (52), and (iii) the ions undergo 'leakage' to the much more

Scheme 2. Paquette and Henzel'sls pathway@) for the transformation of 2-cyclooctatetraenylethyl p-bromobenzenesulphonate (15) into dihydronaphthalene (46) and the tetrahydroazulenyl acetate (47).

stable homotropylium ions (53) and (54) before conversion into the cation (50). It is also possible that the two homotropylium ions might be formed directly without passing through species such as (48) and (49). Winstein and coworkers36 have calculated that a planar cyclooctatetraenyl cation has a free energy which is 93 a2 kJ mol-I higher than that of a homoaromatic monohomotropylium ion.

Another interesting aspect to consider is the possibility that cyclization may be preceded by valence isomerization* of the p-nitrobenzenesulphonate (14) as shown in Scheme 3. For ethylcyclooctatetraene (55), a possible model for (14), the free energy of activation for the valence isomerization to (56) is 112 kJ mol-I at At the same temperature, the free energy of activation for the acetolysis of Zcyclooctatetra- enylethyl p-nitrobenzenesulphonate (14) is 114 kJ mol-I. Also of relevance is the

* The role that valence tautomerism may play in solvolytic reactions is well illustrated for 7-cycloheptatrienylcarbinyl derivative^.^^

36 Winstein, S., Kreiter, C . G., and Brauman, J. I., J. Amer. Chem. Soc., 1966, 88, 2047. 37 Warner, P., and Lu, S., Tetrahedron Lett., 1974, 3455, and references therein; Thompson, G. L., Heyd, W. E., and Paquette, L. A., J. Amer. Chem. Soc., 1974, 96, 3177, and references therein. 38 Gream, G. E., and Huisgen, R., unpublished data.

finding by Wiberg and Hiatt3' that 2-(cyclobut-1'-eny1)ethyl p-toluenesulphonate (57) undergoes acetolysis with n-bond participation to give 80 % of cyclized products. It is thus clear that a pathway initiated by valence isomerization of (14) as shown in Scheme 3 may be involved, at least in part, in the acetolysis of (14). In the

Scheme 3. Possible pathway for acetolysis of 2-cyclooctatetraenylethylp-nitrobenzenesulphonate (14).

absence of further information, the number and exact nature of the intermediates and ionic species involved in the acetolysis of 2-cyclooctatetraenylethyl derivatives remains undefined.

3-Cyclooctatetraenylpropyl(19), 3-(Cyclooct-1'-eny1)propyI (31) and 3-Cycloocty&opyl (44) p-Nitrobenzenesulghonates

From Table 3, it can be seen that 3-cyclooctatetraenylpropyl p-nitrobenzenesulphonate (19) and 3-(cyclooct-1'-eny1)propyl p-nitrobenzenesulphonate (31) undergo acetolysis at 100" only 0.6 and 0.79 times, respectively, as fast as the saturated analogue (44). These findings are in harmony with the general behaviour of substrates which have a double bond in the 4,5-position to the leaving group and which solvolyse more slowly than their saturated analogues since n-participation does not occur; consequently cyclized products are not (for an exception, see4'). The rate retardation exhibited by the olefinic ester (31) can be attributed to the adverse inductive effect of the double bond. In the case of the cyclooctatetraenyl counterpart (19), the inductive effect exhibited by the cyclooctatetraene ring is approximately 25 % greater than that of the isolated double bond in (31) due to the presence of the three additional double bonds in the former compound. The adverse inductive effect of the cyclooctatetraene ring is also greater than that of a benzene ring since the cyclooctatetraenyl derivative (19) undergoes acetolysis at 100' at a rate 0.8 1 times that of 3-phenylpropyl p-nitrobenzenesulphonate.

39 Wiberg, K. B., and Hiatt, J. E., J. Ameu. Chem. Soc., 1968, 90, 6495. 40 Berson, J. A,, Donald, D. S., and Libbey, W. J., J. Arner. Chem. Soc., 1974, 96, 5580.


Three components were present (by g.1.c.) in the acetolysis product formed from 3-cycloo~tatetraenylpropyl p-nitrobenzenesulphonate (19). The major component (97.4%) was readily identified as the uncyclized 3-cyclooctatetraenylpropyl acetate (20) formed by direct nucleophilic displacement by the solvent and added acetate ion. One of the remaining components (1 %) in the mixture was identified as allylcyclooctatetraene (58), an authentic sample of which was prepared by treating cyclooctatetraenylmagnesium bromide with allyl bromide in the presence of lithium chloride-copper chloride catalyst (Li2CuC1,) according to the general method of KochLz7 The possibility that (58) may have been formed by elimination of acetic acid from the acetate (20) was discounted when (20) was recovered quantitatively after being subjected to the conditions of the acetolysis.

Table 3. Rates of acetolysis of 3-cyclooctatetraenylpropyl and related p-nitrobenzenesulphonates at 100°C

The solutions were initially 0 . 0 1 ~ in ester and 0 . 0 2 ~ in sodium acetate

The remaining compound (1 -6 %) formed from the acetolysis of (19) could not be definitely identified. Attempts to isolate the compound by preparative g.1.c. and column chromatography were unsuccessful since the separated substance polymerized before it could be satisfactorily characterized by its spectral properties. 1-Phenyl- cyclopentene (59), a possible solvolysis product from (19) based on mechanistic

considerations, was found to have g.1.c. characteristics different to those of the unidentified compound. The mass spectrum (obtainedfrom a g.1.c.-mass spectrometer combination) of the compound indicates that it has a molecular weight of 144 and that it is probably bicyclic and not a monosubstituted cyclooctatetraenyl derivative. With few exceptions, monosubstituted cyclooctatetraenyl derivatives studied in the present and earlier worklg have shown base peaks at mle 117. The unknown compound showed its base peak at mle 144, with the peak at mle 117 occurring in a relative abundance of only 34%. The mass spectrum of the compound was actually very similar to those of bicyclo[6,4,0]dodeca-1,3,5,7-tetraene (65) and bicyclo[6,4,0]dodeca- 2,4,6,1(12)-tetraene (69, each of which had base peaks corresponding to the molecular ions and peaks at mle 117 which were present in relative abundances of only 25 and 37 %, respectively.

On the basis of its mass spectrum, the unidentified compound is probably either bicyclo[6,3,0]undeca-1,3,5,7-tetraene (60) or bicyclo[6,3,0]undeca-2,4,6,l(ll)-tetraene

(61). The former compound* was recently synthesized in small quantities by Paquette and coworkers4' and was apparently sufficiently stable to allow isolation in a pure state by preparative g.1.c. Despite our failure to isolate it in a pure state, we believe that the unidentified compound formed by acetolysjs of (19) is (60) and not the isomeric (61) for the following reasons. Models of 3-cyclooctatetraenylpropyl p-nitrobenzenesulphonate (19) show that a more likely arrangement for participati0n.j-

can be obtained when the carbon atom bearing the leaving group is placed over the 7',S1-double bond (which is in a 5,6-position relative to the leaving group) than over the 11,2'-double bond in the molecule. Participation by the former double bond would lead (in at least a formal sense) to the cation (62). Loss of a proton from the intermediate cation [which might be (62), the planar species (63) or the homotropylium cation (64)] would lead to (60).

4-Cyclooctatetraenylbutyl (25), 4-(Cyclooct-1'-eny1)butyl (34) and 4-Cyclooctylbutyl (45) p-Nitrobenzenesulphonates

Convincing evidence that the ionization of 4-cyclooctatetraenylbutyl p-nitrobenzenesulphonate (25) is assisted during acetolysis comes from both kinetic (Table 4) and product studies (see later). At 100°, the ester undergoes acetolysis 1 .6 times faster than the saturated analogue (45). The cyclooctatetraenyl moiety is however less efficient than a similarly located double bond as an internal nucleophile since 4-cyclooctatetraenylbutyl p-nitrobenzenesulphonate is 53 times less reactive than 4-(cyclooct-1'-eny1)butyl p-nitrobenzenesulphonate (34) in acetic acid. On the other hand, the cyclooctatetraenyl ring is a more efficient internal nucleophile than a phenyl

* Bond shifts can take place in cyclooctatetraenyl derivatives."-l2 There is evidence" that 1,2- disubstituted cyclooctatetraenes exist preferentially, and sometimes even exclusively, as the less crowded tautomer. Paquette and coworker^^^,^^ have represented (60) and (65), however, as the more crowded tautomers. f For n-bond participation to occur, it is necessary that the carbon atom bearing the leaving group be located directly over the double bond in the plane of the n - o r b i t a l ~ . ~ ~

41 Paquette, L. A., Wingard, R. E., and Photis, J. M., J. Amer. Chem. Soc., 1974, 96, 5801. 42 Paquette, L. A., and Wingard, R. E., J. Amer. Chem. Soc., 1972, 94, 4398; Paquette, L. A., Wingard, R. E., and Meisinger, R. H., J. Amer. Chem. Soc., 1971, 93, 1047; Paquette, L. A., ~ e i s i n ~ e r , ? ~ . H., and Wingard, R. E., J. Amer. Chem. Soc., 1972, 94, 2155; 1973, 95, 2230; Russell, R. K., Wingard, R. E., and Paquette, L. A., J. Ameu. Chem. Soc., 1974, 96, 7483. 43 Bartlett, P. D., Trahanovsky, W. S., Bolon, D. A., and Schmid, G . H., J. Ameu. Chem. Soc., 1965, 87, 1314; Chuit, C . , Colaard, F., and Felkin, H., Chem. Commun., 1966, 118; Sargent, G. D., and Mclaughlin, T. E., Tetrahedon Lett., 1970, 4359.


ring since (25) is 1 .8 times more reactive than 4-phenylbutyl p-nitrobenzenesulphonate* in acetic acid at 100".

The percentage (44%) (see later) of cyclized products formed in the acetolysis of 4-cyclooctatetraenylbutyl p-nitrobenzenesulphonate (25) supports the kinetic evidence for n-participation in the rate-determining step. In earlier with substrates undergoing acetolysis with z-bond participation, the percentage of cyclized products was calculated with good accuracy from the relation (see earlier)

assuming that ku = 0+87ks. For the ester (25), it would seem logical that the factor of 0.877 should be modified in order to take account of the larger adverse inductive

TaMe 4. Rates of acetolysis of 4-cyclooctatetraenylbutyl and related p-nitrobenzenesulphonates The solutions were initially 0 . 0 1 ~ in ester and 0 . 0 2 ~ in sodium acetate

p-Nitrobenzene- Temp. lo5k AH$ A S k ~ ~ t / sulphonate P.3 (s - l) (kJ mol-I) (J K-I mol-') k,,,

4-Cyclooctatetraenyl- 60 0.249, (0. 242)A but-1-yl (25) 91 5.76, 5.79

100 12.7, 12.9 100.1 (0.72)B -52.9 (1.9)B 1.59 107 24.0,24.5

4-(Cyclooct-1'-eny1)butyl (34) 51 7.14, 7.24 60 18.2, 18.2 90.5 (0.57)B -45.7 ( 1 ~ 7 ) ~ 67 36.6, 36.6

100 67gA 84 4-Cyclooctylbutyl (45) 100 7.93, 8.16 4-Phenylbutyl 100 6.83, 7.15

A Extrapolated value based on activation parameters. Value in brackets is the standard deviation.

effect of the cyclooctatetraenyl ring when compared with a similarly located double bond. The extent of this increased effect is given by the finding (Table 3) that 3-cyclooctatetraenylpropyl p-nitrobenzenesulphonate (19) is only 0.75 times as reactive in acetic acid as 3-(cyclooct-1'-eny1)propyl p-nitrobenzenesulphonate (31). It would thus seem that the above factor of 0 - 87 should be replaced by O.65(0 -87 x 0.75). Use of this factor in the above expression leads to the prediction that the percentage of cyclized product formed by acetolysis of 4-cyclooctatetraenylbutyl p-nitrobenzenesulphonate should be 59 % (observed value being 44 %). On the other hand, use of the factor of 0.87 unexpectedly predicts that the percentage of cyclized products from (25) should be 45%, close to the observed value.

Analysis by g.1.c. of the product formed by the acetolysis of 4-cyclooctatetraenylbutyl p-nitrobenzenesulphonate showed the presence of four components. The major

* Acetolysis of 4-phenylbutyl p-brornobenzenesulphonate gives a product containing only 5.5 % of cyclized product (tetralin).44 t It should be noted that Bartlett45 has suggested that the factor 0.76 can lead to a more accurate prediction of the percentage of cyclized product formed in the solvolysis of a substrate with a double bond in the 5,6-position to the leaving group.

44 Heck, R., and Winstein, S., J. Amer. Chem. Soc., 1957, 79, 3105. 45 Bartlett, P. D., Nicholson, E. M., and Owyang, R., Tetrahedron, 1966, Suppl. 8, Part 2, 399.

component (56.2 %) and a minor one (2.6 %) were shown to be 4-cyclooctatetraenylbutyl acetate (26) and bicyclo[6,4,0]dodeca-1,3,5,74etraene (65), respectively, by analytical g.1.c. and by some of their spectral properties after isolation by preparative g.1.c.

The identification of the other two components, formed in 0 .9% and 40.3% yield, in the acetolysis mixture proved difficult. A combination of g.1.c.-mass spectrometry showed that their molecular weights were 158 and that they were thus isomeric with (65). Several attempts to isolate the two unknown substances by preparative g.1.c. and column chromatography were unsuccessful because the separated compounds polymerized before meaningful spectral data could be obtained.

There were good reasons to suspect that the two unknown compounds could have been 1-phenylcyclohexene (66), 4-cyclooctatetraenylbut-I-ene (67) or bicyclo[6,4,0]- dodeca-2,4,6,1(12)-tetraene (68). Significantly, one of the compounds gave a mass spectrum which was very similar to that of bicyclo[6,4,0]dodeca-1,3,5,7-tetraene (65), suggesting that it might be (68). The mass spectrum of the other compound showed a prominent molecular ion peak at mle 158 and a base peak at mle 117; this suggests that it was a monosubstituted cyclooctatetraene derivative (see earlier). Prominent peaks at mle 143, 129, 117, 115, 103 and 91 are those which would be expectedlg in the mass spectrum of 4-cyclooctatetraenylbut-1-ene (67).

1-Phenylcyclohexene (66) was eliminated as a possibility when it was found that its g.1.c. characteristics were quite different to those of the unidentified compounds. All attempts to prepare an authentic sample of 4-cyclooctatetraenylbut-1-ene (67) were however unsuccessful. Pyrolysis of 4-cyclooctatetraenylbutyl acetate (26) gave complex mixtures which rapidly polymerized. Treatment of cyclooctatetraenylmagnesium bromide with 4-bromobut-1-ene in the presence of lithium chloride-copper chloride (Li,CuC14) catalyst2' gave only polymeric material. When but-3-enylmagnesium bromide was allowed to react with bromocyclooctatetraene in the presence of ferric ~hlor ide ,~ ' a low yield (6%) of a yellow oil which contained at least five components (by g.1.c.) was obtained. The reaction between cyclooctatetraenyllithium and 4-bromobut-1-ene gave a tar. Kochi2' has shown that the Grignard reagents formed from primary halides undergo coupling reactions more readily than those formed from vinylic halides. When cyclooctatetraenylmethylmagnesium bromide was treated with ally1 bromide in the presence of lithium chloride-copper chloride catalyst, a yellow oil (36 %) which rapidly polymerized was obtained. At this stage, further attempts to synthesize (67) were abandoned.

The third possibility for one of the unknown compounds was the tetraene (68). An attempt to synthesize the compound was however unsuccessful. Treatment of cyclooctatetraene dianion with 1,4-dibromobutane gave bicyclo[6,4,0,]dodeca-2,4,6- triene (69)46 which was obtained by preparative g.1.c. has shown that the

46 Cotton, F. A., and Deganello, G., J. Amer. Chem. Soc., 1972, 94, 2142.


compound is in valence tautomerism with (70) in the ratio of c. 7 : 3 at room temperature and 1 : 1 at 114'. It was envisaged that bromination of (69) would give the bromide (71) which on treatment with base would lose hydrogen bromide to give a mixture of the required olefin (68) and its isomer (65). Attempted brornination of (69) with N-bromosuccinimide however gave black tars from which no pure components could be obtained; as a result, the investigation was not pursued further.

At this stage, it was decided to attempt to overcome the difficulty created by polymerization of the two components formed by acetolysis of 4-cyclooctatetraenylbutyl p-nitrobenzenesulphonate by catalytic hydrogenation of the mixture. After the ester (25) had been acetolysed at 100' for 16 h (c. 10 half-lives), the solution (without isolation of products) was immediately hydrogenated at atmospheric pressure and temperature in the presence of excess of platinum. G.1.c.-mass spectral analysis showed that the hydrogenated mixture contained butylcyclooctane (72) (1 . I%), cis- and trans-bicyclo[6,4,0]dodecane (73) (46.1 %) and 4-cyclooctylbutyl acetate (74) (52.8 %).

The results f;om the hydrogenation of the acetolysis mixture, combined with the mass spectral data (see earlier), suggest very strongly that the two difficult-to-identify compounds in the initial acetolysis mixture were 4-cyclooctatetraenylbut-1-ene (67) (0.9 %) and bicyclo[6,4,0]dodeca-2,4,6,1(12)-tetraene (68) (40.3%). It should be noted that the former compound could not have arisen from initially formed 4-cyclooctatetraenylbutyl acetate by elimination of acetic acid since the acetate was recovered quantitatively when subjected to the conditions of the acetolysis reaction.

A model of 4-cyclooctatetraenylbutyl p-nitrobenzenesulphoaate (25) clearly shows that the carbon bearing the leaving group can be placed directly over the lf,2'-double bond in the plane of the n-orbitals, enabling n-bond participation to occur. It is thus possible that the ion (76) is formed via the transition state (75). At this stage, it is not possible to determine whether (76) 'leaks' to the planar species (77) and/or the homotropylium ion (78) before product formation occurs.

Another possible route to the ions (76) and/or (77) and/or (78) is shown in Scheme 4. It has already been mentioned that the free energy of activation for the valence isomerization of ethylcyclooctatetraene (55), a seemingly good model for the sulphonate ester (25), is 112 kJ mol-I at 100". Since the free energy of activation

for the acetolysis of (25) is 120 kJ mol-I at 100°, a pathway initiated by valence isomerization as shown in Scheme 4 may be involved," at least in part.

In earlier work from these laboratories involving the 9-de~aly l '~~ and 8-hydrindy13 cations, it was argued that the counter-ion in ion-pairs may be involved in product formation. Let us now consider whether such an argument can be used to rationalize the predominant formation of bicyclo[6,4,0]dodeca-2,4,6,1(12)-tetraene (68) from the p-nitrobenzenesulphonate (25). First, it should be noted that cyclization from (25)

Scheme 4. Possible pathway for acetolysis of 4-cyclooctatetraenylbutylp-nitrobenzenesulphonate (25).

can occur in two ways. In (80), the carbon bearing the leaving group lies away from the ring and there are no unusual steric interactions which might hinder cyclization. On the other hand, unfavourable interactions, especially from the 5',6'-double bond, should seriously hinder cyclization via (81) in which the carbon bearing the leaving group lies within the ring. Thus it would seem that only ion-pairs derived from (80) should be considered in an attempt to rationalize the predominant formation of (68).

If the cation ('76) is the direct precursor of the olefins (65) and (68), three ion-pairs (82)-(84) may be initially formed from (25) via (80). In (82) and (83), the newly formed six-membered rings in the cations exist in boat-like conformations. In the ion-pair (82), the counter-ion is suitably located to be able to act as a base in

* Work is in progress (with C. Schuetz) to determine whether n-bond participation occurs during the solvolysis of 4-(cyclobut-1'-eny1)butyl p-nitrobenzenesulphonate, the simple allalogue of (79).


abstracting the axially disposed hydrogen* at C 12 as a proton to give the olefin (68). On the other hand, it is on the wrong side of the cation to allow proton abstraction from C 8 to give (65). In the ion-pair (83), the counter-ion is unable to function directly as a base since the axially disposed hydrogens at C 8 and C 12 are on the wrong side of the cationic species.

The possibility that the initially formed ion-pair may be (84) in which the newly formed six-membered ring has a chair-like conformation must also be considered. Indeed, Closson and Gray4' have argued that the solvolytic cyclization of hex-5-enyl derivatives probably proceeds in general through chair-like conformations of the newly formed six-membered ring. In (84) however, the counter-ion is unable to abstract a proton from C 8 or C 12 since the correctly oriented hydrogens are on the wrong side of the cation. If the ion-pair (84) is initially formed from (25), and if the counter-ion is to be involved as a base (in initially formed tight ion-pairs) jn product formation, it would be necessary for the cationic species in (84) to change to the boat-like conformation in (82). Such a possibility is unlikely, however, since product formation is probably much faster than conformational change?'

It has already been mentioned that product formation from (25) may occur via the planar species (77) and/or the homotropylium ion (78). Again, arguments based on the position of the counter-ion in ion-pairs involving these species would lead to the prediction that the olefin (681, and not (65), should be formed preferentially.

Experimental Melting points were determined on a Kofler hot-stage microscope and are uncorrected, as are

boiling points. Infrared, n.m.r. and mass spectra and microanalyses were determined as previously described.l

Low-boiling light petroleum, and light petroleum, refer to the fractions having b.p. 30-40", and 55-65", respectively. Unless stated otherwise, all organic extracts were dried over anhydrous magnesium suIphate.

In early work, the analytical gas-liquid chromatography was carried out with either a Perkin- Elmer 800 or 881 gas chromatograph. The later work (including all quantitative analyses) was carried out with a Perkin-Elmer 990 instrument connected to an Hitachi recorder equipped with a disc integrator. Preparative g.1.c. was carried out with a Varian A-700 or A-705 Autoprep, or a Pye Unicam 104 instrument. All instruments were equipped with flame ionization detectors. The following columns were used:

(A) 5 % Carbowax 20M on Gaschrom P (80-100 mesh) which had been treated with cold 10% aqueous sodium hydroxide for 10 min, washed thoroughly with distilled water, and dried at 130" for 15 h, 3 . 0 m by 2.1 mm.

(B) 5 % Apiezon M on Varaport 30 (100-120 mesh), 3.0 m by 3.1 mm.

* The possibility of olefin formation is considered only when the carbon-hydrogen bond being broken is coplanar, or almost coplanar, with the vacant p-orbital at the cationic site.

47 Closson, W. D., and Gray, D., J. Org. Ckem., 1970, 35, 3737. 48 Fort? Fi. C.? Hornish, R. E., and Liang, G . A., J. Amer. Chem. Soc., 1970, 92, 7558.

(c) 10 % Carbowax 20M on Chromosorb W (100-120 mesh), 2 .4 m by 1.5 mm. (D) 5% Hyprose SP80 on Varaport 30 (80-100 mesh), I .6 m by 2.0 mm. (E) 0.75 % FFAP on Chromosorb W (100-120 mesh), 4.9 m by 2.0 mm. (F) 5 % Carbowax 20M on Varaport 30 (100-120 mesh), 2.4 m by 2.0 mm. (G) 5 % FFAP on Chromosorb W (100-120 mesh), 3.0 m by 3.1 mm. (H) 3 % SE30 on Chromosorb W (100-120 mesh), 2.0 m by 1.5 mm. (I) 5 % Apiezon M on Varaport 30 (100-120 mesh), 2.4 m by 2.0 mm. (J) 20% FFAP on Chromosorb W (80-loomesh), 2.0 m by 8.0 mm. (K) 20 % Carbowax 20M on Chromosorb W (60-80 mesh), 3.6 m by 6.3 rnm.

The columns A, B, D, F and J were constructed of Pyrex glass, c, E, G, H and I of stainless steel and K of aluminium. For the analytical columns A-I, the flow rate of the carrier gas (nitrogen) was 30 ml/min. For the preparative columns J and K, the flow rate was 150 ml/min.

The gas-liquid chromatography-mass spectrometer combination (at Flinders University, Bedford Park) was made up of a Perkin-Elmer F11 instrument connected to an A.E.I.M.S. 30 mass spectrometer which operated in general at 15 eV. The g.1.c. columns used were: (L) 5 % SE30 on Chromosorb W (100-120 mesh), 1.2 m by 2.0 mm. (M) 10 % Carbowax 20M on Chromosorb W (100-120 mesh), 2 -0 m by 2.0 mm.

2-Cyclooctatetvaenylethanol(13)

An ice-cold colution of excess of ethylene oxide (10 ml) in ether (10 ml) was added rapidly (3 min) to a stirred ethereal solution of cycIooctatetraenylIithium [from bromocyclooctatetraene6 (4.5 g, 0.025 mol)] at -70' under nitrogen. After the dark red solution had been stirred overnight at room temperature, it was poured into ice-water (50 ml), acidified to pH c. 4 with dllute hydrochloric acid and extracted with ether. The ether extract was then washed successively with saturated sodium bicarbonate solution (20 ml), water (20 ml) and brine (20 ml), dried and concentrated to give an orange residue (5.0 g) which was distilled to give 2-cyclooctatetraenylethanol (13) (2.55 g, 70%) as a yellow oil, b.p. 58-6O0/0.02 mm, n i 5 1.5485 (lit.49 b.p. 77-7S0/0.2 mm; ng5 1.5480) (Found: C, 80.5; H, 8.5. Calc. for CloHI2O: C, 81.0; H, 8.2%). v,,, 3300 br, 1640m cm-'; n.m.r.: 6 5.77 (7H, broad singlet), 3.52 (2H, triplet, J 7 Hz), 2.33 (lH, singlet, removed by D 2 0 exchange), 2.23 (2H, triplet, J 7 Hz); mass spectrum: mle 148 (16%), 117 (100). G.1.c. analysis (columns A,

150"; B, 162") showed that the compound was pure. When cyclooctatetraenylmagnesium bromide,12* and not cyclooctatetraenyllithium, was used

in the above reaction, a low yield (21 %) of the required alcohol was obtained. The use of the Grignard reagent often gave impure (13) which could be satisfactorily purified however by (a) extraction with 20% (by weight) aqueous silver nitrate and regeneration from the formed silver nitrate complex with excess of concentrated ammonium hydroxide or (b) column chromatography on neutral alumina (Woelm) or silica gel.

Ethyl 3-Cyclooctatetraenylpvopanoate (18)

A solution of freshly distilled ethyl acrylate (2.6 g) in ether (10ml) was added dropwise (2.5 h) to a stirred ethereal solution (50 ml) of cyclooctatetraenylmagnesium bromide [from bromocyclooctatetraene (4.9 g)] containing cuprous chloride (0.1 g) at 0" under nitrogen. After the yellow- brown mixture had been stirred at room temperature for 2 h, it was treated with saturated ammonium chloride solution and extracted with ether. The ether extract, after being washed with aqueous sodium bicarbonate solution followed by water, was dried and concentrated to give an oil (4.6 g) which was distilled to give ethyl 3-cyclooctatetraenylpvopanoate (18) (1.1 g, 20%) as a yellow oil, b.p. 77-7g0/0.2 mm (Found: C, 76.4; H, 8.1. C13H1602 requires C, 76.4; H, 7.9%). v,,, 1730s, 1635w, 1180s cm-'; n.m.r.: 6 5.73 (7H, broad singlet), 4.05 (2H, quartet J 7 Hz), 2.35 (4H, broad multiplet), 1.23 (3H, triplet, J 7 Hz); mass spectrum: m/e 204 (39%), 117 (100). G.1.c. (column A, 145') showed that the compound was homogeneous.

When cyclooctatetraenyllithium [from bromocyclooctatetraene (3.4 g)] at - 60" was treated with ethyl acrylate (1.8 g) in a manner analogous to that described above, ethyl 3-cyclooctatetraenyl- propanoate (0.2 g, 5%) was obtained. The residue was a black gum.

* It was found that the rate of addition of ethylene dibromide (as initiator) was critical; satisfactory yields were obtained only when the dibromide was added dropwise very slowly (4-8 h).

49 Cope, A. C., and Rugen, D. F., J. Amer. Chem. Sac., 1953, 75, 3215.


(i) From ethyl 3-cyc1ooctatetraenylpropanoate.-A solution of ethyl 3-cyclooctatetraenyl- propanoate (18) (0.12 g) in ether (10 ml) was added dropwise (10 min) to a stirred suspension of lithium aluminium hydride (0.05 g) in ether (10 ml) under nitrogen at room temperature. After the mixture had been stirred at room temperature for 4 h, it was treated dropwise at 0" with water (2 ml) followed by dilute hydrochloric acid, and then extracted with ether. The ether extract, after being washed successively with saturated sodium bicarbonate solution and water, was dried and concentrated to give a residue (0.095 g) which was distilled to give 3-cyclooctatetraenylpropan-1-01 (17) (0.085 g, 90 %) as a yellow oil, b.p. 66" (block)/0.05 mm (lit.50 91-91. SO/O. 15 mm) (Found: C, 81.4; H, 8.7. Calc. for CllH140: C, 81.4; H, 8.7%). vmax 3300s, 1640w, 1060s cm-I; n.m.r.: 6 5.70 (7H, broad singlet), 3.61 (2H, triplet, J 7 Hz), 2.50 (IH, singlet; removed by D,O exchange), 2.11 (2H, triplet, poorly resolved), 1.60 (2H, quintuplet, poorly resolved); mass spectrum: m/e 162 (28%), 117 (100). G.1.c. (column A, 170") showed that the sample was homogeneous.

(ii) From c~clooctatetraenyltlzagnesium bromide and oxetan.-A solution of oxetan (3.2 g, 0.055 mol) in ether (10 ml) was added dropwise (15 min) to a stirred ethereal solution (50 ml) of cyclooctatetraenylmagnesium bromide [from bromocyclooctatetraene (4.9 g, 0.027 mol)] at 0' under nitrogen. After the mixture had been stirred overnight at room temperature, it was refluxed for 1 h. Benzene (30 ml) was added, and ether (c. 20 ml) was carefully removed by distillation through a column (13 cm, packed with glass helices) until the temperature of the vapour reached c. 40". After additional benzene (20 ml) had been added to it, the solution was heated under reflux for 2 h, and then the ether was removed by distillation (bath at 80"). The mixture was cooled, added to saturated ammonium chloride and extracted with ether. After the ether extract had been washed with aqueous sodium bicarbonate and water, it was dried and concentrated to give a residue (6 g) which was distilled to give 3-bromopropan-1-ol(2.2 g, 29 %) as a colourless oil, b.p. 30-33"/0.02 mm, and 3-cyclooctatetraenylpropan-1-01 (17) (2.2 g, 51 %), b.p. 83-87"/0.l mm. G.1.c. (column A,

170") showed that the product was homogeneous. Its spectral properties were identical with those of the sample prepared by method (i) above.

After a mixture of 3-cyclooctatetraenylpropylp-nitrobenzenesulphonate (19) (0.298 g, 0.85 mmol) and sodium cyanide (0.085 g, 1 . 7 mmol) in dry dimethylformamide (5 ml) had been stirred at room temperature under nitrogen overnight, it was added to water and extracted with ether. The ethereal extract was washed with water, dried and concentrated to give a residue which was distilled to give 4-cyclooctatetraenylbutyronitrile (22) (0.146 g, 100%) as a yellow oil, b.p. 95" (block)/O,2mrn, ni5 1.5290 (lit.'O 103-104"/0.5 mm, nA5 1,5300). v,,, 2250m, 1640m, 805s, 695s and 665s cm-'; n.m.r.: 6 5.80 (7H, broad singlet), 2.5-1.0 (6H, complex); mass spectrum: m/e 171 (14%), 117 (100).

4-Cyclooctatetraenylbutanoic Acid (23) A mixture of 4-cyclooctatetraenylbutyronitrile (22) (6.0 g) and sodium hydroxide (9.0 g) in

water (150 ml) was heated under reflux in an atmosphere of nitrogen for 6 h and then worked up in the described manners0 to give (23) as a yellow oil (4.4 g, 67%), b.p. 100" (block)/O.2 mm, n p 1.5300 (liLS0 95-97"/0.02mm, n p 1.5310). v,,, 3400-2600 br, 1700s, 1640w, 805m, 695m, 665m cm-l ; n.m.r.: 6 10.40 (lH, broad singlet), 5.72 (7H, broad singlet), 2.5-1.2 (6H, complex); mass spectrum: mle 190 (21 %), 117 (100).

4-Cyclooctatetraenylbutan-1-02 (21) Esterification of 4-cyclooctatetraenylbutanoic acid (23) (4.4 g, 0.023 mol) with ethereal diazo-

methane gave methyl 4-cyclooctatetraenylbutanoate (24) as a yellow oil (4.7 g, 100%) whose spectral properties were v,,, 1732s, 1640w, 1215m, 1160m, 805m, 695m, 670m cm-l; n.m.r.: 6 5.73 (7H, broad singlet), 3.58 (3H, singlet), 2.5-1.2 (6H, complex); mass spectrum: mle 204 (2273, 117 (100).

A solution of the crude ester (4.7 g, 0.023 mol) in ether (50 ml) was added dropwise to a stirred suspension of lithium aluminium hydride (1.75 g) in dry ether (50 ml) at room temperature. After

5 0 Cope, A. C . , and Pike, R. M., J. Amer. Chem. Soc., 1953, 75, 3220.

the mixture had been stirred under nitrogen for 23 h at room temperature, it was carefully treated with water followed by dilute hydrochloric acid. It was then extracted with ether and the ethereal extract was washed with aqueous sodium bicarbonate, dried and concentrated to give 4-cyclooctatetraenylbutan-1-01 (21) as a yellow oil (3.5 g, 91 %), b.p. 77-78"/0.01 mm (Found: C, 81.75; H, 9.4. C12H160 requires C, 81.8; H, 9.15%). v,,, 3300 br, 1635w, 1050m, 805w, 695m, 660m cm-' ; n.m.r. : 6 5.68 (7H, broad singlet), 3.50 (2H, poorly resolved triplet, J 7 Hz), 2.22 (lH, singlet, removed by D,O exchange), 1.82 (2H, poorly resolved triplet, J 7 Hz), 1.47 (4H, complex); mass spectrum: m/e 176 (2273, 117 (100). G.1.c. analysis (column c, 180'; D, 160") showed that the compound was homogeneous.

A solution of bromine (160 g) in methylene chloride (200 ml) was added dropwise (2.5 h) to a stirred solution of cyclooctene (110 g) in methylene chloride (100 ml) under nitrogen at 0'. Removal of solvent in vacuum gave crude 1,2-dibromocyclooctane (266 g) which was mixed with diethylamine (700 ml), sealed in a glass tube and heated at 100' for 22 h. After the solid (diethylamine hydrochloride) had been filtered off, the filtrate was washed successively with water, dilute sulphuri~ acid, water, aqueous sodium bicarbonate and brine. The dried solution was then concentrated and distilled to give 1-bromocyclooctene (112 g, 60%) as a colourless liquid, b.p. 83-85"/10 mm (lit.25,51 97-98'123 mm, 92-93"/13 mm). G.1.c. (column A, 100") indicated that the product was c. 97% pure.

Prior to use in the metallation reactions, samples of 1-bromocyclooctene were chromatographed on neutral alumina (elution was carried out with low-boiling light petroleum) and fractionally distilled (at c. 10 mm). This procedure gave samples of 1-bromocyclooctene which were c. 99.5% pure as shown by g.1.c. (column B, 100°).

Cyclooctenylmagnesium Bromide and Cyclooctenyllithium Cyclooctenylmagnesium bromide was prepared from I-bromocyclooctene and magnesium in

tetrahydrofuran (freshly distilled from lithium aluminium hydride) by the method of Maitte." Iodine and ethylene dibromide were sometimes used as initiators. When the Grignard solution was filtered, unchanged magnesium (c. 30%) was usually recovered.

CyclooctenyllithiumZ4 was prepared from I-bromocyclooctene and lithium in dry ether, Iodine, methyl iodide and mercuric chloride were sometimes used as initiators.

(i) A solution of ethylene oxide (10 ml) in cold ether (10 ml) was added rapidly to an ethereal solution (50 ml) of cyclooctenyllithium (from I-bromocyclooctene, 5.0 g, 0.026 mol) at 0" under nitrogen. The mixture, after being stirred at 0" for 30 min and then at room temperature for 30 min, was poured onto ice, acidified with dilute hydrochloric acid and extracted with ether. After the ethereal extract had been washed with aqueous sodium bicarbonate, it was dried and concentrated; the residue was distilled to give 2-(cyclooct-1'-eny1)ethanol (27) (0.92 g, 22%) as a colourless oil, b.p. 64"/0.1 mm, which was homogeneous (columns A, 150"; D, 140") (Found: C , 77.6; H, 11.4. CI0Hl80 requires C, 77.9; H, 11.8 %), v,,, 3400s, 1660w, 1080s cm-I ; n.m.r.: 6 5.41 (lH, triplet, J 8 Hz), 3.50 (2H, triplet, J 7 Hz), 2.15 (7H, multiplet, 1H removed by exchange with D20), 1.50 (8H, broad singlet); mass spectrum: mle 154 (2073, 67 (100).

(ii) A solution of ethylene oxide (5 rnl) in cold tetrahydrofuran (5 ml) was quickly added to a solution of cyclooctenylmagnesium bromide [from 1-bromocyclooctene (9.45 g, 0.05 mol)] in tetrahydrofuran (50 ml) at 0' under nitrogen. After the mixture had been stirred at room temperature for 1 h, it was cooled to Oo and treated slowly with saturated aqueous ammonium chloride (50 ml). The mixture was extracted with ether, and the ethereal extract was washed with saturated aqueous sodium bicarbonate, dried and concentrated to give a colourless oil (8.4 g) which was carefully distilled to give pure 2-(cyclooct-1'-eny1)ethanol (27) (2.0 g, 25%), b.p. 64-67"/015 mm.

I-Allylcyclooctanol After a solution of cyclooctanone (1 1 .O g, 0.087 mol) in ether (30 ml) had been added dropwise

(30 min) to a stirred solution of allylmagnesium bromide52 (0.09 mol) in ether (50 ml) at room

51 Kohler, E. P., Tishler, M., Potter, H., and Thompson, H. T., J. Amer. Chem. Soc., 1939, 61, 1057. 52 Grummitt, O., Budewitz, E. P., and Chudd, C. C., Org. Syn., 1963, Coll. Vol. IV, 749.


temperature under nitrogen, the mixture was stirred (at room temperature) for 41 h and then refluxed for 3 h. Workup with dilute hydrochloric acid (5 %) followed by extraction with ether, and removal of solvent in the usual manner, gave an oil (13.2 g) which was distilled to give 1-allylcyclooctanol (8.2 g), b.p. 67-7O0/0 5 mm, which contained cyclooctanone (5 %) as shown by g.1.c. (column A,

150"). Redistillation gave I-allylcyclooctanol (7.0 g, 50 %) as a colourless oil, b.p. 70-7l0/0. 7 mm, which had a purity of > 99% (column A, 150") (Found: C, 78.1 ; H, 11.9. C11H2,0 requires C, 78.5; H, 12.0%). v,,, 3400s, br, 3060w, 1640m, 995m, 900s cm-' ; n.m.r.: 6 6.2-5.4 (lH, multiplet), 5.2-4.7 (2H, multiplet), 2.3-2.1 (3H, multiplet; 1H removed by DZO exchange), 1 .5 (14H, broad singlet).

A solution of oxetan ( 4 6 4 g, 0.08 mol) in ether (10 ml) was added dropwise (15 min) to a stirred solution of cyclooctenyllithium [from I-bromocyclooctene (7.2 g, 0.038 rnol)] in ether (100 ml) under nitrogen at room temperature. After the mixture had been stirred at room temperature for 2 h, it was refluxed for 2 h, then cooled and added to saturated ammonium chloride solution. The mixture was extracted with ether and the ether extract, after being washed with aqueous sodium bicarbonate solution, was dried and concentrated to give a residue which was distilled to give 3-(cyclooct-1'-eny1)propan-1-01 (29) (3.02 g, 48%) as a colourless oil, b.p. 77-78"/005 mm (Found: C, 78.6; H, 12.0. C1,H2,0 requires C, 78.5; H, 12.0%). v,,, 3350 br, 1660w, 1055s cm-'; n.m.r.: 6 5.30 (lH, triplet, J 8 Hz), 3.52 (3H, broad triplet, J 7 Hz; 1H was removed by D 2 0 exchange to leave a sharp triplet), 2.10 (6H, complex), 1.47 (lOH, broad singlet); mass spectrum: m/e 168 (19 %), 96 (100). G.1.c. analysis (columns D, 165"; E, 1 7 0 ; F, 185") showed that the compound was homogeneous.

A solution of cyclooctanone (28.0 g, 0.22 mol) in dry ether (50ml) was added dropwise (30min) to a stirred solution of excess of but-3-enylmagnesium bromide [prepared from 4-bromobut-1-ene (44.5 g, 0.33 mol) and magnesium (8.3 g, 0.34 g-atom) in ether (100 ml)]. After the mixture had been stirred at room temperature for 18 h, it was refluxed for 3 h, then cooled and worked up (with saturated ammonium chloride solution) in the usual manner. The residue (35 g) was distilled to give a fraction (15.6 g), b.p. 79-85O10.25 mm which after two more fractional distillations gave I-(but-3'-enyl)cyclooctanol (35) as a colourless oil (14.0 g, 3573, b.p. 74"/O. 1 mm (Found: C, 79.1 ; H, 12.1. C12H220 requires C, 79.1 ; H, 12.2%). v,,, 3400s, 1640m, 900m cm-'; n.m.r.: 6 6 . 2 4 . 7 (3H, ABC pattern), 2.3-1.2 (19H, complex including a broad singlet (1H) which was removed by D 2 0 exchange). G.1.c. analysis (column G, 180"; c, 190") showed that the compound was homogeneous.

A solution of 1-(but-3'-enyl)cyclooctanol (35) (16 g) and oxalic acid dihydrate (16 g) in dimethyl sulphoxide (30 ml) and water (2.3 ml) was heated under nitrogen at 110" for 24 h. The mixture was cooled, diluted with water (50 ml) and extracted with low-boiling light petroleum. After the organic extract had been washed successively with aqueous sodium bicarbonate solution and water, it was dried and concentrated to give a residue which was distilled to give 4-(cyclooct-1'-eny1)- but-1-ene (13.2 g, 91 %), b.p. 98-99"/10mm (Found: C, 88.1; H, 12.1. CIZHz0 requires C, 87.7; H, 12.3%). Analysis by g.1.c. (column A, 100") revealed the presence of an impurity (15%), probably 4-cyclooctylidenebut-1-ene (37). The spectral properties were: v,,, 3070w, 1635m, 905s cm-'; n.m.r.: 6 6.0-4.7 (4H, includes an ABC pattern for 3H), 2.65 (small broad signal presumably from the two bis-allylic hydrogens in the impurity-integration is consistent with 15 % of the impurity), 2.02 (8H, multiplet), 1.40 (8H, singlet); mass spectrum: mle 164 (673, 81 (100).

To d i~ iamylborane~~ [prepared from 2-methylbut-2-ene (12.6 g, 0.18 mol) and sodium boro- hydride (2.66 g, 0.07 mol) in diglyme (50ml) and boron trifluoride etherate (12.8 g, 0.09 rnol)] at O" in an atmosphere of nitrogen was added rapidly 4-(cyclooct-1'-eny1)but-1-ene (36) (10.0 g, 0.061 rnol). After the mixture had been stirred at room temperature for 16 h, water (10ml) followed

53 Brown, H. C., and Zweifel, G., J. Amer. Chem. Soc., 1961, 83, 1241.

by sodium hydroxide solution ( 3 ~ , 50 ml) was slowly added to it. Hydrogen peroxide (30%, 50 ml) was then added at such a rate as to keep the temperature below 50". The mixture was then stirred at 50" for 33 h and, after being cooled, it was extracted with ether. After the ethereal extract had been thoroughly washed with water and dried, the solvent was removed in vacuum to leave a residue which was distilled to give the required alcohol (7.2 g, 65 %), b.p. 91-94"/0.06 mm (Found: C, 79.2; H, 12.3. Cl2HZ20 requires C, 79.1; H, 12.2%). Analysis by g.1.c. (column F, 190") showed the presence of an impurity (5 %), probably 4-cyclooctylidenebutan-1-01 (38). The spectral properties were: v,,, 3350 br, 1660w, 1060m cm-I; n.m.r.: 6 5.32 (lH, triplet, J 8 Hz), 3.45 (3H, 1H removed by D,O exchange to leave a triplet, J 7 Hz), 2.10 (6H, broad singlet), 1.50 (12H, broad); mass spectrum: mle 182 (12%), 67 (100).

A solution of 2-cyclooctatetraenylethanol (13) (0.3 g) in absolute ethanol (15 ml) containing Adams platinum oxide (0.15 g) was stirred in an atmosphere of hydrogen for 2 h. After the solution had been filtered, it was concentrated in vacuum to give an oil (0.3 g) which was distilled to give 2-cyclooctylethanol(40) (0.26 g, 80%) as a colourless oil, b.p. 100" (block)/O. 15 mm, ni5 1.4821 (lit.49 78-79"/0.2mm, n g 1.4825), which was homogeneous as shown by g.1.c. (column D, 140"). v,,, 3400s, 1060m; n.m,r.: 6 3.58 (2H, triplet, J 7 Hz), 2.2-1.0 (18H, broad singlet; including 1H which was removed by D 2 0 exchange).

A mixture of 3-(cyclooct-1'-eny1)propan-1-01 (29) (1.0 g) and Adams platinum oxide (20 mg) in ethyl acetate (50ml) was stirred in an atmosphere of hydrogen for 2 days at atmospheric temperature arid pressure. After the mixture had been filtered, the filtrate was concentrated to give a residue (0.97 g, 97%) which was distilled to give 3-cyclooctylpropan-1-01 as a colourless oil, b.p. 100" (block)/O.2 mm (Found: C, 77.5; H, 12.8. CllH,,O requires C, 77.6; H, 13.0%). v,,, 3350s, 1030s cm-I ; n.m.r.: 6 3.88 (lH, singlet; removed by D 2 0 exchange), 3.50 (2H, triplet, J 6 Hz), 1.5 (19H, complex and broad).

A mixture of 4-(cyclooct-1'-eny1)butan-1-01 (33) (0.86 g) and Adams platinum oxide (30 mg) in ethyl acetate (50ml) was stirred in the presence of hydrogen at atmospheric temperature and pressure for 4 days. Workup in the usual manner gave 4-cyclooctylbutan-1-01 (42) (0.73 g, 83 %) as a colourless oil, b.p. 140" (block)/O. 4 mm (Found: C, 78.2; H, 13.0. C12H240 requires C, 78.2; H, 13.1 %). v,,, 3310 br, 1050s cm-l; n.m.r.: 6 3.8 (lH, broad singlet, removed by D,O exchange), 3.48 (2H, triplet, J 6 Hz), 1.40 (21H, broad and complex); mass spectrum: mle 184 (1 %), 69 (100).

p-Nitrobenzenesulphonate Esters

In earlier ~ o r k , l - ~ satisfactory yields of p-nitrobenzenesulphonate esters were obtained from a wide range of alcohols by the use of p-nitrobenzenesulphonyl chloride in a very large excess of pyridine. Application of this procedure to the cyclooctatetraenylalkanols [(13), (17) and (21)] gave almost entirely water-soluble derivatives which were not extractable into ether. When a small excess (1.5-2.5-fold) of pyridine was used, the abovementioned alcohols gave satisfactory yields of the desired p-nitrobenzenesulphonate esters. The following procedure for the preparation of 2-cyclooctatetraenylethyl p-nitrobenzenesulphonate (14) was also used to prepare the higher homologues (19) and (25).

p-Nitrobenzenesulphonyl chloride (0.44 g, 2 mmol) was added to a stirred solution of 2-cyclooctatetraenylethanol (13) (0.20 g, 1.35 mmol) in pyridine (0.24 g, 3 mmol) at 0'. After the mixture had been stirred at Oo for 40 min, water (2 drops) was added to it and the stirring was continued for a further 10 min. The mixture was then added to ice-cold water (50 ml) and extracted with ether. After the ether extract had been washed successively with water, aqueous hydrochloric acid (273, saturated aqueous sodium bicarbonate and water, it was dried and concentrated under vacuum at c. 20' to give a yellow residue (0.37 g, 82%) which was recrystallized several times from ether to give 2-cyclooctatetraenylethyl p-nitrobenzenesulphonate (14) as yellow crystals, m.p. 63-64" (Found: C, 57.4; H, 4.5; N, 4.1. Cl6Hl5NO5S requires C, 57.7; H, 4.5; N, 4.2%). v,,,


1640w, 1604w, 1535s, 1180s cm-l; n.m.r. (CDCI,): 6 8.6-7.9 (4H, AA'BB' system centred at 8.40 and 8.10, JAB, = JBA, = 9 Hz, JAa, = JBB, = 1 .5 Hz), 5.75 (7H, broad singlet), 4.13 (2H, triplet, J 7 Hz), 2.43 (2H, triplet, J 7 Hz).

On some occasions, the derivative (14) was obtained as a cream-yellow solid, m.p. 50-60°, which consisted of two crystalline (dimorphic) forms. Although both forms (yellow and cream crystals) could be almost completely separated mechanically, their melting points were not sharp. The spectral properties (infrared, n.m.r. and u.v.) of both forms were identical.

3-Cyclooctatetraenylgvopyl p-nitrobenzenesulphonate (1 9), m.p. 69-70" (Found : C, 58 8 ; H, 4.9; N, 3.8. C17H17N05S requires C, 58.8; H, 4.9; N, 4.0%). v,,, 1535s, 1360s, 1180s, 950s cm-I ; n.m.r. (CDCI,): 6 8.5-8.0 (4H, AA'BB' system centred at 8.39 and 8.09, J A B , = JBA, = 9 Hz, JAAg = JBB' = 1 .5 HZ), 5.73 (7H, broad singlet), 4.22 (2H, triplet, J 7 Hz), c. 2 .0 (4H, complex multiplet).

4-Cyclooctatetraenylbut-1-yl p-nitrobenzenesulphonate (25), m.p. 61-62" (Found: C, 59.6; H, 5.3; N, 4.1. C ~ S H ~ ~ N O ~ S requires C, 59.8; H, 5.3; N, 3.9%). v,,, 1530s, 1365s, 1190s, 960s cm-l; n.m.r. (CDCI,): 6 8.6-7.9 (4H, AA'BB' system centred at 8.41 and 8.06, J A B , = JBA' = 9 HZ, JAA, = JBB, = 1.5 HZ), 5.70 (7H, broad singlet), 4.10 (2H, triplet, J 6 Hz), 2.03 (2H, triplet, J 6 Hz), 1.50 (4H, complex).

By the method described for the conversion of 3-(2'-methylenecyclohexyl)propan-1-01 into its p-nitrobenzenesulphonate ester,2 the following derivatives were prepared.

2-(Cyclooct-If-eny1)ethyl p-nitvobenzenesulphonate (28), m.p. 66-67' (from ether) (Found: C, 56.7; H, 6.4; N, 4.1. Cl6HZ1NO5S requires C, 56.6; H, 6.2; N, 4.1 %). v,,, 1610w, 1545m, 1180s cm-l; n.m.r. (CDCI,): 6 8.55-7.60 (4H, AA'BB' system centred at 8.34 and 8.02, JAB. = JBA, = 9 Hz, JAA* = JBB, = 1 .5 HZ), 5.35 (IH, triplet, J 8 Hz), 4.12 (2H, triplet, J 7 Hz), 2.37 (2H, triplet, J 7 Hz), 2.07 (4H, singlet), 1.45 (8H, broad and complex).

3-(Cyclooct-1'-eny1)propyl p-nitrobenzenesulphonate (31), m.p. 71-72" (etherllow-boiling light petroleum) (Found: C 57.8; H, 6.7; N, 3.8. CI7HZ3NOSS requires C, 57.8; H, 6.6; N, 4.0%). v,,, 1615w, 1550m, 1355m, 1180s, 960s cm-l; n.m.r. (CDCI,): 6 8.6-7.9 (4H, AA'BB' system centred at 8.40 and 8.11, J A B , = J B A , = 9 HZ J A A ' = J B B , = 1 ' 5 Hz), 5.23 (IH, triplet, J 8 Hz), 4.15 (2H, triplet, J 6 Hz), 1.96 (6H, broad and complex), 1.45 (IOH, broad and complex).

4-(Cyclooct-1'-eny1)but-I-yl p-nitrobenzenesulphonate (34), m.p. 61-63' (ether) (Found: C, 58.8; H, 7.15; N, 3.7. ClsHZ5N05S requires C, 58.8; H, 6.9; N, 3.8%). v,,, 1610w, 1540m, 1175s, 955s cm-'; n.m.r. (CDCI,): 6 8.6-8.0 (4H, AA'BB' system centred at 8.41 and 8.10, JAB, = JBA, = 9 HZ, JAA, = JBB, = 1.5 HZ), 5.27 (lH, triplet, J 8 Hz), 4.12 (2H, triplet J 6 Hz), 2.08 (6H, broad and complex), 1.45 (12H, broad and complex).

2-Cyclooctylethyl p-nitvobenzenesulphonate (43), m.p. 62-64' (ether) (Found: C, 56.5; H, 7.0; N, 4.0. C16H2,N05S requires C, 56.3; H, 6.8; N, 4.1 %). v,,, 3080w, 1608m, 1548s, 1180s cm-l; n.m.r. (CDCI,): 6 8.6-7.9 (4H, AA'BB' system centred at 8.40 and 8.10, JAB, = JBA, = 9 Hz, JAA, = J B B , = 1 .5 Hz, 4.12 (2H, triplet, J 6 Hz), 1.42 (17H, broad and complex).

3-Cyclooctylpropyl p-nitrobenzenesulphonate (44), m.p. 68-70" (etherilow-boding light petroleum) (Found: C, 57.3; H, 7.3 ; N, 4.1. C17HZ5N05S requires C, 57.5; H, 7.1; N, 3.9 %). v,,, 1535s, 1180s, 942 cm-I ; n.m.r. (CDCI,): 6 8.5-8.0 (4H, AA'BB' system centred at 8.38 and 8.05, JAB* = JBA, = 9 HZ, JAA, = J B B , = 1 .5 HZ), 4.08 (2H, triplet, J 6 Hz), 1.53 (17H, broad and complex).

4-Cyclooctylbutyl p-nitvobenzenesulphonate (45), m.p. 60-62' (ether) (Found: C, 58.5; H, 7.4; N, 4.1. ClsH27N05S requires C, 58.5; H, 7.4; N, 3.8%). v,,, 1535m, 1185s, 950s cm-'; n.m.r. (CDCI,): 6 8.5-7.9 (4H, AA'BB' system centred at 8.37 and 8.06, J A B = J B A , = 9 Hz, JAA, = JBB, = 1.5 Hz), 4.10 (2H, triplet, J 6 Hz), 1.53 (21H, broad and complex).

3-Phenylpropyl p-nitrobenzenesulphonate, m.p. 58-59" (ether) (Found: C, 56.4; H, 4.8; N, 4.5. C15H15N05S requires C, 56.1; H, 4.7; N, 4.4%). v,,, 1540m, 1170s, 950s cm-'; n.m.r. (CDCI,): 6 8.6-7.9 (4H, AA'BB' system centred at 8.40 and 8.10, J A B , = JBA, = 9Hz, JAA. = JBB. = 1 .5 Hz), 7.10 (7H, complex multiplet), 4.11 (2H, triplet, J 7 Hz), 2.64 (2H, triplet, J 7 Hz), 2.00 (2H, quintuplet, J 7 Hz). 3-Phenylpropan-1-01 was prepared by catalytic (Pt) hydrogenation of trans-cinnamyl alcohol.

4-Phenylbutyl p-nitrobenzenesulphonate, m.p. 63-65" (ether) (Found: C, 57.1 ; H, 5.1 ; N, 3.9. CI6Hl7NO5S requires C, 5 7 3; H, 5.1 ; N, 4.2%). v,,, 1600m, 1530s, 1355s, 1 l7Os, 942s cm-l; n.m.r.: 6 8.5-7.9 (4H, AA'BB' system centred at 8.34 and 8.08, JAB' = JBA* = 9 HZ, JAAt = JBB, = 1 .5 HZ), 7.16 (5H, multiplet), 4.11 (2H, triplet, J 6 Hz), 2.52 (2H, triplet, J 6 Hz),

1.60 (4H, multiplet). 4-Phenylbutan-1-01 was prepared from 4-phenylbutanoic acid by reduction with lithium aluminium hydride in ether.

Acetyl Derivatives

By the general procedure previously described1 for the conversion of primary alcohols into their acetyl derivatives, the following acetates were prepared as homogeneous (by g.l.c., column A)

yellow oils in high yields: 2-Cyclooctatetraenylethyl acetate (16), b.p. 77" (block)/0.05 mm (Found: C, 75.5; H, 7.4.

C12H1402 requires C, 75.8; H, 7.4%). v,,, 1735s, 1635w, 1255s, 1050m, 815w, 700m, 670w cm-l; n.m.r. : 6 5.73 (7H, broad singlet), 4.05 (2H, triplet, J 7 Hz), 2 32 (2H, triplet, J 7 Hz), 1.98 (3H, singlet).

3-Cyclooctatetraenylpropyl acetate (20), b.p. 75' (block)/0.08 mm (Found: C, 76.3; H, 7.8. Ci3Hi602 requires C, 76.4; H, 7.9 %). v,,, 1735s, 1640w, 1245s, 1045m, 815w, 700m, 670w cm-I ; n.m.r.: 6 5.72 (7H, broad singlet), 4.05 (2H, triplet, J 7 Hz), 2.2-1.5 (7H, including a sharp singlet at 1.95); mass spectrum: mle 204 (30%), 129 (100).

4-Cyclooctatetraenylbutyl acetate (26), b.p. 105" (block)/O.O7mrn (Found: C, 77.1; H, 8.7. C14Hi8O2 requires C, 77.0; H, 8.3 %). v,,, 1735s, 1630w, 1225s, 1020s cm-' ; n.m.r.: 6 5.73 (7H, broad singlet), 4.00 (2H, triplet, J 7 Hz), 2.02 (2H, triplet, J 7 Hz), 1.95 (3H, singlet), 1 a53 (4H, complex); mass spectrum: mle 218 (13 %), 117 (100).

4-Cyclooctylbutyl acetate (74), b.p. 130" (block)/O.8 mm (Found: C, 14.7; H, 11.4. C14H26O2 requires C, 74.3; H, 11.6%). v,,, 1735s, 1240s cm-I; n.m.r.: 6 4.10 (2H, triplet, J 7 Hz), 2.3-1.8 (9H, multiplet including a singlet at 2.05), c. 1.5 (15H, multiplet, broad); mass spectrum: mle 226 (3 %), 55 (100). The acetate was prepared (93% yield) by the catalytic hydrogenation (Pt) of 4-cyclooctatetraenylbutyl acetate in ethyl acetate at atmospheric pressure and temperature.

Allylcyclooctatetraene (58)

After a mixture of bromocyclooctatetraene (4.9 g, 0.27 mol) and magnesium (1.51 g, 0.063 g-at.) in tetrahydrofuran (25 ml) had been stirred under nitrogen at 0" for 1 . 5 h, it was filtered and the filtrate was added to a solution of allyl bromide (18 g, 0.149 mol) in tetrahydrofuran (25 ml) at O". A solution of lithium chloride (0.085 g) and copper chloride (0.135 g) in tetrahydrofuran (1 ml) was added to the mixture which was then stirred at room temperature for 2.5 h. After the mixture had been added to saturated ammonium chloride solution, it was extracted with low-boiling light petroleum. The organic extract, after being washed with water and sodium bicarbonate solution, was dried and concentrated to give a residue (3.23 g) which was distilled to give allylcyclooctatetraene (58) (1.6 g, 44%) as a yellow oil, b.p. 89-9a0/20 mm (Found: C, 91.7; H, 8.4. CllHi2 requires C, 91.6; H, 8.4 %). v,,, 1638m, 1437m, 1000m, 920m, 805m, 700m cm-' ; n.m.r. : 6 6.1-4.9 (lOH, complex and broad, includes a singlet at 5.79), 2.78 (2H, doublet, J 6 Hz); mass spectrum: m/e 144 (40%), 103 (100). G.1.c. (column F, 135") showed that the product was homogeneous.

Butylcyclooctane (72)

A mixture of 4-(cyclooct-1'-eny1)but-1-ene (36) (0.2 g) and Adams platinum oxide (10 mg) in ethyl acetate (20 ml) was stirred in an atmosphere of hydrogen for 12 h at atmospheric pressure and temperature. The mixture was filtered and the filtrate was concentrated and distilled to give butylcyclooctane (72) (0.18 g, 88 %) as a colourless oil, b.p. 90" (block)/lO mm, nis 1.4588 (lit.54 b.p. 89"/10mm, ni5 1.4590). Its infrared spectrum was identical with that reported.54 G.1.c. analysis (columns H, 120"; I, 130") showed that the sample was homogeneous.

Bicyclo[6,4,0]dodecane (73)

A solution of bicyclo[6,4,0]dodec-1(8)-en-9-one55 (0.36 g) ' in ether (70 ml) saturated with hydrogen chloride was treated with zinc powder (3.5 g, added in portions over 30 min) at 0" under nitrogen. After the mixture had been stirred at 0" for 1 h, it was poured into ice-cold water (50 rnl) and extracted with ether. The ether extract, after being washed with aqueous sodium bicarbonate, was dried and concentrated to give a residue (0.34 g) which was chromatographed on Florisil (15 g). Elution with light petroleum gave a colourless oil (0.21 g) which was dissolved in acetic acid

54 Cope, A. C., and Van Orden, H. O., J. Amer. Chem. Soc., 1952, 74, 175. 5%irano, S., Hiyama, T., and Nozaki, H., Tetrahedron Lett., 1973, 1331.


(50 ml) containing platinum oxide (20 mg). After the mixture had been stirred in an atmosphere of hydrogen at atmospheric pressure and temperature for 1 day, it was added to water and extracted with low-boiling light petroleum. The organic extract was washed with aqueous sodium bicarbonate, dried, and concentrated to give bicyclo[6,4,0]dodecane (73) (0.20 g, 59%) as a colourless oil, b.p. 100" (block)/l2 mm, nL5 1.4839 (lit.56 ni5 1.4885 for cis-isomer); n.m.r.: 6 1.66 (12H, complex and broad), 1.32 (lOH, complex and broad).

G.1.c. analysis (columns A, 120"; I, 130') showed the presence of two components in the approximate ratio of 1 : 2, respectively, in increasing order of their retention times. The two components, the cis- and trans-isomers of (73) (but not necessarily in that order) were not separated. G.1.c.-mass spectrometry (column M, 100') showed that the mass spectra of both components (each with M at mle 166) were identical.

Preparative Acetolysis of 2-Cyclooctatetraenylethyl p-Nitrobenzenesulphonate (14) A solution of 2-cyclooctatetraenylethyl p-nitrobenzenesulphonate (0.27 g, 0 .8 mmol) and dry

sodium acetate (1.35 g, 16 mmol) in acetic acid (containing acetic anhydride (1 %)) (50 ml) was heated in a sealed tube under nitrogen at 85" for 12 h. After the cooled solution had been poured into ice-cold water (100 ml), it was extracted with low-boiling light petroleum (c. 100 ml). The petroleum extract was washed successively with cold water, cold 5% sodium bicarbonate solution and water, dried and then carefully concentrated by distilling almost all of the solvent through a column (30 cm) packed with glass helices while the temperature of the bath was maintained at 45-50". Preparative g.1.c. (column J, 170") of the concentrate (c. 3ml) gave 1,2-dihydronaphthalene (46), 2-cyclooctatetraenylethyl acetate (16) and bicyclo[5,3,0]deca-l,3,5-trien-8-yl acetate (47). 1,2-Dihydronaphthalene (46) and 2-cyclooctatetraenylethyl acetate (16) were identified by their g.1.c. characteristics (i.e, retention times and by 'spiking'), mass spectra, and in the case of (16) the infrared spectrum, with those of authentic samples.

The acetate (47), obtained as a blue liquid (due to the presence of traces of azulene which was identified by its g.1.c. behaviour) contained 2-cyclooctatetraenylethyl acetate (16) (5%) (identified by g.1.c.). Preparative thin-layer chromatography (equal mixture of Merck Kieselguhr G and H F 254, 20% ether in light petroleum) gave the acetate as a colourless oil; the acetate (16) was however still present. The compound exhibited the following spectral properties: v,,, 2950w, 1 7 3 0 ~ ~ 1640w, 1435m, 1380m, 1240s, 1040m, 705m cm-'; n.m.r.: (CDCI,): 6 6.55 (2H, multiplet, complex), 6.21 (2H, multiplet, broad), 5.40 (lH, multiplet, broad), 5.10 (lH, multiplet), 2.62 (2H, multiplet, broad), 2.2-1.9 (6H, multiplet, broad, including a singlet at 2.03) (small signals arising from the acetate (16) were also present in the spectrum); mass spectrum: mle 190 (absent), 131 (100%). The n.m.r. spectrum of the acetate (47) was identical with that obtained by Paquette and Henzell8 from (15).

Preparatiue Acetolysis of 3-Cycloocfatetraenylpropyl p-Nitrobenzenesulphonate (19) (i) A solution of 3-cyclooctatetraenylpropyl p-nitrobenzenesulphonate (19) (4.7 g, 13.5 mmol)

and dry sodium acetate (2.2 g, 27 mmol) in acetic acid (50 ml, containing 1 % acetic anhydride was heated in a sealed tube at 100' for 40 h. The solution was then worked up in the usual manner to give a yellow residue (in low-boiling light petroleum) which was immediately chromatographed on neutral alumina (Woelm, 100 g). Elution with low-boiling light petroleum gave a yellow oil which polymerized before satisfactory spectral data could be obtained. Further elution with low-boiling light petroleum containing 20% ether gave 3-cyclooctatetraenylpropyl acetate (20) (2.5 g, 92%) which was identified by its g.1.c. characteristics and infrared and n.m.r. spectra.

(ii) Repetition of the above acetolysis again gave a product (in low-boiling light petroleum) which was chromatographed (column N, 180"). The first two components were obtained as yellow gums. The third component was obtained as a yellow oil which was identified as 3-cyclooctatetraenylpropyl acetate as in (i) above.

Mass Spectrum of Compound Tentatiuely Identified as Bicyclo[6,3,0]undeca-1,3,5,7-triene (60)

When the mixture (in low-boiling light petroleum) obtained by acetolysis of 3-cyclooctatetraenylpropyl p-nitrobenzenesulphonate (19) (c. 36 mg) in acetic acid (5 ml) [which contained acetic anhydride (1 %) and sodium acetate (12.4 mg)] at 100' for 40 h was subjected to analysis on the

56 Cram, D. J., and Allinger, N. L., J. Amer. Chem. Soc., 1956, 78, 2518.

g.l.c.(column L, 80°)/mass spectrometer combination, the mass spectrum of the compound tentatively identified as (60) showed peaks at mle 144 (loo%), 143 (55), 142 (15), 141 (21), 138 (2), 130 (lo), 129 (89), 128 (63), 127 (16), 118 (4), 117 (33), 116 (19), 115 (39), 104 (2), 103 (8), 102 (4), 91 (12), 89 (4), 79 (2), 78 (5), 77 (7), 71 (2), 65 (4), 57 (5), 56 (2), 55 (2), 43 (7).

Prepavative Acetolysis of 4-Cyclooctatetraenylbutyl p-Nitrobenzenesulphonate (25)

(i) A solution of 4-cyclooctatetraenylbutyl p-nitrobenzenesulphonate (25) (1.0 g, 2.76 mmol) and dry sodium acetate (0.33 g, 4 mmol) in acetic acid (containing 1 % acetic anhydride) (50 ml) was heated in a sealed tube under nitrogen for 16 h. The solution was worked up in the usual manner to give a yellow concentrate which was shown by g.1.c. (column F, 180") to contain four components. The concentrate was immediately chromatographed on neutral alumina (Woelm, 25 g). Elution with low-boiling light petroleum separated the first three components (monitored by g.1.c.) as yellow solutions. Removal of the solvent in vacuum from these fractions gave yellow gums. Elution with low-boiling light petroleum c~ntaining 10-50% ether gave 4-cyclooctatetraenylbutyl acetate (26) (339 mg, 56%) which was identified by comparison of its infrared spectrum with that of an authentic sample.

(ii) Repeat of the above acetolysis gave a concentrate which was subjected to preparative g.1.c. (column K, 180"). The first and third components were obtained as yellow gums. The second component was obtained as a yellow oil which was identified as bicyclo[6,4,0ldodeca- 1,3,5,7-triene (65) by comparison of its g.1.c. characteristics and mass spectrum with those of an authentic ample.^^^^^ The fourth component was readily identified as 4-cyclooctatetraenylbutyl acetate (26) by comparison of its infrared spectrum with that of an authentic sample.

G.1.c.-Mass Spectvogvaphic Analysis of Products of Acetolysis of 4-Cyclooctatetraenylbutyl p-Nitrobenzenesulphonate (25)

After a solution of (25) (36 mg, 0 .1 mmol) and sodium acetate (12.3 mg, 0.15 mmol) in acetic acid (containing 1 % acetic anhydride) (5 ml) had been heated in a sealed tube under nitrogen at 100" for 20 h, it was worked up in the usual manner to give a concentrate (in low-boiling light petroleum) which was subjected to g.1.c.-mass spectroscopic analysis on column L at 80" (to elute olefins) followed by 120" (to elute the acetate). The mass spectra of the second and fourth components (in order of elution) were identical with those of bicyclo[6,4,0]dodeca-1,3,5,7-tetraene (65) and 4-cyclooctatetraenylbutyl acetate (26), respectively. The spectrum of the first component showed peaks at mle 158 (90%), 157 (25), 156 (16), 155 (lo), 154 (22), 153 (IS), 152 (20), 147 (15), 144 (12), 143 (55), 142 (15), 141 (25), 133 (26), 132 (IS), 131 (24), 130 (53), 129 (97), 128 (45), 127 (26), 119 (5), 118 (25), 117 (loo), 116 (41), 115 (86), 105 (8), 104 (44), 103 (95), 102 (20), 94 (15), 93 (lo), 92 (16), 91 (60), 90 (6), 81 (5), 80 (14), 79 (30), 78 (45), 77 (40), 73 (24), 71 (20), 70 (20), 69 (7), 67 (15), 66 (12), 65 (25), 64 (IS), 60 (23), 59 (9), 58 (lo), 57 (38), 56 (30), 55 (32), 54 (lo), 53 (6), 52 (lo), 51 (19), 50 (6). In the spectrum of the third component, peaks were shown at m/e 158 (loo%), 157 (16), 156 (5), 155 (2), 154 (2), 145 (2), 144 (5), 143 (35), 142 (6), 141 (5), 132 (3), 131 (13), 130 (67), 129 (74), 128 (IS), 127 (4), 119 (I), 118 (5), 117 (37), 116 (12), 115 (23), 105 (3), 104 (ll) ,

Hydrogenation of Products fvom Acetolysis of 4-Cyclooctatetraenylbutyl p-Nitrobenzenesulphonate (25)

After a solution of 4-cyclooctatetraenylbutyl p-nitrobenzenesulphonate (25) (35.8 mg, 0 .1 mmol) and sodium acetate (12.3 mg, 0.15 mmol) in acetic acid (5 ml) had been heated in a sealed tube under nitrogen at 100' for 16 h, it was transferred quantitatively (with the aid of some more acetic acid) to a flask (50 ml) containing Adams platinum oxide (0.3 g). The mixture was then stirred in an atmosphere of hydrogen at atmospheric pressure and temperature for 1 day and then worked up in the usual manner using low-boiling light petroleum as solvent. The concentrate (with 1-methylnaphthalene as internal standard) was analysed by g.1.c. (column F, 130" for 6 min, then raised to 180') in the usual way.

G.1.c.-mass spectroscopic analysis of the concentrate on column M at 100' (to elute hydro- carbons) and 150" (to elute the acetate) resolved four components with molecular ion peaks at m/e 168, 166, 166 and 225, respectively, in increasing order of retention times. The mass spectra (and retention times) of these components were identical with those of butylcyclooctane (72), the


two isomers of bicyclo[6,4,0]dodecane (73, cis and trans) and 4-cyclooctylbutyl acetate (74), respectively.

Analyses of Products from the Various Acetolyses

The general procedure for the analyses of the products formed by acetolysis of the three substrates (14), (19) and (25) reported in the present work has been outlined earlier.' Low-boiling light petroleum, and not ether, was used to extract the products from the diluted acetic acid in the working-up process. The petroleum extract, was carefully concentrated by distilling almost all the solvent through a column (30cm) packed with glass helices while the temperature of the bath was maintained at 45-50". 1-Methylnaphthalene was used as the internal standard, and the responses (to the g.1.c. flame ionization detector) of the authentic compounds (where available) with respect to the internal standard were determined. For those compounds which were not available in sufficient quantities, or states of purity, to allow their response ratios to be determined, the ratios were assumed to be the same as that of the most closely related compound for which the value was known. Both the qualitative and quantitative analyses were carried out by g.1.c. on columns F and I.

Kinetic Determinations

The rates of acetolysis of the various p-nitrobenzenesulphonates (Tables 1, 3 and 4) were determined titrimetrically by the ampoule technique as previously described.' Bromophenol blue was generally used as an indicator. For the various w-cyclooctatetraenylalkyl derivatives, crystal violet was used; this indicator overcame difficulties with the yellow colour inherently associated with cyclooctatetraenyl derivatives and the blue colour (due to traces of azulene being formed) which developed during the acetolysis of 2-cyclooctatetraenylethyl p-nitrobenzenesulphonate. Values of AH* and AS$ and their standard deviations, as well as extrapolated values of rates constants, which appear in Tables 1, 3 and 4, were determined on a CDC 6400 computer by use of QCPE Program No. 79.

Acknowledgments

We are extremely grateful to Dr S. Hirano (Kyoto University), Professor L. Paquette (Ohio State University) and Badische Anilin- and Soda-Fabrik A.G. for gifts of bicyclo[6,4,0]dodec-1(8)-en-9-one, bicyclo[6,4,0]dodeca-1,3,5,7-tetraene and cyclooctatetraene, respectively. Professor Paquette's assistance in the identification of bicyclo[5,3,0]deca-1,3,5-trien-8-y1 acetate is also gratefully acknowledged. We are also indebted to Dr M. Thompson and Mr K. Steer (both of Flinders University) for their generous assistance with the gas-liquid chromatography-mass spectral determinations.

One of us (M.M.) acknowledges with gratitude the receipt of a Commonwealth Postgraduate Award. Finally, we thank the Australian Research Grants Committee for financial support of the work.

Manuscript received 27 March 1975

Documents

Australian Journal of Chemistry (1975), 28(10), 2227-54